Challenges and Opportunities For Safe Water Supply in

Challenges and Opportunities For Safe Water Supply in Mozambique
Matsinhe, Nelson
Published: 2008-01-01
Link to publication
Citation for published version (APA):
Matsinhe, N. (2008). Challenges and Opportunities For Safe Water Supply in Mozambique Department of Water
Resources Engineering, Lund Institute of Technology, Lund University
General rights
Copyright and moral rights for the publications made accessible in the public portal are retained by the authors
and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the
legal requirements associated with these rights.
• Users may download and print one copy of any publication from the public portal for the purpose of private
study or research.
• You may not further distribute the material or use it for any profit-making activity or commercial gain
• You may freely distribute the URL identifying the publication in the public portal ?
Take down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove
access to the work immediately and investigate your claim.
L
UNDUNI
VERS
I
TY
PO Box117
22100L
und
+46462220000
ACKNOWLEDGEMENTS
During the five years of work that has resulted in this thesis many people have contributed in
one way or another and I am grateful to all of you; but there are some who deserve much
gratitude for their support and guidance.
First of all, I would like to thank Professor Kenneth M. Persson, my supervisor, for guidance,
encouragement and advice throughout the thesis work. Thank you for always taking time for
discussion of the work even, when you had none.
I also thank Professors Lars Bengtsson and Rolf Larsson for giving me the opportunity to join
the Department of Water Resources Engineering of Lund University and for facilitating the
cooperation between Eduardo Mondlane University and Lund University.
The field work and laboratory experiments used in this thesis have been possible thanks to
numerous people that have not only helped with the experiments but also with interesting
discussions and suggestions to keep the work on despite some difficulties encountered. Maria
das Dores for the most of water chemistry analysis performed and Alfredo Manuel for always
fixing the equipment deserve grateful appreciation. I am also particularly grateful to Malin Kajsa,
Anna Lindquist, Emelie Arnoldsson and Maria Bergman for having conducted research work
that helped produce two of the six papers of this thesis.
All colleagues from the Department of Water Resources Engineering are thanked for making me
feel at home in Lund and share their happy mood at coffee breaks and pubs.
Special thanks to my friends Juizo and Luuk Rietveld, for all fruitful discussions and
suggestions.
This research was conducted as part of an ongoing research project jointly implemented by
Eduardo Mondlane University and Lund University which is funded by the Swedish International
Development Cooperation Agency-SIDA-SAREC. I am grateful to their support.
I dedicate this thesis, to Paula and Malik, my wife and son. Thanks for your love and
encouragement during this journey.
Nelson Matsinhe
I
ABSTRACT
In Mozambique, despite considerable investments in the promotion of drinking water supplies, the access
to quality water of sufficient quantity for the majority of people is still far from optimal. Current official
figures report that nearly half of the country´s population and about 45 per cent of urban residents do not
have access to safe water. As a result of poor performance or absence of public water services, in most
areas, informal private operators supply water to the most of underserved populations. Management of
drinking water quality is inadequate and is affected by limitations at production and distribution level.
This research included an analysis of drinking water supply aspects of Mozambique with the view to
contribute to the understanding of the factors affecting present drinking water supply and the challenges
facing the water governance sector in developing and maintaining sustainable drinking water supplies.
Emphasis was put on identifying critical factors affecting production and management of drinking water
quality.
The analysis of water quality aspects shown that present limitations in water safety and water quality are
due mainly to lack of adequate treatment, inadequate management of distribution and lack of knowledge
among operators. The quality of water sources used for drinking water production is very similar to that of
many other parts of the world and the methods used for water treatment are, suited for production of
excellent treated water quality. However, poor knowledge and inefficient operation of treatment processes
causes drinking water production to be ineffective.
Methods of improving drinking water treatment were also investigated. For reasons of sustainability, low
cost treatment methods were selected. Pre-treatment with up-flow roughing filtration and use of natural
coagulants (Moringa Oleifera) for water treatment were the methods tested. The results proven that if
properly incorporated in the drinking water treatment strategy of the country, these methods can provide a
viable and sustainable alternative for improved drinking water production.
Service quality aspects of informal private operators were also analysed. It was concluded that they
provide a reliable alternative for access and for expansion of service delivery to areas lacking piped water
supply. It was also concluded that present human health risks for consumers relying on these services
are comparable to that of formal water supplies. However, the lack of an inclusive regulatory framework to
this type of service providers limits the possibilities for regulation of their activities. Therefore, regulation
aspects around formal and informal service providers formed part of the research and a proposal for
expanding the existing regulatory framework was presented. Licensing and regulatory functions needed
are presented.
The main conclusion of this study is that two major factors affect drinking water supply in Mozambique
specifically; limited service coverage and; limitations in water safety and water quality caused by lack of
adequate treatment, poor management of water distribution and lack of knowledge among operators. The
main contribution of this study is to the water governance sector of Mozambique and it refers to the
various possibilities offered by methods tested during this study, for sustainable improvement of drinking
water production. In addition, the findings of the discussion of the drinking water supply situation looking
not only at quantity but also at quality aspects of service delivery as was done in this study, will hopefully
help the sector redefine its strategy of addressing drinking water supply in Mozambique.
II
CONTENTS
ACKNOWLEDGEMENTS .................................................................................................................... I
ABSTRACT .......................................................................................................................................... II
APPENDED PAPERS ......................................................................................................................... V
RELATED PUBLICATIONS .................................................................................................................VI
1.
2.
3.
INTRODUCTION ......................................................................................................................... 1
1.1
Background and identification of the Problem .......................................................... 1
1.2
Objectives of the research ............................................................................................. 2
1.3
Overall methodology ...................................................................................................... 3
1.4
Thesis structure and appended papers ....................................................................... 3
OVERVIEW OF THE DRINKING WATER SUPPLY SECTOR OF MOZAMBIQUE....................... 5
2.1
Background on the study area...................................................................................... 5
2.2
Overview of the legal and institutional framework .................................................... 6
2.3
Drinking water supply options and choice of technology........................................ 8
2.4
Drinking water treatment methods and quality criteria ............................................ 9
2.5
Alternative water supplies.............................................................................................10
METHODOS..............................................................................................................................12
3.1
Field work .........................................................................................................................12
3.2
Sources of data and choice of water quality variables..........................................12
3.3
Situation analysis with respect to drinking water treatment and efficiency .......13
3.4
Situation analysis with respect to drinking water quality .........................................13
3.5
Laboratory testing ..........................................................................................................14
III
4.
RESULTS AND DISCUSSION.....................................................................................................15
4.1
5.
6.
Source water quality and methods of treatment.....................................................15
4.1.1
Surface water quality .............................................................................................15
4.1.2
Groundwater quality ..............................................................................................18
4.2
Drinking water treatment : current practices and treatment efficiency .............20
4.3
Coverage and service quality .....................................................................................23
4.3.1
Coverage and service quality provided by formal service providers...........23
4.3.2
Coverage and service quality provided by informal service providers........25
CHALLENGES AND OPPORTUNITIES FOR SAFE WATER SUPPLY IN MOZAMBIQUE .........27
5.1
Meeting targets in drinking water production and coverage...............................28
5.2
Meeting drinking water quality targets ......................................................................29
5.3
Management of drinking water quality .....................................................................34
5.4
Role of alternative service providers...........................................................................36
5.5
Regulation and legal aspects around formal and informal service providers....37
SUMMARY AND CONCLUSIONS...........................................................................................39
REFERENCES ....................................................................................................................................42
Appendix. Papers ..........................................................................................................................48
IV
APPENDED PAPERS
This thesis is based on the following papers, which will be referred to in the text by their Arabic
numerals. The papers are appended at the end of the Thesis.
Paper 1
Matsinhe, N.P.; Juizo J.; Persson, K.M. and Rietveld L.C. (2008). Challenges
facing drinking water production in Mozambique – A review of critical
factors affecting treatment possibilities. Submitted to journal Water Science
and Technology.
Paper 2
Matsinhe N.P.; Juizo J. and Persson, K.M. (2008). The effect of intermittent
supply and household storage on the quality of drinking water in Maputo.
Submitted to Water Science and Technology – Water supply.
Paper 3
Matsinhe, N.P.; Juizo J.; Rietveld L.C. and Persson, K.M. (2008). Water services
with independent providers in peri-urban Maputo: Challenges and
opportunities for long-term development. Water SA, Vol. 34, No. 3, pp 411420.
Paper 4
Matsinhe N.P.; Juizo J.; Macheve, B. and dos Santos, C. (2007). Regulation of
formal and informal water service providers in peri-urban areas of Maputo,
Mozambique. Journal Physics and Chemistry of the Earth, Vol. 33, No. 8-13, pp
841-849.
Paper 5
Matsinhe N.P. and Persson, K.M. (2008). Hydraulic flocculation with up-flow
roughing filtration for pre-treatment of surface water prior to conventional
rapid sand filtration. Submitted to Environmental Science and Engineering.
Paper 6
Arnoldsson, E.; Bergman, M.; Matsinhe N.P. and Persson K.M. (2008).
Assessment of drinking water treatment using Moringa Oleifera natural
coagulant. Watten, Vol.64, pp 137-150.
V
RELATED PUBLICATIONS
Conference Proceedings
Matsinhe N.P.; Quessoni J. (2005): Influência da Reserva Domiciliar na qualidade da Água
consumida na cidade de Maputo. 4th Congresso Luso-Moçambicano de Engenharia, 29 August
-1 September 2005, Maputo, Mozambique.
Matsinhe N.P.; Persson K.M. (2006): Opportunities to improve the drinking water quality of
Maputo-Mozambique. International Congress on Integrated Water Research Management and
Challenges of the sustainable development, 23-25 May 2006, Marrakech, Marroco.
Juizo D.; Matsinhe N.P. (2006): Water Services in peri-urban areas of Maputo with private
sector participation. 7th WARFSA/WATERNET/GWP-SA symposium, 1-3 November 2006,
Lilongwe, Malawi.
Matsinhe N.P.; Juizo J.; Berta M and Clara dos Santos (2007): Regulation of formal and
informal water service providers in peri-urban areas of Maputo, Mozambique. 8th
WARFSA/WATERNET/GWP-SA symposium, 31 October - 2 November 2007, Lusaka, Zambia.
Supervised Master Thesis
Malin Kajsa and Anna Linquidist (2004). The effect of intermittent supply and storage in the
quality of drinking water in Maputo. Department of Water Resources Engineering, Faculty of
Engineering, Lund University, TVRL 2005: 7.
Anna Gustafsson and Maria Johansson (2006). An Investigation of nutrient levels along the
Mbuluzi River- A background for sustainable water resources management. Department of
Water Resources Engineering, Faculty of Engineering, Lund University, TVRL 2006:9.
Arnoldsson Emelie, Bergman, M. (2007). Assessment of drinking water treatment using Moringa
Oleifera natural Coagulant. Department of Water Resources Engineering, Faculty of
Engineering, Lund University, TVRL 2006: 7
VI
Challenges and Opportunities for Safe Water Supply in Mozambique
ȱ
2008
1. INTRODUCTION
1.1
Background and identification of the Problem
The provision of adequate water supplies is essential in order to meet basic human needs and
to address poverty, and promote economic development, health and hygiene. Water supply has
a long history in this respect, and the rationale for its improvement has always been the need to
protect public health, to reduce mortality and morbidity in the population, and to promote
economic development, especially in the developing world.
During the last four to five decades, many cities in the developing world, particularly African
cities, have experienced rapid population growth but, water and sanitation infrastructure has
lagged behind. Despite considerable investments made in the development of water supply
infrastructure, the progresses that have been registered in many of those places have been
slow. Achieving the target set in the United Nations Millennium Declaration of halving the
proportion of the population without access to potable water supplies by 2015 remains therefore
a challenge (OECD, 2007). The Global Water Supply and Sanitation Assessment 2000 1
estimated, for instance, that in 2002, about 1.1 billion people worldwide lacked access to
adequate water supplies, and that those lacking services were chiefly in Asia and Africa. In
absolute terms, Asia has the highest number of underserved citizens, but proportionally this
group is larger in African countries (WHO/UNICEFF, 2006) where, an estimated 27% of the
population did not have access to potable water. Levels of service by world population, defined
as access to piped water was estimated to be 47% while, in Africa, this figure was only 24%.
The situation in Mozambique does not differ significantly from that reported in many other
African countries. Coverage estimates obtained from the Mozambique Demographic and Health
Survey (WHO/UNICEF, 2006); indicate that in 2003, 72% of the urban population,
corresponding to about 37% of the country’s total population, had access to adequate drinking
water supplies. This figure in rural areas was only 28%. The proportion of residents with access
to piped water through household connections was far less, corresponding to only 18% of the
urban population and to less than 2% in rural areas (WHO/UNICEF, 2006).
Despite massive investments in the development and extension of the infrastructure for drinking
water supply in Mozambique, service provision to a large number of urban settlements is still far
from optimal, mainly because the investments have focused on increasing water production and
water availability, while little attention paid to other, equally important aspects of service delivery
such as, the quality of the services and water quality. Added to this, lack of adequate strategies
for strengthening, institutional capacity and human resources, and for securing financial
arrangements needed to guarantee the long-term sustainability of the infrastructure, had
worsened the situation. The consequences of this situation are many, and include:
ƒ
1
Inadequate coverage of piped water supplies in most (if not all) urban centres of the
country. Poor quality of piped water supplies due to factors such as intermittency, flow
and pressure fluctuations, and frequent discontinuities of service provision;
Later updated by UNICEFF in September 2002
1
Challenges and Opportunities for Safe Water Supply in Mozambique
ȱ
2008
ƒ
difficulties in ensuring water of acceptable quality due to a lack of, or malfunctioning of
,treatment facilities, and sub-optimal operation of treatment processes at existing
treatment facilities due to factors such as technological and operational constraints and
lack of supplies;
ƒ
difficulties in maintaining acceptable treated water quality during transport, storage and
distribution; and
ƒ
Proliferation of alternative service providers in the informal water market in an
environment where licensing and regulatory tools for service providers are still lacking.
The aim of this study was to contribute to the overall understanding of the status of the drinking
water supply in Mozambique and the challenges facing the water supply governance sector in
developing and maintaining sustainable drinking water supplies. The study focuses on the
quality aspects of urban water supplies with the emphasis on the identification of critical factors
affecting water treatment, the sustainable production of drinking water, and the management of
drinking water quality by formal and informal service providers. The starting point for this work
was the recognition that investments in the development of urban water supplies have generally
focused in increasing water availability, rather than other dimensions of water supply services,
such as water quality, service continuity, equity and reliability, which are now recognized as
important factors in the long-term planning of investments in this sector.
1.2
Objectives of the research
The main objective of this research was to analyse the drinking water supply system in
Mozambique, based not only on the quantitative aspects but also on qualitative aspects of
drinking water supply. Some of the critical factors affecting drinking water production and the
problems and challenges facing the sector in supplying potable water in sufficient quantities in
the long term are discussed, and methods of improving the situation evaluated. Given the wide
range of aspects to be covered, the following research questions were defined to help address
the objectives of the study.
1. Which key factors currently affect drinking water production and the management of
drinking water quality in the context of urban water supplies in Mozambique?
2. Given the present status of urban water supplies, what alternatives are there for
subserviced consumers to overcome the inconveniences caused by a lack of, or
inadequate, piped water supplies? In addition, which factors affect the current and longterm sustainability of service provision through alternative service providers?
3. Given the prevailing situation regarding drinking water production and the management
of drinking water quality, which options exist to improve the efficiency of drinking water
production in order to secure safe and sustainable drinking water supplies in the long
term?
2
Challenges and Opportunities for Safe Water Supply in Mozambique
1.3
ȱ
2008
Overall methodology
In order to address the research questions outlined above, the study was divided into a number
of steps as follows.
1. The identification of factors affecting the overall production of drinking water in the
country. This included the investigation of source and treated water quality, the
evaluation of treatment methods and performance efficiencies in a number of
waterworks, as well as the evaluation of procedures used to assess raw and treated
water quality and to control processes during water treatment. This step helped address
the first research question.
2. The assessment of the quality of water supplied by formal and informal service
providers, and the identification of factors contributing to water quality deterioration
during distribution. The quality of drinking water provided by the formal network in
Maputo and alternative service providers operating in the various neighbourhoods of
Maputo was used in this investigation to help address the first and second research
questions.
3. The assessment of the main characteristics of water supply services in peri-urban
Maputo, the quality and organization of the management of services and the challenges
facing alternative service providers in providing water of sufficient quality and quantity in
the long term. In addition to this, the possibility of expanding the reach and influence of
the existing regulatory framework to cover alternative service providers was also
evaluated. The results of these assessments were used to address the second research
question.
4. The evaluation of methods to improve drinking water treatment and the sustainability of
drinking water production through low-cost treatment technologies. The results of pilotplant experiments and laboratory testing were used to address the issue of the
sustainable development of drinking water treatment, i.e., the third research question.
1.4
Thesis structure and appended papers
This thesis consists of a summary and six appended papers. In the summary the overall status
of the drinking water supply in Mozambique is discussed based on a review of past studies,
reports and related papers. This review, which focused on the organization of the drinking water
sector and on the methods used for the production and distribution of drinking water in
Mozambique, provided the background necessary to address the research questions explored
in this study. The major findings obtained during individual studies conducted as part of this
research are also presented in the summary.
This thesis describes research mainly concerning the water supply services in the city of
Maputo. However, the results are used to infer the overall situation of drinking water supply in
the country and to discuss the main challenges facing the drinking water governance sector in
developing and maintaining safe and sustainable water supply services in the long term. The
individual studies focused on quantity and quality aspects of the formal and informal water
supply services, drinking water production methods and factors affecting water treatment and
possible methods for improvement. Research results have also been presented at international
conferences as well as in the appended papers.
3
Challenges and Opportunities for Safe Water Supply in Mozambique
ȱ
2008
In Paper 1 a critical review is presented of the main factors affecting current drinking water
production in Mozambique. Given that most drinking water supplies use surface water for
drinking water production, the water quality of six rivers used for drinking water production was
analysed and the critical factors affecting treatment possibilities identified. The performance
efficiency of three major waterworks in the country was also analysed.
In the study presented in Paper 2, the quality of drinking water in the supply network of Maputo
was analysed, and the factors causing water quality deterioration during distribution
investigated. Field work was conducted in an area of central Maputo to investigate the overall
quality of drinking water in the Maputo water supply network.
Paper 3 and Paper 4 describe the analysis of the quantity and quality of water supply services
provided by formal and informal service providers. Paper 3 addresses the main challenges
facing alternative service providers, notably small-scale independent providers (SSIPs) in
supplying water of sufficient quality in sufficient quantities in the long term. Potential human
health risks associated with the consumption of water provided by SSIPs were also
investigated. The study described in Paper 4 is based on the findings presented in Paper 3
concerning the activities of alternative service providers in peri-urban Maputo. Paper 4
discusses aspects of service quality in the informal water market and the possibility of
expanding the reach and influence of the regulatory framework for formal water services to also
cover SSIPs.
Paper 5 presents the results of experimental work conducted to investigate methods of reducing
turbidity and improving suspended solids removal with the use of up-flow roughing filters for
hydraulic flocculation and conventional rapid sand filtration for final treatment. In Paper 6 a
comparative study of coagulation efficiency using extracts of the natural coagulant, Moringa
oleifera (M. oleifera) and aluminium sulphate in the treatment of surface water is presented. In
this study, standard jar-test experiments were used to investigate the potential of using seeds of
M. oleifera instead of aluminium sulphate, and to assess the optimal dose and conditions for
their use. The use of M. oleifera together with coarse to fine sand filtration, and the possibility of
using M. oleifera at small-scale treatment plants were also investigated.
4
Challenges and Opportunities for Safe Water Supply in Mozambique
ȱ
2008
2. OVERVIEW OF THE DRINKING WATER SUPPLY SECTOR OF MOZAMBIQUE
2.1
Background on the study area
Mozambique is located in the south-eastern part of Africa bordering the Mozambique Channel,
between South Africa and Tanzania. It lies between coordinates 18q 15cS and 35q 00cE (Figure
1) within a typically tropical/subtropical region, where the climate is under the influence of the
equatorial low-pressure zone with a NE monsoon in the warm season. The climate according to
the Köppen classification system is tropical rain savannah, with mean average temperatures
between 25 qC and 26 qC in the low-lying coastal areas and lower temperatures in the higher
areas.
Figure 1 Map of Mozambique showing geographic location of the country and the location of major cities.
Source: Google maps.
Because of the influence of the NE monsoon, rainfall is mainly restricted to the warm season
between October/November and April. The rainfall varies considerably within the country,
ranging from over 2400 mm/year in the northern parts of the country to only 300 mm/year in the
southern parts, particularly in the Chicualacuala area near the border with Zimbabwe in the
Limpopo River basin (Ferro & Bouman, 1987). In terms of water balance, the region is generally
classified as having a negative water balance, according to the Budyko aridity coefficient 2 (DNA,
1999).
The case study area described in this thesis concerns the Mozambican capital city of Maputo
and the adjacent city of Matola, both of which are served by the same water supply system. The
capital city is characterized by three kinds of districts, namely an area with high-rise buildings,
the so-called ‘cement city’, a few inner suburbs built before independence in 1975, and the outer
neighbourhoods of the peri-urban areas consisting mainly of informal settlements built during
the unstable periods that followed independence.
2
Synopsis of Water Resources of Mozambique (1999)
5
Challenges and Opportunities for Safe Water Supply in Mozambique
ȱ
2008
In similarity to many other parts of the developing world, the peri-urban areas of the city of
Maputo have developed without proper urban planning, due mainly to rapid population growth
and lack of financial resources for investments in infrastructure and, as a result, many of these
areas today face severe difficulties regarding access to public utilities such as water, sanitation
and electricity. While the residents of the neighbourhoods located near the cement city have
access to piped water through overstretching the formal network, their water supply is
unreliable due either to a lack of pressure in the nearby grid, or because they are located
beyond the reach of the formal network. In many of these areas, the lack of an adequate supply
of piped water has prompted the emergence of a multiplicity of alternative providers, among
which small-scale independent providers and household water resellers, presently constituting
the most reliable source of water for a large proportion of un-connected residents. Taking the
city of Maputo as an example, SSIPs presently account for the water supply to nearly 32 % of
households in the city, compared with roughly 62 % of households that relies on the formal
network (Gumbo, 2004; Seureca & Hydroconseil, 2005). The remaining 6 % rely on other kinds
of supply.
The situation in the city of Maputo is very similar to that in other cities in the developing world,
where alternative service providers are reported to play an important dominant role in serving
unconnected residents. It has been estimated, for example, that SSIPs supply as much as half
the urban population in some countries of Asia, and nearly a quarter of the urban population of
Latin America (Kariuki & Schwartz, 2005; OECD, 2007). In African countries, nearly half of
urban dwellers are believed to rely on alternative service providers for at least a portion of their
drinking water (Collingnon & Vézina, 2000). The list of examples in Africa includes the cities of
Bamako, Cotonou, Conakry and Dar es Salaam, where alternative service providers are the
main source of drinkable water for more than 60 % of households, and cities such as Abidjan,
Nairobi and Ouagadougou, where they are reported to serve 22 % to 28 % of unconnected
households (Reweta & Sampath, 2000; Collingnon & Vézina, 2000;).
2.2
Overview of the legal and institutional framework
In Mozambique, water supply services are the mandate of the National Directorate of WaterDNA, under the Ministry of Public Works and Housing. DNA is the primary agency responsible
for water resources policy making, planning and management, and for ensuring the provision of
water supply and sanitation services throughout the country.
In 1995 the first National Water Policy (NWP) was approved, which has since then guided water
sector reforms. In line with the objectives of the NWP, the government was relieved of the task
of the actual implementation of water supply services to focus on policy making and planning of
the management of water supply services (DNA, 1995). Special attention is devoted to
improvements in water supply services, the encouragement and regulation of the involvement of
private service providers, and the participation of beneficiaries in the management of water
supply services such as that provided through public standpipes.
As part of the reforms in the drinking water supply sector, the Government of Mozambique
established the Framework for Delegated Management of Water Supply in 1998 (Decree
Numbers 72, 73 and 74:9 of December 1998), which created the legal basis for the delegation
of operation and management of public water supply services to independent private entities
through concessions, leasing or management contracts (Zandamela, 2002; Gumbo et al., 2003;
Sal-consultores, 2005). The main institutions involved in the Framework are: the Ministries of
6
Challenges and Opportunities for Safe Water Supply in Mozambique
ȱ
2008
Public Works and Housing, of Planning and Finance and of State Administration, the National
Directorate of Water, the Water Supply Investment and Assets Fund (FIPAG), the Council for
the Regulation of Water Supply (CRA), the Co-ordinating Forum for Delegated Management,
the municipal authorities and private operators (Figure 2).
Figure 2 Overview of the Framework for Delegated Management of Water Supply in Mozambique.
Source: Sal-Consultores 2005).
The FIPAG was created to take over the fixed assets and the duties and obligations for water
service delivery in five major cities of Mozambique, previously serviced by state water
companies. The CRA was created with the objective of regulating private sector contracts under
the rubric of the framework. As such, CRA is an independent regulating body designed to
ensure a balance between service quality, the degree to which it meets consumer interests, and
the economic sustainability of the water system. It mediates the interests of the main private
service providers and the lessor (FIPAG) aiming to reconcile them through the mechanism of
tariff setting. The geographical limit of CRA’s influence is therefore defined by the boundary of
the area of each contracted concession.
Private sector participation in water supply in Mozambique started in 1999 when the first private
operator, Águas de Moçambique-AdeM, was awarded a 15 years contract to provide services to
the cities of Maputo and Matola, and management contracts for the cities of Beira, Quelimane,
Nampula and Pemba (Zandamela, 2002; Gumbo et al., 2003). In 2004, the government decided
to expand the framework for delegated management to include four southern cities of
Mozambique. The water supply systems of the cities of Xai-Xai, Chokwé, Inhambane and
Maxixe were integrated and delegated to the Dutch company Vitens (OECD, 2007). CRA
increased its ambit to cover also these cities. In 2006, a further expansion of the delegated
7
Challenges and Opportunities for Safe Water Supply in Mozambique
ȱ
2008
management took place in the central region of Mozambique and five more towns were
included.
In areas outside the cities with delegated management of water supply, as well as in rural areas,
DNA still retains full control of water systems (OECD, 2007) but the operation and management
of the water systems is secured by public or state-owned entities. According to the DNA manual
for the management of piped water systems in cities without delegated management, service
regulation is to be the task of district or municipal regulatory commissions, which have yet to be
established. These institutions will be responsible for tariff and service monitoring, implementing
regulations, evaluation and analysis of information and conciliation of consumer interests with
those of the operator and the DNA (OECD, 2007; Seureca & Hydroconseil 2005)
2.3
Drinking water supply options and choice of technology
The drinking water supply of Mozambique is functionally divided in two sub-sectors; the urban
sub-sector supplying piped water in large urban centres, and the rural water supply sub-sector,
which provides rural water supplies mainly through boreholes with hand pumps and some piped
water to small towns and villages.
Rural water supplies are generally groundwater based. According to statistics (WHO/UNICEFF,
2006) nearly 63% of the country's total population have rural water supplies, only 2 % of which
have access to piped water supplies. Most rural water supply is by point water sources, namely
boreholes and hand pumps. Coverage estimates in these areas are generally assessed
according to accessibility to point water sources, a daily water demand of 20 litres per person
per day, and a maximum distance to the nearest water point of 500 m. Piped water supplies in
small towns and villages are mainly by public standpipes and yard connections.
Urban water supplies are generally surface-water based with service provision generally
secured by piped water supplies. Although the coverage for the urban population in
Mozambique is estimated to be 72 %, urban water supplies in Mozambique are available to only
37 % of the total population of the country, of which only 18 % have access to water through
house connections (UNICEFF, 2006). Surface water from rivers is mainly used for drinking
water production. In Table 1, an overview is given of the current situation regarding raw water
sources and treatment methods for the piped water supply in 10 major cities of Mozambique.
Piped water is generally supplied at three levels of service: house connections, yard
connections or public standpipes. The distribution of consumers by level of service varies from
town to town. Taking Maputo as an example, a survey carried out in 2007 indicated that the
proportion of residents with access to piped water through house connections was 42%. The
figure for yard taps was 35% and for public standpipes 23% (Seureca & Hydroconseil, 2005).
8
Challenges and Opportunities for Safe Water Supply in Mozambique
ȱ
2008
Table 1 Typical design of drinking water treatment plants of 10 major towns of Mozambique. (SW =
surface water; GW = groundwater).
Town
Estimated
water
demand by
2008
(m3/day)1
Type of raw water
source
Type of
treatment
Conventional with
pressure filters
Conventional with
cascade aeration
+ direct filtration
Conventionalstandard3
Conventional with
pressure filters
Slow sand
filtration
Conventional with
cascade aeration
+ direct filtration
Conventionalstandard with
lamella type
settling basins
Disinfection with
granular chlorine
Conventional
standard +
pressure filters
Conventional
standard
Lichinga
2 947
SW+ large reservoir
with selective intake
Pemba
9 487
GW–borehole field
Nampula
14 762
Nacala
6 568
Quelimane
13 320
GW–borehole field
Tete
7 116
GW–borehole field
Beira
30 000
SW–direct intake
from a sugar cane
irrigation channel
Xai-Xai
8 600
GW–dune aquifer
Mocuba
2 279
SW–direct intake
from river
240 000
SW–direct intake
from river
Maputo-Matola
SW+ large reservoir
with constant intake
SW+ large reservoir
with constant intake
Operating
capacity
(m3/day)2
2 160
10 000
13 500
5 700
4 200
5 100
19 200
6 000
1 000
172 800
1
According to estimates from DNA.
2
design capacity according to plant operators.
3
conventional standard = pre-chlorination, chemical coagulation, flocculation, sedimentation, rapid sand filtration and disinfection.
2.4
Drinking water treatment methods and quality criteria
Regarding drinking water production, conventional treatment, consisting of chemical
coagulation, flocculation, sedimentation, rapid sand filtration and disinfection, is most commonly
used. In these methods coagulation is used to reduce the repulsive forces responsible for the
stability of colloidal dispersions while flocculation is used to enhance particle transport and
aggregation, and the eventual formation of suspensions suitable for separation with
sedimentation and filtration. Sedimentation does the bulk of the liquid-solid separation while
filtration, usually deep bed filtration, is used as a polishing step. Particle destabilization and
aggregation therefore largely determine therefore, the efficiency of water purification (Hammer
et al., 2004; Polasek, 2007; Lawrence et al., 2007). Disinfection, usually with the help of
9
Challenges and Opportunities for Safe Water Supply in Mozambique
ȱ
2008
chlorinated compounds is used to kill bacteria and to provide acceptable levels of residual
chlorine to prevent post-contamination (WHO, 2004). Additional processes in conventional
treatment may include pre-chlorination to assist the removal of algae and dissolved organic
matter and conditioning for alkalinisation and corrosion control.
Conventional treatment is generally regarded as the most effective method of producing
drinkable water from surface water sources. However, the efficiency of the method depends on
various factors, among which are the quality of the raw water and its variation. For example,
when flocculation is used for the treatment of water with low turbidity, the process is said to be
kinetically poor if it cannot be accomplished with operational modifications that may include the
use of coagulant aids or coagulation-direct filtration processes (Chuang et al., 1997; Polasek &
Mult, 2005). High values of turbidity affect the efficiency of coagulation processes, and can
cause blockage of filters leading to the demand for frequent cleaning of filters and clarifiers,
which increases the cost of water production.
Conventional treatment is also unsuitable for the removal of high concentrations of organic
matter, algae and impurities resulting from anthropogenic pollution. High concentrations of
natural organic matter-NOM, for instance, influence alum speciation in coagulation processes
(Volk et al., 2000; Viraraghavan & Srinivasa, 2004; Sharp et al., 2005), and NOM is a precursor
of disinfection by-products-DBPs (Ruehl, 1999). The presence of algae can cause several
problems to drinking water production depending on the algal species and their concentration
(Bauldin et al., 2006). These problems can impact the plants operation and the quality of treated
water. Algae impacts for instances, the performance of filters (Polasek & Mult, 2005), reduces
the efficiency of flocculation processes, and can alter the nature of organics, thus influencing
processes designed to remove NOM (Cheng et al., 2003; Leikens, 2004; Qin et al., 2006).
Algae can also be the source of unpleasant taste and odour (WHO, 2004; Bauldin et al., 2006).
Other methods of treating surface water include the use of the so-called non-conventional
methods whereby slow sand filtration is used as the primary treatment process and roughing
filtration (mostly horizontal roughing filters) used for pre-treatment (Huisman, 1984; Smet &
Visscher, 1990; Sanchéz et al., 2006). Most advanced methods include for example, treatment
with membranes and desalinization plants (Hammer et al., 2004; Lawrence et al., 2007). For
groundwater supplies faced with problems of excess iron and manganese, the most commonly
used method of treatment is aeration followed by direct filtration. Water disinfection is also used
mainly to maintain levels of residual chlorine suitable for prevention of post- contamination.
The objectives of drinking water treatment are generally translated into drinking water quality
standards, which should be met at treatment plants. These standards are generally specific to
each country or region, but most countries have based their standards on the WHO Guidelines
for drinking water (WHO, 2004). The Mozambican standards (MISAU, 2004) were also
developed based on the WHO Guidelines.
2.5
Alternative water supplies
Service provision through alternative water supplies is generally by groundwater-based, small
piped water supplies usually developed by small-scale independent service providers.
Consumer access is generally by public standpipes, but many independent providers report to
offer services also through house connections and yard taps. The evolution of household level
water strategies into independent service providers is, in most cases, unplanned. Most owners
of such systems constructed them to provide water for themselves, but at the insistence of
10
Challenges and Opportunities for Safe Water Supply in Mozambique
ȱ
2008
neighbours allow private connections or sell water through yard connections, thus slowly
developing into small-scale water vendors. Also, because the profit from selling water helps
them offset their investment and running costs (break-even in about 2-3 years), many of them
eventually turn into professional service providers.
Per cent of new systems by year of construction
as compared to actual numbers
Taking the city of Maputo as an example, a survey carried out in 2005 (Seureca & Hydroconseil
2005; Boyer, 2006) indicated that around 240 SSIPs existed in Maputo, which were reported to
provide services to as many as 32% of unconnected households in peri-urban Maputo. Around
65% of these SSIPs were reported to have been established since 2001 (Figure 3). During the
same period, service levels were reported to have risen from public taps only to house
connections, yard taps and private standpipes (Sal-Consultores, 2005). Today, some 32% of
SSIPs operating in peri-urban Maputo are reported to have more than 100 house connections.
16
14
12
10
8
6
4
2
0
1975
1980
1985
1990
1995
2000
2005
Year of construction
Figure 3 Evolution of small-scale independent Providers (SSIPs) in Maputo-Mozambique
Most alternative providers, although located within the official boundaries of the main cities and,
in some cases within contracted concession areas, they are currently not formally regulated.
Reasons for that are the lack of legislation and an administrative framework that could be used
to grant licences or regulate their activities.
11
Challenges and Opportunities for Safe Water Supply in Mozambique
ȱ
2008
3. METHODOS
3.1
Field work
This thesis is based mostly on field work and laboratory testing. Field work was carried out to
assess the quality of drinking water from formal and informal service providers, and to assess
the potential of the regional aquifer used by independent service providers operating in the
region of Maputo. Field work was also used to evaluate the quality of water supplied by
independent service providers and to evaluate the efficiency of drinking water treatment in a
number of waterworks used for case studies. Detailed descriptions of the methodology and the
results obtained through field work are presented in Papers 1 to 4. Laboratory work was used
mainly to evaluate the efficiency of drinking water treatment using alternative and low-cost
treatment methods. The low-cost treatment methods tested were hydraulic flocculation with upflow roughing filtration, and the use of extract from M. oleifera seeds for coagulation of surface
water. The methodology and results are presented in detail in Papers 5 and 6.
3.2
Sources of data and choice of water quality variables
The data used in this work were obtained through field work, and from historical sources
available at the DNA, regarding mostly river water flows and water quality. The results of studies
conducted in some of the river courses investigated were used to assess river water quality.
Historical data from plant operators and data collected through field work were used to assess
the quality of drinking water provided by formal and informal service providers in the city of
Maputo, and also to assess the performance of the waterworks at Maputo and Beira and
Nampula.
The parameters used to define overall river water quality were: turbidity, total hardness, pH and
total alkalinity, organic matter content, indicator organisms and phosphates (related to algae
growth). Rivers chosen for the study were: the Limpopo, Umbelúzi and Maputo rivers in south
Mozambique, the Púngue River in central Mozambique and the Monapo and Licungo rivers in
north Mozambique. Turbidity, pH, alkalinity and organic matter were used to describe the
efficiency of drinking water treatment.
Before use, the data on river water quality obtained from the DNA databases were checked for
their analytical accuracy by performing charge balance tests on the data on ionic species.
These tests were used to eliminate erroneous or suspicious observations from the original data
sets. A charge balance error of ±15 % was accepted due to the limited size of available data
sets. Around 53 % of the data sets analysed regarding the Licungo River and 24.5 % from the
Umbelúzi were rejected due to unreliable values of various water quality parameters and charge
balance errors greater than 25 %.
The data used to assess the quality of drinking water in the Maputo network were: residual
chlorine, bacteria, turbidity and solids. Temperature, residence time and the condition and
cleaning of household tanks were used to make the final assessment of drinking water quality
following storage at household level. These data were also tested for statistical significance. In
this way, all pair-wise data collected before and after household storage were checked for
12
Challenges and Opportunities for Safe Water Supply in Mozambique
ȱ
2008
statistical significance using a method that consisted of determining a confidence interval for the
difference between two expected values I(ȝ1- ȝ2); if this interval does not cover zero the
difference is regarded as being significant between two homogeneous groups. A 95 %
confidence interval (p=0.05) was used in all tests. Chemical and physical parameters measured
in samples taken from the network and after household tanks were compared, with an
independent sample T-test. The T-test was also used to compare residual chlorine in samples
with and without bacteria. The T-test is not valid for turbidity, free residual chlorine, total residual
chlorine or nitrate, since the statistical variances of the two groups were not comparable, and
the Mann-Whitney U-test was used instead. The presence and absence of bacteria in samples
taken before and after storage were also compared using the Mann-Whitney U-test.
The presence of faecal bacteria and nitrates, and the electrical conductivity and salinity were the
parameters used to assess the quality of water from alternative service providers. The electrical
conductivity (EC) was measured with the purpose of evaluating the influence of sea water
intrusion on the quality of borehole water. Results of borehole pumping tests were used to
complement this analysis. Emphasis was placed on the potential for groundwater contamination
due to over-exploitation of the aquifer system.
3.3
Situation analysis with respect to drinking water treatment and efficiency of
drinking water production
The situation analysis with respect to drinking water treatment and the efficiency of drinking
water production was based on the evaluation of the results regarding treated water quality from
a number of treatment works and their compliance with internationally accepted drinking water
quality standards. The analysis also included the assessment of operational procedures within
these waterworks, with emphasis on the operation of coagulation-flocculation processes and
procedures for the control of raw and treated water quality during treatment. The waterworks
studied were based on the standard design of conventional treatment, also being the bestequipped waterworks in the country, and also those where records were kept of water quality, or
the quality could be assessed.
3.4
Situation analysis with respect to drinking water quality
Drinking water quality was assessed through field studies conducted in an area of central
Maputo and on the evaluation of the quality of the source water used by alternative service
providers in peri-urban Maputo. Analysis of drinking water quality from the piped network was
motivated by the fact that water distribution in Maputo, and in many other towns of Mozambique,
is generally intermittent, a modus operandi that is often associated with quality problems, due to
the ingress of contamination during periods of low or no pressure in the grid (Tokajian & Haswa,
2003; Totsuka et al., 2004) and reduced disinfection capacity due to long residence times in
pipes and reservoirs (Kiéné et al., 1998; Hua et al., 1999; Powell et al., 2000; Vreeburg &
Boxall, 2007). Also, consumers faced with intermittent supplies often develop their own ways to
cope with this, for example, by constructing household tanks which add to the factors
responsible for drinking water quality deterioration.
In order to carry out this evaluation, water samples collected before and after household tanks
were analysed with regard to their physic-chemical and bacteriological characteristics. The
condition of household tanks was also assessed through observations and interviewing the
13
Challenges and Opportunities for Safe Water Supply in Mozambique
ȱ
2008
owners. Aspects of interest included the overall condition of the tanks, construction materials,
tank dimensions and consumer practices regarding cleaning and disinfection. The results of
time-dependent chlorine decay tests using samples of treated water were used to estimate the
magnitude of chlorine depletion during storage, and to estimate chlorine decay rates for treated
water. The method used, which is often referred to as the “bottle” or jar-test (Powell et al.,
2000), consists of recording the chlorine concentrations, at fixed time intervals, of the water
collected in jars. The results of these tests were compared with the results of calculations
performed with data provided by the service provider and data generated during field work.
Estimates of chlorine decay rates were based on a first-order decay reaction.
The results of the evaluation of the quality of source water used by alternative service providers
were used to identify the present and long-term challenges facing alternative service providers
in supplying water of sufficient quality and quantity in the long term, and to assess possible
human health risks associated with the consumption of water from these providers. Borehole
pumping tests, the results of which were interpreted using the graphical method of Jacob
(Kruesman & Rider, 1991) were used to evaluate not only the potential of the regional aquifer
but also the long-term vulnerability of the aquifer system to external contamination. Samples
from thirty-five wells were used to assess borehole water quality. A total of ten pumping tests
were performed to assess the hydrological potential of the aquifer system.
3.5
Laboratory testing
Laboratory testing was performed with the purpose of investigating methods of improving the
turbidity reduction achieved with conventional treatment. The methods tested were up-flow
roughing filtration for hydraulic flocculation and the use of M. oleifera seed extract for the
coagulation of surface water. Pilot-plant experiments were carried out to determine the optimal
dosage and filtration conditions for hydraulic flocculation with up-flow roughing filters and the
suitability of the suspensions formed for removal by subsequent rapid sand filtration. River water
was used for the experiments. Turbidity reduction, head loss development and velocity
gradients in the roughing filter were the parameters used to evaluate the performance of the
pilot plant.
Coagulation experiments were performed using standard jar test experiments with solutions
prepared from aluminium sulphate and M. oleifera seed extracts, the results of which were used
to compare coagulation efficiencies and the effects on turbidity reduction and the chemistry of
the treated water. The optimal dosage of coagulant was investigated for different levels of
turbidity, and the properties of treated water were monitored. The use of M. oleifera seed extract
together with coarse to fine sand filtration, and the possibility of using M. oleifera at small-scale
waterworks was also investigated.
14
Challenges and Opportunities for Safe Water Supply in Mozambique
ȱ
2008
4. RESULTS AND DISCUSSION
4.1
Source water quality and methods of treatment
4.1.1 Surface water quality
Papers 1 and 3 present a discussion of source water quality and methods of conventional
drinking water treatment. Turbidity, suspended solids, natural organic matter, bacteria, the
presence of algae and other aquatic plants were identified as raw water quality variables of
major concern for drinking water production (Paper 1). Because most rivers used for drinking
water production are of torrential regime, these impurities not only occur at high concentrations
but there are also wide seasonal variations.
The removal of these impurities, particularly turbidity, is essential during water treatment for the
production of safe drinking water. Turbidity itself is not a major health concern, but affects water
acceptability and is generally related to other quality variables (e.g. colour, total suspended
solids, algae and bacteria), some of which are of major concern for human health (WHO, 2004).
The overall condition of water with respect to turbidity therefore influences the levels at which
some of these variables occur in natural waters and also the treatment processes needed to
produce drinking water. When these levels are constantly low, simpler and less expensive
treatment processes can be used, but when these levels are constantly high, more complex and
expensive treatment processes are required. The worst scenario, however, is when the levels
vary greatly, as the treatment process must be sufficiently flexible to treat the varying quality of
the raw water.
Figure 4 shows a typical example of river water turbidity for a number of rivers used for drinking
water production in Mozambique, where it can be seen that river water turbidity is generally high
at all locations, and also that it varies considerably during the year.
Monthly average flows (m3/s) in some of the river courses analysed
Maximum recorded turbidity values at selected sites of some of the rivers
analysed
1400
400
Umbelúzi
Limpopo
Púngué
Licungo
350
1000
Oct.
Nov.
Dec.
Jan.
Feb.
Mar.
Apr.
May.
June
July
Aug.
Sep.
300
Turbidity (NTU)
Monthly average flow (m3/s)
1200
800
600
400
250
200
150
100
200
50
0
0
Oct.
Nov.
Dec.
Jan.
Feb.
Mar.
Apr.
May
June
July
Aug.
Sep.
Umbeluzi site 1 Umbeluzi-site 2
Licungo
Limpopo-site 1 Limpopo site 2
Púngue
Figure 4 Average monthly flows (m3/s) and maximum values of water turbidity (NTU) recorded at
selected sites of rivers used for drinking water production of some cities in Mozambique such as
the city of Maputo (Umbelúzi River), Beira (Púngué River) and Mocuba (Licungo River). Data
source: DNA database.
15
Challenges and Opportunities for Safe Water Supply in Mozambique
ȱ
2008
The relatively low values of maximum turbidity observed at some of the locations investigated
(Umbelúzi and Limpopo-site 1) are due to the effect of major hydraulic works (man-made
reservoirs) located upstream of the sampling locations, where solids and turbidity are reduced
by sedimentation. From the data shown in Figure 4 it can also be seen that river water turbidity
generally shows a good correlation with river water flow, and thus with the hydrology and rainfall
pattern of the region.
Physical parameters such as water pH and temperature do not affect water acceptability, and
are not of concern regarding health, but they have a major influence on water quality
parameters such as bacteria and algae growth. Average river water temperatures in
Mozambique are generally between 20 qC and 25 qC and the raw water pH generally between
6.0 and 9.0. Because high temperatures are generally accompanied by heavy rainfall and high
river water turbidity, the combined effect of these factors provides ideal conditions for bacterial
growth and algae blooming. The common practice of building reservoirs to secure water
availability during periods of reduced flow usually worsens the situation because the hot climate
characterizing tropical and subtropical regions provides the perfect ecological environment for
bacteria and algae blooming (Anderson et al., 1999; Chorus & Bartram, 1999). Added to this
natural climatic effect is the enhanced rate of nutrient input that accompanies the growth of
towns and the development of irrigated agriculture in the catchment areas around the water
courses.
It was concluded that current knowledge concerning river water conditions with respect to the
presence of algae and/or eutrophication problems is poor due to a lack of data and detailed
studies. Besides some measurements taken at irregular intervals by DNA, which relates to
levels phosphorus in river water, the only references that could be found for assessing the
potential for river water eutrophication where, a study on presence and removal of
cyanobacterial toxins in the water from a reservoir supplied by the Umbelúzi river (Bojcevska &
Jergil, 2003), and a further study by Gustafsson & Johansson (2006) on levels of nutrients along
the Umbelúzi river. Other study relate to the Limpopo River where, the river water quality was
also assessed for presence of phosphorous.
Different references found in literature and proposed by different organizations and authors
suggest limits for total phosphorous concentrations in the range 0.1-0.16 mgP/L to avoid
euthrophication in river water which are presented in Table 2.
Table 2 Different limits for phosphorous and nitrogen to avoid euthrophication in rivers. Limits suggested
by different organizations and authors (* all fresh waters; ** Total P = 0.3262 PO43- (mg/L).
Variable
Phosphorous
Limit
Tot. P (rivers draining to lakes- U.S.EPA)
Tot. P (rivers not draining to lakes- U.S.EPA)
Phosphate (Fytianos et al., 2002)
Inorganic P (SAWQG)*
0.05 mgP/L
0.1 mgP/L
0.5 mgPO43-/L = 0.16 mgP/L
0.005 mgP/L
Nitrogen
Inorganic N (SAWQG)**
0.5 mgN/L
16
Challenges and Opportunities for Safe Water Supply in Mozambique
ȱ
2008
From the analysis of the data from DNA database it was found that total phosphorous
concentrations in the Umbelúzi River could vary from as low as 0.01 mg/L to figures as high as
0.63 mgP/L clearly values higher than the indicative limit of 0.16 mgP/L. This suggests that
conditions favourable to euthrophication of Umbelúzi river water may occur from time to time.
The results of the study by Bojcevska & Jergil (2006) gave evidences of presence of high
concentrations of toxic cyanobacteria in the water of the studied reservoir. However, the study
from Gustafsson & Johansson (2006) resulted in much lower values, with average total
phosphorous concentration of all measurements of about 0.054 mgP/L. The sampling period
used in this study covered, however, a much shorter period of analysis (two months from
September to November) and this may explain the difference in results. Yet, the average
concentration of phosphorous at all sites investigated was close or exceeded the guideline
value for rivers draining into lakes as is the case of the Umbelúzi, which supports the findings
from the previous study about the possibility of euthrophic conditions in the river water and, the
possibility of presence of algae and cyanobacteria. Problems with algae at the Umbelúzi River
were also reported in a latter study by Couto (2004) and more recently by managers of the
waterworks of Maputo water supply who report frequent outbreaks of algae at the intake works.
The study on the Limpopo River provided similar results. In this, total phosphorous in the river
water were found to vary from 0.03 mgP/L to values as high as 2.0 mgP/L. Two sampling
locations were considered for this analysis. Given the origin of the problem, this is a situation
that is likely to be encountered in many rivers used for drinking water production, particularly
those where intensively irrigated agriculture is practiced upstream of the major intake works.
Chemical properties such as water hardness and alkalinity generally affect water acceptability
and water treatment efficiency. Problems with hard or soft water are generally site specific
because they depend on the interaction of many factors including the soils and rocks from which
the water is derived, which are generally site or region specific (Stumn & Morgan, 1996; WHO,
2004). The water alkalinity is closely related to the concentration of carbonates, bicarbonates
and hydroxide ions in the water, therefore it is closely related to water hardness. It influences,
for instance, scale formation in the case of hard water and the degree of water corrosiveness of
soft water. During water treatment, the pH and water alkalinity affect processes such as
coagulation and disinfection. For example, when the alkalinity of the raw water is too high, an
excess of destabilizing reagent is generally required to adjust the water pH to optimum reaction
values. In contrast, when the alkalinity is too low, it must be increased in order to optimize
reaction conditions and improve the overall process efficiency (Polasek & Mult, 2005; Velasco et
al., 2007). Knowledge of the water’s pH and alkalinity is therefore useful in evaluating the
optimum conditions for treatment processes, as well as the final condition of treated water with
respect to its corrosive or scaling properties.
Analysis of river water hardness and alkalinity (Paper 1) confirmed the site-specificity of these
variables. As can be seen from the results presented in Paper 1, the river water hardness can
vary from rather soft to moderately hard at some locations, to slightly hard or even hard at other
locations. Related problems for drinking water production are therefore site specific, but in
general river water is classified as suitable for drinking water production.
The presence of high concentrations of organic matter is generally a problem in surface water
treatment. Organic matter affects treated water quality, especially taste and odour, and the
levels of nutrients (e.g. carbon) available for bacterial growth, as well as treatment processes
such as chemical coagulation and disinfection (WHO, 2004; Leiknes et al., 2004; Sharp et al.,
2005). Source of organic matter in fresh water are, the natural processes in fresh water, such as
17
Challenges and Opportunities for Safe Water Supply in Mozambique
ȱ
2008
the decay of organic material from animal and vegetable sources, and human activities, such as
sewage disposal and irrigation of agricultural land (WHO, 2004).
Analysis of the data on the presence of organic matter in river water (Paper 1) showed that
levels of organic matter at all sites investigated were generally high, and at concentrations that
exceed the guideline value for drinking water (2.5 mg C/L measured as DOC). However, the
observed levels of organic matter at the sites investigated are not indicative of organic pollution
of anthropogenic origin. This mean that when river water is used for drinking water production in
Mozambique, water treatment should also include processes to remove organic matter not only
because of the need to comply with guidelines for drinking water, but also to increase the
efficiency of treatment. Because the levels of organic matter are closely related to river water
turbidity and river flow, which are strongly influenced by seasonal variations, problems in water
treatment related to the presence of organic matter in river water will also be influenced by
seasonal variations in river flows.
4.1.2 Groundwater quality
According to the literature (WHO, 2004), most common groundwater quality problems are
associated with contamination by bacteria and nitrates, high levels of salinity caused mainly by
sea water intrusion, iron and manganese, and water hardness, generally associated with
calcium and magnesium dissolved in the water.
The major source of nitrates in groundwater is the natural decay of nitrogen-enriched organic
material and human activities such as sewage disposal, on site-sanitation and the accumulation
of organic material from improper solid waste handling (Boulding & Ginn, 2003; WHO, 2004;
Schmoll et al., 2006). Sewage disposal and on-site sanitation are also important sources of
groundwater contamination with bacteria. In both cases, factors such as hydraulic load, rainfall
patterns, soil type and depth to the water table determine the rate and extent of the transport of
contaminants to the groundwater. Sandy soils, for example, are particularly vulnerable to
bacteria and nitrate leaching into the groundwater because of the limited attenuation they
provide (Lee and Bastmeijer, 1991; Thompson et al., 2007; Dzwairo et al., 2006).
Most groundwater sources contain some amount of dissolved iron or manganese arising from
contact with minerals that contain such minerals (e.g. pyrite). When present in drinking water,
iron and manganese not only cause a bad taste but also staining of pipes and clothing. Calcium
and magnesium, which cause hardness, are found in groundwater that has come into contact
with certain rocks and minerals, especially limestone and gypsum. Drinking water quality
problems caused by hard or soft water have already been discussed in Section 4.1.2 of this
thesis, and include scale formation and the corrosion of pipes and other appurtenances used in
drinking water supply. Problems of high salinity in groundwater are generally the result of sea
water intrusion (coastal aquifers) or high total dissolved solids resulting from contact with certain
rocks and minerals from which the groundwater is derived. The WHO guideline for the EC of
drinking water is 1 500 µs/cm.
Groundwater sources in Mozambique are generally of good quality. However, there are some
well-known cases of groundwater of poor quality, for example, the area north of Inhambane
Province where the regional aquifer is known to have high total dissolved solids, which gives
the water brackish or hard characteristics (Ferro & Bouman, 1987). Also, some alluvial aquifers
are known to have problems arising from excess iron and manganese. The most well-known
18
Challenges and Opportunities for Safe Water Supply in Mozambique
ȱ
2008
example is the aquifer system used for drinking water supply to the city of Pemba in northern
Mozambique, where iron and manganese concentrations in excess of 4.0 mg/L and 0.4 mg/L,
respectively, have been repeatedly reported in several previous studies conducted to assess
drinking water quality in Pemba. Other examples of iron- and/or manganese-enriched aquifers
are the aquifer systems used for drinking water supply to the cities of Quelimane and Tete in
central Mozambique.
The analysis of groundwater quality described in this thesis focused on the quality of
groundwater sources used by alternative service providers in peri-urban areas of Maputo
(Paper 3). The aquifer system used in the case study is part of a large Meso-Cenozoic
sedimentary basin that covers the entire area south of the River Save, which is related to a rift
system extending between Madagascar and Africa. This system extends from Port Dundford in
South Africa to Quelimane in the central part of Mozambique. Karoo basalts and rhyolites,
dated to be Permic and Jurassic, form the basement of the system, while Cretaceous to Tertiary
flat deposits or deposits with nearly horizontal slopes overlay the Karoo sediments. These
deposits are mostly of marine origin and were formed during transgression periods. Sand dunes
or quaternary sand deposits cover the entire study area.
The aquifer is divided into two main units; a sandy aquifer or phreatic aquifer, and a deep
aquifer of sandstones and limestone (Burgeap, 1961). According to findings from studies by
IWACO (IWACO, 1983; IWACO, 1985) later confirmed by Juizo (1995) and SWECO (2004), the
separation between the two aquifers is not clearly defined, and for large-scale exploitation of
groundwater, the two aquifers can be regarded as a single unit.
As discussed in Paper 3, the groundwater sources used by alternative service providers in the
region of Maputo are generally of good quality and virtually free from microbial and organic
contamination. Thirty-five boreholes were used to assess groundwater quality in this aquifer
system. Two major factors were identified as contributing considerably to this situation, namely:
limited hydraulic loads of contaminants due to low population densities (< 100 inh./ha) and the
availability of a relatively thick unsaturated zone (> 30 m), where attenuation of contaminants
still occurs at sufficient levels. Signs of borehole water contamination with coliform bacteria
were, however, found in about 29 % of tested boreholes. About 9 % of tested boreholes
showed the presence of faecal bacteria. Low living standards, poor borehole construction and
somewhat high hydraulic loads were suggested as probable causes of the high incidence of
bacteria in such boreholes, which suggests that the potential for groundwater contamination
due to one or a combination of these factors is high.
Borehole water contamination by nitrates was also found to be low in the case study area.
Although they varied considerably, from values as low as 3 mg/L to about 35 mg/L, the nitrate
concentrations at all sites investigated were found to be below the WHO guideline value of
45 mg/L. Factors contributing to this situation were found to be similar to those determining the
levels of contamination by bacteria, i.e. low hydraulic loads due to low population densities and
the availability of a rather thick unsaturated zone where biological denitrification may occur.
Analysis of the spatial distribution of nitrate concentrations in more densely populated
neighbourhoods located within or adjacent to the case study area (Paper 3) indicated, however,
that nitrate levels in groundwater can reach values as high as 500 mg/L, suggesting that the
potential for groundwater contamination with nitrates may became a real threat to groundwater
exploitation in more densely populated areas.
19
Challenges and Opportunities for Safe Water Supply in Mozambique
ȱ
2008
Although the case study area is located near the coast where the potential for salinity problems
due to sea water intrusion is high, the levels of groundwater salinity (measured as EC) were
found to be low, and within acceptable limits for drinking water supply. Studies conducted in the
past (IWACO, 1985) in the same area indicate, however, that the potential for salinity problems
in this region is high, and more severe in some parts of the case study area (particularly areas
located south-west of the city) when compared to others. Cases of brackish water have also
been reported in some areas located north-east of the city and close to the Maputo bay, which
all fall outside the areas of high potential for brackish water according to results presented by
IWACO (1985). This implies that, while the results of the investigations carried out in this work
suggest that the majority of sites investigated do not experience problems of brackish water, the
situation may change very rapidly due to the high vulnerability of the aquifer system (coastal
aquifer) to sea water intrusion.
4.2
Drinking water treatment methods: current practices and treatment
efficiency
When used for drinking water, surface water must be treated. Water derived from naturally
protected sources (e.g. groundwater), on the other hand, can be used without extensive
treatment provided that it is disinfected at least for safety reasons. The intention of drinking
water treatment is to produce water that is safe for human consumption and suitable for
domestic use. The quality of water at source determines the type and extent of treatment
needed and the input required to perform the task economically and efficiently. Highly polluted
water, as is the case in most surface water sources, is usually unsuitable for human
consumption and requires extensive treatment prior to utilization.
Various aspects of drinking water treatment in Mozambique are discussed in Paper 1. As can
be seen, conventional treatment is the method used for most drinking water production. Most
treatment plants employ pre-chlorination, chemical coagulation, flocculation and sedimentation,
followed by rapid sand filtration and disinfection with chlorine. Modifications to the basic design
also exist, and these involve the use of package units incorporating all the physical processes
(flocculation, sedimentation and filtration) into one single unit, and the used of aeration and
direct filtration mostly for groundwater supplies (see Table 1).
The overall assessment of drinking water production in Mozambique indicates that the situation
today is generally inadequate. The worst situations are found in small and medium-sized water
supplies, where the combined effect of poor water quality at source and operational and logistic
limitations makes drinking water production far more difficult and inefficient. Based on the three
waterworks where the results of drinking water treatment could be analysed (Table 3), it was
concluded that the quality of treated water does not always conform to acceptable drinking
water quality standards, since turbidity and levels of organic matter in the treated water are
generally high and exceed recommended limits.
Although the mean values of turbidity in treated water (Table 3) are generally below the
tolerable limit of 5 NTU, the absolute limit of 1 NTU recommended for instance, for effective
disinfection with chlorine (WHO, 2004; Thompson et al., 2007), is seldom attained. At the
waterworks in Nampula, the situation is even worse, and the turbidity of treated water is
frequently reported to be above the maximum tolerable limit of 5 NTU. Besides limiting the
possibility of effective disinfection, high values of water turbidity are also known to contribute to
the release of undesirable by-products of water treatment such as residual aluminium, which is
20
Challenges and Opportunities for Safe Water Supply in Mozambique
ȱ
2008
associated with human health hazards (Alzheimer’s disease). In order to maintain levels of
residual Al3+ below the health guideline standard of 0.2 mg/L, turbidity residuals of less than
0.15 NTU are generally required (Thompson et al., 2007), and these are clearly not attained in
most drinking water treatment plants in Mozambique.
Table 3 Results of water treatment at three waterworks in Mozambique (Rwater = raw water; Twater=
treated water; -, not measured, Min. = minimum recorded; Max. = maximum recorded; STD =
standard deviation; N= number of samples analysed)
Quality
variable
Turbidity
(NTU)
pH
Org. matter
(mgO2/L)
Res. Cl(mg/L)
Alkalinity
(mg/L)
Umbeluzi waterworks
Rwater
Twater
Rwater
Twater
Rwater
Twater
Rwate
Twater
Twater
r
Min.
2.9
Max.
80.6
Mean
6.7
STD
6.6
N
359
Beira waterworks-old
0.7
7.3
3.1
0.9
358
2.0
6.4
4.0
0.8
351
0.8
4.8
2.7
0.7
350
6.3
8.1
7.4
0.20
357
6.5
7.9
7.4
0.2
358
98
240
160
14.2
353
41
230
152
18.7
351
0.3
2.8
1.4
0.4
333
Min.
13.5
Max.
326
Mean
49.9
STD
31.1
N
310
Beira waterworks-new
0.2
28.0
2.5
3.3
303
-
-
6.4
7.6
6.8
0.1
309
6.2
7.6
6.7
0.2
307
-
-
-
Min.
11.8
Max.
25.9
Mean
17.0
STD
4.6
N
9
Nampula waterworks
2.3
6.5
4.2
1.1
9
4.7
6.4
5.6
0.7
8
2.0
3.2
2.81
0.4
8
6.4
6.9
6.6
0.2
9
6.1
6.3
6.2
9
58
78
64.9
7.4
9
52
82
61
9.6
9
-
6.5
16.8
10.9
3.3
10
4.0
6.0
4.8
0.7
10
3.2
4.6
2.9
0.5
10
7.2
7.9
7.4
0.3
10
6.6
7.8
7.2
0.4
10
26.9
38.4
31.3
3.1
10
25.0
57.6
39.3
9.6
10
-
Min.
Max.
Mean
STD
N
11.7
47.0
25.0
13.8
10
Levels of organic matter in treated water were also found to be high and frequently in excess of
the target level of 2.5 mg/L (WHO, 2004). The treated water on the other hand, is generally
over-chlorinated (see Table 3). High levels of residual chlorine, combined with the high levels of
organic matter and water turbidity mean that taste- and odour-forming compounds, as well as
disinfection by-products, may develop frequently following treatment. The potential of human
health risks due to bacterial growth is also high, as organic matter can act as a source of
nutrients for bacteria during the transport, storage and distribution of water. Chemical
21
Challenges and Opportunities for Safe Water Supply in Mozambique
ȱ
2008
parameters such as pH and alkalinity that affects the corrosive and scaling properties of water
and, thus the suitability of the water for domestic use were generally found within acceptable
limits for corrosion control (pH > 6.5; alkalinity > 61 mg/L).
From the analysis of results of drinking water treatment (Papers) it was concluded that, while
methods currently used for drinking water treatment are generally suited for the production of
water of excellent quality from existing sources, poor quality of the raw water and seasonal
variations in water quality make drinking water treatment generally costly, technically
demanding and heavily dependent on the availability of consumables (e.g. chemicals) and
skilled personnel to operate the plants. Factors identified as having a significant effect on the
performance of water treatment plants are: (i) inadequacy of existing infrastructure and
operational procedures for water treatment, (ii) inconsistent operation of essential chemical
processes due to malfunctioning of equipment or temporary lack of consumables and, (iii) lack
of guidelines for specific contaminants (e.g. DBPs), which mean that water treatment is aimed
only at those variables for which guidelines have been established.
The chemical coagulation and flocculation stage seems to be the unit operation most affected
by the above-mentioned factors, not only because it depends greatly on the availability of
supplies, but also because careful adjustment of the operational parameters, notably the raw
water pH and alkalinity, is generally required ensuring high efficiency. Figure 5 shows annual
records of coagulation processes at the Maputo waterworks, in which it can be seen that the
process is performed with reagent dosages of alum and polymer that vary widely within same
ranges of raw water turbidity and alkalinity.
Treatment with Aluminium vs. raw water alkalinity
at the Umbelúzi waterworks
Treatment with and Aluminium and flocculant
at the Umbeluzi waterworks
60.0
ultrafloc
Alum
80.0
50.0
70.0
Aluminum dose (mg/L)
Aluminum and flocculant dosage
(mg/L)
90.0
60.0
50.0
40.0
30.0
20.0
40.0
30.0
20.0
10.0
10.0
0.0
0.0
0.0
10.0
20.0
30.0
40.0
50.0
60.0
70.0
80.0
90.0
90.0
Raw water turbidity (NTU)
120.0
150.0
180.0
210.0
240.0
Raw water alkalinity (mg/L)
Figure 5:Treatment with aluminium and coagulant aids at the Umbelúzi waterworks and its relation to raw
water turbidity (NTU) and alkalinity (mg/L). Data refers to one full year (2005) operation of
chemical coagulation at the treatment works. Source of data: FIPAG (2005).
Since effective operation of coagulation processes generally requires optimization of the pH,
usually between 6.5 and 7.0, and the addition of hydrolysing coagulants generally alters the
water alkalinity, coagulation using different coagulant doses in the same range of raw water
turbidity and alkalinity mean that deviations from the optimum reaction pH for coagulation may
occur frequently and affect the overall coagulation efficiency. The consequences of inefficient
coagulation are many, and include wastage of chemicals, excessive production of sludge and
22
Challenges and Opportunities for Safe Water Supply in Mozambique
ȱ
2008
increased treatment costs. The most important, however, is the resulting poor performance of
subsequent stages of treatment (sedimentation and filtration) and overall poor quality of the
treated water.
In the discussion of drinking water treatment results (Paper 1) it was also noted that chemical
processes essential to maintain acceptable standards of water treatment are frequently
discontinued, and that the control of raw and treated water quality is rather inconsistent. Proper
control of raw and treated water quality is essential to establish operational procedures and to
assess the performance of drinking water production. Operation in this rather inconsistent way
has certainly had an effect on the efficiency of drinking water production.
4.3
Coverage and service quality
4.3.1 Coverage and service quality provided by formal service providers
Distribution of water supply, in hours in five major
cities of Mozambique
24
Ave. Maputo
Quelimane
Pemba
21
Distribution of water supply, in hours among the five distribution
centres of the water supply of Maputo
24
Ave. Beira
Nampula
Average number of supply hours
in a day
Average number of supply hours
in a day
The quality of water supply services can be defined using various criteria. These include
coverage, service continuity and pressure, water quality, and the degree of responsiveness of
service providers to consumer’s complaints. Service coverage of piped water supplies in
Mozambique is discussed in the introduction of this thesis. The service quality provided by
formal service providers is discussed in Paper 2, using the network of Maputo as an example.
Emphasis was placed on drinking water quality. The reason for choosing this service quality
indicator was the fact that water supply services in Maputo, and many other urban centres of
Mozambique, are intermittent, despite the known inconveniences resulting from such
disruptions.
18
15
12
9
6
21
Matola
Chamanculo
Catembe
Machava
Maxaquene
Feb. Mar. Apr.
May June
18
15
12
9
6
3
3
Jan.
Feb.
Mar.
Apr.
May
June
July
Aug.
Sep.
Oct.
Nov.
Jan.
Dec.
July
Aug.
Sep.
Oct.
Nov.
Dec.
Figure 6 Distribution of water supply, in hours, in five major cities of Mozambique (left) and among
distribution centres of the water supply to the city of Maputo (right). The data on Maputo and
Beira water supplies was taken as average of distribution hours among existing distribution
centres. All data refers to year 2004. Source: FIPAG (2005).
Intermittent supply is employed in many parts of the world as it is believed to help reduce
leakage in old or badly maintained pipe networks, help raise awareness of water conservation
among consumers, and help reduce the per capita demand, compared with continuous
23
Challenges and Opportunities for Safe Water Supply in Mozambique
ȱ
2008
supplies, leaving, in theory, room for savings in investment and operational costs. Many recent
studies (Tokajian & Hashwa 2003; Coelho et al., 2003; Totsuka et al., 2004; Kajsa & Lindquidist,
2006; Vreeburg & Boxall, 2007) indicates, however, that intermittent supplies are linked to a
number of inconveniences to consumers among which, water shortages caused by reduced
flow, low or lack of pressure in the networks, quality problems resulting from the ingress of
contaminated water, and water quality deterioration resulting from prolonged storage.
As seen from figure 6, water supply and distribution in Maputo is also intermittent and the
reason for this is the need to reduce leakage during distribution (Gumbo, 2004; Zandamela,
2000). Methods commonly used by consumers to overcome problems caused by intermittency
are, the reliance on services delivered by alternative providers and the construction of extra
household tanks for those having piped connections. Both alternatives are, however, associated
with considerable water quality problems resulting either from exposure to unsafe water
supplies, the ingress of contamination, or water quality deterioration during storage.
Analysis of drinking water quality in the network of Maputo (Paper 2) showed that drinking water
in Maputo is not always safe for human consumption, due mainly to the occasional presence of
bacteria. Both faecal coliforms and E. Coli were found in reservoirs at distribution centres and
taps in the network. The reasons for this was a combination of factors that includes the
condition of pipes, ingress of contaminated water during periods of low or no pressure, long
retention times in pipes and reservoirs in the distribution network, and the condition and
maintenance of household tanks. Some contamination was found to occur before or at the
reservoirs of the distribution centres, clearly suggesting the ingress of contamination during
periods of low or no pressure.
Storage was found to significantly affect drinking water quality at household level. Samples
collected before household tanks generally had better quality than those collected after the
tanks. Factors contributing to this were found to be long storage times, poor maintenance of
household tanks and the ingress of contaminants. Since residual chlorine, used to maintain a
certain disinfection capacity in water, decreases naturally with time long storage times,
particularly at household level, were found to significantly affect water quality. As can be seen
from the results presented in Paper 2, and in figure 7, storage at household level increased the
risk of the presence of faecal coliforms in water by more than 100 %. These results are in
agreement with those of other researchers (Coelho et al., 2003; Totsuka et al., 2004), who also
found positive correlations between mean bacterial counts, pH, temperature and particularly
storage time. The disinfection capacity in the network was generally found to be high, which
partially explains why bacteria counts in the network were generally low.
24
120
120
100
100
80
80
60
60
40
40
Number of Coliforms
Number of Coliforms
Challenges and Opportunities for Safe Water Supply in Mozambique
20
0
-20
0.0
.2
.4
.6
Total Residual Chlorne (mg/l)
.8
1.0
1.2
ȱ
2008
20
0
House
House
Net
-20
Net
0.0
.1
.2
.3
.4
.5
.6
.7
Free Residual Chlorine (mg/l)
Figure 7: Presence and coliform bacteria in samples taken before and after storage at household level.
Although based on a limited number of samples, sediments in household tanks were also found
to potentially contribute to water quality deterioration. Most household tanks were found to be
over-dimensioned. This, combined with frequent discontinuities in the supply of water to the
household tanks provides the ideal conditions for the settling of turbidity-causing particles and
sediments build-up. The combined effect of sediments in tanks, low water disinfection capacity
and long retention times provides the ideal conditions for bacterial growth in the network and
household tanks. The intermittent mode of operation of the water network in Maputo has,
therefore, been pointed out identified as one of the critical factors affecting service and water
quality in Maputo. The most serious consequence of intermittency in Maputo is the deterioration
of water quality in network reservoirs and household tanks.
4.3.2 Coverage and service quality provided by informal service providers
Coverage and quality of the services provided by alternative service providers are discussed in
Papers 3 and 4. Water supply services offered mainly by small-scale independent providers in
peri-urban Maputo were used as an example. For the purpose of assessing service quality,
services were ranked according to three levels of quality: bad, reasonable and good. The
following three quality indicators were ranked: water price at private standpipes as compared to
public standpipes, promptness of standpipe attendants and neighbourhood authorities to
respond to consumer’s complaints, and access to services by individual consumers as
measured by the number of hours with pressure and open access to consumers at private and
public standpipes. For the assessment of water quality, borehole water used by independent
service providers was investigated regarding bacterial and organic contamination (E. coli and/or
faecal coliforms and contamination with nitrates) and salinity (through EC measurements).
Analysis of the coverage aspects from alternative service providers (Paper 4) gave a clear
picture of the role played by independent providers and their importance in service provision to
unconnected residents in large urban centres. Alternative service providers are known to cater
for as many as 32 % of households in Maputo (Sal-consultores 2005; Seureca & Hydroconseil
2005), while 62 % of households rely on services provided by formal service providers.
25
Challenges and Opportunities for Safe Water Supply in Mozambique
ȱ
2008
Regarding the quality of their services, although alternative providers charge more per unit
volume of water than the formal service provider, their services are generally appreciated by
consumers, mainly because they are more readily available and met their expectations more
rapidly. Regarding the quality of water provided by independent providers (Paper 3) it was found
that today, the quality of groundwater tapped by independent providers is virtually free from
microbial and/or organic contamination and is thus safe for human consumption and domestic
use. However, there were obvious exceptions, for example, high levels of salinity, caused by
sea water intrusion or the vulnerability of the aquifer system to sea water intrusion, making the
water from individual boreholes or specific locations unsuitable for human consumption.
The quality of water that is ultimately delivered to consumers depends not only on the quality of
water at source, but also on the treatment and storage methods applied following abstraction.
External factors, such as those resulting from storage conditions and the condition of pipe
networks, greatly affect the quality of the water. In an investigation of the quality of water from
this type of service providers it was found that the water from a number of systems operated by
independent providers was of inferior quality with respect to its microbial quality (Kajsa &
Lindquist, 2005). About 50 % of samples analysed from a data set consisting of 158 records had
bacteria counts in excess of guideline values.
Given that the quality of water at source is virtually free from contamination as was concluded in
Paper 3, a possible cause of the deterioration of water quality following abstraction is external
contamination, e.g. during storage, and the ingress of contamination during distribution. The
conditions during water treatment and storage following abstraction from the source seem to
contribute greatly to this situation. As noted in the discussion of Paper 2, none of the
independent providers interviewed reported that they regularly treated the water before
distribution. Thus, raw water is only treated when problems are detected during monitoring
checks conducted intermittently by the Ministry of Health via its Water, Food & Hygiene
Department.
26
Challenges and Opportunities for Safe Water Supply in Mozambique
ȱ
2008
5. CHALLENGES AND OPPORTUNITIES FOR SAFE WATER SUPPLY IN
MOZAMBIQUE
Major problems affecting drinking water supply in Mozambique are; inadequate service
coverage, poor reliability and inferior water quality. These problems are common to large and
small to medium size water supplies, but the latter face more serious problems because of
limited financial capacity for investments in new or upgraded infrastructures. The lack of
adequate piped water supplies is viewed as prompting consumers, particularly those living in
the surroundings of the main cities, to find alternatives to cope with water shortages by relying
on services offered by alternative providers, most of which in the form of small-scale
independent providers-SSIPs. These services are presently not formally regulated, yet the
quality and reliability of services seem to be generally better than that offered by formal service
providers (Paper 3). Lack of reliable piped water supply services is also viewed as prompting
consumers to find ways of coping with water shortages by constructing household tanks, most
of which are today claimed to be one of the root causes for water quality deterioration at
household level (Paper 2).
Drinking water production and management of drinking water quality is generally inadequate in
Mozambique. The discussion of source water quality aspects (Paper 1) and those of treatment
methods and efficiencies have indicated that problems exist at almost all systems using surface
water (mostly river water) for drinking water production. Conventional treatment is mostly used
for surface water treatment and drinking water production. The worst situations are found in
small and medium-sized water supplies, where the combined effect of poor water quality at
source and operational and logistic limitations makes drinking water production far more difficult
and inefficient.
Overall, the quality of water at source is suited for drinking water production with conventional
treatment but, factors such as poor water quality at source and seasonal variations and
operational and logistic constraints makes drinking water production difficult, ineffective and to
depend heavily on skilled personnel and reliable supplies to operate efficiently the plants.
Critical variables of surface water quality are turbidity, bacteria, NOM and algae and aquatic
plants. Groundwater sources used for large water supplies (e.g. to the city of Pemba) do not
face problems of organic contamination, but in here, problems with excess iron and/or
manganese and, sometimes, water hardness are the typical problems of large scale drinking
water production from groundwater sources.
Regarding the management of drinking water quality, the intermittent mode of operation of most
piped water supplies in Mozambique is viewed as one of the critical factors affecting and water
quality. The most serious consequence of intermittency is deterioration of water quality due to
ingress of contamination during periods of low or no pressure and prolonged storage in the
network, reservoirs and household tanks (Paper 2). These problems are common to services
offered by alternative service providers as well as those offered through formal piped water
supplies. Regarding the services offered by alternative service providers most, if not all, tap their
water from groundwater sources which are now assessed as virtually free from contamination
(Paper 3). However, this situation may change in future either due to over-exploitation of aquifer
systems or increased hydraulic loads of contaminants caused by increased population
densities.
27
Challenges and Opportunities for Safe Water Supply in Mozambique
ȱ
2008
In view of the above, major challenge facing the drinking water supply sector and institutions of
the water governance framework in Mozambique is, to develop sustainable drinking water
supplies, capable of supplying quantity water of acceptable quality to residents, industry and
commerce in most urban centres of the country.
5.1
Meeting targets in drinking water production and coverage
In sections 4.3.1 and 4.3.2 of this thesis a discussion was held on coverage aspects around
formal and informal service providers from where it was concluded that coverage levels from
either type of services are still far from optimum. In many cases, present operational capacity is
beyond required demands and expected growth. Added to that, existing infrastructure,
particularly the water distribution infrastructure, cannot be extended to reach the rapidly
expanding neighbourhoods of urban centres.
To meet targets in drinking water production and coverage, considerable investment is needed
in the development of additional infrastructure or extension of existing ones. This is an
unrealistic option in many places in Mozambique because of the limited financial capacity to
invest in all components of the drinking water supply infrastructure. Meeting targets in drinking
water production and coverage requires, therefore, a well established strategy whereby, the
priority is given to investments in areas that are more likely to bring short term benefits and
meet expectations of consumers and water governance authorities.
In the discussion of the drinking water production aspects (Paper 1) it was concluded that,
although most drinking water supplies are presently operating beyond required capacities,
actual production levels are close to required demands, suggesting that the priority for
investments is on improvements of water quality, water distribution and service coverage when
compared to bulk water production. Options to improve drinking water production and water
quality requires investments aimed not only to restore and expand partially production
capacities at existing treatment works, but also, to improve operation of treatment processes
and efficiency of drinking water production. Critical aspects around this matter are discussed in
section 4.4.2 of this thesis.
Improvements in water distribution and service coverage can be achieved either by physically
expand existing networks and improve operation and management of water quality during
distribution or by developing small-scale distribution networks designed to provide services to
areas presently located outside the range of coverage of formal distribution networks. The latter
is a strategy currently being explored to expand service coverage in the area of Maputo
whereby, state-funded groundwater based distribution networks are being developed in the
peri-urban areas of the city, for further delegation of operation and management responsibilities
to local-based private operators (Paper 4). SSIPs presently running self-financed water supply
systems will be involved as well.
SSIPs have a long history of acceptance by donors and governmental authorities as viable
alternatives to managing and expanding public services (Paper 3 & 4) and this can be
capitalized by the water supply governance authorities of Mozambique to help expand service
coverage to presently underserved areas. In view of this, water governance authorities in
Mozambique are asked to mobilize resources for developing such alternative water supplies,
and develop mechanisms for the involvement of SSIPs; through management contracts within
specified service areas inclusive those falling within formally contracted service areas under the
28
Challenges and Opportunities for Safe Water Supply in Mozambique
ȱ
2008
rubric of delegated management. Franchising models whereby an official provider (either a
private operator or a public entity) acts as the main franchisor and the independent providers as
the franchisee may also be explored given the potential they have to simultaneously improve
service delivery and local economic development (Wall, 2006). Since most SSIPs are currently
not formally regulated, the challenges facing the water governance authorities in this respect
are also related to formalization and regulatory aspects which are further discussed in section
5.5 of the thesis.
5.2
Meeting drinking water quality targets
Drinking water treatment aspects and quality criteria are discussed in Papers 1, 5 & 6 of this
thesis. From the discussion in Paper 1 it was concluded that, most common source of drinking
water production in Mozambique is surface water and that, most commonly applied method of
drinking water production is conventional treatment. Drinking water quality standards to be met
during water treatment are those developed by the Ministry of Health (MISAU 2004), based on
the WHO Guidelines.
The first place where drinking water quality standards have to be met is at the outlet of water
works. Analysis of drinking water treatment results (Paper 1) have shown, however that treated
water quality from the water works studied did not always conform to acceptable drinking water
quality standards. Turbidity and levels of organic matter in the treated water were generally high
and exceeded recommended limits of the guidelines. The waterworks studied are the bestequipped waterworks in the country, thus the ones where, in theory drinking water treatment
should have to be performed up to required standards. The situation in other, less equipped
waterworks is, therefore, likely to be worse. Factors identified as having a significant effect on
the performance of water treatment plants were: (i) inadequacy of existing infrastructure and
operational procedures for water treatment, (ii) inconsistent operation of essential chemical
processes due to malfunctioning of equipment or temporary lack of chemicals and, (iii) lack of
guidelines for specific contaminants (e.g. DBPs), which mean that water treatment is aimed only
at those variables for which guidelines have been established.
The main challenges facing water governance authorities in this respect is to assure that
currently applied methods of drinking water production are capable of producing treated water
that conforms to national and internationally accepted water quality standards (particularly those
which are health-related) and that consumers are provided with the best possible water quality.
As noted in Paper 1 and in section 5.1 of this summary, improvements of drinking water
production require interventions at the level of source water quality and at design and operation
of drinking water treatment processes. Practical aspects to be considered are:
ƒ
The need to guarantee proper selection, design and operation of drinking water
treatment infrastructure so that, existing or to be constructed infrastructure is capable of
dealing with commonly found raw water quality problems vis a vis current drinking water
quality standards.
ƒ
The need to guarantee that the selection and management of raw water sources is done
in such a way that existing treatment infrastructure is capable of dealing with
corresponding raw water quality aspects,
29
Challenges and Opportunities for Safe Water Supply in Mozambique
ȱ
2008
ƒ
The need to guarantee that the quality of treated water at the outlet of treatment plants
and further on during transport and distribution always conform to specified guidelines
and that consumers are provided with safe water from the physical chemical and
bacteriological points of view. The latter includes the effective removal of bacteria,
viruses and other health-related pathogens.
ƒ
The need to guarantee that water disinfection with chlorine and other chemical
processes used during water treatment do not become source of contaminants from
treatment chemicals such as residual Aluminium and disinfection by-products.
ƒ
The need to develop knowledge and skills to deal with particular raw water sources such
as raw water sources potentially polluted with organic contaminants of anthropogenic
origin.
Various possible methods exist to attain these objectives and these were discussed in Paper 1.
A process approach similar to the illustration in Figure 8 was proposed which is viewed with
potential to bring immediate and significant changes in the actual scenario of drinking water
production.
Nature of the
problem
Possible area of
intervention
Technical or managerial
possibility
ExpectedOutcome
Artificial water improudement,
pre-treatment at source, ground
water
Ineficient operation
of treatment
processes, lack of
installations, high
dependency on
imported supplies,
complexity of
treatment
processes, lack of
skills
Improved source
water selection,
improved
watershed
managment,
Treatment with
low treatment
methods, design
and operational
modifications into
existing
installations,
treatability
studies, training
Treatment at source (riverbank
filtration, river bed filtration ,
dinamic roughing filters)
Treatment with Multi-stage
filtration (slow sand filtration
with roughing filtration for pretreatment)
Treatment with adapted
conventional methods (direct
filtration, hydraulic flocculation
with up-flow roughing filters),
treatment with Moringa
Less polutted
water water,
smaller variations
in raw water
quality
Simpler treatment
methods, better
knoweledge of
treatment processes,
less dependence on
imported supplies,
most reliable
treatment, less costly
Improved drinking water production.
Poor quality of
water at source,
seasonal
variations
Figure 8 Process approach to improve overall drinking water production in Mozambique
Surface water quality in Mozambique is generally assessed as suited for treatment with
conventional methods (paper 1); however, the possibility of abstracting less polluted water from
existing sources may generally help improve drinking water production. Methods that can be
used to meet this objective are, the use of artificial lakes and pre-treatment at source (e.g.
through river bank infiltration) particularly if required flows are small. Problems with large
reservoirs in tropical regions were discussed in Paper 1 and in section 4.1.1 of this summary
and refers mainly to the risk of algae blooms, and presence of other toxic cyanobacteria and,
30
Challenges and Opportunities for Safe Water Supply in Mozambique
ȱ
2008
sometimes, changes in the chemical composition of raw water (e.g. alkalinity) which can turn
drinking water treatment even more complex and costly (Cheng et al. 2003; Baudin 2006).
Possible methods to improve design and operation of drinking water treatment were discussed
in Paper 1 and technical options investigated in Papers 5 and 6. According to the discussion in
Paper 1, improvements in the design and operation of drinking water treatment should aim the
reduction of inefficient operations, which were found to result from variations in the raw water
quality and inadequate operation of treatment processes. For reasons of long term
sustainability, methods proposed for improving drinking water production put emphasis on the
so-called low cost treatment methods, of which, roughing filters and coagulation-direct filtration
processes were the methods investigated.
Extensive research on these methods (Ingallinella et al., 1998; McConnachie et al., 1999;
Polasek & Mult, 2002; Mahwi et al., 2004) have proven that if properly incorporated into existing
treatment schemes, they can help improve treatment efficiency and overall production capacity
at relatively low investment and operational costs. Roughing filters are particularly
advantageous because they can handle large variations in raw water turbidity without
compromising overall treatment efficiency (Mishra & Breemen 1987; Smet & Visscher, 1990;
Sánchez et al., 2006). Coagulation-direct filtration processes are also advantageous because
they combine particle aggregation and separation in one single unit thus, giving room for large
savings in investment costs. Coagulation-direct filtration processes are also advantageous
because they demand lesser amounts of destabilizing reagents when compared to traditional
coagulation-flocculation-sedimentation.
The results of experimental work on roughing filtration, presented in Paper 5, have proven that
use of roughing filtration for hydraulic flocculation following chemical coagulation of surface
water can provide a viable and flexible alternative to improve turbidity and solids removal by
conventional rapid sand filtration. As seen from results presented in Paper 5, use of an up-flow
roughing filters for hydraulic flocculation (contact-filter) resulted in the formation of suspensions
which were completely retained by subsequent rapid sand filtration with minimum head loss
developed in the filter bed of both units (Figure 9). The pilot plant used in the experiments was
tested with the up-flow roughing filter operated at filtration velocities comparable to that of rapid
sand filters yet, without signs of clogging or reduced efficiency of the up-flow filter. Best
performances were attained with the pilot plant run at filtration velocities of about 6.0 m/h which
are suitable for full scale operation of rapid sand filters. Coagulant doses were 25 % less than
the amount required if conventional coagulation-flocculation was used.
31
Challenges and Opportunities for Safe Water Supply in Mozambique
Time dependent behaviour of filtrate water turbidity from the
contact-filter
V=6.3m/h,alum= 1.8 mg/L
v=6.3m/h,alum = 2.5mg/L
V=9.4m/h,alum= 1.8mg/L
V=9.4m/h,alum =2.5mg/L
V=12.7m/h,alum= 1.8mg/L
V=12.7m/h,alum= 2.5mg/L
1.2
1.0
Head loss in the filter bed of the
contact- filter (mm)
Ratio between filtrate and feed water
turbidity
1.4
100.0
0.8
0.6
ȱ
2008
Time dependent behaviour of head losses in the filter
bed of the contact-filter
V=9.4m/h, alum= 1.8mg/l
V=12.7m/h, alum= 1.8mg/l
V=6.3m/h, alum =2.5mg/l
V=12.7m/h, alum= 2.5mg/l
V=9.4m/h, alum=2.5mg/l
V=6.3m/h, alum=1.8mg/l
90.0
80.0
70.0
60.0
50.0
40.0
30.0
20.0
0.4
10.0
0.0
0.2
0
2
4
6
8
10
0
12
1
2
3
4
Filtration time in hours
V=3.2m/h, alum= 2.5 mg/L
3.1
V=6.3m/h, alum=1.8 mg/L
2.8
V=6.3m/h, alum= 2.5 mg/L
V=9.4m/h, alum= 1.8 mg/L
2.5
6
7
8
9
10
11
12
13
14
Time dependent behaviour of head loss in the filterbed of
the rapid sand filter
700.0
V=3.2m/h, alum= 1.8 mg/L
V=3.2m/h; alum= 1.8 mg/L
Head loss in the filter bed of the
rapid sand filter (mm)
Ratio between filtrate and feed water
turbidity
Time dependent behaviour of filtrate water turbidity from the
rapid sand filter
3.4
5
Filtration time in hours
V=9.4ml/h, alum= 2.5 mg/L
2.2
1.9
1.6
1.3
1.0
V=3.2m/h,alum=2.5 mg/L
600.0
V=6.3m/h, alum= 1.8 mg/L
V=6.3m/h, alum= 2.5 mg/L
V=9.4m/h, alum= 2.5 mg/L
500.0
V=9.4m/h,alum= 1.8 mg/L
400.0
300.0
200.0
100.0
0.7
0.0
0.4
0
2
4
6
8
0
10
1
2
3
4
5
6
7
8
9
10
Filtration time in hours
Filtration time in hours
Figure 9 Time dependent behaviour of filtrate turbidity (NTU) and head losses (mm) during filtration
experiments with hydraulic flocculation in a multi-layer up-flow roughing filter followed by single
media rapid sand filtration. Average raw water turbidity was of about 20 NTU. Information about
filtration velocities and alum dosages applied is also shown.
These results were in line with findings from other researchers (Mishra & Breemen, 1987; Mahvi
et al.,2004; McConnachie et al., 1999,) about the possibilities of using roughing filtration for
hydraulic flocculation prior to conventional rapid sand filtration (direct filtration processes). As
seen in the results presented in Paper 5, the quality of feed water transferred to rapid sand filter
following was generally better than that laving the gravel bed of the contact filter which suggests
that apart from a partial removal of particles in the gravel bed, additional removal of particles
took place in the supernatant water above the gravel bed (Figure 10).
This mean that if properly designed a contact filter using roughing filters can perform
simultaneously as a flocculation and sedimentation basin which, in theory, opens room for large
savings in investment costs in the event of construction of rehabilitation of drinking water
treatment plants. Added to that, is the fact that the pilot plant could be operated at dosages of
chemical reagent that were some 25 % less than the amount required if conventional treatment
was used which means that saves are also possible in relation to operational costs of drinking
water production.
32
Challenges and Opportunities for Safe Water Supply in Mozambique
Eff. gravelbed/ Eff.
supernatant
7
Alum dose 1.8 m g/l
Alum dose 2.5m g/l
6
5
4
3
2
2008
12
V=6.3 m /h
Trw = 13-21 NTU(S)
Alum dose 1.8 m g/l
Eff. gravelbed/Eff.
supernatant
8
ȱ
10
Alum dose 2.5 m g/l
V=9.4m /h
Trw =4.0-9.7NTU(N)
8
6
4
2
1
0
0
Filtration runs
Filtration runs
5
Alum dose 1.8 m g/l
5
4
V=9.4 m /h
Trw = 15-20.4NTU(S)
4
Eff. gravelbed/Eff.
supernatant
Eff. gravelbed/Eff.
supernatant
Alum dose 1.8 m g/l
Alum dose 2.5 m g/l
3
2
1
V=12.7 m /h
Trw =14.7-19NTU(S)
3
2
1
0
0
Filtration runs
Figure 10
Alum dose 2.5m g/l
Filtration runs
Relative contribution of gravel bed and supernatant layer in turbidity removal in the contact
filter: (N) = natural turbidity; (S) = synthetic turbidity; Eff. = efficiency (%. Information about
filtration velocities and alum dosages applied is also shown
Because drinking water production in Mozambique is also affected by constraints related to the
temporary lack of supplies essential for effective operation of chemical processes (coagulants),
methods of improving chemical treatment by using natural coagulants (M. Oleifera) as a
substitute to metal salts were also tested. Results obtained (Paper 6) coincide with findings from
other researchers (Ngabigengesere & Narasiah, 1998; Katayon et al., 2005) about the
advantages of using the natural coagulant for drinking water treatment notably that:
x
treatment with M. Oleifera does not affect the chemistry of treated water,
x
treatment with M. Oleifera produces lesser amounts of sludge as compared to
coagulation with metal salts,
x
M. Oleifera has the possibility of production at local level and at relatively low costs.
Use of M. Oleifera for water coagulation has, however the disadvantage of increasing the
concentration of nutrients and COD in the treated water. These findings were confirmed during
the experiments with M. Oleifera the results of which were presented and discussed in Paper 6.
Accordingly most efficient treatment was found when using aluminium sulphate for coagulation
of water, however, treatment with M. oleifera could also produce treated water of excellent
quality. Methods used to extract the active agents from the M. oleifera included: extraction with
tap water, extraction with distilled water and, extraction with distilled water followed by oil
extraction. High turbidity removals were obtained when the active agents were extracted using
tap water. Extraction methods affected therefore M. oleifera performance but not the final
chemistry of the treated water.
33
Challenges and Opportunities for Safe Water Supply in Mozambique
ȱ
2008
Turbidity after treatment
(NTU)
Treatment efficiency
3.5
3
2.5
2
1.5
1
0.5
0
Aluminium sulphate
Standard Moringa
Oil extracted Moringa
Tap water Moringa
5
15
30
50
100
Initial turbidity level (NTU)
Figure 11: Treatment efficiency with Moringa and Aluminium sulphate. Comparison of optimum dosage
of coagulant and treatment efficiency (turbidity removal) for different levels of raw water
turbidity. Results from jar-test experiments.
Treatment of water coagulated with M. oleifera with up- flow roughing filter used for hydraulic
flocculation and rapid sand filtration used final treatment was also tested. The results were
compared to those obtained when alum sulphate was used for coagulation instead. Treatment
with aluminium sulphate resulted in a better effluent quality but, the treated water when M.
oleifera was used was also of acceptable quality. Treated water turbidity when M. oleifera was
used was within acceptable limits for drinking water production (less than 2 NTU) and the
increase of head losses in the filters was not higher for M. oleifera when compared to aluminium
sulphate.
From the results obtained with the experiments with M. oleifera, it was concluded, that treatment
with M. oleifera can be a simple, cheap and sustainable solution to substitute aluminium
sulphate during water coagulation, particularly in small waterworks. Overall, it can be said that
tap water extracted M. oleifera and treatment with flocculation followed by direct filtration
processes are alternatives that can be explored in the event of expansion or construction of
small waterworks.
5.3
Management of drinking water quality
In previous sections of this thesis, a discussion was held on drinking water production aspects
of Mozambique, the critical aspects affecting treatment possibilities and the possibilities of
improving drinking water treatment efficiency through adoption of low cost treatment options.
Methods discussed considered improvements of source water quality (e.g. through best
watershed management practices and pre-treatment at source) and improvements in design
and operation of drinking water treatment processes.
Drinking water quality management may be established through a combination of protection of
water sources, control of treatment processes and management of the distribution and handling
of water. It has generally five key components:
34
Challenges and Opportunities for Safe Water Supply in Mozambique
ȱ
2008
1. The water quality targets;
2. A system assessment to determine whether the water supply chain (up to the point of
consumption) as a whole can deliver water of a quality that meets the above targets;
3. A monitoring of the control points in the supply chain that are of particular importance in
securing drinking water safety;
4. Management plans documenting the system assessment and monitoring; and describing
actions to be taken under normal and incident conditions. This includes documentation
and communication; and
5. A system of independent surveillance that verifies that the above are operating properly.
Formally water supply agencies have a basic responsibility to provide safe water and would be
expected to develop and implement management plans to address points 2 through 4 above.
These management plans should address all aspects of the water supply, focusing not only on
control of water production and treatment but also on the delivery side of drinking water.
While the priority in most water supplies is the need to comply with drinking water quality
standards at the outlet of the waterworks, the quality of water that finally reaches consumers is
not always the same as that leaving the treatment works. External factors such as postcontamination due e.g. to ingress of contamination, and water quality deterioration due to
prolonged storage in pipes, reservoirs and household tanks, are known to largely impact the
final quality of water delivered to consumers. This mean that, while compliance to drinking water
quality standards at production level is important, the management of drinking water quality
following treatment is equally important in the global effort of meeting drinking water quality
targets. Main challenges facing water governance authorities of Mozambique in this respect, is
the need to secure that treated water is of good quality at any time and location downstream the
treatment facilities and that consumers are provided with the best possible water quality.
Management aspects of drinking water quality in Mozambique were discussed in Paper 2, using
the network of Maputo as an example. Factors identified as having a significant impact on water
quality deterioration in the network were: ingress of contamination during periods of low or no
supply and prolonged storage in pipes and household tanks. The intermittent mode of operation
of the network of Maputo was found to largely contribute to the problems. Because intermittent
supplies are used not only in Maputo but in almost all urban settlements of Mozambique, the
drinking water quality situation of other cities where existing pipe works are in poorer conditions
will be similar or even worse than that encountered in Maputo.
This mean that besides efforts to increase drinking water availability in the main towns, the main
challenges facing water governance authorities in this respect is the need to implement actions
aimed at reducing inconveniences resulting from intermittent supplies thus, eliminating factors
responsible for water quality deterioration following treatment. Since, moving from the present
condition of intermittent supplies to conditions of continuous supplies is a long term goal that
requires massive investment, simpler and less costly actions can be implemented which
includes: increasing public awareness on the health risks associated with practices commonly
used to overcome water shortages caused by intermittent supplies, improve water quality
management practices (e.g. through a review of water re-chlorination strategies) and improve
household water management practices (e.g. through mandatory rules for construction and
location of household tanks).
35
Challenges and Opportunities for Safe Water Supply in Mozambique
5.4
ȱ
2008
Role of alternative service providers
Urban water supply services by alternative service providers were discussed in Papers 3 and 4.
From the discussion in Paper 4, it was concluded that they provide a valuable contribution in
overcoming the problems resulting from inadequate service coverage particularly in peri-urban
areas experiencing rapid population growth. In the discussion, it was also concluded that the
demand for their services is driven not only by demand for services in areas presently lacking or
with unreliable piped water supplies but also by the incapacity of formal providers to respond to
the water demand of their service areas. This situation is unlikely to change in the near future.
Moreover, SSIPs have recently been recognized by water governance authorities as important
partners in the global effort of expanding service coverage with piped water supplies. As noted
in Paper 4, the strategy adopted is based on public private partnership arrangements whereby
the public sector transfers part of its responsibility of service delivery to self-financed or formally
delegated private actors, among which SSIPs, while keeping the political responsibility for the
services.
The typical design of water systems constructed by SSIPs or those constructed by water
governance authorities for further delegation of operational responsibility to private operators is
based on groundwater sources which mean that, the potential of regional aquifers used to
develop such systems and the associated quality problems are key elements for the long term
planning of service delivery expansion with alternative service providers. Moreover, most
systems of this type are constructed within moderate to densely populated areas where, sewers
do not exist and sanitation is mainly provided through septic tanks, cesspits and dry-pit latrines
(Paper 3). Under these conditions, seepage from on-site sanitation represents the most serious
source of widespread pollution (both point and diffuse) to the aquifer systems. Construction and
completion details of the boreholes that will be connected to such systems are also crucial
factors in that they may increase the risk of groundwater contamination by creating localized
pathways for ingression of pathogens (Schmoll, 2006; Godfrey, 2005) or by shortening the
distance and time required for pathogens to reach the groundwater table (Argoos, 2001). The
extent and risk of groundwater contamination will depend, however, on many factors among
which, the degree of attenuation of contaminants during percolation through the unsaturated
zone and, eventually, through the aquifer system (Lewis et al., 1981; Sugden, 2006).
The expansion of service provision with the involvement of self-financed or formally delegated
alternative providers is, therefore, associated to quality and quantity problems that may
undermine the long-term sustainability of the proposed strategy. These problems were
discussed in Paper 3 and possible measures to diminish risks of failure of the proposed strategy
identified. One of such measures is the establishment of clear regulatory tools that will enable
water governance institutions frame the activity of alternative providers and enforce the adoption
of more stringent protective measures for boreholes constructed to provide public services and
for construction and management of small-scale water supply systems. This include for
example:
x
Regulated procedures for borehole design and location in order to minimize risks of
groundwater contamination. Emphasis should be put on aspects such as wellhead
protection, positioning of filter screens and the location of boreholes in relation to
existing pit latrines. A minimum radius of influence of 25 m from pit latrines is generally
accepted in Mozambique.
36
Challenges and Opportunities for Safe Water Supply in Mozambique
x
x
ȱ
2008
Mandatory rules for direct protection of boreholes used for drinking water supply (e.g. a
5x5 m surrounding fence)
Mandatory rules for all type of alternative providers regarding chlorination of the water
before distribution.
In the discussion of urban water supply services (Paper 4), it was also concluded that, as for
today, the activity of alternative providers is not formally regulated and that reforms are needed
in the existing legal and regulatory framework so that matters concerning service quality and
protection of consumer´s interests can be addressed. Critical aspects around this matter are
discussed in the next section of this thesis.
5.5
Regulation and legal aspects around formal and informal service providers
Legal and regulatory aspects of urban water supplies in Mozambique are discussed in section
2.2 of the thesis and in Paper 4. In Paper 4, the main aspects of regulation of water supply
services in an environment where formal and informal service providers coexist were discussed
using the city of Maputo as an example. It was concluded that, while regulation aspects around
formal water supply services are clearly defined, the situation concerning alternative service
providers is still unclear. This was found to limit the possibilities of expanding the existing
regulatory framework in a way to make it inclusive of all forms of service provision in urban
environments. Given that alternative providers generally operate in areas where the majority of
the urban poor lives, the existing situation of regulation of urban water supply services was also
found to limit the possibilities of establishing mechanisms to protect unconnected consumers
with regard to matters such as water pricing and water quality thus, undermining the ultimate
goal set by the water governance framework of servicing the poor.
As noted in Paper 4, in cities with delegated management of water supply, a management
model for peri-urban water supplies has been developed by the water governance authorities
that, in theory, should address pro-poor aspects in the existing regulatory framework. This
management model was, however, found to have a number of limitations to meet its objectives
because:
x
The model is based on a contractual agreement between standpipe attendants of the
public service and the main service provider. As such, the model is not applicable to
other forms of service provision (e.g. from the informal water market). The model does
not open room for the establishment of mechanisms to protect unconnected consumers
in aspects related for instance to water pricing and water quality.
x
There is lack of legal basis for issuing licenses to authorize other types of service
providers to operate within the lease area and hence their inclusion in the standpipe
management model.
x
There is lack of clarity concerning the roles of municipalities and neighbourhood
authorities in the management of water services in areas covered by the framework.
This result in a lack of clarity and possible conflicts of interest of the role they have to
play as system managers and system regulators.
x
Lower water governance institutions of the framework (e.g. e.g. municipal and
neighbourhood level authorities) have a weakened position to perform in an unbiased
37
Challenges and Opportunities for Safe Water Supply in Mozambique
ȱ
2008
fashion their regulatory tasks. Neighbourhood level authorities are also involved in the
selection and nomination of standpipe attendants and in the management of the
finances of the standpipes therefore, tasks of management, supervision and regulation
are generally not separated.
The most serious challenge facing water governance authorities in this respect is therefore the
need to establish mechanisms that will allow the expansion of the existing regulatory framework
to cover also the informal water market. As noted in Paper 4, the first step in this respect is the
definition of a clear licensing framework with which, all actors involved in the provision of water
supply services will be obliged to comply. Aspects to be taken into consideration in this respect
are also discussed in Paper 4 and include:
x
The licensing framework should be inclusive of all forms of alternative service providers
regardless of their position in relation to formally contracted concessions.
x
The financial and social sustainability of services provided through public standpipes
should be maintained because it is through this type of service provision that the majority
of urban poor get access to piped water. Actors of the presently informal water market
should therefore be complementary and not competitors to the public service,
x
The issuing of licenses for water resale activities should follow a real demand as
expressed by consumers,
x
To reduce the risk of inconsistent and unsustainable services due to managerial
constraints, the issuing of licences for actors of the presently informal water market
should follow proven managerial capacity of potential candidates.
Since one of the major weaknesses identified in the existing regulatory framework is the lack of
clarity concerning the position and role played by institutions of the lower water governance
framework in the management of water supply services, the proposed licensing framework
should also consider a decentralization of certain regulation activities to institutions of the lower
level of the water governance framework. Emphasis should be put on separating the
supervision and regulatory roles played by these actors. With CRA playing the role of normative
agency, and the Municipalities playing the role of the licensing authority, the water governance
framework will have greater leverage to ensure compliance. In areas without delegated
management of water supply the normative role will be played by district or municipal regulatory
commissions. Furthermore, at neighbourhood level, clarification of communication lines from the
community level upwards to the institutions at higher level of the governance framework is
required. Partnerships with Non Governmental Organizations-NGOs, Community based
organizations-CBOs and the municipality are also needed to develop information channels and
communication campaigns to inform consumers of their rights and options for recourse or
assistance.
38
Challenges and Opportunities for Safe Water Supply in Mozambique
ȱ
2008
6. SUMMARY AND CONCLUSIONS
This thesis had two main purposes: to analyze water supply services in Mozambique and critical
factors affecting provision of safe and reliable drinking water supply services and, to investigate
methods of improving the situation and challenges facing the sector to secure safe and
sustainable drinking water supply in the long run. This was done by analyzing drinking water
supply aspects based not only on quantitative aspects but also on qualitative indicators such as
the quality of drinking water, and the reliability of water supply services. Critical factors affecting
drinking water production were also analysed and methods to improve the situation evaluated.
Legal and regulatory aspects of drinking water supply were also discussed.
From the analysis presented and the results obtained, the following answers were given to the
research questions of the thesis.
Which key factors currently affect drinking water production and the management of
drinking water quality in the context of urban water supplies in Mozambique?
Analysis of drinking water production aspects have shown that as for today, drinking water
treatment is not done up to the standards due to a combination of factors that includes; poor
water quality at source, lack of functional installations and poor performance of processes within
existing installations. Most commonly used source of raw water for drinking water production is
river water and for treatment, conventional methods are mostly used.
The quality of water from rivers in Mozambique is very similar to that encountered in many other
parts of the world, where drinking water production is done up to standards with treatment
methods that are similar or even simpler than those used in Mozambique. Methods used for
drinking water treatment and sources used for drinking water production are, therefore, suited
for production of water of excellent quality.
Factors identified as having a significant effect on drinking water production in Mozambique
were ranked in two categories namely: source water quality related factors and, operational and
technical related factors. Source water quality factors refers mainly to the seasonal variations in
river water quality which makes drinking water treatment to be generally costly and technically
demanding because the treatment process must be sufficiently flexible to treat the varying
quality of the raw water. Chemical coagulation and flocculation seem to be unit operations
mostly affected by such variations because for best efficiencies, careful adjustment of
operational parameters is generally required. Inefficient performance of coagulation-flocculation
generally leads to inefficient performance of subsequent clarification processes thus, to
inefficient production of drinking water. Technical and operational related factors refers mainly
to: (i) inadequacy of existing infrastructure and improper operation of treatment processes due
to lack of knowledge among plant operators, (ii) inconsistent operation of essential chemical
processes due to malfunctioning of equipment and temporary lack of consumables and, (iii) lack
of guidelines for specific contaminants (e.g. DBPs), which means that water treatment is aimed
only at those variables for which guidelines have been established. Overall, technical and
operational related factors have more impact on drinking water production.
Management of drinking water quality is also deficient. Present limitations in water safety and
water quality results from lack of adequate treatment and inadequate strategies for operation of
39
Challenges and Opportunities for Safe Water Supply in Mozambique
ȱ
2008
water distribution (intermittent operation) which leads to frequent deterioration of treated water
quality. Factors identified as having a significant impact on water quality deterioration were;
ingress of contamination and water quality deterioration due to prolonged storage in pipes and
reservoirs. The intermittent mode of operation of water distribution has, therefore, been
identified as one of the critical factors affecting water quality following treatment. Besides
leading to frequent fluctuations in pressure, intermittent supplies are known to force consumers
to find ways of coping with intermittency by constructing extra household tanks. Pressure
fluctuations contribute largely to ingress of contamination while, household water storage
contributes to water quality deterioration due to prolonged storage.
Given the present status of urban water supplies, what alternatives are there for
subserviced consumers to overcome the inconveniences caused by a lack of, or
inadequate, piped water supplies? In addition, which factors affect the current and longterm sustainability of service provision through alternative service providers?
Methods found by subserviced consumers to meet their water demands comprehends the
reliance on services provided by alternative service providers. Service provision with alternative
providers is generally assessed as good in matters related to coverage, service reliability and
accessibility to consumers. The quality of water from sources used to provide services is also
assessed as safe for human consumption and domestic use. Yet, the quality of water that is
ultimately delivered to consumers depends not only on the quality of water at source, but also
on the treatment and storage methods applied following abstraction. External factors, such as
those resulting from storage conditions and the condition of pipe networks, greatly affect the
quality of the water. These are common problems to water services provided through formal
water supplies which mean that consumers relying on alternative providers are exposed to the
same level of human health risks as those relying on formal water supply services.
Service provision with the help of alternative providers is not likely to change in the near future
mostly because of increasing demand for their services. Moreover, alternative service providers
have recently been recognized as strategic partners for service provision expansion in areas
presently lacking formal water supplies. The typical design of water systems run by alternative
providers is based on groundwater flow, therefore, the potential of regional aquifers used to
construct such systems and the associated quality problems are key elements for the long term
planning of service delivery expansion through alternative providers. Factors such as overexploitation of the aquifers and increased hydraulic loads of contaminants due to increased
population densities will certainly impact their capacity of providing quality water of sufficient
quantity in the long term.
The long term sustainability of service provision expansion with the help of alternative providers
is also faced with possible threats associated to their actual legal status. Main challenges facing
water governance authorities in this respect are the need to establish mechanisms that will
allow the formalization of alternative providers and the expansion of the existing regulatory
framework to allow for the enforcement of more stringent protective measures around water
supply services offered through this segment of providers.
40
Challenges and Opportunities for Safe Water Supply in Mozambique
ȱ
2008
Given the prevailing situation regarding drinking water production and the management
of drinking water quality, which options exist to improve the efficiency of drinking water
production in order to secure safe and sustainable drinking water supplies in the long
term?
The evaluation of drinking water production aspects of Mozambique has indicated that drinking
water treatment is deficient. Poor quality at source, inadequate infrastructure and, inconsistent
or inadequate operation of water treatment processes were factors identified as affecting largely
drinking water production. Possible methods to improve the situation were investigated and
found to require improvements both in the management of source water quality and in the
design and operation of treatment processes within the actual context of skills, logistics and
managerial capacity. For reasons of sustainability, low cost treatment technologies, among
which, roughing filtration and coagulation-direct filtration processes with up-flow roughing
filtration used for hydraulic flocculation were investigated given their potential to increase overall
production capacity and treatment efficiency at relatively low investment and operational costs.
These methods were found to be particularly advantageous because they can handle large
variations in raw water turbidity without compromising overall treatment efficiency. Also,
because they generally combine particle aggregation and separation in one or few unit large
savings in investment costs can be attained through elimination of sedimentation and
flocculation basins usually required in conventional treatment. Coagulation-direct filtration
processes are also advantageous because they demand lesser amounts of destabilizing
reagents when compared to traditional coagulation-flocculation-sedimentation.
Use of natural coagulants such as the seeds of the M. oleifera tree for water coagulation at
small scale treatment plants was also investigated and the results compared to those obtained
when using aluminium sulphate for the same purpose. Although most efficient treatment was
found when using aluminium sulphate, treatment with M. oleifera could also produce treated
water of excellent quality. Use of M. oleifera for water coagulation was found therefore an
realistic alternative to conventional methods in small to medium size water supplies, presuming
that an adequate amount of plantations are established. Use of M. oleifera and treatment with
flocculation followed by direct filtration processes are, therefore, alternatives that can be
explored in the event of expansion or construction of small waterworks.
41
Challenges and Opportunities for Safe Water Supply in Mozambique
ȱ
2008
REFERENCES
Anderson A., Haeck, P., and Hagström Å. (1994). Effect of temperature and light on the growth
of micro- nano and pico-plankton: impact on algal succession. Mar Biol, Vol. 120, No 4, pp.
511-520.
Argoss (2001). Guidelines for assessing the risk to groundwater from on-site sanitation. British
Geological Survey Commissioned Report, CR/01/142.97pp.
Baudin, I. Cagnard, O, Grandguillaume, J.J., and Z. Do-Quang (2006). Algae and associated
toxins & metabolites: methodology for risk assessment and risk management. Water
Practice & Technology Vol 1, No 4, doi: 10.2166/WPT.2006081.
Bojcevska, H and Jergil, E. Removal of cyanobacterial toxins (LPS endotoxin and microcystin)
in drinking-water using the BioSand household water filter (2003). Evolutionary Biology
Centre, Uppsala University, Sweden. Report Minor field study
Boyer, Alyssa (2006). The regulation of Peri-urban Areas of Maputo, Mozambique. A Survey of
Household water resale activity in Peri-urban Maputo- Preliminary discussion findings.
Collaboration. Msc. Thesis, Columbia University-USA.
Boulding, J.R. and Ginn J.S. (2004). Practical Handbook of Soil, Vadose Zone and Groundwater
Contamination: Assessment, Prevention and Remediation. (2nd edn.) Lewis Publisher (CRCPress), Boca Raton, Florida-USA.
Burgeap (1961). Hydrogeologie du sud de Save. Dir. dos Serv. de Geol e Minas. Technical
report.
Chuang, C. J., and Li K.Y. (1997). Effect of coagulant dosage and grain size on the
performance of direct filtration. Separation and Filtration Technology, 12, pp. 229-241.
Chorus, I., and Bartram, J. (1999). Toxic Cyanobacteria in Water. A guide to their Public Health
Consequences, Monitoring and Management. World Health Organization. E&FN Spon,
London.
Coelho, S.T., James, S., Sunna, N., Abu Jaish, A. and Chatila, J. (2003). Controlling water
quality in intermittent supply systems. Water Science and Technology- Water Supply Vol.
3 (1-2), pp. 119-125.
Collingnon B., and Vézina M. (2000). Independent water and Sanitation Providers in African
cities. Full report of a ten cities study. UNDP/World Bank.
DNA (1987). Hydrologeological Map of Mozambique. Government of Mozambique, Maputo.
DNA (1995). National Water Policy. Govenrment of Mozambique. Maputo.
42
Challenges and Opportunities for Safe Water Supply in Mozambique
ȱ
2008
DNA (1999). Gestão Delegada de Abastecimento de Agua nas cidades de Maputo, Beira,
Quelimane, Nampula e Pemba. Unpublished report. Departamento Nacional de Aguas,
Maputo.
Dzwairo B., Hoko Z., Love D., and Guzha E. (2006). Assessment of the impacts of pit latrines
on groundwater quality in rural areas: A case study from Marondera district, Zimbabwe.
Physics and Chemistry of the Earth, vol. 31, 15-16, pp. 779-788.
Ferro, B.P., and Bouman, D. (1987). Hydrogeological map of Mozambique. Government of
Mozambique, Ministry of Construction and Water/UNICEF.
Godfrey S., Timo F. and Smith M. (2005). Relationship between rainfall and microbiological
contamination of shallow groundwater in Northern Mozambique. Water SA vol. 31, No 4, pp.
609-614.
Gumbo, B., Juizo, D., and van der Zaag, P. (2003). Information as a pre-requisite for water
demand management: experiences from four cities in southern Africa. Physics and
Chemistry of the Earth, vol. 28, pp. 827-837.
Gumbo, B. (2004). The status of water demand management in selected cities of Southern
Africa. Physics and Chemistry of the Earth, vol. 29, pp. 1225-1231.
Gustafsson A. and Johansson M. (2006). An Investigation of nutrient levels along the Mbuluzi
River- A background for sustainable water resources management. Department of Water
Resources Engineering, Faculty of Engineering, Lund University, TVRL 2006:9.
Hammer, M.J. and Hammer Jr., M. J. (2004). Water and Wastewater Technology, 5th edn.,
Pearson Education International, ISBN 0-13-191140-6.
Hua F., West J.R., Barker A. Barker and Foster C. F. (1999). Modelling of Chlorine Decay in
Municipal Water Supplies. Water Resources, Vol. 33 No. 12, pp 2735-2746.
Huisman, L. (1984). Rapid Filtration. Delft University of Technology, Faculty of Engineering,
Delft, The Netherlands, pp. 162-174.
INE (1999). Resultados do II Recenseamento Geral da População 1997 (Final results of the
1997 population census). Instituto Nacional de Estatística- INE, Maputo, Mozambique.
IWACO, (1983). Study of Groundwater Potential of Aquifer System North of Maputo, DNA.
Technical report.
IWACO (1984). Study of Groundwater Potential of Aquifer System North of MaputoHydrogeological report, DNA. Technical report.
IWACO (1985). Study of Groundwater Potential of Aquifer System North of Maputo-Natural
Recharge Estimates. DNA. Technical report.
43
Challenges and Opportunities for Safe Water Supply in Mozambique
ȱ
2008
Ingallinella A. M., Stecca L. M. and Wegelin, M. (1998). Up-flow Roughing Filtration:
Rehabilitation of a water treatment plant in Tarata, Bolívia. Wat. Sci. Tech. vol. 37, No.9,
pp. 105-112.
Juizo, D. (2008). Methods for Transboundary Water resources Management in Water Stressed
Regions- A case study: Southern Africa. Ph.D Dissertation, Department of Water
Resources Engineering, Lund University, Sweden.
Juizo, D. (1995). Modelação Inversa para a Determinação da Recarga dos Aquíferos a Norte de
Maputo. Bachelor-thesis report, Eduardo Mondlane University- Maputo, Mozambique.
Kajsa M. and Linquidist A. (2005). The effect of intermittent supply and storage in the quality of
drinking water in Maputo. Department of Water Resources Engineering, Faculty of
Engineering, Lund University, TVRL 2005: 7.
Karium M. and Schwartz J. (2005). Small Scale Private Service Providers of water supply and
electricity. A review of incidence, structure, pricing and operating characteristics. World
Bank, Policy and Research Working Paper 3727.
Katayon, S., Megat Mohd Noor M.J., Asma M., Abdul Ghani L.A, Thamer A.M., Azni I., Ahmad
J., Khor B.C., and Suleyman A.M. (2006). Effects of storage conditions of Moringa
Oleifera seeds on its performance in coagulation. Bioresource Technology, Vol 97(13), pp.
1455-60.
Katayon S., Megat Mohd Noor M.J., Asma M., Thamer A.M., Liew Abdullah A.G., Idris A.,
Suleyman A.M,, Aminuddin M.B. and Khor B.C. (2004). Effects of storage duration and
temperature of Moringa Oleifera stock solution on its perfomance in coagulation.
International Journal of Engineering and Technology, Vol. 1(2), pp. 146-151.
Kiéné L., Lu W., and Lévi Y. (1998). Relative Importance of the phenomena responsible for
chlorine decay in drinking water distribution systems. Wat. Sci. Tech., vol 38, No 6, pp.
219-227.
Kruesman G., and Ridder P.A. (1991). Analysis and Evaluation of Pumping Test Data,. ILRI
Publication 47, International Institute for Land Reclamation and Improvement, Wageningen,
The Netherlands.
Lawrence K. Wang, Hung Yung-Tse and Shammas Nazih K. (2007). Handbook of
Environmental Engineering, Vol. 3: Physicochemical Treatment Processes. The Humana
Press Inc. Totowa, NJ. Pp 103-127.
Lee M.D. and Bastemeijer T.F. (1991). Drinking Water Source Protection, A review of
environmental factors affecting community water supplies. IRC International Water and
Sanitation Centre The Hague, The Netherlands.
Leiknes T., Hallvard O., and Havard M. (2004). Removal of natural organic matter (NOM) in
drinking water treatment by coagulation-microfiltration using metal membranes. Journal of
Membrane Science, 274 (2004), p.p. 47-55.
44
Challenges and Opportunities for Safe Water Supply in Mozambique
ȱ
2008
Lewis W.J., Foster D., Read G.H. and Schertenleib R. (1981). The need for an integrated
approach to water-supply and sanitation in developing countries’. ‘Sci. Total Environ. 21
pp.53-59.
Mahwi, A.H., Moghaddam, M.A, Nasseri, S. and Naddafi, K. (2004). Performance of direct
Horizontal roughing filtration (DHRF) system in treatment of high turbid water. Iranian J.
Env. Health Sci Eng. Vol. 1, No 1, pp. 1-4.
McConnachie G.L., Folkard G.K., Mtawali M.A. and Sutherland J.P. (1999). Field trials of
Appropriate Hydraulic Flocculation Processes. Water Resources vol. 33, No. 6, pp. 14251434.
MISAU (2004). Regulamento sobre a Qualidade de Água para o Consumo Humano; Boletim da
República de Moçambique, I Série, Nr. 37, Diploma Ministerial n0 180/2004 de 15 de
Setembro (Guidelines for drinking Water quality in mozambique; Boletim of the Republic of
Mozambique, Series I, No 37, Ministerial decree 189/2005 of September 15). Ministry of
Health, Maputo, Mozambique.
Mishra K.K. and Breemen A.N. (1987). Communication on the Sanitary Engineering & Water
Management: Gravel bed flocculation. Delft University of Technology, Faculty of Civil
Engineering, Delft, The Netherlands, pp. 11-70.
Ndabigengesere, A., Narasiah, K.S. (1998). Quality of water treated by coagulation using
Moringa oleifera seeds. Water Research, Vol. 32 (3), pp. 781–791.
Ndabigengesere A, Narasiah K. S., Talbot B.G. (1995). Active agents and mechanism of
coagulation of turbid waters usng Moringa Oleifera. Water Research, Vol 29 (2), pp. 703710.
OECD (2007). Private sector participation in water and sanitation infrastructure- background
paper for the regional roundtable on strengthening Investment climate assessment in
NEPAD countries, Lusaka, Zambia.
Polasek P. and Mutl S., (2002). Cationic Polymers in Water Treatment-Part 2: Filterability of
CPF-formed suspension. Water SA Vol. 28 No 1, pp. 83-88.
Polasek P. and Mult S., (2005). Optimization of reaction conditions of particle aggregation in
water purification-back to basics. Water SA Vol. 31 No 1, pp. 61-72.
Polasek P. (2007). Differentiation between different kinds of mixing in water purification -back to
basics. Water SA Vol. 33 No 2, pp. 249-251.
Powel J.C., Hallam, N. B. Hallam, West J. R. Chistopher F. and Simms J. (2000). Factors which
control bulk chlorine decay rates. Wat. Sci. Tech., vol 34, No 1, pp. 117-126.
45
Challenges and Opportunities for Safe Water Supply in Mozambique
ȱ
2008
Qin J.J, Maung H.O., Iran A. K., Knops F., and Miller P. (2006). Impact of coagulation pH on
enhanced removal of natural organic matter in treatment of reservoir water. Separation
Purification Technology, 49 (2006), pp. 295-298.
Reweta W.S.J., Sampath R.K. (2000). Performance Evaluation of Urban Water Supply in
Tanzania: The Case of Dar Es Salaam City. Water Resources Development, Vol. 16, No.
3, pp. 407–423.
Ruehl. K. (1999). Easing the pain of meeting the D/DBP rule with enhanced coagulation. Water
Engineering & Management; 146, 1, ABI/INFORM Global pp. 22.
Smet, J.E.M and Visscher, J.T. (1990). Pre-treatment Technologies for community water supply:
An overview of techniques and present experience, 2nd ed. International Water and
Sanitation Centre, The Hague, The Netherlands, pp. 109-124.
SAL-CONSULTORES (2005): Adapting Regulation to the Needs of the Poor- Mozambique
Case Study. Evaluation Report. Technical report.
Sánchez, L.D., Sánchez, A., Galvis, G and Latorre, J. (2006). Multi-stage Filtration. International
Water and Sanitation Centre, The Hague, The Netherlands, pp. 11-18.
Schomoll O., Howard G., Chilton, J. and Chorus I. (2006). Protecting groundwater for Health.
WHO. IWA publishing, London. Seattle.
http://www.who.int/water_sanitation_health/publications/PGWsection1.pdf
SEURECA & HYDROCONSEIL (2005). Unaccounted for Water Feasibility Study for Maputo
Water Supply Scheme. FIPAG. Technical report.
Sharp E. L., Parsons S. A., and Jefferson B. (2006). Seasonal variations in natural organic
matter and its impact on coagulation in water treatment. Science of the Total Environment,
363, pp. 183-194.
Stumn W. and Morgan James J. (1996). Aquatic Chemistry: Chemical Equilibria and rates in
Natural water. John Wiley & Sons, Inc. 3rd edn. New York, USA, pp. 818-866.
Thompson T., Fawell J., Kunikane S., Jackson D., Appleyard S., Callan P., Bartan J. and
Kingson P. (2007). Chemical safety of drinking water: assessing priorities for risk
management. WHO, Geneva Switzerland.
Tokajian S., Hashwa F. (2003). Water quality problems associated with intermitent water supply.
Wat. Sci.Tech., vol 47, No 3: pp. 229-239.
Totsuka N., Trifunovic N., and Vairavammorth K. (2004). Intermittent urban water supply under
water starving situations. 30th WEDC International Conference, Vientnane, Lao PDR,
2004.
Velasco J. R., Uriate J. I. A., and Velasco U. I. (2007). Enhanced coagulation under changing
alkalinity-hardness conditions and its implications on trihalometane precursor removal and
46
Challenges and Opportunities for Safe Water Supply in Mozambique
ȱ
2008
relationship with UV absorbance. Separation Purification Technology, 49 (2006), pp. 295298.
Viraraghavan T. and Srinivasa P.T. (2004). Influence of Natural Organic matter (NOM) on the
speciation of aluminum during water treatment. Water, Air and Soil Pollution 152, pp. 3554.
Volk C, Kimberly B., Ibrahim E, Verges D, Amy G, and Lechevallier M. (2000). Impact of
Enhanced and Optimized coagulation on removal of organic matter and its biodegradable
fraction in drinking water. Water Research vol. 34 (12), pp. 3247-3257.
Vreeburg J.H.G, Boxall J.B. (2007). Discolouration in potable water distribution systems: A
review. Water Research, No 41, pp. 519-529.
Whittington, D (1992). Possible Adverse Effects of Increasing Block water tariffs in Developing
Countries. Economic Development and Cultural Change, Vol.41 (1), pp. 77-88. USA,
University of Chicago.
WHO (2004). Guidelines for Drinking Water Quality. Vol. 1, Recommendations. 3rd edn, WHO,
Geneva, Switzerland.
WHO/UNICEFF, (2006). Basic Needs and Right to Health-Part III: Challanges to life and well
being.
Zamxaka M, Pironcheva G and Muyima NYO (2004). Microbiological and physico-chemical
assessment of thequality of domestic water sources in selected rural communities of the
Eastern Cape Province, South Africa. Water S.A.vol. 30 (3), pp. 333-340.
Zamxaka M, Pironcheva G and Muyima NYO (2004). Bacterial community patterns of domestic
water sources in the Gogogo and Nkonkobe areas of the Eastern Cape Province, South
Africa. Water S.A vol. 30 (3), pp. 341-346.
Zandamela, H. (2002). Lessons from Mozambique-The Maputo Water Concession. Full detailed
report. http://www.kosa.org/english/privatisation.htm (10 January 2008).
47
Challenges and Opportunities for Safe Water Supply in Mozambique
Appendix. Papers
48
ȱ
2008
I
CHALLENGES FACING DRINKING WATER PRODUCTION IN MOZAMBIQUE- A
REVIEW OF CRITICAL FACTORS AFFECTING TREATMENT POSSIBILITIES
Matsinhe N. P*1,2. Dinis Juízo1, Kenneth Persson2, L.C Rietveld3
1
Eduardo Mondlane University, Faculty of Engineering, Av. de Moçambique km 1.5, C. Postal
257, Maputo, Mozambique, Phone:+258-21- 315161, (E-mail:
matsinhe@zebra.uem.mz;juizo@hotmail.com)
2
Lund University, Dept of Water Resources Engineering, P O Box 118, SE-221 00 Lund, Sweden,
Phone: +46-46- 222 41 65, (Email: Kenneth.persson@tvrl.lth.se)
3
Delft University of Technology- Dept of Water Management, 2628CN Delft, The Netherlands
Submitted for publication in the Journal Water Science and Technology
ABSTRACT
This paper discusses the drinking water production situation of Mozambique and critical factors
affecting treatment possibilities. The River water quality of six important rivers is analysed and
some of the critical factors affecting treatment possibilities identified. Treated water quality of
three major waterworks of the country is also analyzed and the results used to infer from the
overall situation of the country. High turbidity values occurring typically during rainy periods,
presence of bacteria, natural organic matter and, sometimes, presence of algae were identified as
water quality variables of major concern for drinking water production. Seasonal variations in
river water quality make drinking water production costly and technically demanding. Treatment
processes should therefore be designed for the worst water quality case. Technologically and
operationally related aspects also impact drinking water production. The situation is critical in
small to medium size water supplies where lack of financial capacity limits the possibilities for
investments in costly production methods. Methods of improving the situation are also discussed.
This includes pre-treatment at source and technical improvements through incorporation of low
cost treatment processes into already existing plants. If properly incorporated, these methods are
proven to increase overall production capacity at relatively low investment and operational costs.
Keywords: drinking water production, surface water; conventional treatment; roughing filtration;
coagulation -direct filtration
INTRODUCTION
Drinking water production in Mozambique depends mostly on river water. Most rivers are of
torrential regime, thus associated with seasonal flow and water quality variations. When used
for drinking water production surface water must be treated. Turbidity, suspended solids,
natural organic matter (NOM), and pathogenic bacteria are generally the quality variables of
major concern. Impurities resulting from human activities can also be present and impact the
raw and treated water quality and sometimes, the complexity of drinking water production
(WHO, 2004). Conventional treatment is the most widely used method of removing impurities
from surface water supplies. In its simplest form, it involves chemical coagulation,
flocculation, sedimentation and filtration followed by disinfection (Hansen, 1988; WHO
1
2004). Additional treatment may include pre-chlorination to assist the removal of algae and
dissolved organic matter and conditioning for alkalinisation and corrosion control.
Conventional treatment consisting of chemical coagulation, flocculation, sedimentation
filtration and chlorination is also used in Mozambique. Because of many reasons, however,
most drinking water treatment plants are presently incapable of producing treated water of
acceptable quality. This is particularly true in small to medium sized water supplies where the
combined effect of poor water quality at source, operational and logistic constraints and
limited financial capacity makes drinking water production far more ineffective.
This paper examines the overall situation of drinking water production in Mozambique and
critical aspects affecting treatment possibilities. The raw water quality of six important rivers
used for drinking water production is analyzed. Results of drinking water production from
three major waterworks of the country are analyzed and used to infer from the overall situation
of drinking water production in the country. Methods of improving the situation are also
discussed with emphasis put on small to medium size water supplies.
METHODS
Selected rivers
Six rivers were selected rivers for the study. These are, the Limpopo, Maputo and Umbelúzi
Rivers located in South of Mozambique, the Púngué River located in central Mozambique and
the Monapo and Licungo Rivers located north of Mozambique. The Umbelúzi, Púngué and
Monapo rivers are used for drinking water production of three largest cities of Mozambique
namely, Maputo, Beira and Nampula cities.
Selected waterworks
Results of treated water quality from the waterworks of Maputo, Beira and Nampula were
used in the study. These are the waterworks where records of water quality determinations
exist or could be assessed. The three waterworks are constructed based on conventional
methods.
Source of data and selection of water quality parameters
Data used in the study were obtained from three major sources namely historical data gathered
by the national water authority (DNA) for the Licungo and Umbelúzi rivers and results from
controlled studies on the Limpopo and Maputo rivers. Records of raw and treated water from
plant operators at Beira and Umbelúzi waterworks were also used. Field work conducted
between March and April 2008 was used to gather additional data on Púngué and Monapo
rivers.
Before use, the data obtained from DNA databases were checked for their analytical accuracy
through charge balance tests performed with the data on ionic species. These tests were used
to eliminate faulty or suspicious observations from the original data sets. A charge balance
error of ±15% was accepted due to the limited length of available data sets. Around 53% of
the data sets analyzed for the Licungo River and 24.5% from the Umbelúzi were rejected due
2
to unreliable figures of different water quality parameters and charge balance errors higher
than 25%.
Turbidity, hardness, pH and alkalinity, organic matter, indicator organisms, phosphate (related
to algae growth) were used to assess the overall condition of river water. To assess drinking
water treatment results, turbidity, pH, alkalinity and organic matter were used. Turbidity was
selected because it impacts treated water acceptability and the performance of unit operations
during treatment. Turbidity is also related to presence of microorganisms (WHO, 2004). The
water hardness was selected because it impacts water acceptability and water treatment. The
water pH and alkalinity were selected because they impact chemical processes (e.g.
coagulation, chlorination) during water treatment. The water alkalinity also helps assess the
degree of water corrosiveness following treatment.
Organic matter was selected because it influences the quality of treated water, especially the
levels of nutrients available for bacteria growth. Organic matter can also interfere with
treatment processes. The guideline value for treated water is 2.5 mgC/Ɛ measured as DOC
(WHO, 2004). Presence of bacteria was selected for obvious reasons of human health
protection and algae were selected because they impact treatment processes (e.g., filtration)
and have the potential to release toxins of health concern.
RESULTS AND DISCUSSION
Typical river water quality
Typical values/ranges of river water quality variables compiled from existing studies and data
bases of selected rivers are presented in Table 1. For comparison, the Mozambican guidelines
(MISAU, 2004) for drinking water are included.
Physic-chemical quality
For all six water sources analyzed, turbidity and total suspended solids concentrations are
generally found within ranges that exceed the guideline values for drinking water even when
they occur at their lowest concentrations. The river water turbidity also experience seasonal
variations typical of river courses of torrential regime. Figure 1, gives an example of typical
variations of river water flow and turbidity in four out of the six rivers analyzed. As seen, the
river water turbidity can vary from figures lower than 10 NTU during the dry season, to
figures above 300 NTU during rainy periods.
The relatively low values of maximum turbidity at Umbeluzi and Limpopo (site 1) rivers are
due to the sampling sites being located downstream major hydraulic works (man-made
reservoirs) where most solids and turbidity is reduced due to sedimentation. Analysis of data
shown in Table 1 suggests a similar pattern with respect to suspended solids concentrations.
Data used in this respect is from Limpopo, Púngué and Umbelúzi Rivers.
3
Montly average flows (m3/s) in some of the river courses analyzed
Maximum recorded turbidity values at selected sites of some of the rivers
analysed
1400
1200
Umbelúzi
Limpopo
Púngué
Licungo
350,0
1000
300,0
T u r b id ity (N T U )
m o n th ly a v e r a g e flo w (m 3 /s )
400,0
800
600
400
Oct
Feb
Jun
Nov
Mar
Jul
Dec
Apr
Aug
Jan
May
Sep
Umbeluzi site 1 Umbeluzi-site 2
Licungo
250,0
200,0
150,0
100,0
200
50,0
0
0,0
Oct
Nov
Dec
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Limpopo-site 1 Limpopo site 2
Púngue
Figure 1 Average monthly flows (m3/s) and maximum values of water turbidity (NTU) recorded at selected
sites of rivers used for drinking water production of some cities in Mozambique such as the city of
Maputo (Umbelúzi River), Beira (Púngué River) and Mocuba (Licungo River). Data source: DNA
database.
Physical parameters such as water pH and temperature do not affect water acceptability, and
are not of concern regarding health, but they have a major influence on water quality
parameters such as bacteria and algae growth. Average river water temperatures in
Mozambique are generally between 20 qC and 25 qC and the raw water pH generally between
6.0 and 9.0. Because high temperatures are generally accompanied by heavy rainfall and high
river water turbidity, the combined effect of these factors provides ideal conditions for
bacterial growth, algae blooming and the release of taste- and odour-forming compounds
(Zamxaka et al., 2004; Chorus & Bartram, 1999). Added to this natural climatic effect is the
enhanced rate of nutrient input that accompanies the growth of towns and the development of
irrigated agriculture in the catchment areas around the water courses.
Chemical properties such as water hardness and alkalinity generally affect water acceptability
and water treatment efficiency. Problems with hard or soft water are generally site specific
because they depend on the interaction of many factors including the soils and rocks from
which the water is derived, which are generally site or region specific (Stumn & Morgan,
1996; Lawrence et al., 2007). The water alkalinity is closely related to the concentration of
carbonates, bicarbonates and hydroxide ions in the water, therefore it is closely related to
water hardness. It influences, for instance, scale formation in the case of hard water and the
degree of water corrosiveness of soft water (Polasek & Mult, 2005; Polasek, 2007; Velasco et
al., 2007). During water treatment, the pH and water alkalinity affect processes such as
coagulation and disinfection. Knowledge of the water’s pH and alkalinity is therefore useful in
evaluating the optimum conditions for treatment processes, as well as the final condition of
treated water with respect to its corrosive or scaling properties.
From the results in Table 1, it is clear that the river water quality can be change from
conditions of rather soft to moderately hard (Total hardness between 10 and 100 mg/Ɛ) in
some rivers (e.g. Licungo, Maputo and Púngué) while, in others (e.g. Limpopo and Umbelúzi)
it varies from conditions of slightly hard to hard (100> total hardness < 440 mg/Ɛ). Therefore,
4
the type and magnitude of river water quality problems with respect to these quality variables
is site specific. River water quality is however within acceptable limits for drinking water
production.
Analysis of river water quality with respect to presence of organic matter could only be done
for the Licungo, Limpopo and Umbelúzi Rivers. As seen in Table 1, levels of organic matter
in river water are generally higher than the guideline value of 2.5 mgC/Ɛ of the guidelines
(MISAU, 2004). Also seen in Table 1 is that the river water quality also experience seasonal
variations in respect to this quality variable. These were attributed to variations in river flow
and water temperature. Yet, observed levels are not indicative of organic pollution of
anthropogenic origin.
5
Bioch.O. Demand
Dissolved Organic C
Total organic Carbon
pH
Temperature
Total dissolved solids
Total suspended solids
Turbidity
Total Hardness
Total Alkalinity
E. Conductivity
Manganese
Total iron
Nitrate
Nitrite
Total Phosphorous
Total coliforms
E.Coli
Faecal coliforms
TDS (mg/l )
TSS (mg/l )
NTU
mgCaCO3/l
mgCaCO3/l
EC (ms/m)
Mn (mg/l )
Fe (mg/l )
NO 3 (mg/l )
NO 2 (mg/l )
(mgP/l )
cfu/100 ml
cfu/100 ml
cfu/100 ml
TOC
(mgC/l )
(mg/l
(mgC/l )
0C
-
Reference
0.5-7.8
-
5.3-8.1
22-26.5
5.5-234
8.0-30.0
20-120
120-410
0.01-2.86
0.2-0.33
0.02-0.32
-
Licungo
(intake)
0.5-3.6
0.7-10.1
0.7-9.2
6.7-9.7
19 -32.0
149-1930
7.3-384.0
5.8-308.0
107-446
26-337
0.05-0.36
0.1-14.6
0.03
0.03
0.05-2.0
66-250
-
Limpopo
(site 1)
0.5-2.5
0.7-8.8
0.7-15.2
6.7-9.3
22-32.8
123-1052
3.6-288
10.1-362
108-340
25-185
0.02-0.31
0.14-12.5
0.03
0.03
0.03-1.98
32-870
-
Limpopo
(site 2)
6
-
-
8.2
16.6 (7.0)
115.6 (101.5)
24.8 (7.9)
5.8 (4.4)
0.06 (0.06)
-
Maputo
(site 1)
-
-
8.0
18.2 (19.5)
95.5 (27.8)
39.0 (11.5)
6.9 (3.9)
0.05 (0.06)
-
Maputo
(site 2)
Typical value/range
-
-
6.0-8.5
18-35
1.7-106
5.0-33.0
16-227
24-43
40-88
0.02-1.07
0.03-4.16
0.03-0.24
-
Púngue
(intake)
2.0-6.4
-
6.7-8.7
21-26.5
370-410
3.8 -173.0
137-165.0
139-160.0
550-630.0
0.01-0.03
0.04-0.18
0.01-0.5
0.01-0.03
0.01-0.63
>1000
36-350
18-300
Umbeluzi
(intake)
-
2.5
6.5-8.5
1000
5
500
1500
0.1
0.3
50
3
1-10
None
None
Mozambican
guidelines
(Misau 2004)
Typical values of river water quality in Mozambique and drinking water standard requirements (MISAU, 2004). For Maputo river
mean values of data sets used are presented with standard deviations values indicated between brackets.
Water quality
constituent
Table 1
Microbial quality
Data on microbial quality of river water could only be assessed for Limpopo and Umbelúzi
rivers. Average monthly values of total coliforms and faecal bacteria are presented in Table 1
for the Umbelúzi river, whereas for the Limpopo River, results of E.Coli determinations are
presented. The data regarding faecal bacteria counts in Umbelúzi river water was further
plotted against turbidity and time (Figure 2). As seen from Table 1 and Figure 2, presence of
faecal bacteria in river water is constant and generally high during the whole year. The data on
presence of faecal bacteria in river water were also found to show good correlation to river
water turbidity (0.5 < R < 0.8) and less with water temperature (0.3 < R < 0.9).
Relationship bacteria counts and river water turbidity
400
1000
E. Coli Limpopo-site 1
E. Coli Limpopo-site 2
Feacal Coliforms umbeluzi-intake
800
Feacal coliforms
350
Bacteria counts (cfu/ 100 mƐ)
Bacteria counts (cfu/ 100 mƐ)
900
Annual distribution of average monthly values of total and feacal
coliforms counts at the intake of Umbelúzi water works
700
600
500
400
300
200
Total coliforms (x10)
300
250
200
150
100
50
100
0
0
0
Figure 2
5
10
15
20
25
30
35
Turbidity (NTU)
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Dec
Monthly records of data related to contamination with bacteria at the intake of the Umbeluzi River
and relationship with river water turbidity (left). Also illustrated is the pattern concerning incidence
of bacteria in river water during the year (right).
Biological characteristics
Presence of algae and aquatic plants were the parameters used to assess the biological quality
of river water. Studies to investigate presence of algae in river water could only be found for
the Umbeluzi and Limpopo rivers. The two studies assessed during this work involved the
Umbeluzi (Bojcevska & Jergil, 2003; Gustafsson & Johansson (2006)) and the Limpopo
rivers. The study from Bojcevska & Jergil, (2003), focused on the presence and removal of
cyanobacterial toxins in water taken from an impounding reservoir supplied by the Umbelúzi
River. The study from Gustafsson & Johansson (2006), focused on presence of nutrients,
among which phosphorus, as an indicator of the potential for river water eutrophication. In the
study on the Limpopo River the water quality was assessed for presence of phosphorous.
Some results of measurements of levels phosphorus in river water taken by DNA at irregular
intervals were also used.
Average phosphorous concentrations as measured by Gustafsson & Johansson (2006) in the
Umbelúzi River are of about 0.054 mgP/L whereas; the data from DNA database indicates
figures in the range 0.01-0.63 mgP/L. The study on the Limpopo River provided values of
7
phosphorous in the river water in the range 0.03 mgP/L-2.0 mgP/L (see Table 1). Different
references found in literature and proposed by different organizations and authors suggest
limits for total phosphorous concentrations in the range 0.1-0.16 mgP/L (Table 2) to avoid
euthrophication of river water.
Table 2 Limiting concentrations of Phosphorous and nitrogen to reduce risk of river water
euthrophication. Limits suggested by different organizations and authors (* all fresh
waters; ** Total P = 0.3262 PO43ā (mg/L).
Constituent
Phosphorous
Tot. P (rivers draining to lakes- U.S.EPA)
Tot. P (rivers not draining to lakes- U.S.EPA)
Phosphate (Fytianos et al., 2002)
Inorganic P (SAWQG)*
Nitrogen
Inorganic N (SAWQG)**
Limit
0.05 mgP/L
0.1 mgP/L
0.5 mgPO43ā/L = 0.16 mgP/L
0.005 mgP/L
0.5 mgN/L
As seen from Table 1, phosphorous concentrations in the Umbelúzi River could reach values
higher than the indicative limit of 0.16 mgP/L. This suggests that euthrophic conditions might
occur from time to time in the Umbelúzi River. The study from Bojcevska & Jergil (2006)
gave evidences of presence of toxic cyanobacteria present in the reservoir water.
The study from Gustafsson & Johansson (2006) provided lower values. However, this study
had a sampling period covering a shorter period of analysis (from Sept. to Nov.) which can
explain the difference in the results. Average values measured by Gustafsson & Johansson
were however close to or exceeded the critical value proposed in Table 2 for rivers draining
into lakes as is the Umbelúzi River which support the statement made before that euthrophic
conditions might occur frequently in the river. Problems with algae at the Umbelúzi River
were been reported in a past study by Couto (2004) and more recently by managers of the
waterworks of Maputo water supply who report frequent outbreaks of algae at the intake
works.
The results for the Limpopo River are similar to the results on the Umbelúzi river (Tot.
phosphorous concentrations in the range 0.03-2.0 mgP/L). Two sampling locations were used
for this analysis. Given the origin of the problem, similar conditions are expected in many
other rivers used for drinking water production, particularly those where intensively irrigated
agriculture is practiced upstream the major intake works.
Typical design of drinking water treatment in Mozambique
In Table 3, a summary is given of the most common design of drinking water treatment plants
in Mozambique. As seen, most water supplies rely on surface water for drinking water
production and conventional methods for water treatment. Most treatment plants are built with
pre-chlorination, chemical coagulation, flocculation, sedimentation, rapid sand filtration and
disinfection with chlorine. Modifications to the basic design comprehend the use of package
units incorporating all physical processes (flocculation, sedimentation and filtration) in one
8
single unit and the used of aeration and direct filtration processes mainly for ground water
supplies.
Table 3 Typical design of drinking water treatment plants of 10 major cities of Mozambique
Town
Estimated
water demand
by 2008
(m3/day)1
Type of raw water
source
Type of treatment Operating
capacity
(m3/day)2
Conventional with
pressure filters
Conventional with
cascade aeration +
direct filtration
Conventionalstandard3
Conventional with
pressure filters
Slow sand filtration
Conventional with
cascade aeration +
direct filtration
Conventionalstandard with
lamella type settling
basins
Disinfection with
granular chlorine
Conventional
standard +
pressure filters
Conventional
standard
Lichinga
2 947
SW+ large reservoir
with selective intake
Pemba
9 487
GW–borehole field
Nampula
14 762
Nacala
6 568
Quelimane
13 320
SW+ large reservoir
with constant intake
SW+ large reservoir
with constant intake
GW–borehole field
Tete
7 116
GW–borehole field
Beira
30 000
SW–direct intake
from a sugar cane
irrigation channel
Xai-Xai
8 600
GW–dune aquifer
Mocuba
2 279
SW–direct intake
from river
240 000
SW–direct intake
from river
Maputo-Matola
2 160
10 000
13 500
5 700
4 200
5 100
19 200
6 000
1 000
172 800
1
According to estimates from DNA.
design capacity according to plant operators.
3 conventional standard = pre-chlorination, chemical coagulation, flocculation, sedimentation, rapid sand filtration
and disinfection.
2
As it is seen from Table 3 most waterworks are operated below or very close to required
demands. The only exception in the Table, are the waterworks for Pemba.
Typical Results of drinking water treatment
Water quality
In Table 4 an example is given of treated water quality from the waterworks of Maputo, Beira
and Nampula. The data used is based on annual records from year 2004.
9
Table 4 Results of water treatment at three waterworks in Mozambique (Rwater = raw water;
Twater= treated water; -, not measured, Min. = minimum recorded; Max. = maximum
recorded; STD = standard deviation)
Quality
variable
Turbidity
(NTU)
Org. matter
(mgO2/L)
pH
Res. Cl(mg/L)
Alkalinity
(mg/L)
Umbeluzi waterworks
Rwater
Twater
Rwater
Twater
Rwater
Min.
2.9
0.7
2.0
Max.
80.6
7.3
6.4
Mean
6.7
3.1
4.0
STD
6.6
0.9
0.8
Nr. Of records
359
358
351
Beira waterworks-old treatment plant
0.8
4.8
2.7
0.7
350
6.3
8.1
7.4
0.20
357
-
-
2.3
6.5
4.2
1.1
9
4.7
6.4
5.6
0.7
8
6.5
16.8
10.9
3.3
10
4.0
6.0
4.8
0.7
10
Min.
13.5
0.2
Max.
326 28.0
Mean
49.9
2.5
STD
31.1
3.3
Nr. Of records
310
303
Beira waterworks-new treatment plant
Min.
11.8
Max.
25.9
Mean
17.0
STD
4.6
Nr. Of records
9
Nampula waterworks
Min.
Max.
Mean
STD
Nr. Of records
11.7
47.0
25.0
13.8
10
Twater
Rwater
Twater
Twater
6.5
7.9
7.4
0.2
358
98
240
160
14.2
353
41
230
152
18.7
351
0.3
2.8
1.4
0.4
333
6.4
7.6
6.8
0.1
309
6.2
7.6
6.7
0.2
307
-
-
-
2.0
3.2
2.81
0.4
8
6.4
6.9
6.6
0.2
9
6.1
6.3
6.2
9
58
78
64.9
7.4
9
52
82
61
9.6
9
-
3.2
4.6
2.9
0.5
10
7.2
7.9
7.4
0.3
10
6.6
7.8
7.2
0.4
10
26.9
38.4
31.3
3.1
10
25.0
57.6
39.3
9.6
10
-
As seen in the Table, treated water quality not always conform to the guidelines (WHO)
particularly in relation to turbidity and organic matter concentrations. The mean values of
treated water turbidity of Maputo and Beira waterworks are generally below the tolerable limit
of 5 NTU, but the absolute limit of 1 NTU of the guidelines is hardly attained. Treated water
turbidity from Nampula waterworks, has always been recorded higher than the absolute limit of
5 NTU. Mean values of organic matter in the treated water were, in all cases, reported higher
than the target limit of 2.5 mg/l. When looking at results from the Umbelúzi waterworks, it is
clear that the treated water is frequently over chlorinated. Other chemical parameters such as
treated water pH and alkalinity were generally found within acceptable limits for corrosion
control (pH > 6.5; alkalinity > 61 mg/Ɛ) at Umbelúzi waterworks and slightly out of the range
in Beira and Nampula.
10
Operation of drinking water treatment
Table 5 gives an example of operational data of chemical processes at the Umbelúzi
waterworks. The data in Table 5 comprehend the range and average dosages of chemicals
applied and the average number of days in a month when chemical treatment was not
performed. As can be seen, chemical processes within this plant are frequently discontinued.
Reasons pointed for that are: the malfunctioning of dosing equipment or the lack of chemicals.
In Figure 3 an example of annual (2004) records of operation data on chemical coagulation at
Umbelúzi waterworks is presented from where it is seen that the process was performed with
reagent dosages (alum and polymer) that vary considerably within the same range of raw
water turbidity and alkalinity. An almost linear relationship is observed between raw water
turbidity and chemical reagent dosages which suggest that, the selection of reagent dosages
was based on a linear relationship with the river water turbidity.
Treatment with and Alum and flocculant
at umbeluzi water works
90.0
50.0
70.0
A lu m d o se (m g /l)
A lu m an d flo c d o sag e (m g /l)
80.0
Treatment with alum dose Vs raw water alkalinity
at Umbelúzi water works
60.0
ultrafloc
Alum
60.0
50.0
40.0
30.0
40.0
30.0
20.0
20.0
10.0
10.0
0.0
0.0
0.0
10.0
20.0
30.0
40.0
50.0
60.0
70.0
80.0
90.0
90
Raw w ater turbidity(NTU)
Figure 3:
120
150
180
210
240
Raw w ater alkalinity (mg/l)
Treatment with aluminium and coagulant aids at the Umbelúzi waterworks and its relation to raw
water turbidity (NTU) and alkalinity (mg/L). Data refers to one full year (2005) operation of
chemical coagulation at the treatment works. Source of data: FIPAG (2005).
From the data shown in figure 3 it is clear that alum doses were somewhat unrelated to both
raw water turbidity and alkalinity and also that, they varied significantly within the same range
of raw water quality values. The raw water pH and alkalinity had also varied considerably
during the period of analysis, from conditions of rather soft (Total alkalinity < 100 mg/Ɛ) to
slightly alkaline (Total alkalinity > 240 mg/l).
11
Chemical coagulation
Alum dosage (mg/l)
Average
Range
No. days disruption
Polymer dosage (mg/l)
Average
Range
No. days disruption
Pre-chlorination
HTH dosage (mg/l)
Average
Range
No. days disruption
Post-alkalinisation
Lime dosage (mg/l)
Average
Range
No. days disruption
Chemical Process
17.5
4.1-57.7
2
1.5
1.2-2.5
18
2.8
0.8-3.6
8
3.9
3.3-5.3
24
1.7
1.2-2.8
22
2.8
1.5-3.8
0
3.6
3.4-3.9
25
Feb.
22.3
6.4-70.0
4
Jan.
3.7
3.5-4.1
23
0.8
0.3-1.6
17
0.5
0.4-1.0
0
12.6
7.8-53.2
0
Mar.
2.7
1.4-4.5
17
1.4
0.3-2.7
18
0.6
0.3-1.1
0
19.5
11.2-42.4
0
Apr.
12
1.9
1.4-4.1
10
0.8
0.3-1.6
21
0.5
0.5-0.6
0
8.7
7.6-11.8
0
May
1.7
1.4-2.9
13
0.7
0.3-1.6
22
0.5
0.3-0.9
0
8.0
7.1-10.2
0
June
1.7
0.7-3.7
16
1.1
0.3-1.8
21
0.5
0.3-0.6
0
10.2
8.1-12.4
0
July
1.4
1.4-1.5
23
1.1
0.3-2.2
22
0.5
0.4-0.6
0
10.0
8.1-11.4
0
Aug.
1.7
1.4-2.7
16
0.8
0.3-1.4
26
0.5
0.2-0.6
0
10.0
7.4-11.8
0
Set.
2.0
1.4-4.8
20
0.5
0.4-0.5
29
0.5
0.3-0.8
0
11.2
9.9-24.4
0
Oct.
1.4
1.4-1.5
19
30
0.3
0.1-0.5
0
8.4
4.8-10.4
0
Nov.
Table 5 Range and average values of dosages applied for chemical treatment at Umbelúzi water works. Also indicated, is the average
number of days in a month during which chemical treatment was disrupted due to lack of supplies or mal-functioning of
equipment.
30
30
0
4.8-10.4
0
Dec.
Consequences for drinking water production
Treatment possibilities with conventional treatment
Treatment possibilities by conventional treatment depend on a number of factors
including:
x
x
x
x
the concentration of the chemical in the raw water;
control measures employed throughout the drinking-water system;
nature of the raw water; and
treatment processes already installed.
If a guideline value cannot be met with the existing system, then additional treatment may
need to be considered, or water should be obtained from alternative sources. From the six
water sources studied, it is clear that river water is generally of inferior quality, thus
unsuitable for consumption without prior treatment. Table 6, adapted from literature
(WHO, 2004) summarizes the treatment achievability by conventional treatment built
according to the standard design from where it is seen that with the commonly found
designs of drinking water treatment plants in Mozambique, conditions exist for the
production of excellent treated water quality from existing surface water sources.
Table 6 Treatment achievability by conventional treatment a,b
Unit process
Typical raw water
quality variable
Turbidity & solids
Faecal Bacteriac
Natural O. matter
Algae cells
Cyan bacteria cells
Algae derived toxins
Irond
Manganesed
Drawbacks
Pre-chlorination
Coagulation/Flocculation
Sedimentation
Rapid Sand
Filtration
Chlorination
No eff.
>80%
50% or more
No eff.
No eff.
Lim. eff
No eff.
No eff.
>80%
Lim. eff
50% or more
50% or more
>80%
No eff.
50% or more
50% or more
>80%
Lim. eff
No eff.
Lim. eff
No eff.
No eff.
>80%
>80%
No eff.
>99.9%
No eff.
No eff.
No eff.
>80%
No eff.
>80% (<0.05)
Insoluble
Al,
turbidity
breakthrough if inefficient,
NOM and extraneous organics
impacts efficiency
Performance related to
efficiency of previous
processes
DBPs,
DBPs
Adapted from WHO, 2004
b The table includes only those quality variables of concern for drinking water production in Mozambique.
c 3 log removal (>99.9% removal) by rapid filtration with coagulation and sedimentation:
d pre-oxidation or aeration required:
As seen from the table, treatment processes arranged according to the design of drinking
water treatment plants in Mozambique can lead to the production of treated water of
excellent quality from existing surface water sources. It is also learn from the table that, if
properly designed and operated, chemical coagulation, flocculation and sedimentation are
the unit operations of the standard design of conventional treatment that does the bulk of
the removal of impurities such as turbidity and solids, organic matter, algae and
cyanobacteria cells. These are raw water quality variables identified as having a
13
significant impact on drinking water production in Mozambique. Critical processes in the
treatment chain are coagulation-flocculation. Poor performance of these processes results
generally in poor performance of subsequent stages (sedimentation and filtration) and
thus, of overall drinking water production. Effective operation of coagulation and
flocculation is therefore essential to meet desired standards of drinking water treatment.
Factors affecting treatment possibilities
Quality variables of major concern for drinking water production were identified as
turbidity and suspended solids, dissolved organic matter, bacteria and in some cases,
presence of algae (Table 1). Iron and sometimes manganese may also be present and pose
some challenges for drinking water production. It is also seen from Figure 1 that most
surface water sources experience seasonal flow variations that impacts both the physicchemical and microbiological quality of river water.
Because most turbidity causing particles have their stability defined by surface charge,
their removal from water is best attained with coagulation processes based on chargeneutralization mechanisms (Polasek, 2005; Stumn & Morgan, Lawrence et al., 2007).
When the raw water turbidity is high the optimum pH-range for coagulation is between
4.5 and 5.5 but when the raw water turbidity is low, the unavailability of high
concentration of particles to promote inter-particle contacts requires the formation of
amorphous hydroxides onto which the fine particles will enmesh (seep coagulation). The
optimum pH-range for coagulation in this case is between 5.5 to or 6.0 or even higher
(Polasek & Mult, 2005).
Most constituent of organic matter are of an acidic character, therefore their removal by
destabilization-aggregation processes should also be performed at an acidic pH-range,
preferably between 4 and 6.5 (Polasek & Mult, 2005; Sharp et al., 2005; Qin et al., 2006).
Optimum coagulation conditions for removal of natural organic matter-NOM are,
therefore, similar to those required for low turbid waters however, at higher reagent
dosages and lower pH-values (Velasco et al., 2007; Volk et al., 2000; Ruehl, 1999;
Eikebrook, 1999).
Algae generally behave like turbidity therefore their removal from water can be made
effective through destabilizing-aggregation mechanisms similar to those applied for
removal of turbidity (Polasek & Mult, 2005). At low concentrations, algae can be
removed by interception onto precipitates of the destabilizing reagent i.e. at an optimized
pH-reaction for coagulation between 5.5 and 6.5. At high concentrations, immobilization
or even destruction by means of pre-oxidation is required.
The effective purification of water takes place at a particular pH which depends on the
nature of impurities to be removed and the particular condition of raw water with respect
to its pH and alkalinity. Therefore, it is necessary to establish the optimised pH value at
which all kinds of impurities present in the water will be removed with the highest
attainable efficiency. Looking at data presented in Table 1, it is seen that the river water
pH and alkalinity can vary significantly, suggesting that, careful adjustment of the water
pH is required to maintain effective operation of coagulation-flocculation processes.
14
The addition of chemical coagulants usually alters the water pH and alkalinity, however,
the extent to which the optimum pH-reaction for effective coagulation is attained depends
on various factors among which the raw water natural alkalinity and the amount of
chemical reagent applied. When the raw water turbidity is high and the water pH is just
above neutral this can be easily be accomplished thanks to the high amounts of reagent
usually required however, when the raw water turbidity is low, the optimum pH-reaction
may, in many occasions, be unattained due to the buffer effect of alkalinity water.
The two factors that significantly influence the overall efficiency of coagulation in water
treatment are, therefore the dosage of the destabilizing reagent and the reaction pH. When
the purification process is aimed at removing a single pollutant, for instance turbidity,
then a dosage at which maximum turbidity removal is attainable, is the optimum dosage
for operation of the process. This is the common practice used in Mozambique for
selection of operational dosages. When, on the other hand the purification process is
aimed at removing a mixture of different impurities (e.g. when removal of organic matter
and algae is also required), then the purification process must take place under optimised
reaction conditions at which residual concentration of each contaminant is below its
permissible limit.
As noted in the previous discussion, the optimum range for each impurity requires
different dosages and different reaction pH values which means that the optimized
operational dosage should be selected from within ranges that are coincident for all
individual contaminants. This requires from plant operators not only a clear knowledge of
the condition of the raw water but also enough skills because, if incorrectly done, the
removal of some of impurities of the water may be negatively affected. For example,
during periods of low turbidity the optimum dosage for operation of coagulation can be
easily adjusted to incorporate also removal of organic matter without compromising
significantly turbidity removal, however, during periods of high turbidity and particularly
when the river water alkalinity is low, the operation of coagulation based on optimum
coagulant dosages for removal of NOM (sweep coagulation) mean that the efficiency in
turbidity removal may be adversely affected.
Looking at the data presented in figure 3, it is clear that the operation of coagulation is
often performed with reagent dosages that vary considerably from within the same range
of turbidity and pH-alkalinity in the raw water. This suggests, that very often, the
optimised pH-reaction for effective operation of coagulation is not attained. This, impacts
overall coagulation efficiency and consequently the performance of subsequent
processes. Looking at data presented in Tables 5, it is also clear that, very often
coagulation is discontinued due, to lack of chemicals or mal-functioning of equipments
which also undermine effective drinking water treatment. The consequences for drinking
water production are; ineffective operation of treatment processes and the production of
treated water of inferior quality.
The data presented in Table 4 is very illustrative of the situation. Although the treated
water turbidity is generally below the tolerable limit of 5 NTU, the absolute limit of
1 NTU recommended in the guidelines is seldom attained. It is also seen from Table 4,
15
that the treated water is generally over chlorinated. This combined to high levels of
organic matter in the treated water means that, taste and odour forming compounds as
well as disinfection-by-products-BDPs may develop following treatment. Also, presence
of high levels of organic matter in the treated water may increase the potential for human
health risks due to bacteria growth in distribution and storage facilities.
Methods of improving the situation
The main challenges facing drinking water production in Mozambique have been
identified as; poor water quality at source and frequent variations and the need for rather
complex and costly operations to maintain standards of drinking water production.
Improvements to the situation require therefore, improvements in the quality of raw water
and in the design and operation of drinking water treatment processes. Various
possibilities exist to attain these objectives which can be summarized as depicted in
Figure 4 below.
Nature of the
problem
Possible area of
intervention
Technical or managerial
possibility
ExpectedOutcome
Artificial water improudement,
pre-treatment at source, ground
water
Ineficient operation
of treatment
processes, lack of
installations, high
dependency on
imported supplies,
complexity of
treatment
processes, lack of
skills
Improved source
water selection,
improved
watershed
managment,
Treatment with
low treatment
methods, design
and operational
modifications into
existing
installations,
treatability
studies, training
Treatment at source (riverbank
filtration, river bed filtration ,
dinamic roughing filters)
Treatment with Multi-stage
filtration (slow sand filtration
with roughing filtration for pretreatment)
Treatment with adapted
conventional methods (direct
filtration, hydraulic flocculation
with up-flow roughing filters),
treatment with Moringa
Less polutted
water, smaller
variations in raw
water quality
Simpler treatment
methods, better
knoweledge of
treatment processes,
less dependence on
imported supplies,
most reliable
treatment, less costly
Improved drinking water production.
Poor quality of
water at source,
seasonal
variations
Figure 4: Process approach for sustainable improvement of drinking water production in
Mozambique.
Because turbidity and bacteria are partially reduced in large reservoirs, abstraction of raw
water from large reservoirs (natural or man-made) is one feasible option to get less
polluted water for drinking water production. Large reservoirs suffer, however, the threat
of eutrophication due to high temperatures. Care should therefore be taken when opting
for this solution to minimize risks of increasing production costs and complexity of water
treatment. As a rule, water containing organic contamination resulting from algae
propagation is more difficult to treat with traditional methods of conventional treatment.
When required flows are small (e.g. small to medium size water supplies), river bank
16
infiltration also provides a feasible alternative to get less polluted water for drinking
water production.
Improvements or modifications to existing designs should be aimed at two major
objectives: (i) smooth variations in raw water quality and (ii) reduce inefficient
operations within treatment plant designs. Examples of methods technically suitable are;
roughing filtration and coagulation-direct filtration processes. Extensive research
conducted on this methods (Smet & Visscher, 1990; Ingallinella et al., 1998; Bauer et al.,
1998) have proven that if properly incorporated into existing plants they can improve
treatment efficiency and production capacity at relatively low investment and operational
costs. Coagulation-direct filtration processes are particularly advantageous because they
combine particle aggregation and separation mechanisms in one single unit.
CONCLUSIONS
Turbidity, bacteria, natural organic matter, and sometimes algae and aquatic plants were
identified as river water quality variables of major concern for drinking water production.
These impurities often occur in high concentrations and experience large seasonal
variations that make treatment methods costly complex and technically demanding.
Conventional treatment consisting of chemical coagulation, flocculation, sedimentation
filtration and chlorination is mostly used in Mozambique. While suited for production of
high quality drinkable water from surface water sources, in Mozambique, poor water
quality at source, compounded by seasonal variations in river water quality and poor or
inadequate operation of treatment processes leads to the production of treated water that
often fail to comply to required standards.
The first place where drinking water quality standards have to be satisfied is at the outlet
of the treatment plants. However, in a large number of treatment plants of Mozambique
(inclusive Maputo waterworks) these standards are not fully achieved. Factors
contributing to that includes: (i) inadequacy of existing infrastructure and operational
procedures for water treatment (ii) inconsistent operation of essential chemical processes
due to mal-functioning of equipments or temporary lack of supplies and, (iii) lack of
guidelines for specific contaminants (e.g. DBPs) which makes water treatment practices
to target only those variables for which, guidelines have been established.
Methods of improving the situation require either improvements in the quality of raw
water or improvements in the design and operation of treatment processes. Though linked
to potential problems with algae, abstraction of raw water from large reservoirs is an
option technically suited to get less polluted water for drinking water production. When
required demands are small, river bank infiltration also helps reduce loads of
contaminants carried with the raw water.
Improvements/modifications to the existing design of the water treatment plants with the
view of smoothening variations in raw water quality can also help minimize inefficient
operations within treatment plants. Roughing filters, and coagulation-direct filtration
17
processes are all technical options potentially suited to improve treatment efficiency and
increase overall production capacity at relatively low investment and operational costs.
ACKNOWLEDGEMENTS
This paper contains results of a study conducted under a wider research program funded by the
Swedish International Development Cooperation Agency (SIDA/SAREC) jointly implemented by
Eduardo Mondlane University and the Lund Institute of Technology of Lund University-Sweden.
The authors wish to acknowledge the valuable support of the grant agency and to thanks all
people and institutions that provided support, data and information used in the study.
REFERENCES
B. Eikebrokk (1999): Coagulation-direct filtration of soft, low alkalinity humic waters. Wat.Sci. Tech. Vol.
40, N0. 9, pp. 55-62
Chorus, I. And Bartram, J. (1999). Toxic Cyanobacteria in Water. A guide to their Public Health
Consequences, Monitoring and Management. World Health Organization. London, E & FN Spon.
Chuang, C. J., and and K.Y. Li (1997). Effect of coagulant dose and grain size on the performance of
direct filtration. Separation and Purification Technology 12 (1997) 229-241.
Hansen, Singurd P. (1988). Filtration Technologies for Water Treatment. Water Engineering & Management,
Aug. 1988, 135, 8. ABI/Inform Global, pg. 16
Ingallinelia Ana Maria, Stecca L. M., and Wegelin, M. (1998). Up-flow Roughing Filtration: Rehabilitation
of a water treatment plant in Tarata, Bolívia. Wat.Sci.Tech. vol. 37 (9), pp 105-112
Juizo, D. (2008). Methods for Transboundary Water resources Management in Water Stressed Regions- A
case study: Southern Africa. Ph.D Dissertation, Department of Water Resources Engineering, Lund
University, Sweden.
Karen Ruehl. (1999). Easing the pain of meeting the D/DBP rule with enhanced coagulation. Water
Engineering & Management; Jan 1999; 146, 1; ABI/INFORM Global pg. 22.
Lawrence K. Wang, Hung Yung-Tse and Shammas Nazih K. (2007). Handbook of Environmental
Engineering, Vol. 3: Physicochemical Treatment Processes. The Humana Press Inc. Totowa, NJ. Pp
103-127.
Leiknes T., Hallvard O., and Havard M. (2004). Removal of natural organic matter (NOM) in drinking
water treatment by coagulation-microfiltration using metal membranes. Journal of Membrane Science,
274 (2004), p.p. 47-55
MISAU (2004). Regulamento sobre a Qualidade de Água para o Consumo Humano; Boletim da República
de Moçambique, I Série, Nr. 37, Diploma Ministerial n0 180/2004 de 15 de Setembro (Guidelines for
drinking Water quality in mozambique; Boletim of the Republic of Mozambique, Series I, No 37, Ministerial decree
189/2005 of September 15). Ministry of Health, Maputo, Mozambique
Mussagy, A., and Cronberg G. Effects of nutrient addition on the freswater plankton community in the
Pequenos Libombos Reservoir, Mozambique. - International conference held at Jinja, Uganda- Lake
Victoria 2000
Polasek P., and Mult S., (2005). Optimization of reaction conditions of particle aggregation in water
purification-back to basics. Water SA Vol. 31 No 1, pp 61-72.
Polasek P. (2007). Differentiation between different kinds of mixing in water purification -back to basics.
Water SA Vol. 33 No 2, pp 249-251.
Qin J.J, Maung H.O., Iran A. K., Knops F., and Miller P. (2006). Impact of coagulation pH on enhanced
removal of natural organic matter in treatment of reservoir water. Separation Purification Technology, 49
(2006), pp. 295-298
18
Sharp E. L., Parsons S. A., and Jefferson B. (2006). Seasonal variations in natural organic matter and its
impact on coagulation in water treatment. Science of the Total Environment, 363(2006), pp. 183-194
Shen Y. H. (2005). Treatment of low turbidity water by sweep coagulation using bentonite. J.Chem
Technology and Biotechnology, 80 (2005) pp. 581-586
Smet, J.E.M and Visscher, J.T. (1990). Pre-treatment Technologies for community water supply: An
overview of techniques and present experience, 2nd ed. International Water and Sanitation Centre, The
Hague, The Netherlands, pp 109-124.
Stumn W. and Morgan James J. (1996). Aquatic Chemistry: Chemical Equilibria and rates in Natural
waters. John Wiley & Sons, Inc. 3rd ed. New York, USA, pp 818-866.
Thompson T., Fawell J., Kunikane S., Jackson D., Appleyard S., Callan P., Bartan J. and Kingson P. (2007).
Chemical safety of drinking water: assessing priorities for risk management. WHO, Geneva
Switzerland
Velasco J. R., Uriate J. I. A., and Velasco U. I. (2007). Enhanced coagulation under changing alkalinityhardness conditions and its implications on trihalometane precursor removal and relationship with
UV absorbance. Separation Purification Technology, 49 (2006), pp. 295-298
Viraraghavan T. and Srinivasa P.T. (2004). Influence of Natural Organic matter (NOM) on the speciation
of aluminum during water treatment. Water, Air and Soil Pollution 152 (2004) pp. 35-54
Volk C, Kimberly B., Ibrahim E, Verges D, Amy G, and Lechevallier M. (2000). Impact of Enhanced and
Optimized coagulation on removal of organic matter and its biodegradable fraction in drinking
water. Water Research vol. 34 (12), pp. 3247-3257.
WHO (2004) Guidelines for Drinking Water Quality. Vol. 1 Recommendations (3rd edn.) Geneva, Switzerland.
Zamxaka M, Pironcheva G and Muyima NYO (2004). Microbiological and physico-chemical assessment of
thequality of domestic water sources in selected rural communities of the Eastern Cape Province,
South Africa. Water S.A vol. 30 (3), pp. 333-340.
Zamxaka M, Pironcheva G and Muyima NYO (2004). Bacterial community patterns of domestic water
sources in the Gogogo and Nkonkobe areas of the Eastern Cape Province, South Africa. Water S.A
vol. 30 (3), pp. 341-346.
19
II
THE EFFECTS OF INTERMITTENT SUPPLY AND HOUSEHOLD STORAGE
IN THE QUALITY OF DRINKING WATER IN MAPUTO
Matsinhe, N.P.1, D.L. Juízo1, Kenneth M. Persson2
1
Eduardo Mondlane University, Faculty of Engineering, Av. de Moçambique km 1.5, C. Postal 257,
Maputo, Mozambique, Phone:+258-21- 315161, (E-mail: matsinhe@zebra.uem.mz;Juizo@hotmail.com).
2
Lund University, Dept of Water Resources Engineering, P O Box 118 SE-221 00 Lund Sweden, (E-mail:
Kenneth.persson@tvrl.lth.se)
Submitted to the journal Water Science & Technology-Water Supply
Abstract A serious problem arising from intermittent supplies is the associated high level of contamination
and public health hazards resulting either from ingress of contaminated water or from prolonged storage.
This paper discusses the overall condition of the drinking water quality of intermittent water supplies in
Mozambique. The network of Maputo is used as an example. Records of water quality determinations from
different locations in the network are used to assess the final quality of water at consumer’s taps. Chlorine
residual levels measured in reservoirs and results of bulk chorine decay tests performed with samples of
treated water are used to estimate chlorine decay constants (kb) at different locations of the network and to
predict the influence of retention time in the final quality of water. Presence of bacteria, low disinfection
capacity and long residence times in the network and reservoirs are the main factors affecting the final
quality of distributed water. Post-contamination due to ingress of contaminated water and prolonged storage
in reservoirs is high. The intermittent operation of the distribution network is therefore inadequate to
guarantee safe drinking water.
Keywords: intermittent water supply, household storage, chlorine decay, water quality Maputo,
INTRODUCTION
The provision of safe drinking water is essential in maintaining the quality of life in all
communities. During the last decades, the demand for improved supplies in the Third World has
been increasing considerably due to rising per-capita incomes, rising standards of living, and
population increase (Reweta & Sampath, 2000). While the pressure on existing supplies
increased with time, the development of additional sources and/or extension of existing supplies
have been an unrealistic option for many places of the developing world due to financial
constraints.
The solution frequently applied in such places is the adoption of intermittent supplies.
Intermittent supplies, however, are associated with quantity and quality problems, occasionally
linked to fatal health hazards (Totsuka et al., 2004). Despite considerable negative impacts,
intermittent supplies are used in many parts of the world, especially in arid and densely populated
areas of the developing world.
Totsuka et al., (2004) point out that more than 90% of the population served by piped water
supply in South Asian countries receives water during less than 24 hours/day. In most African
countries, conditions are worse. According to the same author, in Zaria (Nigeria), only 11% of
consumers with a piped supply receive water one day in two while, in Mombasa (Kenya), the
1
average number of service hours is of about 2.9 hours/day. For Dar-es Salaam (Tanzania),
Reweta & Sampath (2000) indicate that the main supply to the town provides less than 1% of the
required demand, forcing residents to depend heavily on alternative supplies. The long list of
examples also includes many countries in The Middle-East.
The most critical consequence of intermittent supply is the risk of water contamination due to
ingress of contaminated water and the consequent public health hazards. Other consequences
include, inequitable water distribution, inconveniences to consumers and added costs of water
supply due to the need of additional facilities such as storage tanks and pumps (Totsuka et al.,
2004).
Unequal distribution of water forces consumers to find their own ways to cope with intermittency
by constructing household tanks. However, contamination of drinking water in household tanks is
a second important health risk. Totsuka et al., (2004) report water quality tests carried out in
Istanbul (Turkey) which revealed that 24% of samples taken from consumer storage tanks were
found positive for coliforms compared to only 4% positive samples taken from the pipe network.
Tokajian & Haswa (2003), reporting results from a controlled study in Beirut (Lebanon), have
found a positive correlation between mean bacterial counts and pH, temperature and storage time.
Intermittent supplies can also promote bacterial regrowth in the network during stagnant hours
and consequent biofilm detachment when the supply is restablished. These events were found to
greatly impact the water quality distributed in a controlled study run in a suburb of Nablus
Palestine, Coelho et al., (2003). Bacteria counts were about eight times higher during the first
five minutes of supply compared to the overall water quality.
Water supply and distribution in a large number of Mozambican cities and villages is also
intermittent. Existing transport and distribution networks are old; suffer from high levels of
leakage, limited hydraulic capacity and limited coverage, caused by city demand increases and
city growth. The average number of supply hours in the majority of the cities is of less than 12
hours (Gumbo et al., 2003). The purpose of this work is to evaluate the effect of intermittency
and household storage on the quality of drinking water distributed in Maputo.
MATERIALS AND METHODS
The study Area
Maputo is the largest city of Mozambique and the city capital of the country. The town is
supplied by piped water from a system consisting of a single source in the Umbeluzi River,
treatment by conventional treatment, transport through a 28 km transmission main, and
distribution through a reticulated network of approximately 840 km (Gumbo, 2004). Five
distribution centres (DCs) exist of which three are located in series along the main supply line
(Matola at about 10 km from the water works, Chamanculo at 20 km and Maxaquene at 28 km).
The others are Machava DC located some 17 km from the water works, and Alto-Maé DC
located at about 24 km.
The supply from the water works is done 24 hours a day, but the distribution to the different
consumption zones is intermittent because of low pressure in the system and the need to
2
minimize losses (Gumbo et al., 2003). Many consumers of the city of Maputo have therefore,
built extra household tanks to cope with water shortages. A large proportion of these tanks are
built on ground level from where stored water is further pumped to roof tanks or supplied directly
to taps in the households.
Data collection
The drinking water quality in Maputo was investigated through analysis of data provided by the
service provider and data collected through fieldwork. The data from the service provider
covered the period 2001-2004 but only data from 2004 was used for the analysis. Due to
incompleteness of the data with respect to bacteriological data, the analysis of bacteriological
aspects of distributed water was done using data from 2003.
Residual chlorine, bacteria, turbidity and solids are the parameters used for the analysis of the
water quality in the network. Temperature, residence time and the condition and treatment of
reservoirs and household tanks are used to make the final assessment of the quality of drinking
water in Maputo. Reference locations are: the water works (treated water), distribution centres
(DC), household tanks and taps in the network.
Fieldwork took place during 7 weeks in November/December 04. Water samples taken from
different locations in the network were analyzed for physic-chemical and bacteriological
characteristics. During the fieldwork, samples were collected twice a week (Mondays and
Wednesdays) usually between 7.30 a.m. and 2.00 p.m. Four sampling points were visited each
day where, water samples were collected pair wise on taps located before and after household
tanks. The taps used to collect samples for bacteriological analysis were cleaned with cotton and
ethanol before sampling. In addition, the taps were left open for about 1 minute before sampling.
During sampling of household reservoirs, the condition of the tanks was assessed through
observation and interviews to owners. Aspects of interest included, the overall condition of the
tanks, construction materials and dimensions, consumers practices regarding the cleaning of
reservoirs and consumer’s concerns about water quality.
Time dependent chlorine decay tests were performed with samples of treated water. The method
used (often referred to as ‘bottle” or “jar test”, (Powel et al., 2000)) consisted of recording
chlorine concentrations at fixed time intervals from bottles previously filled with sample water.
The tests were used to estimate the magnitude of chlorine depletion with time and to estimate
chlorine decay rates (Kb) for treated water. The results of the tests are compared to results of
calculations done with data provided by the service provider and that generated during the
fieldwork. Existing or measured data on residual chlorine was used to estimate chlorine decay
rates at selected locations of the network with emphasis put on household tanks and reservoirs of
distribution centres. Chlorine decay rates estimations where based on a first order decay reaction.
Estimates of the total residence times in the network, reservoirs and household tanks are used to
discuss the results of calculations of chlorine decay rates.
Measurements
Temperature, pH, TDS, conductivity, free and total residual chlorine, was measured directly in
the field while turbidity, bacteria and solids were measured in the laboratory. All Temperature
3
readings were taken with a standard mercury thermometer (accuracy of ± 10C) and TDS, pH and
Electrical conductivity (EC) were measured with pocket Wagtech digital meters. Free and Total
residual chlorine (FRC, TRC) were measured with the DPD (diethyl-p-pheneylene diamine
tablets) colorimetric method with colour measurement through a portable digital photometer
(Wagtech 5000) calibrated at 520 nm wavelength.
Turbidity and alkalinity were measured a few hours later in the laboratory. A Hach turbidity
meter DR 2500 was used for turbidity readings. Alkalinity was measured with a simplified
titration method described in the Standard Methods for the Examination of Water & wastewater,
(20th edn., 1998). All bacteriological analyses were done at the laboratory of the Ministry of
Health using the membrane method described in the standard Methods for the Examination of
Water & Wastewater. The parameters measured were total bacteria, faecal coliforms and E.coli.
Sediments sampled in at least three reservoirs were analyzed for solids (total, volatile and fixed
solids). The method used for analyzing solids followed the Standards Methods for the
Examination of Water & Wastewater (20th edn., 1998).
All pair wise data was checked for its statistical significance. The method used consisted of
making a confidence interval for the difference of two expected values I(ȝ1- ȝ2); if this interval
does not cover the zero it is regarded to be a significant difference between two homogeneous
groups. A 95% confidence interval (p=0.05) was used in all tests.
Chemical and physical parameters measured in samples taken from the network and after
household tanks were compared, with an independent sample T-test. The T-test is a parametric
comparative test used to show if a difference exists between two homogenous groups. For
turbidity, free residual chlorine, total residual chlorine and nitrate the test was not valid since the
variance of the two groups were not equal. The Mann-Whitney U-test was used instead. Presense
and absense of bacteria in samples taken before and after storage were also compared with a
Mann- Whittney U-test. The T-test was also used to compare residual chlorine in samples with
and without bacteria.
RESULTS AND DISCUSSION
Residual Chlorine
For assessing chlorine residual levels in the network of Maputo, about 1192 records of the
operator database and 26 records of determinations done during the fieldwork were investigated.
The records from the operator database contained only values for TRC while that of the fieldwork
included also data for FRC. The results of the assessment for TRC, indicate levels of residual
chlorine in the network and water leaving reservoirs of DCs, mostly above the lower limit
0.25mg/l (Table 1).
Typically, the allowable minimum is 0.2 mg/l (WHO, 2004). The Mozambican standards for total
residual chlorine are 0.25 -1.0 mg/l in the net and 1.5 mg/l at the water works (MISAU, 2004). In
the specific case of Maputo water supply where a private company (Águas de MoçambiqueAdeM) has been awarded a 15 years lease contract to manage the water supply for the city, the
contract standards indicate limits only for free residual chlorine between 0.2-1.0 mg/l.
4
338
1192
27
320
345
343
343
345
26
Net (db)
Net (fw)
DC-Matola (db)
DC-Machava(db)
DC-Chamanculo(db)
DC4-Alto Maé(db)
DC5-Maxaquene(db)
H-reservoirs (fw)
Nr.
records
Treated water (db)
Location
0.24
0.63
0.81
0.98
0.97
0.98
0.50
0.69
1.58
Ave.
0.65
1.9
2.5
2.5
3.0
3.0
1.1
2.0
2.49
Max
0.06
0.11
0.15
0.15
0.15
0.20
0.08
0.13
0.87
Min
0.14
0.30
0.36
0.41
0.40
0.44
0.25
0.32
0.37
Std
Total Residual chlorine
5
0.06
-
-
-
-
-
0.17
-
0.29
-
-
-
-
-
0.68
-
Max
0.01
-
-
-
-
-
0.01
-
Min
0.06
-
-
-
-
-
0.14
-
STD
Free Residual chlorine
Ave.
Residual Chlorine
30
342
340
341
341
320
30
-
338
Nr.
records
1.40
1.76
2.30
2.31
2.39
2.52
2.00
-
3.22
Ave.
3.50
6.73
7.36
6.99
9.19
11.20
6.00
-
7.64
Max
Turbidity
0.7
0.79
1.08
0.78
0.81
1.15
0.60
-
1.81
Min
0.6
0.59
0.58
0.67
0.73
0.73
1.50
-
0.59
STD
Table 1. Total and Free residual chlorine and Turbidity in the network, reservoirs of DCs and household tanks of Maputo water supply
The data on TRC levels of water leaving reservoirs of DCs also show average TRC levels
decreasing as the distance from the water works increases even though, the water is rechlorinated at some DCs located along the main supply line (Matola and Chamanculo).
Analysis of records for TRC and FRC using sampled data reveals much lower values for FRC in
the network and household reservoirs with 77% of samples falling below the target limit of 0.2
mg/l. TRC levels were in general 0.26 mg/l lower after reservoirs while, FRC was 0.11mg/l
lower. A statistically significant difference (p < 0.01) between levels measured before and after
household tanks has been observed.
Analysis in different parts of the network thus suggests that the disinfection capacity in the
network is rather high when evaluated with TRC levels. The same analysis done on basis of FRC
suggests, however, a different scenario with chlorine residual levels (0.17 mg/l) very close to the
lower limit recommended for an effective disinfection of the water.
Turbidity
The results of the analysis (see Table 1) suggest mean values of about 3.22 NTU for treated
water and mean values between 1.8-2.5 NTU for water leaving DCs. The data from the operator’s
database did not contain data on turbidity levels in the network but the results of the sampled data
suggest mean values of 2.0 NTU. From Table 1 it is clear that at all measured sites turbidity
levels were most of the times above the WHO desired limit of 1.0 NTU and that in few occasions
the WHO limit of 5 NTU was exceeded. The absolute limit of 20 NTU stated in the contractual
limits of the water company of Maputo was, however, never exceeded.
Overall, turbidity levels at all investigated sites show rather large variations between extreme
values with occasional increases of turbidity levels of the water leaving some of the reservoirs of
DCs. This was reported in the reservoirs of the DCs of Chamanculo and Alto Maé but not in the
reservoirs of Maxaquene DC.
This suggests the possibility of occasional loads of turbidity-causing particles entering the
reservoirs, causing sediments to build up during periods of low demand and further release with
the water leaving the reservoir during periods of high demand. Sources of turbidity causing
particles can be pipe and fitting corrosion, lining erosion, biological growth, chemical reactions
and external contamination that may occur during operations such as pipe repairs (Vreeburg &
Boxall, 2007). The formation and growth of particles is, however, a very complex process, which
is currently poorly understood but factors such as contact times, contact surface, and hydraulic
conditions are likely to play an important role in controlling these processes (Vreeburg & Boxall,
2007).
Bacteria
Presence/absence of bacteria was investigated using records from both the service provider and
the additional field survey. The operator’s database had records of samples of treated water,
reservoirs at DCs and reservoirs and taps of the network. Samples taken during the fieldwork
considered two locations at each sampling point namely taps before and after household
reservoirs. Around 503 records taken from the operator’s database (148 from reservoirs at DCs
6
and 355 from taps and household tanks) were investigated for presence of bacteria. The fieldwork
produced another 60 records.
The results suggest that both faecal coliforms and E.coli were found frequently in reservoirs of
distribution centres, network taps and household tanks. The records on treated water did not
contain data on presence/absence of bacteria. In the reservoirs of DCs, bacteria were found in
about 24% of the records investigated.
The levels of contamination were generally low (<10cfu/100 ml), but occasionally reached high
values (up to 165cfu/100ml). Analysis of records from reservoirs of DCs also suggests that the
DCs located first along the main line had the lowest incidence of bacteria counts when compared
to DCs located further along (Figure 1a.).
60
Nr. records with positive/negative counts
Number of positive records
12
10
8
6
4
2
0
50
Records with negative
counts
40
Records with positive counts
30
20
10
0
Matola
Machava
Chamanculo
Alto-Maé
Maxaquene
Jan
Figure 1.a
Feb Mar
Apr
May
Jun
Jul
Aug Sep Oct
Nov
Dec
Figure 1.b
Figure 1. Incidence of bacteria in reservoirs of DCs (Fig. 1a) and taps of the network (Fig. 1.b).
The data from the operator’s database suggest, additionally, that on the network between the DCs
and the households tanks, around 20% of the investigated samples were found contaminated with
bacteria. The annual distribution of cases with positive counts of bacteria over the total number of
records investigated in the distribution network is show in Figure 1b.
The incidence of positive samples is evenly distributed throughout the year, which suggests that
contamination with bacteria is a rather persistent problem in the network. A similar analysis done
with records of the fieldwork suggests that coliform bacteria were found in thirteen out of the
sixty samples collected (22% of samples taken). Six out of the thirteen samples found with
bacteria had bacteria counts over 100 cfu/100ml. Coliform bacteria was generally found in
samples collected after household tanks (Figure 2). Ten out of the thirteen samples found with
bacteria were collected after household tanks.
7
120
100
100
80
80
60
60
40
40
Number of Coliforms
Number of Coliforms
120
20
0
-20
0.0
.2
.4
.6
.8
1.0
20
0
House
House
Net
-20
1.2
0.0
Total Residual Chlorne (mg/l)
Net
.1
.2
.3
.4
.5
.6
.7
Free Residual Chlorine (mg/l)
Figure 2. Presence of Coliform bacteria in samples taken before and after household tanks.
Because these samples were collected from seven different locations in the network, in three
cases bacteria were found at least twice during the study. The results of the fieldwork also
indicate that bacteria were only found in samples collected from reservoirs with rather low
concentrations of residual chlorine (see Figure 2). TRC was generally low in the influent of
distribution centres in those days when faecal coliforms were found. E.coli was found three times
in the network when the concentration of FRC was between 0.05 and 0.19mg/l. These levels of
FRC should be sufficient for disinfection (WHO, 2004). This suggests that recent contamination
of the drinking water took place.
One of the major reasons for bacterial contamination in pipes and reservoirs of distribution
networks is insufficient disinfection capacity due to insufficient residual chlorine. Since chlorine
decays over time, increased retention time by either storage or prolonged periods of interruption
of supply increases the risk of occurrence of bacteria at points of water consumption. Biofilm
formation and bacteria growth may also be occurring during periods of low or no pressure and
further biofilm (containing bacteria) detachment when the supply is restart (first flush effect).
The network of Maputo is left under pressurized more often than the transport main between
distribution centres (at least once a day), therefore, ingress of contaminated water may occur
more frequently. The network is however, left under pressurized only few hours/day so, the time
for biological growth between flushes is short. Problems with biofilm formation play, therefore, a
minor role with respect to contamination with bacteria.
Sediments
Sediments may increase the rate of chlorine decay in pipes and reservoirs, decrease the
disinfection capacity and act as a source of nutrients for bacterial growth (WHO, 2004).
Sediments resulting from particle accumulation are in fact know to have a relation with biological
activity since, one-12% of the organic matter in the particles may consist of bacterial biomass.
This make sediments resulting from particle accumulation an important factor in hygienic safety
of drinking water (Vreeburg & Boxall, 2007).
Sediments taken from few tanks in the network were investigated for factors known to influence
chlorine decay rates in the reservoirs and to contribute for microbial growth (Table 2).
8
Table 2. Solids concentration in water collected from household reservoirs
Reference
Total solids
Volatile solids Fixed solids
VS/TS
site
(mg/l)
(mg/l)
(mg/l)
3
411.0
133.0
278.0
0.32
7
631.0
166.0
465.0
0.26
3
464.0
112.5
352.0
0.24
The VS/TS fraction is used as an indication for the organic matter content. The results (see Table
2) show amounts of 24-32% of organics present in the sediments. The organics can potentially
serve as nutrients for microbial growth or exert an extra demand of chlorine in tanks and
reservoirs. According to findings from literature, the existence of large variations in turbidity will
be the predominant causes of sediments build up in reservoirs.
Apart from increasing the rate of chlorine decay, sediments build up in reservoirs decreases the
disinfection capacity as bacteria can hide inside the particles and escape from the effect of
chlorination. The fact that the water stays relatively long in reservoirs may also increase the
potential for biofilm formation (on the walls of reservoirs) and bacterial growth.
Chlorine decay & retention time
Chlorine decay constants calculated according to first order decay models for different parts of
the network are resumed in Table 3. Mean values of total residual chlorine in the influent and
effluent of reservoirs were used in the calculations. The values of retention time in reservoirs of
DCs were calculated from average figures of water consumption as provided by the operator and
the net volume of reservoirs of the specific distribution centers.
For household storage tanks, retention time was calculated based on an average daily
consumption equivalent to an average of 4,4 persons/house (INE, 1999) and a per capita water
demand of 120 l/p.d as pescribed in the national standards (DNA, 2003) for water consumption.
Though the vary significantly in sizes, household tanks were assumed to have a capacity of 1.0
m3. Results of calculations of average retention time in reservoirs and estimated constants for
residual chlorine decay are resumed in Table 3. Chlorine decay rates vary significantly in
different parts of the network which suggests that the extent of chlorine decay varies in the
system.
9
Table 3: Chlorine decay constants (Kb (l/h)) in different parts of the network.
Location
Average Chlorine Average Chlorine
Average
concentration
concentration
retention time
inlet C0(mg/l)
effluent Ce(mg/l)
(hours)
Treated watera
1.58
DC-Matola
1.17
0.98
12.5
DC-Chamanculo
1.20
0.97
37,2
DC-Alto Maé
0.95
0.81
11.8
DC-Maxaquene
0.887
0.63
10.1
H-reservoirs (TRC)
0.5
0.24
48
H-reservoirs (FRC)
0.17
0.06
48
Note: a obtained from time dependent chlorine decay test R = 0.903
Kb
(1/h)
0.603
0.016
0.006
0.015
0.035
0.015
0.022
The results of chlorine residual tests performed with samples of treated water are also resumed in
Table 3. Compared to the results obtained for different reservoirs in the network, the rate of
chlorine consumption in the treated water is much higher than that observed in reservoirs of DCs
and households. This is expected pattern since, the treated water is re-chlorinated in two points
along the main line (Matola and Chamanculo) before reaching the reservoirs at DCs located
downstream. According to findings from Kiéné et al., (1998) and Fang et al., (2000), chlorine
decays more rapidly in freshly chlorinated water when compared to water that has been rechlorinated.
The rate of chlorine decay in DCs and household reservoirs appears to be generally high and to
vary significantly (between 0.006 and 0.035 l/h). This suggests that there may be external factors
(e.g. biomass builds up, particles mixed with turbidity) that influence the decay rates.
Condition and treatment of household tanks
The materials mostly used for construction of household storage tanks is concrete, plastic (prefabricated black-PVC tanks) and asbestos cement. The capacity of concrete and asbestos cement
tanks ranges from 0.25 m3-4.5 m3, while that of plastic tanks ranges from 0.5m3-1.5m3 (Table 4).
The maintenance of the tanks is often poor, cleaning and disinfection is hardly done and in many
of them lids was missing or was locked for long periods meaning that they are hardly opened for
cleaning, maintenance and repair work.
Most of the tanks are oversized for the average demand/family, which, according to estimations
based on an average of 4.4 persons/household @ 120 l/p.day, lies between 0.5 and 0.6 m3/day.
This result in relatively long storage times, excessive depletion of residual chlorine and the
possibility of bacterial growth and biofilm formation due to insufficient disinfection capacity and
bad condition of tanks.
10
Location
Ground
Roof
Roof
Roof
Under ground
Underground
Roof
Ground
Roof
Roof
Roof
Ground
Ground
Ground
Ground
Ground
Roof
Ground
Ground
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
9.0
(6x1.5 m3 e.a.)
2.0
(2x1.0 m3 e.a.)
1.0
0.5
0.5
1.0
0.5
2.0
(2x1.0 m3 e.a.)
1.0
1.0
1.0
3.0
1.0
4.5
6.0
(2x3.0 m3 e.a.)
1.0
0.5
2.0
0.5
Net volume
(m3)
2.8
6
1.7
2.5
30
5
0.7
1.2
1.4
0.7
1.7
1.4
1.2
1.4
5.0
1.7
6
7
6
6
10
6
7
6
5
5
3.8
6.3
8
10
1.4
0.4
2.7
0.8
Average
retention
time (days)
6
7
6
5
Average
number of
users
Table 4: Location, condition and treatment of household tanks
Ref.
house
11
Asbestos cement
Asbestos cement
Concrete
Plastic
Asbestos cement
Asbestos cement
Asbestos cement
Asbestos cement
Asbestos cement
Asbestos cement
Plastic
Concrete
Asbestos cement
Concrete
Concrete
Plastic
Plastic
Asbestos cement
Plastic
Material
Good
Good
Fairly good
Fairly good
Good
Fairly good
Good
Good
Good
Good
Bad
Good
Fairly Good
Good
Good
Fairly good
Good
Physical
condition
Cover missing. Exposed to ingress of contaminants
Tank cover broken. Ingress of contaminants possible.
Lid locked. Access for cleaning and maintenance
limited
Lid locked. Access for cleaning and maintenance
limited
Lid locked. Access for cleaning and maintenance
limited
One tank with a broken lid. Second tank with lid
locked. Access for cleaning and maintenance limited.
Ingress of contaminated water possible.
Lid locked. Access for cleaning and maintenance
limited
None.
Tank cover heavily corroded. Ingress of pollutants and
contaminated water possible
Lid locked. Access for cleaning and maintenance
limited.
None
Dirty inside the tank. Biofilm developing on the walls
of the tank
Cover missing. Exposed to ingress of contaminants
Cover missing. Exposed to ingress of contaminants
One tank with a broken cover. Second tank with cover
locked. Access for cleaning and maintenance limited.
Ingress of contaminated water possible.
Tank cover heavily corroded. Ingress of contaminants
possible
Lid locked Access for cleaning and maintenance
limited
Access ladder heavily corroded. Access for cleaning
and maintenance limited
Covers missing. Exposed to ingress of contaminants.
Encountered maintenance problem
According to results of this study, the average free residual chlorine levels downstream
household tanks showed a decrease (Figure 3) of about 60% during storage (from 0.167mg/l to
0.061mg/l).
1.2
.8
1.0
.6
.8
.4
.6
Free Residual Chlorine (mg/l)
Total Residual Chlorine
.4
.2
0.0
-.2
N=
27
26
Net
House
.2
0.0
-.2
N=
27
26
Net
House
Figure 3: Total and free residual chlorine before (net) and after storage in household tanks
(o=outliers, *=extremes).
Free residual chlorine in the majority of household reservoirs is therefore far lower than
recommended limits. This finding is in line with conclusions from earlier studies by Coelho et al.
(2003) and Tokajian & Haswa (2003), who found a strong relationship between storage time and
bacterial growth as residual chlorine decreased. The combined effect of intermittency and
household storage usually increases retention times, with the observed consequences for chlorine
concentrations and occurrence of bacteria.
CONCLUSIONS
The results of this study suggest that the drinking water quality in Maputo is generally not safe
for human consumption due to the presence of bacteria. The water quality deteriorates gradually,
from the treatment works to the distribution centres, further on into the network, and finally in the
household tanks. The reason is a combination of factors such as condition of pipes, ingress of
contaminated water when the network is without pressure, long retention times in the network
and reservoirs, and condition and treatment of storage tanks.
Both faecal coliforms and E.coli were found frequently in reservoirs of DCs and in the network.
Apparently, some contamination occurs before or at reservoirs of DCs, which suggests
contamination due to ingress of contaminants during periods of low or no pressure. The
intermittent mode of operation of the network of Maputo water supply is therefore, pointed as the
most critical factor causing contamination in the network.
Storage is influencing the water quality either because of retention time or because of poor
management and ingress of contaminants. Long storage times seem to be the major factor of
water quality deterioration, particularly in household tanks. Storage increased the risk of
occurrence of faecal coliforms in the water with more than 100%.
12
Though based on a limited number of samples, sediments in household tanks are found to
potentially contribute to water quality deterioration. Turbidity of the water entering the reservoirs
is generally high. Because the majority of household reservoirs are over- dimensioned and the
network is operated intermittently, conditions exist for the settling of turbidity-causing particles
and sediments build up.
The combined effect of sediments and low disinfection capacity and long retention time, results
in favourable conditions for bacterial growth.
ACKNOWLEDGEMENTS
This paper contains research results from a M.Sc. project by Kajsa Lindqvist and Malin
Wicklander who conducted their Minor Field Studies (MFS) in Mozambique. The study was
funded by the Swedish International Development Cooperation Agency (SIDA/SAREC) and is
part of an ongoing research project on Integrated Water Quality/Quantity Management on
Tansboundary rivers jointly implemented by Eduardo Mondane University and the Lund Institute
of Technonoly of Lund University-Sweden. The authors wish to thank all people and institutions
that provided support and data and information used in present study.
REFERENCES
Coelho, S. T., James, S., Sunna, N., Abu Jaish, A. and Chatila, J. (2003). Controlling water
quality in intermittent supply systems. Wat.Sci.Tech.: Water Supply 3 (1-2), pp 119-125
Fang Hua, J.R. West, R. A. Barker and C. F. Forster (1999). Modelling of Chlorine Decay in
Municipal Water Supplies. Water Resources, Vol. 33 No. 12, pp 2735-2746.
Gumbo, B., Juizo, D.,van der Zaag, P. (2003). Information as a pre-requisite for water demand
management: experiences from four cities in southern Africa. Physics and Chemistry of the
Earth, vol. 28, pp 827-837.
Gumbo, B. (2004). The status of water demand management in selected cities of Southern Africa.
Physics and Chemistry of the Earth, vol. 29, pp 1225-1231.
INE (1999). Resultados do II Recenseamento Geral da População 1997 (Final results of the 1997
population census). Instituto Nacional de Estatística (INE), Maputo, Mozambique.
James C. Powell, Nicholas B. Hallam, John R. West, Chistopher F. Forter and John Simms.
(2000). Factors which control bulk chlorine decay rates. Wat. Sci. Tech., vol 34, No 1, pp
117-126.
L. Kiéné, W. Lu, Y. Lévi. (1998). Relative Importance of the phenomena responsible for chlorine
decay in drinking water distribution systems. Wat. Sci. Tech., vol 38, No 6, pp 219-227.
Regulamento de Sistemas Públicos de Distribuição de Água e de Drenagem de Águas Residuais
(2003) (Mozambican Guidelines for design and operation of public water supply and
distribution systems and for drainage of wastewater 2003). DNA, Maputo, Mozambique.
Regulamento sobre a Qualidade de Água para o Consumo Humano; Boletim da República de
Moçambique, I Série número 37, Diploma Ministerial n° 180/2004 de 15 de Setembro
(2004) (Guidelines for drinking water quality in Mozambique; Boletim of the Republic of
13
Mozambique, Series I, No 37, Ministerial decree 180/2004 of September 15). Ministry of
Health (MISAU), Maputo, Mozambique.
Reweta W.S.J., Sampath R.K. (2000). Performance Evaluation of Urban Water Supply in
Tanzania: The Case of Dar Es Salaam City. Water Resources Development, Vol. 16, No. 3,
pp 407–423.
Standard Methods for the Examination of Water and Wastewater (1998). 20th edn, American
Public Health Association/American Water Works Association/Water Environment
Federation, Washington DC, USA
Tokajian S., Hashwa F. (2003). Water quality problems associated with intermitent water supply.
Water Sci.Technology, vol 47, No 3: pp 229-239.
Totsuka N., Trifunovic N., and Vairavammorth K. (2004). Intermittent urban water supply under
water starving situations. 30th WEDC International Conference, Vientiane, Lao PDR, 2004.
Vreeburg J.H.G, Boxall J.B. (2007). Discolouration in potable water distribution systems: A
review. Water Research, No 41, pp 519-529.
WHO (2004). Guidelines for Drinking Water Quality. Vol. 1, Recommendations. 3rd edn, WHO,
Geneva, Switzerland.
14
III
Water services with independent providers in peri-urban
Maputo: Challenges and opportunities for long-term
development
Nelson P Matsinhe1*, Dinis Juízo1, LC Rietveld2 and Kenneth M Persson3
1
Universidade Eduardo Mondlane, Faculdade de Engenharia, Av. de Moçambique km 1.5, C. Postal 257, Maputo, Mozambique
2
Delft University of Technology, Dept of Water Management, 2628CN, Delft, The Netherlands
1
Lund University, Dept of Water Resources Engineering, P O Box 118, SE-221 00 Lund, Sweden
Abstract
Water service delivery to most residents of peri-urban areas of greater Maputo depends largely on alternative service providers, mostly in the form of small-scale independent providers (SSIPs). This paper discusses the present and long-term chalOHQJHVIDFLQJ66,3VLQVXSSO\LQJTXDOLW\ZDWHURIVXI¿FLHQWTXDQWLW\LQSHULXUEDQ0DSXWRDQGSRVVLEOHKXPDQKHDOWKULVNV
associated with the consumption of water provided by SSIPs. Extensive water sampling and analyses were conducted to evaluate the physicochemical and bacteriological quality of water provided by independent providers and the associated human
KHDOWKULVNV%RUHKROHSXPSLQJWHVWVWKHUHVXOWVRIZKLFKZHUHLQWHUSUHWHGXVLQJWKHJUDSKLFDOPHWKRGRI-DFREZHUHXVHG
to evaluate the regional aquifer potential, the long-term impacts of its exploitation and the aquifer vulnerability to external
contamination. From the results of borehole pumping tests it was concluded that the present yields are in average 33% lower
than estimated safe yields and that larger than present yields therefore can be exploited. The aquifer vulnerability to external
contamination (e.g. by E. coliDQGQLWUDWHVLVORZPDLQO\EHFDXVHRIORZK\GUDXOLFORDGVDQGWKHH[LVWHQFHRIDUDWKHUWKLFN
WRPVDQG\XQVDWXUDWHGVWUDWXPZKHUHEDFWHULDGLHRIIDQGELRORJLFDOGHQLWUL¿FDWLRQSUREDEO\RFFXUV+RZHYHUWKHDTXLIHUYXOQHUDELOLW\WRVHDVHDZDWHULQWUXVLRQLVKLJK&XUUHQWO\WKHKHDOWKULVNVSRVHGWRFRQVXPHUVUHO\LQJRQVHUYLFHVSURYLGHG
by SSIPs are small; even so, 13 out of 35 controlled boreholes had either total coliform or faecal coliform levels higher than
WKH:+2VWDQGDUG,QWKHORQJUXQ66,3VPD\IDFHPRUHVHULRXVZDWHUTXDOLW\SUREOHPVGXHHLWKHUWRRYHUH[SORLWDWLRQRIWKH
aquifer system or increased hydraulic loads resulting from increased population density.
Keywords: water supply services, peri-urban areas, small-scale independent providers, water quality,
public health
Introduction
Most cities of developing countries are characterised by two distinct set-ups, namely the formally built ‘cement areas’ and the
nearly rural types of neighbourhoods, the so-called peri-urban
settlements, of metropolitan areas. The latter are usually slums,
ODFNLQJHYHU\IRUPRIXUEDQSODQQLQJZKHUHWKHPDMRULW\RIWKH
urban poor lives. The development of this type of settlements is
RIWHQQRWLQFOXGHGLQWKHRI¿FLDOSODQVIRUXUEDQGHYHORSPHQWRI
WKHPDLQFLWLHVDIDFWWKDWRIWHQFDXVHVGLI¿FXOWLHVLQWHUPVRI
planning, implementation, provision and maintenance of public
VHUYLFHV%ROD\DQG5DELQRYLFK7RDFFHVVSRWDEOHZDWHU
supplies, most residents of such neighbourhoods rely on alternative service providers such as small-scale independent providers
(SSIPs), often acting as investors, developers and/or managers of
ZDWHUNLRVNVDQGVPDOOVFDOHSLSHGZDWHUV\VWHPV
2IWKHYDULRXVIRUPVRIVPDOOVFDOHVHUYLFHSURYLVLRQ66,3V
have long been accepted by donors and governmental authorities as a viable alternative to developing, managing and expanding service coverage in remote and underserved areas. In the
¿HOGRIZDWHUVXSSO\66,3VDUHHVWLPDWHGWRUHDFKDVPXFKDV
* To whom all correspondence should be addressed.
+258-21475330; fax: +258-21475311;
e-mail: matsinhe@zebra.uem.mz; nelson.matsinhe@tvrl.lth.se
Received 12 December 2007; accepted in revised form 23 May 2008.
Available on website http://www.wrc.org.za
ISSN 0378-4738 = Water SA Vol. 34 No. 3 July 2008
ISSN 1816-7950 = Water SA (on-line)
KDOIRIWKHSRSXODWLRQLQVRPHFRXQWULHV.DULXNLDQG6FKZDUW]
2005), while up to a quarter of the urban population of Latin
America and nearly half of urban dwellers in African countries
UHO\RQ66,3VIRUDWOHDVWDSRUWLRQRIWKHLUGULQNLQJZDWHUVXSSO\
(Collingnon and Vézina, 2000). The list of examples in Africa
LQFOXGHV WKH FLWLHV RI %DPDNR &RWRQRX &RQDNU\ DQG 'DU HV
Salaam, where SSIPs are reported to be the main source of
GULQNDEOHZDWHUIRUPRUHWKDQRIKRXVHKROGVDQGFLWLHVOLNH
$ELGMDQ1DLURELDQG2XDJDGRXJRXZKHUH66,3VDUHUHSRUWHG
to reach 22% to 28% of unconnected households (Collingnon
and Vézina, 2000). The city of Maputo is no exception.
SSIPs have operated in Maputo since the beginning of the
1980s, but they have only become widely established as from
2000. Estimates from a survey carried out on SSIPs in greater
Maputo indicate that by 2005, more than 240 groundwaterbased small piped systems run by SSIPs existed in the municiSDOLWLHVRI0DSXWRDQG0DWROD6HXUHFD+\GURFRQVHLO
$URXQGZHUHUHSRUWHGWRKDYHEHHQHVWDEOLVKHGZLWKLQWKH
last 5 to 8 years with roughly 43% built between 2002 and 2005
)LJ'XULQJWKHVDPHSHULRGVHUYLFHOHYHOVURVHIURPSXElic taps only to services provided also through house connections, yard taps and private stand pipes (Sal-Consultores, 2005).
Today, some 32% of SSIPs are reported to have more than 100
house connections.
Most SSIPs, even though located within the boundaries of
greater Maputo and furthermore within the lease area of a private operator – Águas e Moçambique (AdeM) – contracted to
provide services to the city of Maputo, are currently not for-
411
3HUFHQWRIQHZV\VWHPVDVRPSDUHGWR
DFWXDOQXPEHU
\HDURIFRQVWUXFWLRQ
Figure 1
Statistical distribution of SSIPs in Maputo and Matola
Municipalities by year of construction (adapted from: Seureca
& Hydroconseil, 2005)
PDOO\UHJXODWHG7KLVLVSDUWLDOO\GXHWRWKHODFNRIOHJDOEDVHV
RUOHJLVODWLRQIUDPHZRUNVWKDWFRXOGEHXVHGWRLVVXHOLFHQFHVRU
regulate their activities.
The evolution of household-level water strategies into SSIPs
has in most cases been unplanned. Most owners of systems had
built those to provide water for themselves, but because of insistence from neighbours they ended up allowing for some private
FRQQHFWLRQVRUVHOOLQJZDWHULQVPDOOTXDQWLWLHVWRƐDQG
thus slowly developing into water vendors. Also, because the benH¿WVRIVHOOLQJZDWHUKHOSHGSURYLGHUVRIIVHWWKHLQYHVWPHQWDQG
UXQQLQJFRVWVEUHDNHYHQH[SHFWHGZLWKLQWR\HDUVPDQ\RI
them eventually turned into professional service providers.
The rapid increase in SSIP number in the neighbourhoods of
greater Maputo has always been driven by demand, particularly
LQDUHDVZKHUHDIRUPDOQHWZRUNLVODFNLQJ,QVXFKQHLJKERXUhoods, SSIPs have the dominant role in service provision and are
reported to reach as many as 32% of unconnected households
6HXUHFD+\GURFRQVHLO6DORPRQ7KLVVLWXDWLRQ
LVQRWOLNHO\WRFKDQJHLQWKHIXWXUHPDLQO\EHFDXVHWKHSK\VLFDO
H[SDQVLRQRIWKHIRUPDOQHWZRUNLVXQOLNHO\WRHYHUPDWFKWKH
speed at which new suburbs emerge in the city. Moreover, SSIPs
KDYHUHFHQWO\EHHQUHFRJQLVHGDQGDFFHSWHGDVNH\SDUWQHUVLQ
WKH UHFHQWO\ ODXQFKHG µ0DSXWR :DWHU 6XSSO\ 3URMHFW¶
which among other aspects foresees the development of complementary groundwater- based systems and the involvement of
small local private operators in the management of service provision.
In this study, the long-term challenges facing SSIPs with
UHJDUGV WR VXSSO\LQJ TXDOLW\ ZDWHU RI VXI¿FLHQW TXDQWLW\ WR
unconnected residents of peri-urban Maputo have been assessed
based on an analysis of the present situation with respect to the
quality of water, the hydrogeological set-up of the aquifer system used by SSIPs and the potential for quality and quantity
problems associated with the proliferation of SSIPs. The focus
of the study was on the source water quality, potential yields
vis-à-vis present and future demands, and the best strategy to
locate and monitor boreholes under a scenario of the continuous
growth of SSIPs.
Materials and methods
The study area
Maputo is the capital of Mozambique and is situated on the Indian
2FHDQFRDVWOLQH7KHFLW\LVFKDUDFWHULVHGE\WKUHHGLVWLQFWVHW
412
ups, namely the area with high-rise buildings of the so-called
old ‘cement city’, a few inner suburbs built before independence
from the Portuguese in 1975, and the outer neighbourhoods consisting mainly of informal settlements.
'XULQJWKHXQVWDEOHSHULRGDIWHULQGHSHQGHQFHH[DFHUEDWHG
by the civil war that raged in the country for almost a decade,
PDQ\IDPLOLHVZKRÀHGWKHUXUDODUHDVLQVHDUFKRIVHFXULW\LQ
WKHFLWLHVZHUHDOORZHGWRVHWWOHLQWKHRXWVNLUWVRI0DSXWRZKHUH
SURSHU XUEDQ SODQQLQJ ZDV ODFNLQJ :LWK WKH LQWURGXFWLRQ RI
municipal reforms in 1998, the city of Maputo was divided into
two municipalities namely, Maputo and Matola. Furthermore,
the administrative boundaries of the newly created municipalities were re-drawn and most of the neighbourhoods previously
considered as informal settlements now became part of the new
municipalities.
Without proper planning and investments, however, most of
the new neighbourhoods today face severe limitations concerning access to adequate municipal/public services such as water,
sanitation and electricity. When it comes to piped water supply, while the residents of the neighbourhoods located near the
‘cement city’ can still access the piped grid through overstretchLQJ WKH IRUPDO QHWZRUN WKH UHVLGHQWV RI WKH RXWHU QHLJKERXUKRRGVIDFHPRUHUHVWULFWLRQVLQSLSHGZDWHUDFFHVVGXHWRODFN
RISUHVVXUHLQWKHQHDUE\JULGRUEHFDXVHWKHQHWZRUNFDQQRWEH
VXI¿FLHQWO\H[WHQGHGWRUHDFKWKHLUQHLJKERXUKRRGV
These problems have not only led to enormous disparities
concerning the access to piped water supply but also prompted
the emergence of alternative service providers (e.g. SSIPs) who
presently constitute the most reliable sources of water for the
PDMRULW\RIWKHXQGHUVHUYHGXUEDQSRRU7RGD\URXJKO\RI
households in Maputo rely either on SSIPs (32%) or other types
RIDOWHUQDWLYHVRXUFHVIRUWKHLUZDWHUVXSSO\FRPSDUHGWR
WKHURXJKO\RIKRXVHKROGVVXSSOLHGWKURXJKWKHIRUPDOQHWZRUN*XPER6HXUHFD+\GURFRQVHLO
Pumping tests and water quality measurements
Pumping tests
%RUHKROHSXPSLQJWHVWVZHUHSHUIRUPHGWRDVVHVVWKHK\GURJHRlogical potential of the aquifer system used by SSIPs. A total of
10 pumping tests were performed on an equal number of boreholes distributed within the study area. Attempts were made to
establish an evenly distributed grid of test boreholes covering
the entire study area. Some problems arose while performing
WKLVWDVNLH
‡ Access limitations 2ZQHUV ZHUH UHTXHVWHG WR LQWHUUXSW
WKHLU VHUYLFHV WHPSRUDULO\ WR DOORZ IRU LQVWDOODWLRQ RI ÀRZ
meters and other equipment required for the tests, and to run
WKHDFWXDOWHVWV'XHWRSRVVLEOHGLVWXUEDQFHVWRWKHLUVHUYices, some owners were not willing to participate; thus, new
ERUHKROHV KDG WR EH LGHQWL¿HG QHDU WR SUHYLRXVO\ VHOHFWHG
boreholes in order to maintain the desired level of coverage
of the study area.
‡ Poor system condition. In many cases the pipe casing
ZDV SRRUO\ GRQH PDNLQJ LW GLI¿FXOW DQG FRVWO\ WR LQVWDOO
the equipment for running the pumping test. Therefore, all
tests were carried out using pumps already installed in the
boreholes. This also limited the possibilities of running the
tests with a three-stage pumping rate as is common practice.
Instead, a single-stage pumping rate followed by a recovery
test was used.
‡ Reliability of power supply. The area suffered from frequent power outages. The borehole locations thus had to
Available on website http://www.wrc.org.za
ISSN 0378-4738 = Water SA Vol. 34 No. 3 July 2008
ISSN 1816-7950 = Water SA (on-line)
MATOLA GARE
MATOLA GARE
%
%
%
%
SINGATELA
CICUAMA
%
DHLAVELA
%
% %
%
%
%
%
%
%
B. MACHAVA
%
%
ZONA VERDE
%
%
%
%
%
%
%
%
%
%
%B. MAHOTAS
B. PESCADORES
MAHLAZINE
B. GEORGE DMITROV
%% B. 3 DE FEVEREIRO
%
%
%
% %
% %
% %%
B. DA COSTA DO SOL
%
%
%%
%%B. LUALANE
%
%
%%
HULENE A
%
25 DE JUNHO B
%
%
% % %% %%%
25 DE JUNHO A
%
%
%
ACORDOS DE LUSAKA
%
%
%
B. FERROVIARIO
%
AEROPORTO B % %
INHAGOIA A
%%
HULENE
BAIRRO TRIUNFO
%B
%
%
%
%
%
%MACHAVA VALE DE INFULENE
%
INHAGOIA B
% %
B
%
% %MAVALANE
% %%
%%
%%F.P.L.M.
BAIRRO DO JARDIM
%
%%
MAVALANE
A
%
%
%
% % %
% %
%
BAGAMOIO
MACHAVA CENTRAL
B. DA LIBERDADE
%
B. DA MATOLA J
TREVO
B. DA MATOLA H
%
B. PATRICE LUMUMBAINFULENE UNIDADE D
MOSSUMBULUKU
B. UNIDADE 7
FOMENTO
MATOLA B
%
%SAO DAMASSO
%
%
%
% %
%%
ALBAZINE
%
%
%
%
%
%
%
%%
%
MAGOANINE
%%
%
%
%
%
MAXAQUENE
% D
B. LUIS CABRAL
POLANA CANICO B
AEROPORTO A
%
B. DA URBANIZACAOMAXAQUENE A
MAXAQUENE BPOLANA CANICO A
CHAMANCULO C
CHAMANCULO DXIPAMANINE MUNHUANA MAXAQUENE C
B. DA MATOLA D MATOLA F & G
%
%
%
KM 15
TSALALA
%
%
CONGOLOTE
ZIMPETO
%
CHAMANCULO B MINKADJUINE
B. MALANGA
MAFALALA
CHAMANCULO AMALHANGALENE B COOP
%
%
SOMMERSHIELD
ALTO MAE A
C.F.M
MALHANGALENE A
ALTO MAE B CENTRAL A
B. DA MATOLA C
B. DA MATOLA SEDE
CENTRAL B
/(*(1'
POLANA CIMENTO A
Water quality measurements
®
%
SSIP Systems
AdeM Network
ADMINISTRATIVE BOUNDARY
Maputo Municipality
CENTRAL C
POLANA CIMENTO B
'HVSLWH WKH DERYHPHQWLRQHG FRQstraints, 10 different sites for the
SXPSLQJ WHVWV ZHUH LGHQWL¿HG 7KH
pumping tests consisted of a well
drawdown test with a single stage
pumping rate followed by recovery
tests. Water levels were measured
by divers in all tested wells with an
DFFXUDF\RI“FP'LVFKDUJHUDWHV
ZHUH GHWHUPLQHG XVLQJ D ÀRZ PHWHU
connected to the rising main of each
well. The test results were used to
determine the aquifer transmissivity,
WKHERUHKROHVSHFL¿F\LHOGDQGERUHhole productivity.
B. CUMBEZA
be chosen on the basis of power
UHOLDELOLW\ %RUHKROHV UHSRUWHG WR
EHORFDWHGZLWKLQDQDUHDVXEMHFW
to a high rate of power cuts, as
indicated by the owners, were not
included in the list of test sites.
Matola Municipality
10
5
0
10
Kilometers
7KLUW\¿YH VDPSOLQJ ZHOOV ZHUH
used to assess borehole water qualFigure 2
ity. Samples of borehole water were
Distribution network of greater Maputo and distribution of water systems run by SSIPs
collected and further analysed for
nitrates and bacteria (E. coli and faecal coliforms).
(& HOHFWULFDO FRQGXFWLYLW\ DQG S+ ZHUH PHDVXUHG LQ WKH Demand for services provided by SSIPs
¿HOGXVLQJKDQGKHOGGLJLWDOPHWHUVIURP:DJWHFK,QWHUQDWLRQDO
Ltd., and temperature was measured using a standard mercury
)URP)LJLWLVHYLGHQWWKDWWKHODFNRIIRUPDOVHUYLFHVLQODUJH
type thermometer. EC measurements were carried out with the
areas of peri-urban Maputo has prompted the proliferation of
SXUSRVHRIHYDOXDWLQJWKHLQÀXHQFHRIVHDZDWHULQWUXVLRQRQWKH private service providers who operate either as the sole servquality of borehole water.
ice providers in their neighbourhoods or in competition with
1LWUDWHVDQGEDFWHULDZHUHDQDO\VHGDWWKH$GH0ODERUDWRU\ AdeM. In a survey conducted as part of this study, some 187
following procedures described in Standard Methods (1995).
SSIPs of Maputo and Matola, who were reported to be respon%DFWHULRORJLFDODQDO\VHVZHUHFDUULHGRXWXVLQJWKHPHPEUDQH sible for some 192 small-scale piped systems, were interviewed.
method. Although a total of 35 boreholes were used to investigate
About 84.4% of providers interviewed said they offered services
borehole water quality with respect to nitrates, the results from
WKURXJK KRXVH FRQQHFWLRQV DQG \DUG WDSV ZKLOH RIIHUHG
MXVW ERUHKROHV ZHUH XVHG GXH PDLQO\ WR XQUHOLDEOH UHVXOWV services through standpipes.
obtained from one of the laboratories involved. The nitrate conThe density and distribution of SSIPs in peri-urban areas of
FHQWUDWLRQVUHSRUWHGE\$YLJQRQ+\GURJHRORJLFDO/DERUDWRU\LQ the city more or less follows the population density (see Fig. 2).
France were the results used in this study.
The highest population densities are in the north-eastern part of
WKHFLW\LQWKHQHLJKERXUKRRGVRI$OED]LQH0DKRWDV+XOHQH
Results
Laulane and Mavalane) and in the new expansion zones located
QRUWKZHVWRIWKHFLW\QDPHO\LQ=LPSHWR1GKODYHOD%XQKLoD
Present condition of peri-urban water supply
Singatela and Tsalala.
services
The quality of services provided by SSIPs is highly appreciated by residents of peri-urban areas; therefore, the demand for
The Maputo water supply system is presently run by AdeM,
WKHLU VHUYLFHV KDV DOZD\V EHHQ KLJK 7KLV KDV EHHQ FRQ¿UPHG
a private operator rendering services through a 15-year lease
WKURXJKVWXGLHVFDUULHGRXWIRU&5$LQ6HXUHFD+\GURVLJQHGLQDVSDUWRIWKHLPSOHPHQWDWLRQRIWKHIUDPHZRUN FRQVHLO DQG IXUWKHU XSGDWHG LQ 6DORPRQ for delegated management of water supply (Zandamela, 2002;
these studies indicated that more than 75% of consumers inter*XPERHW DO*XPER$VDFRQVHTXHQFH),3$* YLHZHGZHUHVDWLV¿HGZLWKWKHVHUYLFHVRIIHUHGE\66,3V
(Fund for the Investment and Property of Water supply) was creDWHG WR WDNH RYHU DV RZQHU RI WKH ¿[HG DVVHWV RI ZDWHU VXSSO\ SSIPs source water potential
RI PDMRU FLWLHV RI WKH FRXQWU\ DOVR &5$ WKH :DWHU 6XSSO\
5HJXODWLRQ&RXQFLOZDVFUHDWHGIRUUHJXODWLRQRISULYDWHVHFWRU Geological settings
FRQWUDFWVZLWKLQWKHIUDPHZRUN7KHOHDVHDUHDRI$GH0VSDQV
across the entire area of the two municipalities while the existThe study area is part of the large Meso-Cenozoic sedimentary
LQJQHWZRUNLVOLPLWHGWRWKHPRVWXUEDQLVHGDUHDV)LJ
EDVLQZKLFKFRYHUVWKHDUHDVRXWKRIWKH6DYH5LYHUDQGLVUHODWHG
Available on website http://www.wrc.org.za
ISSN 0378-4738 = Water SA Vol. 34 No. 3 July 2008
ISSN 1816-7950 = Water SA (on-line)
413
(discharge for a 1 m3/h yield) values calculated for the 10 tested
boreholes are also presented in Table 1. These were further compared to historical data of the same boreholes and also from
nearby boreholes in order to identify possible yield reductions
ZLWKWLPH$FFRUGLQJWRWKHUHVXOWVLQ7DEOHHVWLPDWHGVSHFL¿F
yields for all tested boreholes were found to be within the range
of 0.2 to 4.0 m3/h·m; these are rather high values for sand aquifers and for low-cost boreholes probably constructed without a
JUDYHOSDFN7KHORQJWHUPSURGXFWLYLW\RIWKHWHVWHGERUHKROHV
FDOFXODWHGIURPWKHDTXLIHUWUDQVPLVVLYLW\DQGERUHKROHVSHFL¿F
yields is also shown in Table 1. An estimated borehole lifetime of
approximately 3 years and a continuous pumping rate of 24 h/d
with a maximum drawdown tolerance of ± 10 m were assumed.
The results indicate that all tested boreholes can be exploited to
DKLJKHU\LHOGWKDQWKHSUHVHQWRQHZLWKRXWUXQQLQJLQWRULVNVRI
over-exploitation of the aquifer system. The total potential yield
for the 10 tested boreholes was 1 772 m3/d. Eighty per cent of the
tested boreholes could be exploited within the range of 100 to
300 m3/d.
to the rift system between Madagascar and Africa. This system
H[WHQGVIURP3RUW'XQGIRUGLQ6RXWK$IULFDWR4XHOLPDQHLQWKH
central part of Mozambique. Karoo basalts and rhyolites, dated
3HUPLF DQG -XUDVVLF IRUP WKH EDVHPHQW RI WKH V\VWHP &UHWDFHRXVWR7HUWLDU\ÀDWGHSRVLWVRUGHSRVLWVZLWKQHDUO\KRUL]RQWDO
slopes overlay the Karoo sediments. These deposits are mostly
of marine origin and were formed during transgression periods.
Sand dunes or quaternary sand deposits cover the entire study
area.
Characterisation of the aquifer system
The aquifer system of the region of Maputo is divided into two
PDMRU XQLWV WKH VDQG\ DTXLIHU RU SKUHDWLF DTXLIHU DQG D GHHS
aquifer of sandstones and limestone with fresh- water, regarded
DVKDYLQJJRRGK\GUDXOLFSRWHQWLDO%XUJHDS$WDORFDO
level, the aquifer potential is substantially different for the two
aquifers; however, the deep aquifer is the best in terms of strategic groundwater exploitation. The separation between the two
DTXLIHUVLVQRWFOHDUO\GH¿QHGEXWIRUODUJHVFDOHH[SORLWDWLRQRI
groundwater, the two aquifers can be regarded as a single unit
DFFRUGLQJWR¿QGLQJVRI,:$&2ODWHUFRQ¿UPHGE\VWXGLHV E\ ,:$&2 -XL]R DQG 6ZHFR $VVRFLDWHV
(2004).
Water quality assessment
Microbial contamination
$WRWDORIVDPSOHVZHUHWDNHQIURPZDWHUSURYLGHGE\66,3V
DQGDQDO\VHGIRUPLFURELDOFRQWDPLQDWLRQ5HVXOWVRIEDFWHULRlogical analysis for those sites where bacteria counted in excess
of recommended guidelines are resumed in Table 2. Accordingly, 14 out of the 35 sites investigated revealed total coliform
counts above prescribed guidelines for human consumption (<3
FIXPƐZKLOHDWIRXUVLWHVWRWDOFRXQWVRIE. coli and faecal
coliforms indicated moderate to high levels of contamination by
faecal bacteria. In these, faecal bacteria counts were reported to
UDQJHIURPDVORZDVFIXPƐWRDVKLJKDVFIXPƐ
while E. coli was found to be present at least at three sites.
The sites with the highest incidence of contamination with
faecal bacteria were located in two neighbourhoods north-west
RI WKH VWXG\ DUHD 6LJDQWHOD DQG 6mR 'DPDVR EXW DOVR LQ RQH
neighbourhood located north-east of the study area (Albazine).
The results of the microbial analysis with respect to total coliform counts were further mapped (Fig. 3) to indicate the spatial dispersion of contaminated boreholes within the study area.
Accordingly, the borehole at Singatela was the one showing the
most critical situation in terms of bacterial contamination.
Groundwater potential
The results of the borehole pumping tests are presented in
Table 1. The pumping test results were interpreted using the
JUDSKLFDO PHWKRG RI -DFRE .UXHVPDQ DQG 5LGGHU 7KH
REMHFWLYHZDVWRLQIHUWKHIROORZLQJSDUDPHWHUVFKDUDFWHULVWLFRI
the groundwater potential: aquifer transmissivity, borehole speFL¿F\LHOGVDQGERUHKROHSURGXFWLYLW\
For all tested boreholes, the plotted drawdown curve
DGMXVWHG WR WKH VHPLORJ ODZ 7DEOH 7KLV LQGLFDWHV WKDW WKH
DTXLIHU V\VWHP VKRXOG EH FODVVL¿HG DV D SRURXV PHGLD DTXLIHU
in which fractures do not play an important role in the overall
aquifer permeability. The aquifer transmissivity was found to be
between 40 and 520 m 2/d, which is a reasonably high value for a
single sandy aquifer with a depth to the water table of between
10 and 30 m (Table 1).
%RUHKROHVSHFL¿F\LHOGGHSHQGRQWKHDTXLIHUVHWXSEXWDOVR
on the borehole equipment (length and other characteristics of
¿OWHUVFUHHQVDQGWKHTXDOLW\RIWKHJUDYHOSDFN6SHFL¿F\LHOG
TABLE 1
Productivity and potential yield of tested boreholes
Ref. site
Site 1
Site 2
Site 3
Site 4
Site 5
6LWH
Site 7
Site 8
Site 9
Site10
Total
depth (m)
25
30
55
30
30
48
35
48
44
45
Present yield (m3/d)
a
414
Depth to
water table
(m)
Discharge
testing
(m3/d)
9.9
12.3
17.3
27.4
27.1
18.5
32
74
77
55
59
28
89
30
575
Measured
drawdown
¨6P
Spec. yield
103 sec
(S1000s)
Trans-missivity
(m2/d)
5.35
1.43
3.58
2.81
0.24
2.50
1.07
0.75
44
85
113
115
122
147
232
243
332
518
0.133
0.137
0.120
0.123
0.083
0.073
0.022
0.017
0.023
Potential yield (m3/d)
a
Qexpl 10 m
(m3/d)
41
145
245
104
130
304
211
131
355
1 772
Safe yield for 10 m drawdown after 3 years of continuous pumping
Available on website http://www.wrc.org.za
ISSN 0378-4738 = Water SA Vol. 34 No. 3 July 2008
ISSN 1816-7950 = Water SA (on-line)
TABLE 2
Microbial contamination analysis results of boreholes from which positive counts were reported
Ref. site
Location
%DFWHULDFRXQWVFIXPƐ
Total coliforms
Site 1
Site 2
Site 3
Site 4
Site 5
6LWH
Site 7
Site 8
Site 9
Site 10
Site 11
Site 12
Site 13
5RPmR
Albazine
Abazine
Albazine
Albazine
Zimpeto
'ODYHOH
'ODYHOH
6'DPDQVR
Singatela
Tsalala
M. Socimol
Khongolote
9
4
43
4
15
39
20
150
240
4
23
4
Guideline standard (MISAU, 2004)
Faecal coliforms
0
0
9
28
0
0
0
0
28
240
0
0
0
E. coli
0
0
0
4
0
0
0
0
1
13
0
0
0
Total coliforms
<3
Faecal coliforms
0
Figure 3
Bacteriological contamination analysis
results (total coliforms)
of the tested boreholes
Contamination with nitrates
To assess the degree of borehole water contamination with
nitrates, results from 12 tested boreholes were used. The locations of sampled boreholes were mapped in Fig. 4. Accordingly,
nitrate concentrations in all sampled boreholes were rather low
DQGEHORZWKHWDUJHWOLPLWRIPJƐVHWLQWKHJXLGHOLQHVIRU
GULQNLQJZDWHUTXDOLW\:+25HSRUWHGFRQFHQWUDWLRQV
UDQJHGIURPYDOXHVDVORZDVPJƐWRDPD[LPXPRIPJƐ
for all tested wells.
The highest nitrate concentrations were reported in boreholes
located in areas long established as residential areas (suburbs) of
the city of Maputo and thus having moderate to high population
densities. These are areas where sanitation is mostly provided
by means of cesspits and dry-pit latrines. The area north of the
Available on website http://www.wrc.org.za
ISSN 0378-4738 = Water SA Vol. 34 No. 3 July 2008
ISSN 1816-7950 = Water SA (on-line)
dotted line in Fig. 4 shows the new expansion zones of the cities of Maputo and Matola, and is therefore characterised by low
SRSXODWLRQGHQVLWLHV+HUHVDQLWDWLRQLVDOVRRIIHUHGLQWKHIRUP
of cesspits and pit latrines. The analysis of the spatial distribution of nitrate concentrations as indicated in Fig. 4 suggested the
existence of three distinct areas, namely:
‡ $QDUHDZLWKWKHORZHVWOHYHOVRIQLWUDWHVPJƐORFDWHG
mainly in neighbourhoods with low population densities (the
northern part of the study area, consisting of the neighbourKRRGVRI&RQJRORWH=LPSHWR6LQJDOHWD6mR'DPDVVRDQG
KM 15). This also includes areas reported to have a rather
shallow groundwater table where the unsaturated zone is
relatively thin.
‡ $Q DUHD ZLWK PHGLXP OHYHOV RI QLWUDWHV WR PJƐ
located mainly in neighbourhoods with moderate to high
415
B. CUMBEZA
GUAVA
!(
MATOLA GARE
CONGOLOTE
!(
ZIMPETO
!(
!(
!(
ALBAZINE
MAGOANINE
KM 15
SINGATELA
DHLAVELA
!(
(! !(
!(
SAO DAMASSO
!(
B. MAHOTAS
MAHLAZINE
B. GEORGE DMITROV
B. MACHAVA
B. 3 DE FEVEREIRO
ZONA VERDE
BAGAMOIO
B. DA COSTA DO SOL
B. PATRICE LUMUMBAINFULENE UNIDADE D
B. LUALANE
HULENE A
25 DE JUNHO B
25 DE JUNHO A
MOSSUMBULUKU
MACHAVA CENTRAL
ACORDOS DE LUSAKA
INHAGOIA A
B. DA LIBERDADE
MACHAVA
HULENE B
VALE DE INFULENE
MAVALANE A
TREVO
B. UNIDADE 7
FOMENTO
MATOLA B
B. LUIS CABRAL
BAIRRO TRIUNFO
MAVALANE B
BAIRRO DO JARDIM
B. DA MATOLA H
F.P.L.M.
MAXAQUENE D
POLANA CANICO B
AEROPORTO A
B. DA URBANIZACAO MAXAQUENE A
CHAMANCULO CCHAMANCULO D MAXAQUENE BPOLANA CANICO A
XIPAMANINEMUNHUANA MAXAQUENE C
B. DA MATOLA DMATOLA F & G
CHAMANCULO BMINKADJUINE
B. MALANGA
MAFALALA
CHAMANCULO AMALHANGALENE B COOP
SOMMERSHIELD
ALTO MAE A
C.F.M
MALHANGALENE A
ALTO MAE B CENTRAL A
B. DA MATOLA C
B. DA MATOLA SEDE
Figure 4
Nitrate concentraWLRQVPJƐZLWKLQ
the study area
B. FERROVIARIO
AEROPORTO B
INHAGOIA B
B. DA MATOLA J
B. PESCADORES
CICUAMA
!(
!(
TSALALA
CENTRAL B
POLANA CIMENTO A
CENTRAL C
®
5
0
population densities (namely the north-eastern neighbourhoods of Magoanine and Albazine), in the aquifer of which a
UHODWLYHO\WKLFNXQVDWXUDWHG]RQHDQGGHSWKWRWKHZDWHUWDEOH
are reported to exist.
‡ $QDUHDZLWKWKHKLJKHVWQLWUDWHOHYHOVWRPJƐORFDWHG
PDLQO\ LQ GHQVHO\ SRSXODWHG QHLJKERXUKRRGV 'ODYHOD DQG
Tsalala).
Electrical conductivity and water salinity
Sea-water intrusion is a very common problem when it comes
to groundwater extracted in coastal areas. When aquifers are
depleted by high-yield exploitation, sea-water intrusion may
occur and the quality of water with respect to its taste is thus
compromised. When sea water invades coastal aquifers, the
LQWHUIDFHSRVLWLRQEHWZHHQIUHVKDQGEUDFNLVKZDWHULVDIXQFWLRQ
of aquifer properties, pumping rate and aquifer recharge potential. In order to assess the levels of salinity of borehole water
used by SSIPs, EC measurements were carried out on samples
from 35 boreholes located further north of the study area. The
results were mapped and are presented in Fig. 5. Accordingly,
only 3 sites were reported to show EC levels in excess of the
JXLGHOLQHIRUGULQNLQJZDWHURIPJƐDQGKHQFHSRWHQWLDOO\
WRFRQWDLQEUDFNLVKZDWHU7ZRVLWHVZHUHORFDWHGQRUWKZHVWRI
the study area while the third was located to the north-east.
Discussion of results
%HFDXVHRIWKHIRUPDOZDWHUQHWZRUNQRWUHDFKLQJPRVWRISHUL
urban Maputo, SSIPs have, over the past few years, become an
integral part of the supply chain of services to the suburbs of
greater Maputo. They play the predominant role in service provision to such areas, and the quality of their services is highly
appreciated by consumers. Most independent providers operat-
416
10
Kilometers
POLANA CIMENTO B
10
ing in greater Maputo are currently unregulated, but important
VWHSV KDYH DOUHDG\ EHHQ WDNHQ WR FKDQJH WKDW VLWXDWLRQ 7KHVH
LQFOXGHVWXGLHVFRPPLVVLRQHGE\&5$IRUWKHGHYHORSPHQWRI
UHJXODWRU\ WRROV WR IUDPH WKHLU DFWLYLWLHV DQG WKH MRLQW HIIRUWV
E\&5$DQGPXQLFLSDODXWKRULWLHVIRUWKHIRUPDOLVDWLRQRILQGHSHQGHQWSURYLGHUV2(&'66,3VKDYHDOVREHHQLQFOXGHG
DVNH\SDUWQHUVLQWKHUHFHQWO\ODXQFKHGµ0DSXWR:DWHU6XSSO\
3URMHFW¶7KHGHPDQGIRUVHUYLFHVSURYLGHGE\66,3VLVWKHUHIRUH
QRW OLNHO\ WR GHFUHDVH LQ WKH QHDU IXWXUH 4XDOLW\ DQG TXDQWLW\
SUREOHPVDVVRFLDWHGZLWKIXWXUHH[SDQVLRQDUHOLNHO\WRFRQVWLtute the main challenges for the long-term sustainability of SSIP
activities.
The typical design of water systems run by SSIPs is based
on groundwater abstraction. Consequently, the potential of the
UHJLRQDODTXLIHUV\VWHPLVDNH\HOHPHQWLQWKHORQJWHUPSODQning of intervention of SSIPs. From the results of borehole testing presented in Table 2, it is clear that most boreholes used by
66,3VDUHRIOLPLWHGGHSWKPDQGDOVRWKDWWKH\WDSZDWHU
at relatively shallow depths to the water table (< 30 m), probably from within the sandy aquifer. The pumping test results
also indicate that present yields are generally lower than estimated potential yields (0.2 to 0.4 m3/h.m); consequently, boreKROHZDWHUFDQEHH[SORLWHGDWHYHQKLJKHU\LHOGVZLWKRXWULVNLQJ
over-exploitation of the aquifer system. This implies that, within
reasonable limits of expansion of small-scale service providers
DQGSURSHUORFDWLRQRIERUHKROHVWKHULVNVRIDQHYHQWXDOGHSOHWLRQRIWKHDTXLIHUV\VWHPDUHVPDOO$OVREHFDXVHWKHWZRPDMRU
units comprising the aquifer system of Maputo can be regarded
DVRQHXQLW,:$&2,:$&2-XL]R6ZHFR
Associates, 2004), the yield properties in case of larger abstraction rates will also be determined by the deeper aquifer, which
is regarded as having better hydraulic properties and being suitable for large-scale exploitation. In fact, the deeper aquifer is
regarded as the main supplementary source for the water supply
Available on website http://www.wrc.org.za
ISSN 0378-4738 = Water SA Vol. 34 No. 3 July 2008
ISSN 1816-7950 = Water SA (on-line)
GUAVA
!(
!( !(
!( !(
!(
!( !( !(
CONGOLOTE
ZIMPETO
!(
!(
ALBAZINE
MAGOANINE
KM 15
SINGATELA
(!
!( !(
!(
TSALALA
DHLAVELA
!( !(
CICUAMA
!(
SAO DAMASSO
!( !(
!(
B. MACHAVA
(!
!( !(
!(
B. PESCADORES
B. MAHOTAS
MAHLAZINE
B. GEORGE DMITROV
B. 3 DE FEVEREIRO
ZONA VERDE
BAGAMOIO
B. DA COSTA DO SOL
B. PATRICE LUMUMBAINFULENE UNIDADE D
B. LUALANE
HULENE A
25 DE JUNHO B
25 DE JUNHO A
MACHAVA CENTRAL
ACORDOS DE LUSAKA
INHAGOIA A
B. DA LIBERDADE
MACHAVA
HULENE B
B. UNIDADE 7
FOMENTO
B. LUIS CABRAL
F.P.L.M.
MAXAQUENE D
POLANA CANICO B
AEROPORTO A
B. DA URBANIZACAOMAXAQUENE A
CHAMANCULO CCHAMANCULO D MAXAQUENE BPOLANA CANICO A
XIPAMANINE MUNHUANA MAXAQUENE C
B. DA MATOLA DMATOLA F & G
CHAMANCULO B MINKADJUINE
B. MALANGA
MAFALALA
CHAMANCULO AMALHANGALENE BCOOP
SOMMERSHIELD
ALTO MAE A
C.F.M
MALHANGALENE A
ALTO MAE B CENTRAL A
B. DA MATOLA C
MAVALANE A
TREVO
MATOLA B
BAIRRO TRIUNFO
MAVALANE B
BAIRRO DO JARDIM
B. DA MATOLA H
B. FERROVIARIO
AEROPORTO B
VALE DE INFULENE
INHAGOIA B
B. DA MATOLA J
MOSSUMBULUKU
B. DA MATOLA SEDE
!(
!(
CENTRAL B
POLANA CIMENTO A
CENTRAL C
®
!(
!(
!(
MATOLA GARE
Figure 5
EC levels within the
VWXG\DUHDȝ6FP
The dotted red line
represents the 1 200
ȝ6FPERXQGDU\$OO
areas located north
of the dotted line are
suitable for borehole
installation.
B. CUMBEZA
POLANA CIMENTO B
10
5
0
10
Kilometers
to the city of Maputo, particularly to the neighbourhoods located
north of the city centre.
The results of groundwater quality assessment suggest that,
as for today, the quality of water tapped by SSIPs generally conIRUPVWRJXLGHOLQHVWDQGDUGV0,6$8:+2ZLWK
respect to the parameters investigated and therefore poses no
constraints on use at domestic level. Yet, none of the independent providers said they treated the water, except when instructed
due to problems detected during monitoring intermittently conGXFWHGE\WKH0LQLVWU\RI+HDOWKYLDLWV)RRG+\JLHQH'HSDUWPHQW,QIDFWWKH0LQLVWU\RI+HDOWK0,6$8KDVLQLWVGDWDEDVH
about 220 boreholes, the water quality of which is to be regularly
monitored; 83 of these are private boreholes run by SSIPs.
The analyses of microbiological quality indicate that, with
IHZH[FHSWLRQVZDWHUIURPWKHPDMRULW\RISULYDWHO\RZQHGV\Vtems is virtually free from faecal contamination, as proven by
the absence of E. coli as well as faecal coliforms in more than
90% of samples investigated. This suggests that as for today,
there is no widespread bacterial contamination of the aquifer
system used by SSIPs.
Most SSIPs are located within moderately to densely populated residential areas, where sewers do not exist and sanitation
LV PDLQO\ SURYLGHG WKURXJK VHSWLF WDQNV FHVVSLWV DQG GU\SLW
latrines; therefore, seepage from on-site sanitation represents
the most widespread and serious source of pollution (both point
and diffuse) to the aquifer system. Since pathogens can survive
for many days while percolating the unsaturated strata and evenWXDOO\WKURXJKWKHDTXLIHUV\VWHP6XJGHQWKHPDMRUFRQcern regarding poor sanitation is direct migration of pathogenic
bacteria and viruses to underlying aquifers and neighbouring
groundwater sources (Lewis et al., 1981).
7KHH[WHQWDQGULVNRIJURXQGZDWHUFRQWDPLQDWLRQGHSHQG
however, on many factors, one of which is the degree of pathogen attenuation during percolation through the unsaturated
Available on website http://www.wrc.org.za
ISSN 0378-4738 = Water SA Vol. 34 No. 3 July 2008
ISSN 1816-7950 = Water SA (on-line)
zone and, eventually, through the aquifer system. Some studies
6XJGHQ$UJRRV6FKPROOHWDOLQGLFDWHWKDW
pathogen attenuation is generally most effective in the uppermost layers of the unsaturated zone, where biological activity
is greatest and where the presence of protozoa and predatory
organisms, rapid changes in soil moisture and temperature as
well as competition from established microbial communities
help reduce the level of pathogens in that particular zone and,
consequently, in the groundwater.
The distance between the base of pit latrines and water
WDEOH DOVR NQRZQ DV VRLO LQ¿OWUDWLRQ OD\HU SOD\V DQ LPSRUWDQW
UROHLQWKHGLHRIIRISDWKRJHQV$VQRWHGE\6XJGHQDQG
6FKPROOHWDODVRLOLQ¿OWUDWLRQOD\HUGHSWKRIEHWZHHQ
DQGPLVVXI¿FLHQWIRUUHGXFLQJFRQWDPLQDWLRQRIWKHZDWHU
table by pathogens to acceptable levels.
%HFDXVH PRVW 66,3V WDS WKHLU ZDWHU DW GHSWKV WR WKH ZDWHU
table of greater than 10 m and pit latrines are built with depths of
WRPVXI¿FLHQWGHSWKLVPDLQWDLQHGWKURXJKWKHXQVDWXrated zone to prevent pathogens from reaching the groundwater
table. The horizontal distance between pit latrines and water
wells will also have an impact on levels of contamination.
The soil type and permeability characteristics of the unsaturated zone and aquifer system are also important factors, since
they impact residence times in the unsaturated zone and prevent
WKHSDVVDJHRIEDFWHULDOSDWKRJHQVLQWRWKHVXEVXUIDFH+LJKVRLO
SHUPHDELOLW\ LV DVVRFLDWHG ZLWK KLJK ULVNV RI SDWKRJHQV UHDFKing the groundwater table. According to the literature (Schmoll,
/HH DQG %DVWHUPHLMHU UHVLGHQFH WLPHV RI DERXW D
month are long enough to free groundwater sources from pathogens naturally. Long residence times in the unsaturated zone are
DVVRFLDWHG ZLWK ORZ ULVNV RI JURXQGZDWHU FRQWDPLQDWLRQ 6HDsonal variations of the groundwater table, e.g. due to rainfall,
LQFUHDVH WKH ULVN RI SDWKRJHQV VHHSLQJ LQWR WKH JURXQGZDWHU
*RGIUH\ HW DO 3XMDUL DQG +RZDUG HW DO 417
KDYH DOO UHSRUWHG VLJQL¿FDQW LQFUHDVHV LQ IDHFDO FRQWDPLQDWLRQ
of groundwater following rainfall events.
Construction and completion details of boreholes are also
FUXFLDOIDFWRUVLQWKDWWKH\PD\LQFUHDVHWKHULVNRIJURXQGZDWHU
contamination by creating localised pathways for ingression of
SDWKRJHQV6FKPROO*RGIUH\RUE\VKRUWHQLQJWKH
distance and time required for pathogens to reach the groundZDWHU WDEOH $UJRRV 7KH GHHSHU WKH ¿OWHU VFUHHQ IRU
example, the longer the time required for pathogens to reach the
aquifer system and the higher the die-off rate.
The results of the geohydrological characterisation of the
aquifer system used by SSIPs in Maputo suggest that the aquifer system does not have any natural barrier against pollution;
this means that contaminants disposed of on the surface or seeping from pit latrines can percolate relatively freely across the
unsaturated stratum and readily reach the groundwater system.
The results of this study indicate, however, that the incidence of
contaminated wells is rather limited, which suggests that either
the load of contaminants is rather limited or that the ingress of
SDWKRJHQVLVVWLOOEHLQJVXI¿FLHQWO\DWWHQXDWHGLQWKHXQVDWXUDWHG
strata underlying the aquifer system. Yet, 13 out of the 35 tested
boreholes had either total coliform or faecal coliform levels
KLJKHUWKDQWKH:+2VWDQGDUGZKLFKVXJJHVWWKDWRWKHUSDWKVRI
contamination may probably have occurred.
The most critical situation with respect to faecal contamination was found to be related to an open hand-dug well (in São
'DPDQVRZKHUHIDHFDOFROLIRUPFRXQWVDVKLJKDVFIX
PƐZHUHUHSRUWHG2SHQKDQGGXJZHOOVDUHXVXDOO\FRQVWUXFWHG
in areas with relatively shallow groundwater tables, where the
proximity of the groundwater table facilitates the transport of
contaminants; consequently, contamination with bacteria could
be widespread. For example, a study conducted in rural ZimEDEZH ']ZDLUR HW DO KDV VKRZQ WKDW SLW ODWULQHV FRQstructed at less than 2.0 to 3.0 m above the water table affected
WKHJURXQGZDWHUTXDOLW\DWODWHUDOGLVWDQFHVRIXSWRP+RZever, it appears as though this is not necessarily true; in the very
few cases where contamination with bacteria was reported, the
problem appeared to be related to borehole construction and siting rather than to percolation of contaminants. This is in line
ZLWK ¿QGLQJV IURP D VWXG\ E\ +RZDUG HW DO RQ ZDWHU
quality variations in shallow protected springs in Kampala,
ZKR FRQFOXGHG WKDW LPSURYLQJ WKH VDQLWDU\ ¿QLVKLQJ RI ZHOOV
and of local environmental hygiene is more important to protect
groundwater quality than controlling wide-spread construction
of on-site sanitation facilities. The situation of the contaminated
ZHOOVLGHQWL¿HGLQWKLVVWXG\VKRXOGWKHUHIRUHQRWEHFRQVLGHUHG
as representative of all SSIPs located within the study area, nor
as a common critical aspect of this type of water systems.
The area located north-east of the city (Albazine) where
moderate to high levels of bacterial contamination were also
reported, is a densely populated area. The depth to the water
table, as measured in nearby boreholes where pump tests were
SHUIRUPHGVLWHVVXJJHVWVWKDWUDWKHUWKLFNVWUDWDRIWKH
unsaturated zone exist (> 25 m). The problem here is attributed
either to poor borehole construction or high hydraulic load from
pit latrines. Improper location of pit latrines may also be a contributing factor. The area also experiences high recharge conGLWLRQVDQGUHODWLYHO\JRRGSHUPHDELOLW\UDWHV,:$&2
hence, there is a high potential for bacterial contamination due
to percolation. Low living standards, poor borehole construction
and high hydraulic contaminant loads are therefore the most
probable causes of the high incidence of bacterial contamination
reported in these boreholes.
The presence of nitrates is a common groundwater prob-
418
OHP :+2 $OWKRXJK QLWURJHQ PD\ RFFXU QDWXUDOO\ LQ
groundwater, the main sources of groundwater pollution are
human activities such as agriculture, sanitation (pit latrines)
and accumulation of organic material from improper solid waste
KDQGOLQJ%RXOGLQJDQG*LQQ6FKPROOHWDO:+2
7KH:+2JXLGHOLQHIRUWKHSUHVHQFHRIQLWUDWHVLQGULQNLQJZDWHULVPJƐ7KHSHULXUEDQDUHDRI0DSXWRLVODFNLQJ
formal waste collection services and sanitation is mainly proYLGHGE\GU\SLWODWULQHV'ZHOOHUVRQWKHRWKHUKDQGXVXDOO\
burn their waste in the garden, thereby reducing the source of
diffuse pollution from solid waste handling and leaving seepage
from pit latrines as the main source of nitrogen that could potentially pollute the groundwater.
+LJK OHYHOV RI QLWUDWH DUH D PDMRU SUREOHP IRU ERWWOHIHG
LQIDQWV DV WKH ULVN RI PHWKDHPRJORELQDHPLD LQFUHDVHV ZKHQ
QLWUDWHFRQFHQWUDWLRQVULVHDERYHPJƐ:+27KRPSson et al., 2007). Factors such as hydraulic loads, soil type and
depth to the water table determine the rate and extent of nitrate
transport into the groundwater. Sandy soils are particularly
vulnerable to nitrate leaching into the groundwater because of
the limited attenuation they provide (Thompson et al., 2007).
5DLQIDOODOVRDIIHFWVQLWUDWHWUDQVSRUWLQWRDTXLIHUV\VWHPV,IWKH
ZDWHUWDEOHLVWRRVKDOORZWKHUHLVDJUHDWULVNRIKLJKFRQFHQWUDtions of nitrate occurring after a relatively short time, particularly after heavy rainfall.
The results of nitrate analyses for sampled boreholes (Fig. 4)
VXJJHVWWKDWQLWUDWHOHYHOVZHUHDOZD\VEHORZWKH:+2JXLGHOLQHYDOXHRIPJƐIRUGULQNLQJZDWHU$FFRUGLQJO\WKUHHGLVtinct areas can be distinguished:
The area north of the study area where the lowest nitrate
FRQFHQWUDWLRQVPJƐZHUHUHSRUWHGFRYHUVPDLQO\QHLJKbourhoods with low population densities and thus with relatively
low hydraulic loads potentially harmful to the groundwater. It
includes zones where the unsaturated zone is rather thin (less
than 10 m) and so features a limited capacity for attenuating any
nitrate loads leaching from pit latrines. This leaves hydraulic
loads as the main factor determining the levels of nitrates in the
groundwater, thereby supporting the idea that, presently, levels
of groundwater contamination in this area are determined principally by population density.
7KH DUHD ZLWK PRGHUDWH FRQFHQWUDWLRQV WR PJƐ LV
located mainly in neighbourhoods with moderate to high populaWLRQGHQVLWLHVDQGZKHUHDUHODWLYHO\WKLFNVWUDWXPRIWKHXQVDWXrated zone exists. The moderate levels of nitrates observed in
this area are probably the result of moderate hydraulic loads
from pit latrines, combined with a rather limited capacity of
the unsaturated zone for preventing the ingress of pollutants
due to its relatively high permeability. The area with the highHVWQLWUDWHFRQFHQWUDWLRQVWRPJƐLVORFDWHGLQGHQVHO\
SRSXODWHGQHLJKERXUKRRGV+LJKK\GUDXOLFORDGVIURPFHVVSLWV
DQG SLW ODWULQHV DUH FHUWDLQO\ WKH PDMRU QLWURJHQ FRQWULEXWRU WR
the groundwater system.
The spatial distribution of nitrate concentrations in the study
area thus suggests that population density and the characteristics
of the unsaturated zone are the main factors determining present
nitrate levels in the groundwater. Present contamination levels
were still below critical levels for safe use of the groundwater, a
state of affairs that was believed to result from:
‡ 6WLOO OLPLWHG K\GUDXOLF ORDGV GXH WR ORZ SRSXODWLRQ GHQVLties (< 100 persons/ha) even in areas considered as densely
populated
‡ 5HODWLYHO\ WKLFN VRLO LQ¿OWUDWLRQ OD\HUV LQ DUHDV ZKHUH WKH
PDMRULW\ RI ERUHKROHV ZHUH ORFDWHG DQG ZKHUH ELRORJLFDO
GHQLWUL¿FDWLRQPD\SRVVLEO\RFFXU
Available on website http://www.wrc.org.za
ISSN 0378-4738 = Water SA Vol. 34 No. 3 July 2008
ISSN 1816-7950 = Water SA (on-line)
Figure 6
Nitrate concentrations in the neighbourhoods of
Maputo based on
historical data
An analysis of the spatial distribution of nitrate concentrations based on historical data collected in other, more densely
SRSXODWHG QHLJKERXUKRRGV RI 0DSXWR )LJ LQGLFDWHV WKDW
QLWUDWHOHYHOVLQJURXQGZDWHUFDQEHDVKLJKDVPJƐZKLFK
means that contamination by nitrates is a real threat in some
QHLJKERXUKRRGV 7KLV LV DOVR FRQ¿UPHG E\ UHVXOWV RI VLPLODU
VWXGLHV GRQH LQ RWKHU SDUWV RI WKH VWXG\ DUHD ,:$&2 -Xt]RZKHUHQLWUDWHOHYHOVRIDVKLJKDVPJƐLQWKH
groundwater were reported.
Sea-water intrusion is a common groundwater quality probOHP LQ FRDVWDO DUHDV 7KH :+2 JXLGHOLQH IRU (& LQ GULQNLQJ
water is 1 500 as/cm. Studies with the purpose of assessing the
problem of sea-water intrusion in the Maputo and Matola areas
,:$&2 KDYH FRQFOXGHG WKDW VXFK SUREOHPV DUH PRUH
severe in the south-western part of the city (Matola, Fomento,
Cicuama) than in other parts of the city. The occurrence of
EUDFNLVK ZDWHU KDV KRZHYHU EHHQ UHFRUGHG LQ VRPH RWKHU
areas located north-east of the city and close to the Maputo bay,
QDPHO\ WKH µEDLUURV¶ 3HVFDGRUHV &RVWD GR 6RO DQG +XOHQH $
While the results of the present study suggest that, as for today,
WKHPDMRULW\RIVXUYH\HGVLWHVGRQRWIDFHSUREOHPVZLWKEUDFNish water, the nature of the geological formations suggests that
the area is vulnerable to sea-water intrusion, particularly if the
aquifer system should become over-exploited.
Conclusions
SSIPs have become an integral part of the water services delivery chain to the peri-urban areas of greater Maputo, where they
are reported to reach as much as 32% of the population. They
are presently expanding rapidly to cover new or already established neighbourhoods, not only because of increasing demand
UHVXOWLQJ IURP WKH ODFN RI IRUPDO VHUYLFHV LQ VXFK DUHDV EXW
also because they have recently been recognised as important
partners for expanding service coverage to reach unconnected
residents. The expansion of services with the help of both self¿QDQFHGDQGIRUPDOO\GHOHJDWHGVPDOOVFDOHVHUYLFHSURYLGHUV
however, comes with a number of issues deserving urgent atten-
Available on website http://www.wrc.org.za
ISSN 0378-4738 = Water SA Vol. 34 No. 3 July 2008
ISSN 1816-7950 = Water SA (on-line)
tion, among which their capacity for providing quality water of
VXI¿FLHQWTXDQWLW\LQWKHORQJUXQ
An analysis of the potential of the aquifer system from which
SSIPs tap water to render their services suggests that present
yields in general are lower than estimated safe yields and that
borehole water can be exploited at higher than the present yields
ZLWKRXW UXQQLQJ LQWR ULVNV RI RYHUH[SORLWDWLRQ RI WKH DTXLIHU
system. This implies that, within reasonable limits of expansion and proper location of boreholes, the expected increase in
the number of SSIPs will not lead to an eventual depletion of
WKH H[LVWLQJ DTXLIHU V\VWHP <HW FDUH PXVW EH WDNHQ WR HQVXUH
that quality problems will not arise due to increased abstraction
rates.
So far, the quality of water used by SSIPs to provide services has been virtually free from microbial and organic conWDPLQDWLRQ7ZRPDMRUIDFWRUVFRQWULEXWHWRWKLVQDPHO\OLPLWHG
hydraulic loads due to low population densities and relatively
WKLFNVWUDWDRIWKHXQVDWXUDWHG]RQHZKHUHDWWHQXDWLRQRIFRQWDPLQDQWVVWLOORFFXUVDWVXI¿FLHQWOHYHOV<HWRXWRIWHVWHG
ERUHKROHV KDG WRWDO FROLIRUP OHYHOV DERYH WKH :+2 UHFRPmended standard and 4 had faecal coliforms in excess of the
standard. Low living standards, poor borehole construction and
high hydraulic loads were the probable causes of the high incidence of bacterial contamination in those boreholes.
1LWUDWHFRQFHQWUDWLRQVLQDOOERUHKROHVWHVWHGZHUHEHORZWKH
UHFRPPHQGHGVWDQGDUGRIPJƐ+RZHYHUWKHVLWXDWLRQPD\
change in future, either due to over-exploitation of the aquifer
system or to increased hydraulic loads resulting from increased
population density.
1R SUREOHPV ZLWK EUDFNLVK ZDWHU ZHUH UHSRUWHG LQ ODUJH
parts of the studied area, but in future this could become a serious problem because of the high vulnerability of the aquifer system (coastal aquifer) to sea-water intrusion.
While the potential for development of SSIPs is high, based
on demand and the characteristics of the aquifer system, the
long-term sustainability of their activities require that efforts be
put in place to speed up the already initiated attempts at formalising small-scale independent providers and to establish regula-
419
tory tools to frame their activities. In doing so, pressure must
be put on authorities to establish mechanisms that will ensure
the adoption of more stringent protective measures for boreholes
constructed to provide public services, including:
‡ 5HJXODWHG SURFHGXUHV IRU ERUHKROH GHVLJQ DQG ORFDWLRQ LQ
RUGHUWRPLQLPLVHULVNVRIFRQWDPLQDWLRQ6SHFLDOHPSKDVLV
should be put on aspects such as wellhead protection, posiWLRQLQJRI¿OWHUVFUHHQVDQGWKHORFDWLRQRIERUHKROHVLQUHODWLRQWRH[LVWLQJSLWODWULQHV$PLQLPXPUDGLXVRILQÀXHQFH
25 m away from pit latrines is generally accepted in Mozambique.
‡ Mandatory rules for direct protection of boreholes used as
GULQNLQJZDWHUVXSSO\HJD[PVXUURXQGLQJIHQFH
‡ Mandatory rules for SSIPs regarding chlorination of the
water before distribution.
%HFDXVHPRVWLQGHSHQGHQWSURYLGHUVWHQG WR WDS ZDWHU DW UHODtively shallow depths to minimise investment costs, enforcement rules regarding minimum depths of abstraction could also
be put in place.
2YHUDOOLWFDQEHVDLGWKDW66,3VSURYLGHDYDOXDEOHFRQWULEXWLRQWRRYHUFRPLQJWKHSUREOHPVZLWKGULQNLQJZDWHUVXSSO\
to peri-urban areas experiencing rapid growth; however, it is
imperative that the governmental institutions and already established regulatory bodies put in place mechanisms to reduce the
SRVVLELOLWLHV RI IXWXUH SXEOLF KHDOWK ULVN DVVRFLDWHG ZLWK WKHVH
systems.
References
$5*266 *XLGHOLQHV IRU $VVHVVLQJ WKH 5LVN WR *URXQGZDWHU
IURP2Q6LWH6DQLWDWLRQ%ULWLVK*HRORJLFDO6XUYH\&RPPLVVLRQHG
5HSRUW&5SS
%2/$<-&DQG5$%,129,&+$,QWHUPHGLDWHFLWLHVLQ/DWLQ
$PHULFD ULVN DQG RSSRUWXQLWLHV RI FRKHUHQW XUEDQ GHYHORSPHQW
Cities 21 (5) 407-421. Elsevier.
%28/',1* -5 DQG *,11 -6 Practical Handbook of Soil,
Vadose Zone and Groundwater Contamination: Assessment, Prevention and Remediation (2nd HGQ /HZLV 3XEOLVKHU &5&3UHVV
%RFD5DWRQ)ORULGD86$
%85*($3+\GURJHRORJLHGXVXGGH6DYH'LUGRV6HUYGH*HRO
H0LQDV7HFKQLFDO5HSRUW
&2//,1*121%DQG9e=,1$0,QGHSHQGHQW Water and SaniWDWLRQ3URYLGHUVLQ$IULFDQ&LWLHV)XOO5HSRUWRID7HQ&LWLHV6WXG\
81'3:RUOG%DQN
'=:$,52%+2.2=/29('DQG*8=+$($VVHVVPHQWRI
the impacts of pit latrines on groundwater quality in rural areas: A
case study from Marondera district, Zimbabwe. Phys. Chem. Earth
31 *2')5(< 6 7,02 ) DQG 60,7+ 0 5HODWLRQVKLS EHWZHHQ
rainfall and microbiological contamination of shallow groundwater
LQ1RUWKHUQ0R]DPELTXHWater SA 31
*80%2%-8,=2'DQG9$1'(=$$*3,QIRUPDWLRQDVD
pre-requisite for water demand management: experience from four
cities in Southern Africa. Phys.; Chem. Earth 28 827-837.
*80%2%7KHVWDWXVRIZDWHUGHPDQGPDQDJHPHQWLQVHOHFWHG
cities of Southern Africa. Phys. Chem. Earth 29 1225-1231.
+2:$5'*3('/(<6%$55(701$/8%(*$0DQG-2+$/
. 5LVN IDFWRUV FRQWULEXWLQJ WR PLFURELRORJLFDO FRQWDPLQDtion of shallow groundwater in Kampala, Uganda. Water Res. 37
(14) 3421-3429.
,:$&2 6WXG\ RI *URXQGZDWHU 3RWHQWLDO RI $TXLIHU 6\VWHP
1RUWKRI0DSXWR'1$7HFKQLFDO5HSRUW
420
,:$&2 6WXG\ RI *URXQGZDWHU 3RWHQWLDO RI $TXLIHU 6\VWHP
1RUWKRI0DSXWR+\GURJHRORJLFDO5HSRUW'1$7HFKQLFDO5HSRUW
,:$&26WXG\RI*URXQGZDWHU3RWHQWLDORI$TXLIHU6\VWHP1RUWK
RI0DSXWR1DWXUDO5HFKDUJH(VWLPDWHV'1$7HFKQLFDO5HSRUW
-8,=2'0RGHODomR,QYHUVDSDUDD'HWHUPLQDomRGD5HFDUJD
GRV$TXtIHURVD1RUWHGH0DSXWR%DFKHORUWKHVLVUHSRUW(GXDUGR
Mondlane University- Maputo, Mozambique.
.$5,8., 0 DQG 6&+:$57= - 6PDOO 6FDOH 3ULYDWH 6HUYLFH
Providers of water supply and electricity. A review of incidence,
structure, pricing and operating characteristics. World Bank, Policy
and Research Working Paper 3727.
.58(60$1 * DQG 5,''(5 3$ $QDO\VLV DQG (YDOXDWLRQ RI
3XPSLQJ7HVW'DWD,/5,3XEOLFDWLRQ,QWHUQDWLRQDO,QVWLWXWHIRU
/DQG 5HFODPDWLRQ DQG ,PSURYHPHQW :DJHQLQJHQ 7KH 1HWKHU
lands.
/(( 0' DQG %$67(0(,-(5 7) 'ULQNLQJ :DWHU 6RXUFH 3URWHFWLRQ$5HYLHZ2I(QYLURQPHQWDO)DFWRUV$IIHFWLQJ&RPPXQLW\
:DWHU6XSSOLHV,5&,QWHUQDWLRQDO:DWHUDQG6DQLWDWLRQ&HQWUH7KH
+DJXH,7KH1HWKHUODQGV
/(:,6 :- )267(5 ' 5($' *+ DQG 6&+(57(1/(,% 5 The need for an integrated approach to water-supply and sanitation
in developing countries. Sci. Total Environ. 21 53-59.
0,6$85HJXODPHQWRVREUHD4XDOLGDGHGHÈJXDSDUDR&RQVXPR+XPDQR%ROHWLPGD5HS~EOLFDGH0RoDPELTXH,6pULH1U
'LSORPD0LQLVWHULDOQ0 180/2004 de 15 de Setembro (Guidelines
for Drinking Water Quality in Mozambique; Boletim of the Republic
of Mozambique, Series I, No 37, Ministerial Decree 189/2005 of September 15)0LQLVWU\RI+HDOWK0DSXWR0R]DPELTXH
2(&' 3ULYDWH 6HFWRU 3DUWLFLSDWLRQ LQ :DWHU DQG 6DQLWDWLRQ
,QIUDVWUXFWXUH%DFNJURXQG3DSHUIRUWKH5HJLRQDO5RXQGWDEOHRQ
6WUHQJWKHQLQJ ,QYHVWPHQW &OLPDWH $VVHVVPHQW LQ 1(3$' &RXQWULHV/XVDND=DPELD
38-$5,751$127,01,71$:$5(9&.+$5(/$7+$&.(5
13DQG.(/.$536(IIHFWVRIRQVLWHVDQLWDWLRQRQJURXQGwater contamination in basaltic environment – A case study from
India. Environ Monit. Assess. 134 271-278.
6$/&2168/725(6 $GDSWLQJ 5HJXODWLRQ WR WKH 1HHGV RI
WKH3RRU±0R]DPELTXH&DVH6WXG\(YDOXDWLRQ5HSRUW7HFKQLFDO
5HSRUW
6$/2021 0DSHDPHQWR GD 6LWXDomR GH $EDVWHFLPHQWR GH
ÈJXDQR'LVWULWR0XQLFLSDOQU0DSXWR&5$7HFKQLFDO5HSRUW
6(85(&$ +<'52&216(,/ 8QDFFRXQWHG IRU :DWHU )HDVLELOLW\6WXG\IRU0DSXWR:DWHU6XSSO\6FKHPH),3$*7HFKQLFDO
5HSRUW
6&+02//2+2:$5'*&+,/721-DQG&+2586,3URWHFWLQJ JURXQGZDWHU IRU +HDOWK :+2 ,:$ 3XEOLVKLQJ /RQGRQ
Seattle. http://www.who.int/water_sanitation_health/publications/
3*:VHFWLRQSGf
67$1'$5'0(7+2'6(1995) Standard Methods for the Evaluation
of Water and Wastewater (19thHGQ:DVKLQJWRQ$3+$$PHULFDQ
3XEOLF+HDOWK$VVRFLDWLRQ
68*%(167KH0LFURELRORJLFDO&RQWDPLQDWLRQRI:DWHU6XSSOLHV:(//)$&76+((7KWWSZZZLERURDFXNZHOOUHVRXUFHV
fact-sheets-htm/contamination.htm.
6:(&2 $662&,$7(6 7KUHH %DVLQV 6WXG\ 'HWDLOHG
*URXQGZDWHU,QYHVWLJDWLRQV'1$7HFKQLFDO5HSRUW
7+203621 7 )$:(// - .81,.$1( 6 -$&.621 ' $33/(<$5'6&$//$13%$57$1-DQG.,1*6213&KHPLFDO6DIHW\RI'ULQNLQJ:DWHU$VVHVVLQJ3ULRULWLHVIRU5LVN0DQDJHPHQW:+2*HQHYD6ZLW]HUODQG
:+2Guidelines for Drinking Water Quality. Vol. 1 Recommendations UGHGQ*HQHYD6ZLW]HUODQG
=$1'$0(/$ + /HVVRQV IURP 0R]DPELTXH ± 7KH 0DSXWR
Water Concession. Full detailed report.http://www.citizen.org/documents/Lessons%20from%20Mozambique.pdf (14 April 2008).
Available on website http://www.wrc.org.za
ISSN 0378-4738 = Water SA Vol. 34 No. 3 July 2008
ISSN 1816-7950 = Water SA (on-line)
IV
Physics and Chemistry of the Earth 33 (2008) 841–849
Contents lists available at ScienceDirect
Physics and Chemistry of the Earth
j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / p c e
Regulation of formal and informal water service providers in peri-urban areas of
Maputo, Mozambique
Nelson P. Matsinhe a,*, Dinis Juízo a, Berta Macheve b,1, Clara dos Santos b,1
a
b
Universidade Eduardo Mondlane, Faculdade de Engenharia, Av. de Moçambique km 1.5, C. Postal 257, Maputo, Mozambique
CRA-Conselho de Regulação do Abastecimento de Água, Av. Amilcar Cabral Nr. 757, C. Postal 253, Maputo, Mozambique
a r t i c l e
i n f o
Available online 10 July 2008
Keywords:
Peri-urban
Water supply
Service provider
a b s t r a c t
Service delivery to large areas of peri-urban Maputo depends largely on alternative informal service providers. These providers are located within the limits of Maputo, in a water supply area that is formally
leased to a private operator. Informal service providers therefore operate within the main regulatory body,
but their activity is presently unregulated. This paper discusses activities of informal alternative providers in peri-urban areas of Maputo, Mozambique, and opportunities to expand the reach and influence of
the main regulatory body to this segment of service providers. The study was commissioned to assist the
main regulatory body to setup a strategy to improve the pro-poor focus of the existing regulatory environment and so improve access to potable water for the majority of the under-serviced urban poor. Results
of field surveys conducted in selected areas of peri-urban Maputo are presented. The surveys focused on
the quality of services, the legal status of independent providers and the organization of water supply
services at neighbourhood level. The results indicate that household water resellers and small-scale independent provides are presently an important and indispensable source of access to water for the majority
of unconnected residents in peri-urban Maputo and that they are reported to cater for as many as 21% of
unconnected households of such neighbourhoods. In the near future, alternative providers will continue
to have a dominant role in service delivery in peri-urban Maputo, therefore their legalization and decentralization of certain regulatory functions to the neighbourhood level is required. A neighbourhood based
management model is proposed for that purpose. The model is based on a standpipe management model
that is broadened to include alternative service providers. The model addresses issues such as water pricing, bidding and compliance strategies, channels for consumer’s representation and possibilities of creating neighbourhood-based regulation bodies, which will act as extension branches of the main regulatory
body. Sustainability issues around the proposed model are also discussed.
© 2008 Elsevier Ltd. All rights reserved.
1. Introduction
This paper discusses household water resale activities of alternative service providers and small-scale independent providers
(SSIPs) in peri-urban areas of greater Maputo, Mozambique. The
work discusses also the opportunities of expanding the reach and
influence of the main regulatory body to this segment of providers,
who arguably serve the largest portion of Maputo’s urban poor.
The work was commissioned to assist the regulatory body to setup
a strategy to expand its coverage to alternative service providers.
The case paper focuses on peri-urban areas of greater Maputo,
Mozambique, where the informal water market plays a dominant
role in service provision. These areas are located within the lease
* Corresponding author. Tel.: +258 21 315161; fax: +258 21 475312.
E-mail address: matsinhe@zebra.uem.mz (N.P. Matsinhe).
1
Tel.: +258 21 312825; fax: +258 21 312826.
1474-7065/$ - see front matter © 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.pce.2008.06.046
area of a private operator contracted to provide services to the city
of Maputo, therefore, within the official limits of the main regulatory body. Presently, this regulatory function is restricted only to
the main service provider. The main regulatory body has therefore
expressed interest to explore the potential of legalizing alternative
informal service providers and expanding the reach of the body’s
regulatory role and influence to this segment of providers. The ultimate objective is to improve the pro-poor focus of the existing regulatory framework and ultimately improve access to potable water
to the majority of under-serviced urban poor.
2. Background
In Mozambique, water supply services are the mandate of the
National Directorate of Water (DNA), under the Ministry of Public
Works and Housing (MOPH). DNA is the primary agency responsible for water resources policy making, planning and management,
842
N.P. Matsinhe et al. / Physics and Chemistry of the Earth 33 (2008) 841–849
of services provided by formal and informal providers, the organization of management of peri-urban water supply services and the
possibilities of expanding the reach of the regulatory framework to
the neighbourhood level.
The focus of the study is on services provided through connected
household water resellers and standpipes and household taps connected to groundwater based small piped systems of AdeM and
SSIPs. These, generally consist of one or few boreholes connected
to one or few water towers and a small distribution network. AdeM
also runs few stand pipes connected to the main network. A neighbourhood based management model with its scope broadened to
include alternative service providers is proposed to assist the main
regulatory body expand the reach of the regulatory framework to
the neighbourhood level.
and for ensuring the provision of adequate water supply and sanitation services all over the country.
In 1995 DNA approved the first National Water Policy (NWP)
which has since guided water sector reforms. In line with the objectives of the NWP, the government withdrew from the task of direct
implementation of water supply services to focus more on policy
making and planning of the management of water supply services
(DNA, 1995). Among other issues, the policy gives special attention
to improvements of water supply services in urban and peri-urban
areas, the encouragement and regulation of the involvement of private service providers and the participation of beneficiaries in the
management of neighbourhood water supply services such as that
provided through public standpipes.
As a response to the objectives of the NWP, in 1998 the Government established the framework for delegated management of
water supply (decrees number 72, 73 and 74:9 of December 1998),
from which the Water Supply Investment and Assets Fund (FIPAG)
and the Council for Regulation of Water Supply (CRA) were created
(Zandamela, 2002). The framework created the legal basis for the
delegation of operation and management of public water supply
services to independent private entities through concessions,
lease or management contracts. The main institutions comprising
this framework are: the Ministries of Public Works and Housing
(MOPH), of Planning and Finance and of State Administration, the
National Directorate of Water (DNA), and the Co-ordinating Forum
for Delegated Management, FIPAG, CRA, the Municipal authorities,
and the private operator. FIPAG is the institution created to take
over the fixed assets for water supply and the duties and obligations for water service delivery in five major cities of the country
previously serviced by state water companies. CRA is the institution created with the objective of regulating private sector contracts under the rubric of this framework. The geographic limit of
the regulator’s activities is defined by the boundary of the area of
each contracted concession.
The framework has been implemented since 1998 in five largest cities of Mozambique, among them, the city of Maputo (DNA,
1999). In 1999 a private operator, Águas de Moçambique-AdeM,
signed a 15 years lease contract with FIPAG to provide services to
the city of Maputo and a nearby area called Matola (Zandamela,
2002; Gumbo et al., 2003). Since then, service quality in large areas
of peri-urban Maputo improved significantly, however, the situation today, is still far from satisfactory. This has prompted the emergence of a multiplicity of alternative service providers, including
small-scale independent providers (SSIPs) and household water
resellers, the majority of who operate within the lease area of
the main service provider (Seureca and Hydroconseil, 2005). Presently, however, the activity of SSIPs and household water resellers
is unregulated.
4. The study area
Maputo is characterized by three distinct set-ups, namely the
area with high storey buildings of the so-called old “cement city”,
few inner suburbs built before independence in 1975, and the outer
neighbourhoods which consist of informal settlements. During the
period after independence and probably exacerbated by the civil
war that raged the country for almost a decade, there was high
migration of rural citizens to Maputo city. These immigrants were
allowed to settle in the outskirts of the city, where proper urban
planning was lacking. Temporary licenses were issued for construction of temporary houses but after many years of occupation, the
residents of such areas virtually acquired the right to settle and
construct permanent houses. With the introduction of municipal
reforms started in 1998, the city of Maputo was divided into two
municipalities namely, Maputo and Matola. The administrative
boundaries of the newly created municipalities were re-drawn and
most of the neighbourhoods previously considered informal settlements became now part of the new municipalities.
Without proper planning and investments, however, most of
the new neighbourhoods today face severe limitations concerning
access to adequate municipal/public services such as water, sanitation and electricity. When it comes to piped water supply, while
the residents of the neighbourhoods located near the cement city
can still access the piped water through overstretching the formal
network, the residents of the outer neighbourhoods face more
restrictions to access piped water supply, due to lack of pressure in
the nearby network or because the network cannot be sufficiently
extended to reach their neighbourhoods.
Besides enormous disparities concerning the access to piped
water supply, this has favoured the widespread emergence of informal alternative service providers, among them SSIPs, who presently constitute the most reliable sources of water for the majority
of the under-serviced urban poor. Like in many other African cites
(Collignon and Vézina, 2000), SSIPs play a dominant role in service
provision to unconnected consumers of peri-urban Maputo (Table
1). As shown in Table 1, the situation in Maputo is such that roughly
32% of households rely on SSIPs and 6% on other types of alterna-
3. Objectives of the study
The work presented here investigates the main characteristics
of the water supply environment in peri-urban Maputo, the quality
Table 1
Access to drinking water in 10 Africa citiesa and Maputob
Abidjan (Côte
d’Ivoire)
Nairobi city
(Kenya)
Dakar
Kampala Dar es
(Senegal) (Uganda) salaam
(Tanzania)
Source of water for household use by 1999 (percent of household)
House connect
76
71
71
36
Stand pipe
2
1
14
5
SSIPs/traditional
22
27
15
59
sources
a,b
c
31
0
69
Adapted from Collignon and Vézina, 2000 and Seureca and Hydroconseil, 2005.
Figures for Maputo refer to survey done in 2005. All other cities refer to 1999.
Conakry
(Guinea)
Nouakchott
(Mauritania)
Cotonou
(Benin)
Ouagadougou
(Burkina Faso)
Bamako
(Mali)
Maputoc
(Mozambique)
29
3
68
19
30
51
27
0
73
23
49
28
17
19
64
40
22
38
N.P. Matsinhe et al. / Physics and Chemistry of the Earth 33 (2008) 841–849
tive sources (6%) for their water supply, while 62% of households
supplied through the formal network (Gumbo, 2004; Seureca and
Hydroconseil, 2005).
It is also within the peri-urban areas of Maputo city where the
largest proportion of lowest income groups lives and where unconnected consumers are reported to pay on average about three times
more per cubic meter of water than customers with house connections (Sal-consultores, 2005). Yet, these figures are not comparable
to other African cities (e.g. Luanda) where residents of informal settlements are said to pay as much as 10,000 times more for water to
private vendors than those living in the cement city where piped
water is available (Cain, 2004).
The peri-urban areas of the city of Maputo face also limitations
in terms of sanitation since no formal sewer systems exist in such
areas. Estimates from a survey carried out in 2005 (Sal-consultores, 2005) indicates however that almost 89% of households of
peri-urban Maputo have some kind of improved sanitation facility which can be either an improved pit latrines (50%), septic
tanks (39%), whereas the rest of households rely on traditional pit
latrines or share facilities with the nearby households.
5. Methods
This study was based on field work performed in one of the five
urban districts of the Municipality of Maputo. Selection of neigh-
bourhoods for the survey was done with the assistance of the main
regulatory body based on the following criteria:
(1) High prevalence of water resale activities either via connected consumers or via SSIPs;
(2) facility for identification of households where water resale
is practiced;
(3) existence of large variations in coverage conditions by the
main service provider; and
(4) representativity or resemblance of conditions of other periurban neighbourhoods of Maputo concerning coverage levels and existence of alternative service providers acting in
competition with the main service provider.
The neighbourhoods of urban district number 4 were determined to fit the aforementioned criteria and thus chosen as locations for the research. The district consists of 11 neighbourhoods
(Fig. 1) of which only two have access to potable water via the formal network. Projections from the Instituto Nacional de Estatística
– INE (1999) and data obtained from the local administration indicate that at the time of the survey the district’s population was
about 282,300 inhabitants.
The survey was conducted between November and December
2007 and consisted of site visits to identify existing conditions
of access to potable water supplies and interviews with key
843
C
C
C C
C
C C
C
C
C
C C
C
C
C
CC
CCC CCC
CC C
C
C
C
C
C
C
CC CCCC C
C
C C
C
CC CC CC
CCC CC CC C
C
C C
CC C C C
C CC
C
CCC
C
CCC C C C
CC
C CCCCCC C
C CCCC
CCCCCC C
CC
CC
CCC CC
C
C
CCC C
CCC C
CCCCCCCCC C
CCCCCCC CC
C CCC CCCCC
C
C
CC
CC
C CCCCCCCCC
CCCC
C CC C C CC
C
CCCCC
C
CC
CC CC CCCCCC C
CCC C C CC
CC
CC CC C
CCC
C
CCCC
C
C
CCCCC
CC
CCCC
CCCCC
C C
C
CCCC CCCC C
CCC
C
C
C C
C
CC
C
CCC
C
C CC
C C CC C
CC
C
-FHFOE
C
QSJWBUFTNBMMTZTUFNT
"EF.TUBOEQJQFT
"EF.TNBMMTZTUFNT
"EF.OFUXPSL
"ENJOJTUSBUJWFCPVOEBSZ
.VOJDJQBMEJTUSJDU/
Fig. 1. Urban network of district no. 4 of Maputo city and distribution of small piped systems run by Independent providers.
844
N.P. Matsinhe et al. / Physics and Chemistry of the Earth 33 (2008) 841–849
informants to get an insight on the quality of services offered by
different providers, the organization of services and on the level
of consumer’s satisfaction with respect to services provided. Interviewed people included consumers, owners and stand pipe attendants of small piped systems run by SSIPs, managers and stand
pipe attendants of AdeM owned systems and members of water
committees (WC), established at neighbourhood level.
Those interviewed included a total of 11 WCs and 249 stand
pipe attendants of which 55 were for small-scale piped systems
owned by AdeM, 18 for AdeM formal network and 176 of smallscale piped systems owned by SSIPs. In addition, 550 consumers
selected randomly at some of the stand pipes visited were interviewed.
6. Results
6.1. Overview of the water supply situation
6.1.1. Access to water services
The survey identified two major sources from which residents
of surveyed neighbourhoods get access to potable water supplies.
These are standpipes and household taps connected to AdeM formal network and small-scale piped systems, and standpipes and
household taps connected to small piped systems owned by SSIPs.
Table 2 provides a summary of the number and distribution of
standpipes and household taps reported to exist within the surveyed neighbourhood. Table 2 also gives estimates of the proportion of residents depending on the informal water market as compared to those depending on services of the formal provider. As
can be seen, about 45% of residents rely on SSIPs to access water
and about 13% rely on services provided by AdeM (Salomon, 2007).
A large proportion of residents (about 42%) rely on other type of
services which includes private wells and household water resellers. From the results in Table 2, it is clear that the informal market
plays the predominant role in the provision of water for the majority of residents of the surveyed neighbourhoods.
The results in Table 2 do not include data on the number and distribution of households relying on neighbour-to neighbour water
resale to get water. This resulted partially from lack of cooperation
from most of the interviewed residents who did not want to expose
households practicing household water resale. However, findings
from a similar study conducted in a nearby neighbourhood (Boyer,
2006) suggests that neighbour-to neighbour water resale is a longestablished and widely-practiced activity in most neighbourhoods
and also that it occurs regardless of existing conditions with respect
to availability of public standpipes. According to the results of the
same study, roughly 69% of unconnected residents depended on
neighbour-to-neighbour water resale as their primary source of
access to water, while some 57% of residents claimed to have their
neighbours as the only viable source of access to water.
6.1.2. Service quality
Water supply service quality is defined using various criteria.
These include service continuity, water quality, pressure and the
degree of responsiveness to complaints by consumers from service providers. For the purpose of the study, assessment of service
quality was based on how consumers evaluated services offered
by different providers. A criteria that ranked services according to
three levels namely, bad, reasonable and good was used. Ranking
was based on three quality indicators namely, water price at public
and private standpipes, promptness of standpipe attendants and
neighbourhood authorities to respond to consumer’s complaints,
and the accessibility to services measured by the number of hours
with pressure at standpipes and with open access to consumers.
Sixty-nine users of public standpipes and 108 users of privately
owned standpipes were interviewed. The survey included also an
assessment of how consumer complaints are dealt with by standpipe attendants and on how these are channelled to higher level
institutions. The results are shown in Fig. 2 from where it is seen
that consumers generally appreciate the quality of services offered
by both types of service providers. The public service is better
appreciated in terms of water prices and the promptness of standpipe attendants to listen to consumer’s complaints, while the private service is best appreciated in terms of service continuity and
access to water points.
The results of the survey also revealed also that privately owned
standpipes are more readily accessible to consumers, with roughly
90% of them reporting to offer services for more than 8 h per day
as compared to 49% of public standpipes offering the same level
of accessibility. For many consumers, access to stand pipes and
service continuity are the most important aspects to meet their
expectations in terms of access to services, particularly for those
who can only access the services during the early morning hours
or after work.
Private service providers are also viewed by many consumers
as the ones that addresses consumer requests and complaints
more promptly, particular requests and complaints related to matters such as payment modalities (consumers may get water on a
credit basis), negotiation of water prices and the requests for new
connections and extensions to the network. The public service,
Table 2
Number and distribution of small piped systems, public and privately owned standpipes and household/yard taps within neighbourhoods of study area
Name of
neighborhood
No. of small
piped systems
No. of connections
from main service
provider
AdeM
Private
Stand pipes
HC/yard taps
AdeM
SSIPs
AdeM (%)
SSIPs (%)
3 de Fevereiro
Albazine
Costa do Sol1
Ferroviário1
FPLM1
Hulene A
Hulene B
Laulane
Mahotas
Mavalane A
Mavalane B
Total
2
2(1)
0
5(2)
2(1)
3(2)
1
2
3
1
1
22(6)
13
17
0
25
7
13
14(11)
37
51
8
4
189
9
5
18(18)
13
4
3
17
2
11
2
4
88
0
127
0
0
0
0
0
9
23
0
0
159
6
11
0
20
9
22
20
22
29
6
3
148
493
207
0
2281
226
912
1403
1864
2541
453
180
10,560
1.6
1.1
0.0
2.3
0.7
0.5
3.0
0.4
2.0
0.4
0.7
12.7
1.9
2.3
0.0
7.6
2.0
5.5
6.0
7.2
9.6
1.9
0.9
44.9
AdeM stands for the main service provider; HC-house connections.
1 – Neighbourhoods covered by the formal network.
5(2) – In brackets, number of systems/units not functioning.
No. of connections
from SSIPs
Estimated
coverage by
service provider
N.P. Matsinhe et al. / Physics and Chemistry of the Earth 33 (2008) 841–849
$POTVNFSTBQQSFDJBUJPOUPUIFRVBMJUZPGUIFQSJWBUF
TFSWJDF
$POTVNFSTBQQSFDJBUJPOUPUIFRVBMJUZPGUIFQVCMJD
TFSWJDF
(PPE
3FBTPOBCMF
#BE
QSJDFPGXBUFS
SFTQPOTJWFOFTTUP
DPOTVNFSTDPNQMBJOUT
845
(PPE
3FBTPOBCMF
#BE
"DFTTJCJMJUZUP
DPOTVNFST
QSJDFPGXBUFS
SFTQPOTJWFOFTTUP
DPOTVNFSTDPNQMBJOUT
"DFTTJCJMJUZUP
DPOTVNFST
Fig. 2. Consumer’s appreciation to the quality of services offered by public (left) and privately (right) owned standpipes. Service quality indicators based on water price at
consumer’s tap, responsiveness to consumer’s complaints and accessibility to water points.
on the contrary is viewed as having limited options to deal with
consumer’s complaints mainly due to the lack of clarity and legal
definition of CRA’s role in receiving complaints and in resolving
conflicts between consumers and operators/managers. This is compounded by the fact that at neighbourhood level the role of local
authorities and institutions of the framework in relation to water
supply is not yet clear. The result is that most of the complaints on
management issues are taken directly to the standpipe attendants
who often convey these to the main service provider (AdeM).
6.2. Overview of the regulatory framework
of consumer tariffs based on performance indicators agreed to in
the lease or management contracts.
The most important area of action of CRA is consumer protection. In this respect, CRA is expected to carry out surveys on
consumer opinions concerning service quality and work with
consumer associations to study and analyse areas of interest. CRA
has an important role in ensuring that the codes of practice for
customer relations with the main service provider, in particular
customer complaints procedures are reviewed, approved, and complied with and sufficiently publicised to inform consumers of their
content.
6.2.1. General framework
The regulatory tasks of CRA, FIPAG, the operator, municipal
authorities and neighbourhood-based water committees are presented in Table 3. Though the lease contract for Maputo water
supply is signed between FIPAG and the private operator, the regulatory responsibilities are shared between FIPAG, CRA and to a certain extent, the Municipal authorities. Regulations covering quality
of services are defined by FIPAG and compliance is monitored by
FIPAG and CRA.
Monitoring service quality is done via information provided by
operators’ progress reports and annual reports on customer complaints, performance audits, and any other information held by the
operator or “lessor”. The extent of CRA’s regulatory intervention in
the lease contract, apart from conducting periodic reviews, is limited mainly to an advisory and endorsing role principally in areas
related to price regulation and consumer’s protection. For price regulation, CRA receives from FIPAG proposals concerning consumer
tariff levels and approves these for use by the operator. CRA additionally defines and approves alterations to the tariff structure and
6.2.2. Operational situation with respect to serving low
income-groups
The management of peri-urban services in areas covered by
the lease contract is based on a stand pipe management model
shown in Fig. 3. The model is specifically designed for services
provided through public standpipes and is based on a contractual
arrangement between a standpipe attendant and the service provider. As such, the regulations are those provided indirectly via the
Municipal authorities whereby, CRA, as the main regulatory body
defines (in consultation with Municipality) the regulation norms
but the implementation in practice is left to the responsibility of
Municipal authorities (Sal-Consultores, 2005). Further delegation
of authority can be done to more decentralized actors (e.g. water
committees, local associations and other community based organizations-CBOs) via adequate legal instruments such as bye-laws.
At neighbourhood level, the main actors of the standpipe management model are; the water committees (WCs), the standpipe
attendant and a representative of the local authorities. Contrary to
other situations in the country and elsewhere in the world where
Table 3
Regulation responsibilities for which CRA, FIPAG, AdeM and community water committees are responsible
Service provider
Price regulation
Service quality regulation
Competition regulation
Consumer protection
AdeM
CRA for tariff review
and approval
Bidding process
CRA for approval and ensuring
compliance with codes of practice
Small-scale independent
providers
Standpipe attendants
None/competition
CRA/FIPAG, respectively, for defining
and enforcing
penalties and performance standards
Competition
None
None
Water committee/AdeM/Municipality
Selection by water committee
None
None
Water committee, local authority,
Municipality, AdeM, CRA
None
Water resale
Water committee/
Municipalities
None/competition
Source: SAL-Consultores, 2005
846
N.P. Matsinhe et al. / Physics and Chemistry of the Earth 33 (2008) 841–849
4UBOEQJQF.BOBHFNFOU.PEFM
TIPXJOHSFHVMBUJPOSFMBUJPOT
BOEDIBOOFMTGPSDPTUVNFS
WPJDF
.6/*$*1"-*5:
%JTUSJDU"ENJOJTUSBUJPO
/PNJOBUJPO8$T
TUBOEQJQFBUUFOEBOUT
4VQFSWJTJPO
-PDBM"VUIPSJUZ
3FHVMBUJPOWJB
JOTUJUVUJPOTPGUIF
GSBNFXPSL
'*1"(
'FFQBZNFOU
4VQFSWJTJPOQSJDF
DPOUSPM
/PNJOBUJPO
8BUFS$PNJUFF
$3"
4UBOEQJQFBUUFOEBOU
4VQFSWJTJPO
$POUSBDU
SFMBUJPOTIJQ
ÆHVBTEF
.PmBNCJRVF
.VOJDJQBM"VUIPSJUJFT
/FJHICPVSIPPEMFWFM
$POTVNFST
*OTUJUVUJPOTPGUIFGSBNFXPSL
Fig. 3. The existing standpipe management model showing regulation relations and channels for costumer voice. Adapted from Sal-Consultores (2005).
WCs hold full responsibility for water supply in their neighbourhoods which includes collection of user fees and maintenance and
repair of water points, in the standpipe management model WCs
are only given only a supervisory role.
Analysis of the situation within the surveyed neighbourhoods
reveals a number of factors that undermine the full implementation of the standpipe management model and the development of
realistic action plans to expand the reach of the regulatory framework to the urban poor. These include
(1) The model is based on a contractual agreement between
standpipe attendants of the public service and the main service provider, and is therefore not applicable to other forms
of service provision (e.g. neighbour-to-neighbour water
resellers and SSIPs). The model does not open room for the
establishment of mechanisms to protect unconnected consumers with regards to water pricing and water quality.
(2) The lack of legal basis for issuing licenses to authorise other
types of service providers to operate within the lease area
and hence their inclusion in the standpipe management
model.
(3) The lack of clarity concerning the roles of municipalities and
neighbourhood authorities in the management of water services in areas covered by the framework. This can result in
conflict of interest of the role they have to play as system
managers and system regulators.
(4) A weakened position of WCs to perform in an unbiased fashion their regulatory tasks due to the influence of neighbourhood authorities, and frequently their dependence on revenue from standpipes.
(5) A weakened position of WCs to properly regulate standpipes
attendants because they also intervene in the selection and
nomination of standpipe attendants and in the management
of the finances of the standpipes. As a result, the tasks of
management, supervision and regulation are generally not
separated.
(6) Difficulties to maintain WCs members always committed
to their duties because of insufficient revenue from stand
pipes water sale to cover subsidies they are entitled to. In
fact, WC-members are a significant financial burden to standpipe attendants.
7. Discussion
7.1. Developmental perspectives for peri-urban water supply
In the near future, informal service providers, mainly SSIPs, are
expected to continue to play a dominant role in peri-urban water
supply, either because the selling of water is an important source
of income for households or because the expansion of the formal
network is unlikely to match the speed at which the suburbs of the
city will grow.
The current physical expansion of the formal network gives priority to areas with high population densities which are also areas
where the majority of informal suppliers exist. This suggests that
in the near future, while many of the existing private systems will
have to be decommissioned or, otherwise, linked to bulk water supply from AdeM, new market opportunities will certainly develop
in areas where the formal network will still be lacking, thus maintaining the status quo with respect to the dominant role played
by informal service providers in service provision to peri-urban
areas.
Moreover, FIPAG’s strategy to expand the network of Maputo water supply considers the development of complementary
groundwater based distribution networks to serve the northern
part of the city to be contracted out to private operators through
management contracts within the lease area of the official provider
or in a competitive manner. In the latter case, franchising models
whereby the operator acts as the main franchisor and independent providers as the franchisee may also be explored given their
potential to simultaneously improve service delivery and local economic development (Wall, 2006).
N.P. Matsinhe et al. / Physics and Chemistry of the Earth 33 (2008) 841–849
The potential for development of independent service providers is, therefore, highly favoured by the existing demand for services but the long term sustainability of the activity is threatened
by factors such as the lack of a clear regulatory framework and the
potential for future water quality problems due to over-exploitation of the aquifer system (Seureca and Hydroconseil, 2005).
While informal service providers are important as a viable alternative to improve water supply services in peri-urban areas of Maputo, the current status under which most of them operate means
that they are unprotected by law and hence cannot be properly
regulated. Legalization of the informal market and decentralization of certain regulatory activities to the neighbourhood level are
therefore important steps in the efforts to improve service delivery
to the urban poor.
7.2. Key aspects to be addressed
The first step to meet the demands for an expansion of the
reach and influence of the regulatory framework to cover also the
informal water market is the establishment of a licensing framework with which all participants of peri-urban water supply will
be obliged to comply. In doing so the following need to be taken
into consideration:
(i) The licensing framework should be inclusive of all forms of
alternative service providers. This includes household water
resellers, SSIPs depending on bulk water supply from the
main provider and SSIPs depending on their own sources to
render services provided that they operate within the Maputo lease area,
(ii) the financial and social sustainability of services provided
through public standpipes should be maintained. Household water resellers and SSIPs should therefore be viewed as
complementary and not competitors to the public service,
(iii) the issuing of licenses for water resale activities should follow a real demand as expressed by consumers,
(iv) to reduce the risk of inconsistent and unsustainable services
due to managerial constraints, the issuing of licenses for
household water resale and SSIPs should follow proven managerial capacity of potential candidates.
The issuing of licences for service providers relying on bulk
water supply from AdeM shall also consider the problem of the
tariff structure applied for bulk water supply. Increasing block tariffs (IBT) developed with the intent of protecting the poor, discourage wasteful consumption and provide opportunities for crosssubsidies is also known to lead to disadvantageous conditions to
household relying on water resellers since, such households are
forced pay higher unit prices for water than those connected to the
network (Wegelin-Schuringa, 1999). In fact, third parties receiving
bulk water supply from the main provider for further resale are
billed for greater quantities of water at a higher unit cost than nonseller households (Whittington, 1992; Kariuti and Schwartz, 2005;
SAL-Consultores, 2005).
Adoption of promotional tariffs based for instance on singleprice tariff structure for monthly consumptions of say 30 m3 billed,
for instance, at a rate corresponding to the minimum scale in the
tariff structure is more likely to result in affordable prices to the
poor and to maintain acceptable profit margins for those practicing
the activity. If done otherwise a risk exists that most licensed operators will run into financial problems that obviously will undermine
the ultimate goal of servicing the poor. This way of approaching
the tariff structure for third parties receiving bulk water supply
has also been proposed by Whittington (1992) in his analysis of the
water system in Kumasi, Ghana. In that study, Whittington (1992)
847
concluded that the adoption of a single price tariff structure serves
the poor far more equitably than the IBT structure.
Together with the establishment of a licensing framework for
household water resellers and SSIPs, decentralization of certain
regulation activities to the neighbourhood level is required. There
should also be a reconsideration of the official view that the influence of neighbourhood authorities and, frequently, their direct
benefit from standpipe funds is detrimental to the sustainability of
the management of water supply services at neighbourhood level.
Emphasis should be put on separating the supervision and regulatory roles currently played by WCs and neighbourhood authorities.
On the other hand, the role of the neighbourhood WCs which are
conceived as local level arbitration bodies to deal with issues concerning unconnected consumers should be strengthened in order
to ensure that these committees operate in an unbiased and independent manner.
At neighbourhood level clarification of communication lines
from the community level upwards to the institutions at higher
level of the governance framework is required. Partnerships with
NGOs, CBOs and the municipality are also needed to develop information channels and communication campaigns to inform consumers of their rights and options for recourse or assistance.
7.3. Proposed management and regulatory model for peri-urban
water supplies
Fig. 4 shows a modified version of the standpipe management
model with its scope broadened to include informal service providers. This model is proposed for management and regulation of
water supply services in peri-urban areas of Maputo.
The model assumes that all service providers operating within
the lease area will be connected consumers of AdeM who will benefit from bulk water supply from AdeM or, alternatively, be legally
authorised to run their own water sources and piped systems
within the lease area under service and quality criteria defined by
the operator.
CRA will define operational and performance norms while the
Municipality will assume the responsibility for issuing licenses
and implementing the regulatory functions. CRA shall make use of
the opportunities offered in the lease contract2 for Maputo water
supply which states that the “lessor” may authorise third parties to
establish, within the lease area, a system or infrastructure to supply water to areas not already served by the formal network.
Regulation will be conducted directly via the institutions of the
framework and indirectly via the Municipal authorities. For indirect regulation, the model further explores some of the opportunities offered by the institutional reforms currently in place within
municipalities3 which require Municipal authorities to have functional committees for liaison with neighbourhood representatives.
Furthermore, the model explores the opportunities offered by
the framework for delegated management regarding the position
and responsibilities of CRA. The proposed framework foresees the
position of a CRA-officer4 representation in municipalities where
CRA is obliged to perform its regulatory tasks. The CRA-officer is
therefore the key linchpin in the strategically established partnership between CRA and each municipality.
Among other aspects, the officer’s responsibilities include ensuring quality service to the poorly served areas and lower income
groups served by standpipes. The officer is also responsible to
2
Revised concession contract between FIPAG and AdeM (2004) for Maputo and
Matola cities.
3
Municipal legislation concerning water (Municipal laws n° 2/97 and 11/97).
4
Council for regulation of water supply – CRA (decree n.° 74/98, December,
1998).
848
N.P. Matsinhe et al. / Physics and Chemistry of the Earth 33 (2008) 841–849
.BOBHFNFOU.PEFMGPSQFSJ
VSCBOXBUFSTFSWJDFTXJUI
CSPBEFOFESFHVMBUJPOSFMBUJPOT
BOEDIBOOFMTGPSDPTUVNFS
WPJDF
.6/*$*1"-*5:
%JTUSJDU"ENOJOJTUSBUJPO
-PDBM"VUIPSJUZ
EJTUSJDUMMFWFM
*OEJSFDUSFHVMBUJPOUISPVHIBO$3"
%FMFHBUF
3"'
5FDIOJ
DBM
3-$#
-PDBM"VUIPSJUZ
OFJHICPVSIPPEMFWFM
$3"
%JSFDUSFHVMBUJPOWJBJOTUJUVUJPOT
PGUIFGSBNFXPSL
8BUFSDPNJUFF
.POJUPSJOH
TVQFSWJTJPOQSJDF
DPOUSPM
'*1"(
'FFQBZNFOU
$POUSBDUSFMBUJPOTIJQXJUI
.VOJDJQBMBVUIPSJUJFT
/PNJOBUJPO
"QQPJOUNFOU
$$%
QBUIGPSDPTUVNFST
WPJDF
4UBOEQJQFBUUFOEBOUIPVTFIPME
XBUFSSFTFMMFS44*1TXJUICVML
XBUFSTVQQMZ
$POUSBDU
SFMBUJTPOTIJQ
4VQFSWJTJPO
ÆHVBTEF
.PmBNCJRVF
.VOJDJQBMBVUIPSJUJFT
$POUSBDU
SFMBUJTPOTIJQ
/FJHICPVSIPPEMFWFM
$0/46.&34
*OTUJUVUJPOTPGUIFGSBNFXPSL
$$% %JTUSJDUBM$POTVMUBUJWF'PSVN
3"' 'JOBODJBMEFQBSUNFOU
3-$# /FJHICPVSIPPEMJBTPO$BCJOFU
Fig. 4. Proposed model for management of peri-urban water supply with regulation and contractual relations broadened to include alternative service providers.
follow up service provision evaluations in case problems are
reported, for example, by calling for meetings with representatives
from the private sector, FIPAG, municipality and other relevant
government offices in order to verify the reasons for the identified
problems.
Since one of the key aspects of regulation is consumer protection, the WCs, the appointed member of local authorities positioned
within the liaison committee to be created within the municipality
organizational structure and the CRA-officer will provide the necessary platform for hearing consumer’s voice and provide feedback
from and to the institutions of the framework.
therefore develop and formalize a system of payment of licenses
and regulatory fees in order to secure a sustainable environment
of peri-urban water supply regulation based on neighbourhood
WCs.
As for today, it is unrealistic to expect that WCs will be able to
perform effectively all expected duties due to their limited human
capacity. Therefore, long-term support and capacity building is
required to reinforce those elements that will ensure that they act
independently in pursuing community interests. Issues related to
the size of water committees and the possibility of reducing their
composition to a single individual operating more or less like a
CRA officer shall also be considered.
7.4. Sustainability aspects of the proposed model
8. Conclusions
In the proposed model, WCs are given most of the responsibility for regulation of water supply services at neighbourhood level.
The long term sustainability of service regulation based on the proposed model requires that financial means are made available to
secure the payment of subsidies to which WC-members are entitled and also their operations. CRA and the municipalities should
Service delivery to large areas of peri-urban Maputo is largely
dependent on alternative informal service providers. Though
located within the lease area and therefore within the official competence of the main regulatory body, the regulatory framework is
restricted to services of the main service provider. Despite improve-
N.P. Matsinhe et al. / Physics and Chemistry of the Earth 33 (2008) 841–849
ments being done to expand the formal network, in the near future,
alternative service providers will continue to play an important
role in service delivery in peri-urban Maputo either because the
selling of water is an important source of income for households or
because the expansion of the formal network is not likely to match
the speed at which the suburbs of the city will grow. Legalization of
informal alternative providers and decentralization of certain regulatory activities to the neighbourhood level is therefore a must.
Formalization of alternative service providers will require a definition by CRA and the municipalities of a framework for issuing
licenses and for ensuring the correct implementation of regulatory
functions. With CRA playing the role of normative agency, and the
Municipality playing the role of the licensing authority, the water
governance framework will have greater leverage to ensure compliance.
The long-term sustainability of peri-urban water services regulation based on neighbourhood water committees requires that
CRA and the municipalities formalize a system of payments of
license and regulatory fees to ensure the long-term functioning of
institutions created for the purpose.
References
Boyer, Alyssa, 2006. The regulation of peri-urban areas of Maputo, Mozambique. A
survey of household water resale activity in peri-urban Maputo. Preliminary discussion findings. Collaboration, Msc. Thesis, Columbia University-USA, Unpublished.
Cain, A., 2004. Livelihoods and the informal economy in post-war Angola. In: Clover, J., Cornwell, R. (Eds.). Supporting Sustainable Livelihoods: A Critical Review
849
of Assistance in Post-conflict Situations. khttp://www.iss.co.za/pubs/Monographs/No102/chap5.html (Chapter 5).
Collignon, B., Vézina, M., 2000. Independent water and sanitation providers in African cities. Full report of a ten cities study, UNDP/World Bank.
DNA National Water Policy, August 1995. Direcção Nacional de Aguas, Maputo,
khttp://www.dna.mz/natwpol.html (10.1.08).
DNA, 1999. Gestão Delegada de Abastecimento de Agua nas cidades de Maputo,
Beira, Quelimane, Nampula e Pemba. Departamento Nacional de Aguas, Maputo, Unpublished report.
Gumbo, B., Juizo, D., van der Zaag, P., 2003. Information as a pre-requisite for water
demand management: experiences from four cities in Southern Africa. Physics
and Chemistry of the Earth 28, 827–837.
Gumbo, B., 2004. The status of water demand management in selected cities of
Southern Africa. Physics and Chemistry of the Earth 29, 1225–1231.
INE, 1999. Projecções Anuais da População por Província e Área de Residência 1997–
2010. Instituto Nacional de Estatisticas, Maputo.
Kariuti, M., Schwartz, J., 2005. Small scale private service providers of water supply
and electricity. A review of incidence, structure, pricing and operating characteristics. The World Bank, Working paper, 3727.
Wall, Kevin, 2006. An investigation of the franchising option for water services operation in South Africa. Water SA 32 (2).
SAL-Consultores, 2005. Adapting Regulation to the needs of the poor-Mozambique
Case study. Evaluation report, Unpublished.
Salomon, 2007. Mapeamento da Situação de Abastecimento de Água no Distrito
Municipal nr. 4-Maputo – CRA. Technical report, Unpublished.
Seureca and Hydroconseil, 2005. Unaccounted for water feasibility study for Maputo water supply scheme, FIPAG, Unpublished report.
Wegelin-Schuringa, M. 1999. Water Demand Management and the urban poor. In:
Final Proc. Int. Symposium on Efficient Water Use in Urban Areas- Innovative
Ways of Finding Water for Cities. UNEP. khttp://www.unep.or.jp/ietc/Publications/ReportSeries/IETCRep9/4.paper-F/4-F-wege1.aspl (10.1.08).
Whittington, D., 1992. Possible adverse effects of increasing block water tariffs
in developing countries. Economic Development and Cultural Change 41 (1),
77–88.
Zandamela, H., 2002. Lessons from Mozambique – the Maputo water concession. Full detailed report. khttp://www.citizen.org/documents/Lessons%20from%20Mozambique.pdfl (14. 4. 2008).
V
HYDRAULIC FLOCCULATION WITH UP-FLOW ROUGHING FILTERS FOR
PRE-TREATMENT OF SURFACE WATER PRIOR TO CONVENTIONAL
RAPID SAND FILTRATION.
Matsinhe, N.P.1, Kenneth M. Persson2
1
Eduardo Mondlane University, Faculty of Engineering, Av. de Moçambique km 1.5, C. Postal 257,
Maputo, Mozambique, Phone:+258-21- 315161, (E-mail: matsinhe@zebra.uem.mz;Juizo@hotmail.com).
2
Lund University, Dept of Water Resources Engineering, P O Box 118 SE-221 00 Lund Sweden,
(E-mail: Kenneth.persson@tvrl.lth.se)
Submitted to the journal Environmental Science Engineering
ABSTRACT
Hydraulic flocculation has been used in many different ways in drinking water treatment for many
years. In this paper, the results of experimental work using an up-flow roughing filter for hydraulic
flocculation prior to treatment with conventional rapid sand filtration are presented. The objective
was to evaluate optimum flocculation conditions with up-flow roughing filtration and the quality of
formed suspensions with respect to their filterability. River water was used for the experiments.
Turbidity removal, head losses development and velocity gradients in the roughing filter were the
parameters used to evaluate the effectiveness of pilot plant processes. Overall turbidity removal in
the pilot plant was between 84% and 97%. Removal efficiencies in the up-flow filter were between
18% and 73% and at the rapid sand filter between 77% and 92%. Both units operated under positive
pressure. Irrespective of operational conditions established, G-values between 45-190 s-1 were
attained in the up-flow filter. Best performances were attained when the up-flow filter was operated
at the lowest filtrations velocities and alum doses. Overall, the up-flow filter performed both as
particle aggregation and separation unit and the quality of formed suspensions was suitable for
removal by rapid sand filtration. The method can therefore provide a rather versatile technique for
pre-treatment of turbid water prior to conventional rapid sand filtration.
Keywords: water treatment; conventional treatment, roughing filtration; contact-flocculationfiltration;
1
NOMENCLATURE
d0
Eff.
G
Gt
g
Km
L
Lf
NTU
P
Q
t
Tf
Tres
Trw
V
vf
¨H
¨Ht
Ʊ
Ƨ
Ƨt
Ƭ
ƭ
ƨs
filter medium grain size (m)
efficiency (%)
average shear rate or velocity gradient (s-1)
Camp number [-]
gravitational constant (m s-2)
filter medium constant [-]
effective length of filtration (m)
depth (height) of filtration layer (m)
Nephelometric turbidity units
power dissipated (watts)
volumetric flow rate (m3 s-1)
time (s)
mean residence time of fluid flow in filtration layer (s)
residual turbidity (NTU)
turbidity raw water (NTU)
volume (m3)
filtration velocity (m h-1)
head loss (m)
head loss at time t through filtration layer length Lf (m)
density (kg m-3)
porosity filter bed [-]
porosity filter bet at time t [-]
absolute viscosity of water (kg m-1 s-1)
kinematic viscosity of water (m2 s-1)
filter media grain sphericity factor [-]
INTRODUCTION
The conventional methods of removing turbidity and solids from raw water generally consist of
coagulation-flocculation followed by sedimentation and rapid sand filtration. In these methods
chemical coagulation is used to reduce the repulsive forces responsible for the stability of colloidal
dispersions while flocculation is used to enhance particle transport and aggregation, and the eventual
formation of settleable/filterable suspensions. Filtration (deep bed filtration) is used as a polishing
step (Chuang & Li, 1997).
Following destabilization with chemical coagulation, the rate of particle aggregation (flocculation) is
governed by the possibility and frequency of collisions between destabilized particles, the efficiency
of such contacts and the existence of transport mechanisms (mixing) to get particles close to each
other, collide and eventually become attached (Stumn & Morgan, 1996; Lawrence et al., 2007).
Fluid motion for flocculation can be induced either by mechanical stirring or by the energy derived
from hydraulic head loss. Mechanical flocculation provides high efficiency and flexibility of operation
but is relatively costly in operation and maintenance along with its dependence on the availability of
supplies and skilled labour. Hydraulic mixing on the other hand is less costly, can be operated by
2
relatively unskilled personnel but has the restriction of being less flexible to mixing intensity and to
flow and water quality variations (McConnachie et al., 1999; Polasek, 2007).
Hydraulic flocculation has been used in water treatment since many years and is particularly well
suited for situations of limited financial capacity and skilled labour such as those prevailing in most
developing countries. Methods of providing hydraulic flocculation include the use of baffled
flocculation channels (Mishra & Breemen 1987; McConnachie et al., 1999), filtration through fixed
granular media (McConnachie et al., 1999), and filtration through buoyant media (Vigneswaran &
Ngo, 1995).
The methods relying on filtration through fixed granular media are the basis of the so-called
flocculation supported filtration processes whereby, coagulant is introduced directly to the raw water
inflow immediately prior to the filter inlet (Huisman, 1984; Hansen, 1988; McConnachie et al., 1999).
The induced fluid shear resulting from the sinuous flow of the water through the interstices of the
filter medium promotes the transport of destabilized particles from the suspension to the grain
surface of the filter medium where they eventually become attached by mechanisms of
sedimentation, adsorptions and interception (Mishra & Breemen, 1987; Hansen, 1988; Chuang & Li,
1997).
A common design of treatment plants using this concept is the so-called up-flow-down flow
filtration. In this method, an up-flow roughing filter (also known as contact filter) is used for
hydraulic flocculation prior to filtration with conventional gravity rapid sand filters (Mishra &
Breemen, 1987). The primary potential advantage of up-flow/down-flow process is the reduction of
capital and operational costs of water treatment which results from the elimination of settling basins
and the elimination or significant reduction of dimensions of flocculation tanks (Mishra & Breemen,
1987; Vigneswaran & Ngo, 1995; Chuang & Li, 1997). Other advantages include the reduction in
coagulant dosages, decreased sludge production, reduced operation and maintenance needs, and the
possibility of maintaining flocculation efficiency regardless of flow and turbidity variations in the raw
water.
Previous studies (McConnachie et al., 1999; Ingallinella et al., 1998; Mahvi et al., 2004) concerning
the use of roughing filters for hydraulic flocculation, which emphasized the understanding of their
performance as compared to conventional paddle flocculation, have demonstrated that the method
can provide a rather versatile pre-treatment process capable of handling wide fluctuations in raw
water turbidity and operating conditions such as coagulant doses, and filtration rates. McConnachie
et al., (1999) reporting results from pilot studies conducted with an up-flow roughing filter operated
with M. oleifera seed solutions as coagulant, concluded that the unit could treat effectively raw with
turbidity as high as 50NTU with minimum head losses generated in the filter. Ingallinella et al.,
(1998) reporting results of similar studies but using roughing filters operated with aluminium salts as
coagulant, concluded that removal efficiencies as high as 90% could be achieved even for raw water
with initial turbidity as high as 340 NTU.
The work presented here follows a similar approach. Experimental work was conducted to assess the
performance of a pilot plant consisting of an up-flow roughing filter used for hydraulic flocculation
and a rapid sand filter used for final treatment in the removal of turbidity from river water. Turbidity
removal, head losses development, and velocity gradients in the up-flow filter were the parameters
used to assess the pilot plant performance. River water taken from the same source used for drinking
water production at a full scale treatment plant servicing the city of Maputo was used to run the pilot
3
plant experiments. Experimental results are therefore compared to those obtained at the full scale
treatment plant.
The aim of this paper is to present the main results of the pilot plant experiments. The term ‘contact
filter’ is used in this paper to describe the up-flow roughing filter operated as a hydraulic flocculator.
BACKGROUND
In conventional treatment, mechanical or hydraulic flocculation is used to promote the formation of
ideal suspensions in respect to their settling properties or filterability. Suspensions of four different
properties can be formed (Polasek & Mult, 2002):
(i)
Suspensions that are completely retained in the filter bed at the expense of high head loss
during filtration;
(ii)
Suspension which generates low head losses but which are poorly retained in the filter bed;
(iii)
Suspensions which are poorly retained and generates high head losses and,
(iv)
Suspensions which are completely retained in the filter bed and generates a minimum head
loss. Suspensions of this type represent the ideal suspensions, the formation of which should
be aimed at.
Whether settleable of filterable suspensions are envisaged, the key principle is to induce fluid motion
to cause velocity gradients enough to promote particle contacts and aggregation. Two major factors
govern the efficiency of flocculation processes (Lawrence et al., 2007; Polasek & Mult, 2005): the
intensity and duration of agitation. The mixing intensity is expressed in terms of mean velocity
gradient G(s-1), which expresses the energy input into the system. The standard expression for G is
G = (P/Ƭ*V)
(1)
Where P is the power input into the system, Ƭ is the absolute viscosity of the water and V, the
volume of liquid in the reactor. For the work done by the water flowing through a system where
hydraulic head loss is involved, the energy input P is given by
P = Ʊ*g*Q*ƅH
(2)
4
with Ʊ, the water density, g the gravitational constant, Q the flow rate and ƅH the head loss across
the system. Combining equations (1) and (2) results in:
G
'H * g
v *t
(3)
Were ƭ is the kinematic viscosity of the water.
The relationship for G in a porous media is derived from the following equation (Polasek, & Mult,
2002; Lawrence et al., 2007):
G
'H t * g
'H t * g * v f
Q * T f *Kt
Q * L f *K t
(4)
The mean residence time through the filter media is based on the model of length over velocity.
Since the approach to fluid flow in a packed bed is based on an idealized capillary model based on
which the packed media is regarded as a bundle of capillary tubes, to account for the tortuous path
of the flow through the filter bed, the effective length of the idealized capillary tubes is related to the
porosity of the filter bed and can be calculated as (Huisman, 1984; Chuang & Li, 1997):
L = Lf*Ƨ
(5)
The symbols in the right side of Eq. 5 have the same meaning as described in Eq. 4. The mean
residence time can therefore be calculated according to the following expression (Chuang & Li,
1997):
Tf = L/vf = Lf*Ƨt/vf
(6)
The porosity of the clogged filtration layer Ƨt is obtained from Carman-Kozeney equation based on
the head loss at specific time t. With Km taken as 5.0, the expression for the porosity of the clogged
filter layer is, according to Polasek & Mult (2002):
5
36 * K m * v f * v * L f * (1 K t ) 2
'H t
2
2
g * T s * d 0 *K t
(7)
3
MATERIALS AND METHODS
Pilot plant description
The experiments were carried out at the laboratory
ry of Hydraulics of the Department of Civil
engineer of Eduardo Mondlane University in Maputo. The pilot plant consists of two Perspex
columns, 2.75 m high with an internal diameter of 90 mm. One column was used as a contact filter
and the other as a rapid gravity filter. The plant arrangement is depicted in Figure 1.
The contact filter was provided with a 1.25 m filter bed, consisting of three layers of gravel placed in
the following manner: bottom layer: broken gravel, 0.25 m high 19.05 mm effective size and porosity
of about 55.2 %; middle layer, coarse gravel, 0.55m high, 12.5 mm effective size and porosity of
about 54.6% and upper layer, 0.45 m high, fine gravel with 2.78 mm effective and porosity of about
52%. The depth of water above the gravel bed was set at 0.75 m. The column was further provided
with a false floor consisting of a metal plate provided with evenly spaced 5 mm diameter holes onto
which the gavel bed rested.
C o lu m n 1
C o lu m n 2
10
Legend
1 . G rav el b e d m e dia (h o ttom la y e r, b ro k e n g rav e l 3 /4 ")
2 . G rav el b e d m e dia (m idle la y e r, c o a rse g rav el 1 /2 ")
3 . G rav el b e d m e di (u p p e r la y e r, fin e g rav el 3 /8 ")
4 . T u b e - ty p e pie z om e tric p a n n e l
5
3
5
4
5
5
4
6
1
5 . Inlin e flo w m e te rs (ro tam ete rs)
6 . F ilte r b e d ra p id sa n d f ilte r (riv e r sa n d 1 .1 m m )
7 . D o sin g p u m p (c o a g u la n t)
10
8 . R a w w a te r p u m p
2
10
10
10
10
9 . B a c k w a sh p u m p
1
10
Figure 1
1 0 . C o n tro l v a lv e s
7
8
9
10
Schematic diagram of the pilot plant arrangement. Column 1: multi-layer up flow roughing filter;
column 2: single media gravity rapid sand filter
6
The gravity rapid sand filter was provided with a 1.05 m filter bed consisting of river sand with an
effective diameter of 1.10 mm, and a porosity of about 44 %. The depth of water above the filter bed
was of about 0.95 m. The filtration column was also provided with a false floor consisting of a metal
plate drilled with evenly spaced 1 mm diameter holes onto which the filter bed rested.
Both columns were provided with diametrically opposed connections located 100 mm apart over the
height of the column used for water sampling and piezometric head losses readings. The sampling
ports consisted of stainless steel tubes extended some 5 mm into the filter bed onto which flexible
draw-off tubes where fixed which allowed continuous head loss measurements (via a tube-type
pressure gauging) and periodic collection of water samples for turbidity measurements. Roller type
clamps were provided on the flexible tubes to allow interruption of flow during periods of no
measurement.
Alum prepared as 10% solution of Al2(SO4)3.18H2O was used for coagulation purposes. The
chemical was dosed from a reagent tank to the inlet pipe of the contact filter with the help of a
positive displacement pump. Homogenization of the added chemical was achieved by turbulence
generated by means of a throttled inlet valve.
All experiments were run at constant filtration velocities maintained through manual flow control
attained via inline flow meters (rotameters). The contact filter was however run at filtration rates
higher than those used in rapid sand filter therefore; excess water was wasted through overflow pipes
located at the top of the columns.
Filtration experiments
A total of thirty four experiments were performed during a period of approximately 6 months from
April to October 2007. The raw water inlet to the plant was arranged via a 600 ȱ raw water reservoir
connected to a positive displacement gear type pump and a manually operated inline flow controller.
The feed water to the contact filter was prepared from two scenarios of coagulant addition (1.8 mg/ȱ
and 2.5 mg/ȱ) and raw water turbidity. The effluent from the contact filter constituted therefore the
feed water to the rapid sand filter.
During the period of experiments, the river water turbidity was generally low (less than 10 NTU). In
order to test the pilot plant also for higher values of raw water turbidity some experiments were
conducted with synthetic turbidity water prepared by adding clay to the raw water until levels of
turbidity larger than 15 NTU were attained. The pilot plant was further run at filtration velocities of
6.3 m.h-1, 9.4 m.h-1 and 12.7 m.h-1 in the contact filter and of 3.2 m.h-1, 6.3 m.h-1 and 9.4 m.h-1 in the
rapid sand filter.
Jar test experiments
Standard jar tests using a Janke & Kunkel jar test apparatus were used to determine the optimum
alum doses for the raw water which showed a dosage rate of 2.5 mg Al3+/ȱ as the optimum dosage
for maximum turbidity removal if conventional flocculation sedimentation were to be used. The pilot
plant was tested also at a dosage rate of about ¾ the optimum dosage.
7
Sampling and analytical methods
Turbidity and head losses were the main parameters used to assess the performance of the plant.
Samples of water for turbidity analysis were taken at different depths of the filter columns at regular
time intervals of 45 minutes. The termination criterion was defined as turbidity breakthrough or
maximum utilization of the permissible head loss, but because of logistic restrictions all experiments
were interrupted after 9 to 10 hours of filtration.
Besides turbidity and head losses, temperature, pH and alkalinity were also used to analyse the raw
water quality. These parameters were measured through analytical methods. Temperature, pH and
alkalinity were measured prior to the initiation of the experiments. Temperature readings were taken
with a standard mercury thermometer (accuracy of ± 10C), pH was measured with a handheld digital
meter from Wagtech International Ltd., and turbidity via a Hach turbidity meter DR 2500. Alkalinity
was determined using a simplified titration method described in the Standard Methods, (APHA, 1998,
2nd ed.).
Head loss readings
Head loss readings were taken from both columns using a tube-type differential pressure gauge.
Head loss readings were also taken at regular time intervals of 45 minutes.
RESULTS AND DISCUSSION
Raw water physicochemical characteristics
During the experiments the river water turbidity (Trw) was between 4.0 and 9.7 NTU, the pH
between 8.0 and 8.4, the total alkalinity between 115.6 and 122 mgCaCO3/ȱ and the temperature
between 19 and 31.5 0C. The quality of the feed water to the filter columns was slightly different, first
because some experiments were run with synthetic turbidity water and secondly because the source
water was stored for about a day in a closed room before experiments took place. This slightly
lowered the water temperature. The main characteristics of the raw and tested water are resumed in
Table 1.
Table 1
The range and average (N > 25) values of raw water quality used in this study.
Parameter
Turbidity (NTU)
Temperature (0C)
pH
Alkalinity (mg CaCO3/ȱ)
1
Range
4.0-22.91
19.0-31.5
8.2-8.6
115.6-122.0
synthetic turbidity
8
Average raw
water quality
4.7
23.4
8.1
118.0
Average test
water quality
13.7
21.2
8.4
122.2
Overall performance of the pilot plant
A summary of the results for turbidity removal in the pilot plant is presented in Table 2. For each
filtration run, mean values of influent and effluent turbidity are presented. As can be seen from Table
2, turbidity removal in the pilot plant was generally high and reached figures between 84% and 96%.
This was independent of the quality of feed water or operational conditions (coagulant dose and
filtration rates) established during individual runs.
The rapid sand filter had shown also very good performances. The filtrate quality was in general
below the desirable limit of 1.0 NTU of the guidelines for treated water (WHO, 2004) and terminal
head losses were, in general below the maximum permissible head loss defined on the basis of the
depth of supernatant water (0.95 m) above the filter bed. The rapid sand filter could therefore have
been run for longer periods without running into problems of negative pressures in the filter bed.
From Table 2, it is also seen that the filtrate from the rapid sand filter had little variations during
individual runs but increased slightly to a mean value of 1.9 NTU when the unit was operated at the
highest filtration velocity (9.4 m.h-1) and alum doses in the pre-treatment of 2.5 mg/ȱ (runs 19, 21,
22). However, the absolute limit of 5.0 NTU of the guidelines (WHO, 2004), was never exceeded.
The contact filter behaved slightly different. Turbidity removal in this unit was generally lower than
in the rapid sand filter (18.7-73.4 %) and the filtrate turbidity experienced large variations during
individual runs (see Table 2). This occurred particularly when the unit was run at a filtration rate of
9.4 m.h-1 and alum doses in the feed water of 2.5 mg/ȱ (runs 6-12). During these runs, turbidity
breakthrough could be observed frequently. Because the sampling port used to tap effluent water
from the contact filter was placed some 10 cm above the top of the filter bed, the frequent increases
in effluent turbidity were attributed to the effect of gravitational sedimentation that occurred in the
supernatant water above the gravel bed. This had a straining effect on the surface above the gravel
bed that caused the concentration of flocs to reach its highest values.
This phenomenon, which is similar to the processes taking place in sludge-blanket type clarifiers
started to develop right from the beginning of the filtration runs and gradually develop into a thicker
and concentrated cloud of particles positioned few centimetres above the gravel bed. In subsequent
experiments, the sampling port was lowered to ±1-2 cm above the top of the gravel bed. This
allowed the collection of samples not affected by differential settling, hence of lower turbidity values.
The quality of feed water to subsequent rapid sand filtration was in general better than that laving the
gravel bed of the contact filter (Figure 2). This suggests that apart from a partial removal of particles
in the gravel bed, additional removal of particles took place in the supernatant water above the gravel
bed.
Proportion wise the gravel bed performed better than the supernatant layer in the removal of
aggregates formed during flocculation. However, for low values of raw water turbidity the effect of
gravitational settling had a higher impact in the removal of aggregates particularly when alum dosages
of 2.5 mg/ȱ, were applied. This resulted probably from the presence of a large amount of aluminium
hydroxide aggregates which were thin enough to flow through the relatively coarse media of the
gravel bed, but large and in concentration enough to rapidly develop a sludge blanket in the
supernatant above the gravel bed.
9
The influence of the contact filter to the quality of suspensions transferred to the rapid sand filter
appears therefore to have been that of particle aggregation and separation whereby, the gravel bed
contributed mostly with particle aggregation and partial separation through mechanisms of particle
bridging, and the supernatant water with partial separation through mechanisms of gravitational
settling.
10
9.4/6.3
4.0-4.21
15.4-20.4
14.7-18.9
15.0-17.1
6÷9
10÷13
14, 23÷27
19÷21
12.7/9.4
12.7/9.4
2.5
1.8
2.5
2.5
1.8
1.8
2.5
1.8
Alum
dose
(mg/Ɛ)
14.2-19.8
12.2-17.3
19.7-24.8
9.0-12.8
14.3-19.5
8.8-10.7
9.8-17.3
16.6-17.4
water
6.1-7.0
8.1-12.1
4.1-9.6
12.1-20.6
5.0-9.7
5.5-13.5
4.6-5.5
3.7-6.3
57.7-73.4
5.3-6.2
29.4-41.6
7.0-10.4
4.0-5.9
6.3-7.0
-3
34.6-73.2
3.7-5.8
4.0-6.0
3.9-5.4
3.0-4.8
-3
18.7-65.4
21.4-36.1
63.7-73.2
11
1.1-1.9
0.6-1.0
0.5-0.7
0.5-0.7
0.5-0.7
0.6-0.8
0.7-0.8
0.7-0.8
Filtrate
77-84
82-87
91-92
86-91
84-87
78-87
78-85
87
Removal
efficiency
(%)
Feed
water
Removal
efficiency
(%)
Feed
Filtrate
Average turbidity (NTU) rapid
sand filter
Average turbidity (NTU) contact
filter
Tests run with natural turbidity
2
6.3/3.2; filtration velocity in the contact filter and rapid sand filter respectively
3
turbidity breakthrough in the contact filter
4
filtration experiments 1 and 2 run continuously without filter cleaning totalling 15 hours of continuous operation
1
9.4/6.3
15.1-16.6
15÷18
9.4/6.3
9.4/6.3
5.7-9.71
1÷ 4
6.3/3.2
6.3/3.2
16.6-22.9
13.0-20.9
28 ÷32
Filtration
velocity2
(m.h-1)
33 ÷ 34
Raw water
turbidity
(NTU)
Run test
Table 2
Summary results of turbidity removal in the pilot plant, contact filter and rapid sand filter.
8.25-9
7.5-9
8.25-9
9
6.75-9
9-154
9
9
Run
duration
(hr.)
670
583
308
320
272
937
204
269
Maximum
terminal
head loss
rapid sand
filter ( mm)
88-93
95-96
96-97
84-89
95-97
88-92
94-97
96
Overall
removal
efficiency
pilot plant
(%)
Eff. gravelbed/ Eff.
supernatant
7
Alum dose 1.8 m g/l
Alum dose 2.5m g/l
12
V=6.3 m /h
Trw = 13-21 NTU(S)
Alum dose 1.8 m g/l
Eff. gravelbed/Eff.
supernatant
8
6
5
4
3
2
10
Alum dose 2.5 m g/l
V=9.4m /h
Trw =4.0-9.7NTU(N)
8
6
4
2
1
0
0
Filtration runs
Filtration runs
5
Alum dose 1.8 m g/l
5
4
V=9.4 m /h
Trw = 15-20.4NTU(S)
4
Eff. gravelbed/Eff.
supernatant
Eff. gravelbed/Eff.
supernatant
Alum dose 1.8 m g/l
Alum dose 2.5 m g/l
3
2
1
Alum dose 2.5m g/l
V=12.7 m /h
Trw =14.7-19NTU(S)
3
2
1
0
0
Filtration runs
Filtration runs
Figure 2 Relative contribution of gravel bed and supernatant layer in turbidity removal in the contact filter: (N) =
natural turbidity; (S) = synthetic turbidity; Eff. = efficiency (%. Information about filtration velocities
and alum dosages applied is also shown.
The performance of the pilot plant was also compared to treatment results obtained at the full scale
treatment plant of Maputo water supply (Figure 3). In this plant, conventional
coagulation/flocculation sedimentation is used for pre-treatment and rapid sand filtration is used for
final treatment. The plant is operated with two parallel production lines. Sludge-blanket clarifiers
operated at surface hydraulic loads of 1.6 m.h-1 (line 1) and 2.4 m.h-1 (line 2) are used for
flocculation/sedimentation purposes. The rapid sand filters are operated at filtration rates of 5.2 m.h1
and 7.1 m.h-1 respectively. The data used for the comparison was taken from the operator’s
database and comprehend results of filtrate turbidity when the pilot plant was operated under similar
conditions of raw water turbidity and alum doses used for pre-treatment. As seen in Figure 3, the
pilot plant had removal efficiencies comparable to that obtained at the full scale plant and produced
a filtrate of better quality.
Performance of the contact filter
Head losses and filtration runs
In Figure 4, the development of pressure drop in the gravel bed of the contact filter is illustrated.
The pressure drop increased linearly along with filtration rates and alum dosages which clearly
indicate a time-dependent reduction of the gravel bed porosity and an increase in the inter-pore shear
stress due to accumulation of particles.
The pressure drop in the contact filter was in all cases of a few centimetres and well below the
maximum allowed head loss calculated from the available depth of supernatant water.
12
Performance pilot plant as compared to convenional treatment
Filtrate Turbidity (NTU)
12
Raw w ater
Filtrate pilot plant
Filtrate conventional plant (line 1)
Filtrate conventional plant (line 2)
10
8
6
4
2
0
April-14
Figure 3
April-18
April-20
April-26
May-14
May-17
June-06
June-08
Performance pilot plant as compared to conventional treatment with coagulation flocculation,
sedimentation and rapid sand filtration.
In up-flow filters, maximum head loss is limited by the danger of uplifting the filtering material
which occurs when the soil pressure equals the water pressure (Huisman, 1984). The filtering
material properties such as porosity, specific gravity and thickness of the filter bed set the limits.
Because the contact filter was provided with a top layer made up of the finest gravel the maximum
head loss was limited by the properties of this layer and was calculated as 0.35 m based on the
following properties of the filtering material: porosity 52 %; specific density 2.6 kg.m-3; and thickness
of about 0.45 m. Accordingly, the contact filter operated under positive pressure during all
experiments which means that it could have been run for longer periods and also with a much lower
(almost 50%) depth of supernatant water.
Head loss development contact filterfeed water with syntetic turbidity
Head loss development contact-filterfeed water with natural turbidity
100
Alum 1.8mg/l,V= 9.4m/h
90
35
alum 1.8mg/l,V=12.7m/h
Alum 1.8 mg/l, V=9.4 m/h
Alum 2.5mg/l,V=6.3m/h
30
Alum 2.5mg/l,V=12.7m/h
Alum 2.5 mg/l, V=9.4 m/h
Alum 2.5mg/l,V=9.4m/h
70
P ressu re d rop (m m )
Pressure drop (m m )
80
Alum 1.8mg/l,V=6.3m/h
60
50
40
30
25
20
15
10
20
5
10
0
0
0
Figure 4
1
2
3
4
5
6
7
8
9
10
11
12
13
Run time (hr.)
14
0
2
4
6
8
10
12
14
Run time (hr.)
Time dependent behaviour of head losses (mm) in the contact-filter for experimentsrun with syntetic
turbidity (left) and natural turbidity (right). Average raw water turbidity wasin the range 4.0-9.5
NTU. Syntetic feed water had a turbidity of about 20 NTU.
13
Analysis of head losses developed when the unit was run at a filtration velocity of 12.7 m.h-1 indicates
that much higher values were observed and also that the head losses developed more rapidly. This
resulted from high fluid shear stress established when the unit was run at such high filtration
velocities which may also have promoted high rates of particle aggregation within the gravel bed,
eventually associated with high rates of solids retention. This observation coincides with findings
from other researchers (Chuang & Li 1997; Ingallinella et al., 1998 and McConnachie et al., 1998)
who concluded, that shear stress affects flocculation processes and head loss development. As noted
by the same authors, associated with increases in the rate of particle aggregation within porous
media, an increase in head losses is expected. The magnitude depends on the induced shear stress
but also on the rate of solids deposition/detachment.
The head losses at a filtration velocity of 12.7 m.h-1 developed much faster than at 6.3 m.h-1 or 9.4
m.h-1 but, in contrast, the head losses at 9.4 m.h-1 developed slightly lower than at 6.3 m.h-1. This
unexpected behaviour was attributed to a possible predominance of thin aggregates that could flow
easily through the relatively coarse media of the filter bed, thus limiting the rate of solids deposition
and consequently the increase in head losses. This also explains the high filtrate turbidity observed
with the plant operated at 9.4 m.h-1 as compared to operation of the plant at 6.3 m.h-1, and similar
conditions of raw water turbidity and alum doses. In fact, since flocculation in porous media is
predominantly under ortokinetic conditions (Mishra & Breemen, 1987; Chuang & Li, 1997), the
conditions with lower filtration velocities and longer retention times resulted in better conditions for
the formation of larger aggregates and for increased rate of solids deposition that explains the
relatively large head losses in the gravel media.
The quality of feed water with respect to its turbidity seems to have had little influence on head loss
development in the contact filter. The differences shown in Figure 4 seem to have resulted from
differences in filtration velocities and coagulant dosages applied rather than from differences in the
feed water quality. This suggests that irrespective of the feed water quality, the flocculation
conditions created within the gravel media resulted in suspensions of relatively similar properties
with respect to their filterability. In fact, as noted by Chuang & Li (1997) and Declan et al. (2008) the
value of turbidity in suspensions formed during flocculation in porous media is qualitatively
proportional to the solids content but inverse of the particle size which means that the filterability of
corresponding suspensions is independent of the raw water turbidity.
Turbidity removal
Figure 5 illustrates time-dependent values of filtrate turbidity from the contact filter. To account for
variations in raw water turbidity, readings are plotted on the basis of the ratio between the filtrate
turbidity and that of the feed water. As shown in Figure 5, the filtrate turbidity decreased slightly
during the initial stages of filtration, but soon after that it started deteriorate and to show variations
occasionally associated to turbidity breakthrough with filtration time.
The initial decrease in filtrate turbidity was probably due the high solid retention capacity of the clean
gravel bed which led to a rapid accumulation of particles during the initial stages of filtration. During
subsequent stages, the increase in solids being retained in the gravel bed accompanied by reduction
in gravel bed porosity, led to an eventual increase in the inter-pore shear stress which may have
promoted particle detachment and turbidity breakthrough with the filtrate. As shown in Figure 5, the
highest variations in filtrate turbidity occurred more frequently when the unit was operated at 9.4 and
14
12.7 m.h-1, suggesting that shear stress affects not only flocculation processes as noted previously,
but also particle retention and detachment in porous media.
Time dependent filtrate turbidity contact-filter (feed water with
synthetic turbidity)
1,2
Time-dependent turbidity contact filter (feed water with natural
turbidity)
V=6.3m/h,alum 1.8 mg/l
1,4
v=6.3m/h,alum 2.5mg/l
1,3
V=9.4 m/h, alum 1.8 mg/l
1,2
V=9.4m/h, alum 2.5mg/l
ratio filtrate turbidity / feed water
turbidity
ratio filtrate turbidity / feed water
turbidity
1,4
V=9.4m/h,alum 1.8mg/l
V=9.4m/h,alum 2.5mg/l
V=12.7m/h,alum 1.8mg/l
1
V=12.7m/h,alum 2.5mg/l
0,8
0,6
0,4
1,1
1,0
0,9
0,8
0,7
0,6
0,5
0,4
0,3
0,2
0,2
0
2
4
6
8
10
0
12
run tim e(hr.)
1
2
3
4
5
6
7
8
9
10
11
12
13
Run time (hr.)
Figure 5 Time dependent behaviour of filtrate turbidity (NTU) in the contact-filter, for experimentsrun with
syntetic turbidity (left) and natural turbidity (right). Average raw water turbidity wasin the range 4.0-9.5
NTU. Syntetic feed water had a turbidity of about 20 NTU. Information about filtration velocities and
alum dosages applied is also shown.
As noted from Figure 5, most efficient treatment was obtained when the unit was operated at a
filtration velocity of 6.3 m.h-1 and alum doses of 1.8 mg/ȱ. The low performances observed when
the contact filter was run at 9.4 m.h-1 were associated to poor flocculation conditions resulting from
the combined effect of moderate to low agitation intensities and short residence times which,
eventually resulted in the formation of aggregates that could flow easily through the relatively coarse
media of the contact filter and where not strong enough to withstand the induced fluid shear
stresses. Changing the filtration velocity to 12.7 m.h-1 resulted in much higher agitation intensities
and inter-pore shear stress. This eventually resulted in the formation of thin but much stronger
aggregates (despite the short residence times) which, because of their smaller size were mostly
retained through mechanism of particle bridging and attachment on the surface of the gravel media
but were strong enough to withstand high induced shear stresses.
According to results shown in Figure 5, changing the alum dose from 1.8 mg/ȱ to 2.5 mg/ȱ seems
to have had impacted the performance of the contact filter particularly when high filtration velocities
were used. As shown in Figure 5, the poorest performances were mostly observed when using alum
doses of 2.5 mg/ȱ. Possible reasons for this could be that the aggregates formed when using high
alum doses are generally large in size but weak in strength, suggesting that they removal in porous
media is largely affected by mechanisms of particle breakage and detachment from the filter grains.
Chuang & Li (1997) and Ingallinella et al. (1998) have also reported that flocs formed with high alum
doses are generally large in size but weak in strength, thus unable to withstand high inter-pore shear
stresses.
Velocity gradients
Velocity gradients (G) and Gt-values calculated from to Eqs. 4 and 7 and the head losses generated
across the two upper layers of the contact filter are presented in Table 3. As can be seen, velocity
gradients (G s-1) were between 45-120 s-1 in the middle layer of the contact filter and between 88 –
15
190 s-1 in the upper layer. The head loss across the bottom layer of the contact filter was in all cases
negligible therefore, corresponding G and Gt-values are not presented.
Table 3
Initial and terminal values of G (s-1) and Gt generated at the two uppermost layers of the contact
filter.
Filtration velocity
(m.h-1)
Initial G and Gt values
Reference
layer
Terminal G and Gt values
Alum dose (mg/Ɛ)
1.8
2.5
1.8
Alum dose (mg/Ɛ)
2.5
1.8
2.5
1.8
2.5
-1
Velocity gradients G (s ) and Gt values
967
1006
45
56
1482
1819
1779
88
87
2971
6.3
Middle
Upper
25
46
27
44
9.4
Middle
Upper
41
40
34
31
12.7
Middle
55
58
1015
1042
110
123
1659
1784
Upper
64
79
1407
1631
168
192
2867
3128
991
1198
906
997
60
96
56
88
1706
2951
1352
2353
1245
1312
References from text books (Stumn & Morgan 1996; Lawrence et al., 2007) recommend that mixing
for optimum flocculation should generally be of low intensity, with G-values preferably between 20
and 70 s-1 and Gt-values between 2 x 104 and 2 x 105. Below these limits no proper flocculation
occurs while, increasing G and t values beyond these limits results generally in floc breakage and
turbidity breakthrough in subsequent treatment processes. The principle behind these limits is
associated to the belief that low agitation intensity favours the formation of large and readily setleable
aggregates and that, beyond a certain limit of agitation, floc breakage occurs. However, recent studies
from Polasek (2007), arguing the principles behind the so-called customary flocculation suggests that,
while slow mixing promotes the formation of large and readily settleable flocs the end result is in fact
the formation of flocs that are large in size, but of low density, very fragile and with a tendency to
fragment. As noted by the same author, this type of flocs is suitable neither for sedimentation nor for
rapid sand filtration. In contrast if the flocculation is performed under high agitation intensities over
the entire process until optimum flocculation is reached, the formed flocs are generally more
compact and dense. Accordingly, depending on the resultant size of aggregates required, flocculation
can take place under high agitation intensities with G-values preferably above 50 s or low agitation
intensities with G-values below 50 s-1..
-1
The high and low agitation intensities involve the same transport mechanism and differ only by the
agitation intensity (G-value). When micro-flocs are to be formed, high agitation intensities (G-values
between 100 and 500 S-1 are usually preferred while, for large and readily settleable macro-flocs, low
agitation (G-values between 5 and 20 S-1 is generally preferred. The agitation intensity together with
the duration of the process determines the final result. Flocs formed under low agitation and long
retention times are generally larger and denser than those formed with high agitation and short
contact times (Polasek & Mult, 2005; Polasek, 2007).
From the results shown in Table 3, it is seen that G-values in the contact filter ranged from
conditions of moderate to high agitation intensities. Our interpretation to this is that this has
16
favoured the formation of aggregates of different characteristics concerning the size and density but
which were removed by sedimentation and attachment onto the surface of the gravel grains. From
the results of Table 3, it also appears that best flocculation conditions were attained when the contact
filter was run at a filtration rate of 6.3 m.h-1. However, due to the small size of the filtering material
used in the upper layer of the contact filter, relatively large G-values were established (G § 87-96 s-1)
in this layer which, eventually contributed to particle breakage. This was independent of the alum
dose or filtration velocity applied.
At a filtration velocity of 6.3 m.h-1, the combined effect of lower agitation intensities and longer
retention times resulted eventually in the formation of large aggregates hence, the highest removal
efficiencies attained when compared to other filtration velocities. When the contact filter was
operated at 9.4 m.h-1 relatively large aggregates were eventually formed but now, the effect of high
induced inter-pore shear stress and short retention times may have caused formed aggregates to
break and be detached from the grains, thus leading to the highest concentrations of particles in the
filtrate. Increasing the filtration velocity to 12.7 m.h-1 resulted in agitation intensities that favoured
the formation of thin but dense aggregates (Polasek, 2007) which were poorly retained in the gravel
media thus, the lowest performances observed at this filtration velocity. The effect of particle
breakage and detachment from the gravel grains when the contact filter was run at 9.4 m.h-1 and
alum doses of 2.5 mg/ȱ seems, however, to have impacted the filtrate quality more seriously.
Performance of the rapid sand filter
Time-dependent filtrate turbidity and head losses from the rapid sand filter are presented in Figure 6.
As can be seen the filtrate from the rapid sand filter was always of acceptable quality (Tres > 1 NTU).
The terminal head losses was , in all cases, below the maximum permissible head loss of 1.35 m,
calculated from Carman-Kozeney equation (Huisman, 1984), based on the available depth of
supernatant water (0.95 m) and a clogged layer of about 30% the filter bed thickness.
Performance rapid sand filter- Turbidity removal
V=3.2m/h, alum 1.8 mg/l
V=3.2m/h,alum 2.5 mg/l
V=6.3m/h, alum 1.8 mg/l
V=6.3m/h, alum 2.5 mg/l
V=9.4m/h, alum 2.5 mg/l
V=9.4m/h,alum 1.8 mg/l
V=3.2m/h, alum 2.5 mg/l
3,1
600
V=6.3m/h, alum1.8 mg/l
V=6.3m/h, alum 2.5 mg/l
2,8
p ressu re d ro p (m m )
filtra te turbidity (N TU )
Performance rapid sand filter-pressure drop
700
V=3.2m/h; alum 1.8 mg/l
3,4
V=9.4m/h, alum 1.8 mg/l
2,5
V=9.4ml/h, alum 2.5 mg/l
2,2
1,9
1,6
1,3
1,0
500
400
300
200
100
0,7
0
0,4
0
1
2
3
4
5
6
7
8
9
0
10
run time (h)
1
2
3
4
5
6
7
8
9
10
run time (h)
Figure 6 Performance of the rapid sand filter. Time dependent values of filtrate turbidity (left) and head loss
development (right) are presented. Also information on filtration velocities in the rapid sand filter and
alum dosages applied in the contact filter is presented.
17
Exception is made for the filtration runs done with feed water prepared from filtration velocity of
9.4 m.h-1 and alum doses of 2.5 mg/ȱ in the contact filter, during which the filtrate from the rapid
sand filter started deteriorate 3 to 4 h, after the beginning of the experiments. This resulted
eventually from the high load of fine particles being transferred from the contact filter which, at the
corresponding scenario of operation, had the poorest performances as is can be seen from Figure 6.
From analysis of results of head loss development in the rapid sand filter it is seen that the unit
always operated under positive pressure which means that longer filtration runs could have been
established without running into problems of negative pressures. This also indicates that turbidity
breakthrough was the factor determining the duration of filtration runs particularly when filtration
velocities higher than 6.3 m.h-1 were chosen to run the plant.
From the results shown in Figure 6, it is also seen that head losses in the rapid sand filter developed
more or less linearly. This indicates that impurities penetrated uniformly through the depth of the
filter bed. This also indicates that irrespective of the performance of the contact filter, the quality of
suspensions transferred to the rapid sand filter were generally of similar properties in respect to their
filterability. As shown in Figure 6, irrespective of the conditions of operation of the contact filter, the
resulting suspensions were in general completely retained in the rapid sand filter and generated
minimum head losses. As noted by Polasek (2002) suspensions of this type represent the ideal
suspensions, the formation of which should be aimed at during pre-treatment for filtration with
conventional rapid sand filters. The optimum combination seems to have been that of operating the
contact filter and the rapid sand filter at filtration velocities of about 6.3 m.h-1 or lower and alum
doses of 1.8 mg/ȱ.
CONCLUSIONS
Results of this study supports claims made by other researchers that the use of roughing filters for
hydraulic flocculation provides a viable and flexible alternative for improved turbidity and solids
removal by conventional rapid sand filtration.
In this study, the quality of suspensions produced at the contact filter was generally suitable for
removal by subsequent rapid sand filtration independent of the operational conditions established
(feed water turbidity, filtration velocities and alum doses) at individual filtration runs. Best
performances were, however, attained when the contact filter was operated at a filtration velocity of
6.3 m.h-1 and alum doses of about ¾ of the optimum dosage obtained from jar test experiments.
Overall performance of the pilot plant performance was in general above 84%. The filtrate from the
rapid sand filter was of acceptable quality and consistently below 1 NTU and the units operated
under positive pressure during the entire duration of the experiments. Longer than the 9 to 10 hours
duration of filtration could therefore, have been established.
Velocity gradients in the contact filter were within limits of moderate to high agitation intensities and
were, in general within limits recommended in literature for effective flocculation. Formed aggregates
were suitable for removal by mechanisms of sedimentation and particle bridging in the gravel media
of the contact filter and dense enough to be removed by mechanisms of sedimentation in the
supernatant water above the gravel bed. The remaining flocs could be effectively removed through
subsequent filtration.
18
The contact filter used in this study was designed with the filter bed arranged with the gravel size
decreasing in the direction of the flow. This impacted significantly the unit’s performance since floc
breakage and detachment occurred mainly at the upper and finer layer of gravel bed. Further
research is therefore required concerning the optimum composition and arrangement of the gravel
media. The use of a relatively large supernatant layer above the gravel bed helped however; reduce
significantly particle (turbidity) concentration in the filtrate.
Because filtration velocities used to run the contact filter, were much larger than those recommended
for plain roughing filters (Smet & Visscher, 1990; Sánches et al., 2006) large investment and
operational costs can be attained by using up-flow roughing filters as hydraulic flocculators. Saves
can also be attained in relation to costs with chemical reagents. For optimum operation of up-flow
roughing filters used for hydraulic flocculation, the units should however be designed for G-values
between 40 and 90 S-1 and filtration velocities lower that 6.0 to 7.0 m.h-1.
ACKNOWLEDGEMENT
This paper contains results of a study conducted under a wider research program funded by the
Swedish International Development Cooperation Agency (SIDA/SAREC) jointly implemented by
Eduardo Mondlane University and the Lund Institute of Technology of Lund University-Sweden.
The authors wish to acknowledge the valuable support of the grant agency and to thanks all people
and institutions that provided support, data and information used in the study.
REFERENCES
APHA, AWWA, WEF, (1998). Standard Methods for the Examination of Water and Wastewater, 20th ed.
American Public Health Association, Washington, DC.
Chuang, Ching-J. and Li K-Y (1997). Effect of coagulant dosage and grain size on the performance of
direct filtration. Separation and Filtration technology, 12, pp 229-241.
Declan P., Edwin L., E. and Pavelic P. (2008). A new method to evaluate polydisperse Kaolinite clay
particle removal in roughing filtration using colloid filtration theory. Water Research, 42, pp 669676.
Hansen, Sigurd P. (1988). Filtration Technologies for Water Treatment. Water Engineering &
Management, vol. 135, No. 8 ABI/Inform Global, pp. 16-24.
Huisman, L (1984). Rapid Filtration. Delft University of Technology, Faculty of Engineering, Delft, The
Netherlands, pp 162-174.
Ingallinella A. M., Stecca L. M. and Wegelin, M (1998). Up-flow Roughing Filtration: Rehabilitation of a
water treatment plant in Tarata, Bolívia. Wat.Sci.Tech. vol. 37, No.9, pp 105-112.
Lawrence K. Wang, Hung Yung-Tse and Shammas Nazih K. (2007). Handbook of Environmental
Engineering, Vol. 3: Physicochemical Treatment Processes. The Humana Press Inc. Totowa, NJ. Pp
103-127.
Mahwi, A.H., Moghaddam, M.A, Nasseri, S. and Naddafi, K. (2004). Performance of direct Horizontal
roughing filtration (DHRF) system in treatment of high turbid water. Iranian J. Env. Health Sci Eng.
Vol 1, No 1, pp 1-4.
McConnachie G.L., Folkard G.K., Mtawali M.A. and Sutherland J.P. (1999). Field trials of Appropriate
Hydraulic Flocculation Processes. Water Resources vol. 33, No. 6, pp 1425-1434.
Mishra K.K. and Breemen A.N. (1987). Communication on the Sanitary Engineering & Water
Management: Gravel bed flocculation. Delft University of Technology, Faculty of Civil
Engineering, Delft, The Netherlands, pp 11-70.
19
Polasek P. and Mutl S., (2002). Cationic Polymers in Water Treatment-Part 2: Filterability of CPF-formed
suspension. Water SA Vol. 28 No 1, pp 83-88.
Polasek P. and Mult S., (2005). Optimization of reaction conditions of particle aggregation in water
purification-back to basics. Water SA Vol. 31 No 1, pp 61-72.
Polasek P. (2007). Differentiation between different kinds of mixing in water purification -back to basics.
Water SA Vol. 33 No 2, pp 249-251.
Smet, J.E.M and Visscher, J.T. (1990). Pre-treatment Technologies for community water supply: An
overview of techniques and present experience, 2nd ed. International Water and Sanitation
Centre, The Hague, The Netherlands, pp 109-124.
Sánchez, L.D., Sánchez, A., Galvis, G and Latorre, J. (2006). Multi-stage Filtration. International Water
and Sanitation Centre, The Hague, The Netherlands, pp 11-18.
Stumn W. and Morgan James J. (1996). Aquatic Chemistry: Chemical Equilibria and rates in Natural
waters. John Wiley & Sons, Inc. 3rd ed. New York, USA, pp 818-866.
Vigneswaran S., Ngo, H. H. (1995). Application of Floating Medium Filter in water and Wastewater
treatment with contact-flocculation filtration arrangement-Research note. Water Research vol. 29
No. 9, pp 2211-2213.
WHO (2004) Guidelines for Drinking Water Quality. Vol. 1 Recommendations (3rd edn.) Geneva,
Switzerland.
20
VI
VATTEN 64: 137–150. Lund 2008
ASSESSMENT OF DRINKING WATER TREATMENT USING
MORINGA OLEIFERA NATURAL COAGULANT
Värdering av Moringa Oleifera för fällning av dricksvatten
by EMELIE ARNOLDSSON 1, MARIA BERGMAN 1, NELSON MATSINHE 1, 2 and KENNETH M PERSSON 1, 3
1 Department of Water Resources Engineering, Lund University, Sweden
2 Dept. of Civil Engineering, Faculty of Engineering, Eduardo Mondlane University, Maputo, Mozambique
3 SWECO Environment Malmö, Sweden
Abstract
In this study, a comparison between Moringa Oleifera MO and aluminium sulphate was conducted for coagulation of turbide water using jar test. The optimum coagulant dosage was investigated for different levels of turbidity, and the impact on treated water properties monitored. Use of MO together with direct filtration was
also investigated.
Coagulation with aluminium sulphate led to more efficient treatment but MO could also produce water of
acceptable quality. On the other hand, treatment with MO did not change the chemistry of treated water and
was more efficient for high initial turbidities. Highest removal efficiencies were obtained when MO was
extracted using tap water as compared to distilled water and oil extraction. Coagulation with aluminium, followed by direct filtration led also to better performances but in both cases the treated water met WHO water
guidelines. Prolonged sedimentation helped improve MO treatment efficiency for low initial turbidities.
MO is found to be a sustainable solution for coagulation in drinking water treatment. The possibilities of
using MO together with direct filtration are good and provide a realistic alternative to conventional methods
for small to medium size water supplies.
Key words – Moringa Oleifera, drinking water treatment, aluminium sulphate
Sammanfattning
Att minska turbiditeten är en central del vid beredning av dricksvatten från ytvatten, och sker oftast med
koagulering och flockning, där polyvalenta metallsalter som aluminiumsulfat används. Studier indikerar att
metallsalterna kan ersättas av ett naturligt koaguleringsmedel från det tropiska och subtropiska trädet Moringa
Oleifera (MO). Dess frön har använts i småskalig dricksvattenberedning i Indien och Sudan i flera generationer.
Koaguleringsmedlet är troligtvis ett eller flera protein som agerar likt katjoniska polyelektrolyter.
I denna studie har en jämförelse mellan MO och aluminiumsulfat genomförts i bägareförsök. Den optimala
dosen av koaguleringsmedlet undersöktes för olika turbiditetsnivåer, och effekten på vattenkemin kontrollerades. Användning av MO tillsammans med direktfiltrering, samt möjligheterna att använda MO i vattenverk
undersöktes också.
Den mest effektiva beredningen för alla turbiditetsnivåer återfanns vid användning av aluminiumsulfat. MO
påverkade inte de fysikaliska egenskaperna och var mest effektiv för högre grumlighetsnivåer i råvattnet. Extraktionsmetod för MO-extrakt påverkade inte det beredda vattnets kemiska egenskaper. En förlängd sedimentationstid tillsammans med MO förbättrade reningen och försöket med direktfiltrering var lyckat. Rening med
MO och direktfiltrering resulterade i mer effektiv rening än med aluminumsulfat i kombination med direktfiltrering. MO kan anses vara ett förnyelsebart och billigt koaguleringsmedel vid dricksvattenberedning. För
bästa resultat bör MO användas tillsammans med direktfiltrering. Om vattenberedningen sker i liten skala är
MO ett realistiskt alternativ till konventionella metoder, förutsatt att plantager etableras i tillräcklig omfattning.
VATTEN · 2 · 08
137
1. Introduction
When surface water is used for drinking water production, turbidity removal is an essential part of the treatment process. It is generally achieved using coagulation
with metal salts followed by aggregation of particles
through flocculation and separation through sedimentation and filtration. Aluminium (e.g. Al2(SO4)3 .18H2O)
and iron salts are mostly used as coagulant reagents.
Recent studies (Ngabigengesere & Narasiah, 1998;
Katayon et al., 2005) have indicated a number of serious
drawbacks linked to the use of aluminium salts such as
Alzheimer’s disease associated with high aluminium residuals in treated water, excessive sludge production
during water treatment and considerable changes in
water chemistry due to reactions with the OH– and alkalinity of water. In addition, the use of alum salts is inappropriate in some developing countries because of the
high costs of imported chemicals and low availability of
chemical coagulants (Schultz and Okun quoted by
Katayon et al., 2005).
A number of studies have pointed out that the introduction of natural coagulants as a substitute for metal
salts may ease the problems associated with chemical coagulants (Katayon et al., 2005). Using natural coagulants such as the seeds from the Moringa Oleifera MO
tree instead of aluminium salts might give advantages,
such as smaller costs of water production, less sludge
production and ready availability of reagents. There are
also some disadvantages such as increased concentration
of nutrients and COD in the treated water due to the
organic nature of this type of coagulants.
In this paper, the potential of using Moringa Oleifera
as an alternative to aluminium sulphate for drinking
water treatment in Mozambique is evaluated, its limits
analyzed and the optimal use and dosage assessed. Standard Jar-test experiments performed with solutions prepared from aluminium and MO seeds were used to
compare the efficiency in turbidity removal and the impacts on the water chemistry. Filtration experiments in a
coarse to fine media filter were used to compare filtration efficiency of suspensions prepared through coagulation with Al and MO.
The more specific objectives for this study were to:
• Evaluate the optimum dosage of MO for different levels of turbidity, and its removal efficiency at each level.
• Compare the treatment efficiency of MO to that of
aluminium sulphate, regarding both treatment efficiency and influence on water quality and characteristics.
• Find a suitable method of preparation for the MO
coagulant, and establish a procedure manual for the
preparation, use and dosage of MO in order to use it
for drinking water treatment.
138
• Investigate the possibilities of using MO on an industrial scale, regarding availability and reliability of production and distribution.
2. Background
In nature, water is always contaminated with various
types of pollutants. Some of them are harmless and
sometimes even desired in water whereas others need to
be removed before the water can be used for drinking
purposes. Physical properties such as turbidity, colour
and solids impact the aesthetic appearance of water and
its acceptability for consumption. The microbiological
quality has a large effect on the taste and smell of water
and can sometimes be a large problem in river water.
Eutrophication of surface water sources due to nutrients
disposal (e.g., from agriculture and wastewater) and
physical properties such as pH and temperature, favours
algae and bacteria growth and can cause health risks.
Bacteria in water can cause illnesses such as typhoid (Salmonella typhus), cholera (Vibrio cholera), and diarrhoea
(Giardia lamblia). Faecal coliforms and streptococci indicate contamination from human or animal faeces.
The aim of drinking water treatment is to remove impurities and bacteria in order to meet the quality guidelines for drinking water (WHO, 2004). The design of
water treatment process varies between different treatment plants. It also depends on the quality of raw water
and the requirements regarding treated water quality.
Surface water generally requires more treatment than
ground water, since the former is more easily contaminated.
Conventional treatment is mostly used for surface
water treatment. It generally involves chemical coagulation followed by flocculation, sedimentation, filtration
and disinfection. Common coagulants are aluminium
sulphate, ferric chloride, polyaluminium chlorides and
synthetic polymers. All of these coagulants have in common the ability of producing positively charged ions
when dissolved in water, which can contribute to charge
neutralization (Degrémont, 1979; Hammer et al., 2004).
The dosage of coagulant depends on several parameters
such as type and concentration of contaminants, pH
and temperature. The optimal dosage for specific water
is defined as the dosage which gives the lowest turbidity
in the treated water. Dosage beyond the optimum point
will, apart from obvious disadvantages such as increased
aluminium/iron content in the treated water, also lead
to an increase in turbidity.
If the sedimentation step is omitted, and the flocculated water is led directly to filtration, the process is
called direct filtration (McConnachie et al., 1999). A
pre-requisite is that the raw water is of seasonally uniform quality with turbidity routinely less than 5NTU. If
VATTEN · 2 · 08
Figure 1. The Moringa Oleifera tree (left) and
dried and shelled seed (above).
© M. Begman.
otherwise, pre-treatment is required. This includes pretreatment with contact basins, flocculation chambers
and roughing filters with flocculation processes. The latter can be designed either with a horizontal flow direction (HRF), up-flow (UPRF) or down-flow (DFRF). In
all cases, coarse media of decreasing size in the direction
of flow is used (Sánchez et al., 2006).
If flocculated water is filtered first in an up-flow
roughing filter prior to its final treatment in rapid sand
filters the processes is called up-flow-down flow filtration. This method is claimed by researchers (Ingallinela
et al., 1998; McConnachie et al., 1999) as capable of
producing treated water of quality equal to that of standard coagulation-flocculation followed by sedimentation
and filtration, but, at lower investment and operational
costs. Surface water treatment can also be treated by
slow sand filtration-SSF. In this case, pre-treatment is
always needed because SSF cannot handle high turbidity
values of the raw water. For pre-treatment roughing filters are mostly used.
When conventional coagulation is used, chemical
coagulants such as iron and alum salts are needed but,
natural coagulants such as the seeds from the Moringa
Oleifera tree can also be used. Traditionally, surface water
has been treated with the help of herbs as natural coagulants for centuries in India. Ripe seeds of Strychnos potatorum, wiry roots of the rhizome of Vetiveria zizanioides,
seed coats of Elettaria cardamomum and leaves from
Phyllantus emblica have all been recorded to be used for
water treatment in past and present times (Sadgir,
2007).
VATTEN · 2 · 08
The Moringa Oleifera (MO) tree (Figure 1) is a perennial plant that grows very fast, with flowers and fruits
appearing within 12 months after planting. The tree
grows up to a height of 5–12 meters with branches extending between 30 and 120 cm (Lilliehöök, 2005).
With its origin in India and Pakistan the Moringa
Oleifera plant was brought to the Africa continent and
Sudan, in particularly during the colonial era for ornamental purposes. The natural coagulant found in MO is
present in 6 of the 14 species of MO growing in Africa,
Madagascar, India and Arabia. Knowledge that MO
seeds can purify water is not new; the seeds have been
used for water treatment for generations in countries
like India and Sudan. For example, the women of Sudan
have used the seeds from the MO three for water treatment since the beginning of the 20th century (Schwartz,
2000) through a technique that comprehended the
swirling of seeds in cloth bags with water for a few minutes and let it settle for an hour.
Scientifically, the coagulation properties of MO seeds
were first confirmed by the German scientist Samia Al
azharia Jahn (Schwartz, 2000). The active agent is believed to be a protein, but the exact form of the protein
is not yet known. Recent research has identified proteins
of sizes ranging from 3 to 60 kDa, all possessing coagulating ability. The protein(s) act as a cationic polyelectrolyte, which attaches to the soluble particles and creates
bindings between them, leading to large flocs in the water. Stirring and mixing accelerates the electrostatic flocculation, and the flocs condense the contaminants
(Göttsch, 1992).
139
Extraction of the active agents can be done in several
ways but most techniques follow the pattern: dried seeds
grained or powdered with or without shells; powder
mixed with a small amount of water and, the solution is
stirred and filtrated (Ndabigengesere & Narasiah, 1998;
Muyibi & Alfugara, 2003; Ghebremichael et al., 2005).
The filtered solution is called “crude extract” or “stock
solution” and can be used for water treatment without
further preparation. Several studies have shown that the
use of salt water and/or tap water is more efficient as
solvent than e.g. distilled water. Okuda et al. (1999)
showed for example that the crude coagulation capacity
was up to 7.4 times higher when the active agents were
extracted with salty water as compared to distilled water.
The reason for this is assumed to be that the coagulating
protein is more soluble in water with high concentration
of ions (Okuda et al., 2001a). Other studies have focused
on purifying the active agent as much as possible and
producing a stable protein powder without excessive organic matter. Two separate studies show that the active
agents could be purified from the extract using a cation
exchanger column, leading to reduced levels of COD in
the treated water (Ghebremichael, 2005; Ndabigengesere & Narasiah, 1998). However, a more low-tech way
of reducing the organic content is to extract the oil from
the seeds with an organic solvent (Ghebremichael,
2005).
The treatment efficiency is dependent on the turbidity of the raw water, as revealed in previous studies by
Katayon et al., (2004). MO has also been proven to produce significantly less sludge than aluminium sulphate,
which is an advantage especially if the sludge is to be
dewatered or treated in some other way before disposal
(Ndabigengesere et al., 1994). MO can also be used in
combination with other coagulation salts, such as aluminium sulphate (Sutherland et al., 1994).
The coagulation and flocculation ability of the seeds
has been investigated in several different studies around
the world (Ndabigengesere & Narasiah 1998; Bengtsson, 2003; Muyibi & Alfugara, 2003). These studies
have shown that neither pH nor alkalinity nor conductivity was affected during water treatment, but an increase
in COD, nitrate and orthophosphate has been observed.
Other studies have indicated that treatment with MO is
dependent on the pH of the raw water were optimum is
above neutral (Okuda et al., 2001b) whereas others say
it is independent of raw water pH (Schwartz, 2000).
Storage of seeds and extract and its influence on coagulation properties has also been investigated by Katayon (Katayon et al., 2004; Katayon et al., 2006). Seeds
were dried, crushed and stored in different containers
and at different temperatures. These studies concluded
that the temperature and type of container did not have
a significant effect on treatment efficiency but that the
140
duration of storage did. The seeds stored for one month
showed better treatment efficiency than seeds stored for
longer periods (three to five months). Storage was also
found to influence the coagulation properties of the extract and to decrease treatment efficiency with increased
duration (Katayon et al., 2006). The study does not discuss the reason for this but it could be assumed that it is
due to microbial degradation of the proteins. The study
also concluded that the duration of storage of extract
should not exceed 24 hours as degradation of active
agents is assumed to occur within this time. As noted by
Katayon et al., (2004) extract solutions stored for longer
than 2 to 3 days had between 73.6 % and 92.3 % lower
turbidity removals as compared to fresh solutions.
3. Methodos
3.1. Jar tests experiments
The equipment used for jar test experiments was a Janke
& Kunkel jar test apparatus with 5 beakers of 2.0 l capacity each (Figure 2). Each beaker was filled with 1.5 l
of test water with identical turbidity. Different volumes
of coagulant reagent were added to 4 of the five beakers
with the last used as the blank sample. Mixing of the
coagulant with water was provided by flash mixing
during approximately 3 minutes with propellers set at
120 rpm followed by slow mixing at 40 rpm during
approximately 17 minutes. Then the propellers were
stopped and the content of the jars left to settle for approximately 30 minutes. After sedimentation, samples
were taken for water quality determination.
For each coagulant and turbidity level, three identical
jar tests were performed in order to obtain statistically
reliable results. However, some of the parameters were
only measured during one of these three jar tests and/or
Figure 2. Jar test equipment.
VATTEN · 2 · 08
in the jar with the optimal dosage, due to restricted time
and economic means. If the optimal dosage was not
found in the jar test, a new test with new dosage was
carried out until the optimum was found.
3.2. Preparation of extract and test water
Dried and shelled MO seeds were obtained from IIAM
(Agronomic research institute of Mocambique) via the
Department of Chemical Engineering at Faculty of Engineering of Eduardo Mondlane University. The shells
were ground to a fine powder using a mortar. The powder was then weighed and dissolved in distilled water to
make a 50 g/l solution. The solution was stirred for 30
minutes using a magnetic stirrer, and finally filtrated
through a Whatman filter no. 40. A fresh solution was
prepared every day in order to avoid ageing effects. Two
alternative preparation methods were used, one involving oil extraction from the seeds and the other using tap
water instead of distilled water. To extract the oil, the
ground seeds were first dissolved in cyclo-hexane, stirred
for 30 minutes and then filtrated through a Whatman
filter no. 40. The remaining solids in the filter (“press
cake”) were then dissolved in water, stirred and filtrated
according to procedures described previously. The tap
water preparation process was identical to the standard
method, but with tap water as solvent instead.
Raw water from Umbeluzi river was used for all experiments. During the period of the experiments (September and October) the river water had low values of
natural turbidity, therefore most experiments were performed with synthetic turbidity water. This was done by
using ordinary clay, obtained from the Geology department of Eduardo Mondlane University. The clay was
first ground with a mortar to make the particles as fine
as possible, and then added to the water in sufficient
amounts to produce the desired levels of turbidity.
3.3. Measurements and analytical methods
Turbidity, suspended solids, temperature, pH, EC (electrical conductivity), total dissolved solids (TDS), alkalinity, COD (chemical oxygen demand), and bacteria
were the water quality parameters measured. Turbidity
was measured using a 2100P turbidimeter from Hach.
In order to increase reliability of measurements, water
turbidity readings were tripled and that of settled water,
doubled and the average values used as reference values.
Temperature was measured with a standard mercury
thermometer (accuracy of ± 1°C) held for 1 minute in
the water, and the observed values rounded to the nearest integer. pH and EC were measured using handheld
digital meters from Wagtech International Ltd, and
TDS was measured using an handheld digital meter
VATTEN · 2 · 08
TDScan Low-range (0–1990 ppm) from Eutech Instruments. All readings were taken with the digital meters
held for 1 minute (or until a stable value was reached) in
the sample water.
Alkalinity was measured with a simplified titration
described in Standard Methods (1998). The samples were
titrated with hydrochloric acid using an 725 Dosimat
automatic titration equipment from Metrohm. The
added volume of acid was noted at the colour changes
(from pink to transparent for phenolphthalein and from
blue to yellow for the mixed indicator).
COD and bacteriological analyses were conducted by
the laboratory at the Ministry of Health in Maputo. For
the bacteriological analysis, sterilized bottles were provided by the laboratory. During sampling, the bottles
were filled completely to minimise the dissolution of air
oxygen, and thereby aerobic degradation, in the samples. Microbial parameters analysed were total coliforms,
faecal coliforms and faecal streptococci.
Suspended solids were also determined using standard procedures described in the Standard Methods
(1998). The concentration of suspended solids was
measured for 6 different levels of turbidity and a calibration curve plot, in order to provide a relationship between turbidity and suspended solids. This was done
since the measurement of suspended solids is very time
demanding, up to 2 days for each set of measurements,
whereas turbidity is measured in less than a minute.
With the calibration curve, turbidity values could be
easily converted into approximate values of suspended
solids concentrations.
3.4. Filtration experiments
Filtration experiments were performed in a pilot scale
plant consisting of an up-flow roughing filter followed
by single media gravity rapid sand filter. Pre-coagulated
water with MO was used as test water. The test water
turbidity was of about 30 NTU, and the dosage of
MO was set to 50 mg/l. The flow rate was set to 60 l/h
through the roughing filter, and 40 l/h through the
single media filter. Turbidity and head loss measurements were used to assess the pilot plant performance.
Turbidity readings were taken as described previously
and head loss readings were taken using a tube-type differential pressure gauge. Turbidity and head loss readings were taken at regular time intervals of 30 minutes.
3.5. Sources of error
The procedure of the experiments was done consistently
through the whole study to minimise the sources of
error. Possible errors in the study might arise from the
lack of calibration for the equipment used in measuring
141
Table 1. The range and average quality of the raw water used in this study.
Test water1
Parameter
Raw water
Range
(historical)
Range
(study period)
Range
Average
Turbidity (NTU)
Temperature (°C)
pH
Alkalinity (mg CaCO3/l)
TDS (mg/l)
EC.(μS/cm )
3.8–173
21–27
6.7–8.7
140–160
370–410
550–630
4.6–6.8
22–24
8.2–8.4
209.6–211.1
480–490
710–720
5.2–100
21–25
8.2–8.6
206.4–215.2
480–500
700–730
36.1
23
8.4
212
490
720
1
synthetic turbidity water
turbidity, pH, conductivity and TDS. The dosage for
the flocculation was not done at exactly the same time in
each jar, leading to time differences for the mixing of the
water. However, the main factors affecting the jar test
results are believed to be differences in preparation of
raw water and MO extract.
COD and bacteriological analyses were done at the
Health ministry. The time passing between filling the
bottles with samples and the actual analysis affects the
result since organic matter can be degraded if stored too
long. The large variation of the results (and large deviation from the values recorded in 2003 and 2006 in the
Umbeluzi River) suggests that either the sampling
method or analysis procedure were not satisfactory, and
that the results are not reliable.
4. Results and discussion
4.1. Raw and test water
physio-chemical characteristics
The general physio-chemical quality (range and average
values) of the raw and test water is presented in Table 1.
All variables were measured at the laboratory, prior to
each jar tests experiment.
As shown in Table 1, the raw water quality did not
vary much during the period of experiments but analysis of historical data suggests that large variation can be
expected, particularly in relation to turbidity. It is also
seen from Table 1, that measured values of conductivity,
total dissolved solids and alkalinity were generally higher
than given values from previous measurements. Since
raw water samples were collected from a location just
downstream the waterworks where aluminium sulphate
and lime are used for coagulation and neutralization the
observed increase in raw water pH and alkalinity could
be explained by the disposal of sludge and waste water
containing Al(OH)3 from the plant. Also, downstream
142
the waterworks, the river is sometimes affected by saltwater intrusion because of its proximity to the river
mouth. This may explain the observed increases in EC
and TDS. The sludge from the waterworks also contains
large amount of ions, which also explains the increase in
conductivity and TDS.
Most coagulants perform less effective when the water
temperature is low or when the raw water pH and alkalinity experiences fluctuations. Unlike turbidity, these
parameters show very small variations during the year.
Therefore during preparation of test water these parameters were not changed but the raw water turbidity
was. A calibration curve for converting turbidity readings into approximate levels of suspended solids was also
developed and is presented in Figure 3.
Accordingly, an almost linear relationship between
turbidity and suspended solids is observed particularly
when the raw water turbidity is high (>10NTU). The
reason for the bad correlation at low turbidity values is
probably due to the fact that the total mass of suspended
solids in a 200 ml sample when the turbidity is low (less
than 10NTU) is very small, around 1 mg, which is dif-
Figure 3. Estimated relationship between turbidity and suspended
solids.
VATTEN · 2 · 08
Figure 4. Turbidity removal for raw water turbidity values in the range 5 to 100 NTU and aluminium sulphate as coagulant.
ficult to measure with acceptable precision. A larger
sample volume would increase the precision, but the
measurements were already time consuming and unfortunately there was not enough time to filter larger volumes.
solid Al2(SO4)3 .18H20. The coagulant formed large
flocs that settled in 30 minutes and lead to a stable
outgoing turbidity of 0.5–0.6 at the optimum dosage.
The relative standard deviation for outgoing turbidity at
optimum was 10–20 %, corresponding to approximately
0.1 NTU.
4.2. Optimum dosage and turbidity removal
4.2.1. Coagulation-flocculation with Alum sulphate
Results of optimum conditions of turbidity removal
using Aluminium sulphate as the main reagent are presented in Figure 4 from which it is seen that coagulation
with aluminium sulphate resulted generally in high removal efficiencies, irrespective of the raw water turbidity.
The optimum dosage in all cases was between 3.0 and
4.0 mg Al3+/l, corresponding to approximately 0.1g/l of
VATTEN · 2 · 08
4.2.2. Coagulation with Moringa Oleifera
Coagulation-flocculation with Moringa Oleifera was
done with reagent solutions extracted in three different
ways. Moringa Oleifera extracts prepared with distilled
water, was chosen as the standard preparation method to
comply with earlier studies and also to reduce the
number of unknown parameters in the tests. The other
methods considered solutions prepared with distilled
143
Figure 5. Optimum dosage and turbidity removal at different levels of raw water turbidity using Moringa Oleifera extract prepared
according to standard method.
water and oil extraction and solutions prepared with tap
water.
The results of optimum dosages and turbidity removal with MO prepared with the standard procedure
are presented in Figures 5. The coagulant dosage indicates the mass of seeds that were used initially per litres
of raw water, not the actual concentration of MO extract
in the water. This difference is important to note since,
a lot of the seed mass was separated during the filtration
step when preparing the extract. The exact concentration of MO in the crude extract is therefore unknown.
For medium and high turbidity levels in the raw
water (30–100 NTU), the optimum dosage was found
between 40 and 70 mg/l and to more or less increase
144
with increasing raw water turbidity. The outgoing turbidity at the optimum dosage ranged between 1–1.5
NTU. For low values of initial turbidity (5–15 NTU),
the process was less effective since an optimum reagent
dosage was generally not attained and the lowest outgoing turbidity remained around 3 NTU. Coagulation
with MO extracted with distilled water resulted therefore in high removal efficiencies when the raw water turbidity was high but in poor efficiencies when the raw
water had low values of turbidity (5 and 15 NTU). In
the latter case, the treated water still had high amounts
of suspended flocs even after 30 minutes of sedimentation. This indicates that flocs formed were either too
small or not dense enough to settle within the 30 minVATTEN · 2 · 08
Figure 6. Residual turbidity after 30 minutes (filled dots) and 2 h
of sedimentation (empty dots) respectively, using an initial turbidity of approximately 5 NTU. Dotted lines show individual jar test
results while solid lines indicate the average values.
utes chosen for sedimentation. Complementary test
with longer sedimentation time (Figure 6) confirmed
this observation and resulted in significantly better removal efficiencies.
The prolonged sedimentation time resulted in an
average outgoing turbidity of 1.4 NTU, a significant
improvement. The extra sedimentation time indicates
that coagulation of low turbid waters, using MO can
only be effective if accompanied with long sedimenta-
tion times, which, in practical terms means large sedimentation basins.
The results of coagulation with MO prepared with
distilled water and oil extraction as well as with tap water
are presented in Figure 7. Two different levels of turbidity were tested; 15 NTU and 50 NTU. Preparation with
oil extraction was conducted as recommended in previous studies (Ghebremichael, 2005; Narasiah & Ndabigengesere, 1998) which indicate that oil extraction
helps prevent increases in COD of treated water. Preparation with tap water was used because it is more convenient for large scale production.
When solutions prepared with oil extraction were
used, the average outgoing turbidity was 2.4 NTU, for
15 NTU initial turbidity and 1.9 NTU for 50 NTU
initial turbidity at optimum dosage rates of 67 mg/l and
100 mg/l respectively (Figure 7, top). The relative standard deviation at optimum was 10 % for the low turbidity
level, and over 50 % for the high turbidity. The latter
was due to one jar test with much higher results than the
other. The reason for this different result may have been
bad coagulation properties of the seeds used to prepare
the crude extract that day. The same extract was used in
one of the jar tests at 15 NTU, and gave higher outgoing
turbidities in that test as well, although not as extremely
high as in the 50 NTU jar test. Another possibility is
that the extraction process with cyclo-hexane removed
Figure 7. Turbidity removal efficiency with Moringa Oleifera prepared with oil extraction (top) and tap water (bottom) and initial
turbidity in the raw water of 15 and 50 NTU.
VATTEN · 2 · 08
145
some of the coagulating agents as well, thus lowering the
coagulating ability of the crude extract.
When solutions prepared with tap water were used,
the average outgoing turbidity was of 1.85 and 1.3 NTU
for low and medium turbidity respectively (Figure 7,
bottom). Optimum dosage rates were of about 33 mg/l
for water with low (15 NTU) and medium (50 NTU)
initial turbidity. The standard deviation at optimum was
10–15 %.
4.3. Effect on water quality
The effect of the reagents on some of the treated water
quality, notably the water pH, alkalinity and EC is illustrated in Figure 8. As can be seen, coagulation with
aluminium sulphate (Figure 8, right) led to a decrease in
the water pH and alkalinity and to an increase in conductivity. These trends were confirmed statistically at
0.05 level of significance. These effects are also well
known from previous studies (kemira, 2003) and use of
aluminium sulphate throughout the world.
At an initial turbidity level of 15 NTU, the results
from the COD analyses indicated a modest increase in
COD from 1.6 to 2.4 mg O2/l. Bacterial contamination
expressed as the number of total coliforms per 100 ml
were reduced from >100 to 24. Faecal coliforms on contrary, increased from <1 to 18 counts per 100 ml while
faecal streptococci were found to be >100 counts per
100 ml both before and after treatment. At 50 NTU
initial turbidity, the COD decreased from 12.8 to 9.6
Figure 8. Effect of alum (on the left) and MO (on the right) on the pH, alkalinity and Ec. of treated water.
146
VATTEN · 2 · 08
mgO2/l. Altogether, these results are too few and too
diverse to be reliable, and no conclusions can be drawn
regarding the effect of aluminium sulphate on COD
and on the microbial quality of treated water.
Treatment with Moringa Oleifera had no effect on the
pH, alkalinity or conductivity of the treated water. The
impact on COD levels and on the bacteriological quality
of water was evaluated for initial turbidity of 15 and 50
NTU but the results are inconsistent. At 15 NTU, the
COD results indicated an increase from 2 mg O2/l to
2.4 mg O2/l after sedimentation at optimum dosage,
whereas at 50 NTU the result was a decrease from
40 mg O2/l to 7 mg O2/l. Considering the yearly variations of COD in the Umbeluzi river in 2006, where no
value exceeding 9 mg O2/l was reported the high COD
levels found in the samples indicates either that readings
were misleading or that external sources of organic pollution (e.g. from the clay used to make artificial turbidity) existed.
The COD results from the 15 NTU turbidity level
supports however the findings from previous studies
(Ghebremichael, 2005; Ndabigengesere & Narasiah,
1998) that the use of MO leads generally to an increase
in COD levels in the treated water. As for the bacteriological quality, the count in raw water resulted in >100
counts per 100 ml for all three bacteria types at 15 NTU
turbidity level, and <1 counts per 100 ml for the equivalent at 50 NTU turbidity level. The amount of faecal
streptococci were reduced to 3 after treatment at 15
NTU turbidity level, and the faecal coliforms were increased to 7 after treatment at 50 NTU turbidity level.
4.4. Comparison between coagulants and
preparation methods
4.4.1. Optimum dosage and turbidity removal
Figure 9 provides a comparison of turbidity residuals
after treatment at optimum dosage rates for all scenarios
of raw water turbidity and coagulant reagent used. As
can be seen, the efficiency of MO compared to aluminium sulphate was significantly lower in the jar tests.
Aluminium sulphate led to outgoing turbidities of
0.5–0.7 NTU regardless of the initial turbidity, whereas
MO never produced water with turbidity below 1 NTU.
Treatment with aluminium sulphate resulted generally
in a more stable effluent quality as is indicated by the
relative standard deviation which was never more than
20 %, corresponding to 0.1 NTU.
In contrast, treatment with MO resulted in an effluent quality that varied considerably with the relative
standard deviation reaching as high as 50 %. The relative
standard deviation at optimum for high initial turbidity
levels was however lower for MO than for aluminium
sulphate. This could have resulted from the fact that,
VATTEN · 2 · 08
Figure 9. Comparison of optimum dosage for different coagulants.
when using Al for coagulation of water with high initial
turbidity, high amounts of reagent are required which
may lead to the formation aluminium hydroxide precipitates that increase turbidity levels in the effluent.
Overall, MO prepared with tap water was more efficient than the other two methods of preparation. The
outgoing turbidity was lower, especially at high initial
turbidity level, and the dosage needed to reach optimum
was significantly lower. Coagulation with MO resulted
also in smaller and lighter flocs than with aluminium
sulphate. Small and light flocs settle more slowly therefore, they remain longer in the supernatant water. This
explains the higher outgoing turbidity levels when MO
was used as coagulant. Treatment efficiencies with MO
could be improved by prolonging the sedimentation
time to about 2 hours. While at laboratory scale these
improvements are possible, at large scale operation, increasing duration of sedimentation means that larger
investment is needed for construction of larger sedimentation basins. This is generally not feasible. Also, large
scale operations flocs that do not settle during sedimentation will continue to subsequent stages of water treatment (filtration), where they may be removed at the
expenses of higher costs of operation of filters.
4.4.2. Influence on water quality and characteristics
The effect of adding coagulation reagents on water quality and characteristics is illustrated in Figure 9. As can be
seen MO coagulant shows a major advantage compared
to aluminium sulphate; it does not affect neither pH
and alkalinity nor conductivity and TDS, whereas aluminium sulphate influences all of these. No specific conclusion can be drawn regarding the effect on COD level
and bacteriological quality of the water, due to lack of
analysis results and large uncertainties in the existing results.
4.5. Filterability of formed suspensions
In large scale operations, the removal of particles from
suspensions formed during coagulation-flocculation is
147
Figure 10. Time-dependent turbidity residuals in the up-flow filter and single media gravity rapid sand filter and head loss development
in the single media filter.
accomplished by sedimentation followed by filtration in
case of conventional treatment or by direct filtration
when the sedimentation step is skipped. In order to assess the filterability of suspensions formed during coagulation with Moringa Oleifera, filtration experiments were
conducted in a pilot plant consisting of an up-flow
roughing filter used for hydraulic flocculation and a single media sand filter used for final treatmen. The results
are resumed in Figure 10.
Raw water with an average initial turbidity of 30
NTU previously treated with MO extracts was used.
Turbidity removal and head loss development were the
variables used to evaluate the pilot plant’s performance.
The results obtained for the pre-filter (Figure 10-top
left), indicates that turbidity removal was generally low
and in average did not exceed 20 %. The effluent turbidity remained steady during the first 4.0 to 4.5 hours of
filtration run but soon after that it started deteriorate
and to experience variations. Two major factors might
have contributed to this; first the relatively thin aggregates formed when coagulation was performed with
MO as noted previously in this study and secondly the
fact that the up-flow filter as was filled with a relatively
coarse filtering material through which the relatively
thin aggregates could flow without being retained.
The performance of the sand filter was much better.
148
Average turbidity removals were in general above 92 %
and the filtrate quality remained below 2.0 NTU
throughout the entire test. As can be seen from Figure
10 (top-right) most of the particles were retained within
the top 50–60 cm of the filter bed. This indicates that
particles present in the suspensions transferred from the
pre-filter were generally thin and could easily penetrate
through the relatively fine filtering material used in the
rapid sand filter. This also supports the observation
made previously that flocs formed during coagulation
with MO are generally thin and light which means they
are poorly retained in coarse medium filters (e.g. roughing filters) or sedimentation tanks but are effectively removed in rapid sand filters.
The increase in head loss over filters indicates generally the extent of clogging due to particles being retained
in the filter media. As shown in Figure 10, the head loss
developed more or less linearly in the rapid sand filter.
This indicates a time-dependent reduction of the filter
bed porosity as a result of accumulation of particles, and
also that a uniform penetration of particles through the
depth of the filter bed was observed. It is also seen from
Figure 10 that the head losses developed very rapidly
during the initial stage of filtration, a fact that was
attributed to rapid accumulation of particles in the
upper layers of the filter bed.
VATTEN · 2 · 08
The pressure drop in the pre-filter was in all cases of a
few centimetres and in general negligible. This resulted
from the fact that the pre-filter did not contribute for
the retention of aggregates formed during coagulationflocculation.
Similar tests performed with aluminium sulphate as
the main coagulant and with raw water turbidity in the
range 4.0 to 22NTU produced better results. Overall,
filtrate turbidities ranging from 0.4–0.8NTU could be
attained at the end of the treatment train. The increase
of head loss pressure in both units was significant and
generally higher than that observed when using MO.
The results also indicated that aggregates formed during
coagulation were effectively retained at the pre-filter and
rapid sand filter. This is an expected pattern since coagulation with aluminium sulphate generally results in large
and dense aggregates which are easily retained even within coarse filter media such as that used in the pre-filter.
2
3
4.6. Applicability at large scale operation
The optimum dosage of MO, using tap water, was
found between 17 mg/l and 67 mg/l for both low (15
NTU) and medium (50 NTU) initial turbidities. Assuming daily water production of about 5 000 m3/d
which is enough for an average of 50 000 people at an
average of 100l/person/day, an average dosage of about
40 g MO seeds/m3, an average daily supply of about
0.2 tons of MO seeds is required if alum sulphate is
replaced with MO. This corresponds to a plantation of
about 60 ha if an average of 3kg seed kernels/tree is
assumed with a tree spacing of about 3 m (WELL,
1999).
The area required for production of MO seeds is not
entirely unrealistic particularly for small to medium size
water supplies located in rural areas. Moreover, a hypothetical change from aluminium sulphate to MO may
bring about significant reductions in transportation
costs of imported chemicals. Yet, additional investments
will be required concerning facilities for storage, grinding and mixing of the seeds.
The preparation of the extract should be carried out
on-site to minimize transport costs and also because the
extract must not be stored for too long before use. Since
the extracts can be stored for one day without losing coagulation properties, batch processes designed for the
demand of one full production day are probably more
suitable than continuous process.
5. Conclusions
1 MO shows good coagulating properties, and has many
advantages compared to aluminium sulphate; it does
not affect the pH, alkalinity or conductivity of the
VATTEN · 2 · 08
4
5
water and it can be produced locally at low cost. The
optimum dosage of MO, using tap water preparation
method, was found to be between 17 mg/l and
67 mg/l for both low (15 NTU) and medium (50
NTU) turbidity levels.
The extraction of MO should be performed using tap
water as this is the cheapest, most practical and most
efficient method. MO does not show the same efficiency in turbidity removal as aluminium sulphate,
particularly for low turbid waters. In this case, turbidity removal can be increased by increased sedimentation time at the expenses however of large investment
and operational costs.
MO combined to direct filtration is less effective for
turbidity removal than aluminium sulphate with direct filtration. Yet, the treated water turbidity when
MO was used was within acceptable limits for drinking water production (less than 2 NTU) and the increase in head losses over the filters was not higher for
MO than for aluminium sulphate. This suggests that
MO could be a very good substitute for aluminium
sulphate when using this technique.
MO is a method that can be considered as a good,
sustainable and cheap solution for smaller waterworks,
if the supply of MO seeds can be guaranteed. Tap
water extracted MO and treatment with flocculation
followed by direct filtration processes should be considered in the event of expansion or construction of
small scale waterworks. Complementary tests should
however be carried out in order to determine the impact of raw water pH on treatment efficiency.
Overall, the amount of seeds required for production
of MO extract is quite large. On the other hand, the
knowledge about actual production of MO seeds in
Mozambique is limited which means the potential for
large scale use of MO in drinking water production
still needs further research. Aspects concerning pH
dependence, COD increase and optimum direct filtration treatment design should also be further examined before the method is implemented at large scale
basis. However, once plantations are established and
the supply of seeds secured, MO provides a good,
cheap and sustainable alternative to aluminium sulphate which should be considered as a coagulant in
smaller waterworks.
References
APHA, AWWA, WEF, (1998). Standard Methods for the Examination of Water and Wastewater, 20th ed. American
Public Health Association, Washington, DC.
Bengtsson Jessica, (2003). Förbehandling av kommunalt avloppsvatten före anaerob behandling, Avdelningen för
vattenförsörjnings- och avloppstenik, Lunds Tekniska
Högskola.
149
Degrémont (1979). Water Treatment handbook, 5th edn., Halsted Press, John Wiley & Sons Inc.
Ghebremichael K. (2007). Overcoming the drawbacks of natural coagulants for drinking water treatment. Water Malaysia IWA Conference Developing Countries 2007.
Ghebremichael K. A., Gunaratna K. R., Henriksson, H.,
Brumer, H. and Dalhammar, G. (2005). A simple purification and activity assay of the coagulant protein from Moringa oleifera seed. Water Research, Vol 39(11), June 2005,
pp. 2338–2344.
Hammer, Mark J. and Hammer Jr, Mark J. (2004). Water and
Wastewater Technology, 5th edn., Pearson Education International, ISBN 0-13-191140-6.
Ingallinelia Ana Maria, Stecca L. M., Wegelin,M (1998). Upflow Roughing Filtration: Rehabilitation of a water treatment plant in Tarata, Bolívia. Wat.Sci.Tech. vol. 37 (9),
pp 105–112
Katayon S., Megat Mohd Noor M.J., Asma M., Abdul Ghani
L.A, Thamer A.M., Azni I., Ahmad J., Khor B.C. and
Suleyman A.M. (2006). Effects of storage conditions of
Moringa Oleifera seeds on its performance in coagulation.
Bioresource Technology, Vol 97(13), September 2006,
pp. 1455–60.
Katayon S., Megat Mohd Noor M.J., Asma M., Thamer A.M.,
Liew Abdullah A.G., Idris A., Suleyman A.M,, Aminuddin
M.B. and Khor B.C. (2004). Effects of storage duration
and temperature of Moringa Oleifera stock solution on its
perfomance in coagulation. International Journal of Engineering and Technology, Vol. 1(2), pp. 146–151, ISSN:
1823-1039.
Kemira K. (2003). About water treatment, ISBN: 91-6314344-S.
Lilliehöök, H. (2005). Use of Sand Filtration on River Water
Flocculated with Moringa Oleifera. Division of Sanitary
Engineering, Department of Civil and Environmental Engineering, Luleå University of Technology, ISRN: LTUEX-05/177- SE.
McConnachie G.L., Folkard G.K., Mtawali M.A. and Sutherland J.P. (1999). Field trials of Appropriate Hydraulic
Flocculation Processes. Water Resources vol. 33, No. 6,
pp 1425–1434.
Muyibi, Suleyman A. and Alfugara, A. M. S. (2003). Treatment of surface water with Moringa Oleifera seeds and
alum – a comparative study using a pilot scale water treatment plant. International Journal of Environmental Studies,
Vol. 60(6), pp. 617–626(10).
150
Ndabigengesere, A., Narasiah, K.S. (1998). Quality of water
treated by coagulation using Moringa oleifera seeds. Water
Research, Vol. 32 (3), pp.781–791(11).
Ndabigengesere A, Narasiah K. S., Talbot B.G. (1995). Active
agents and mechanism of coagulation of turbid waters usng
Moringa Oleifera. Water Research, Vol 29(2), pp. 703–710.
Okuda T., Baes A.U., Nitshijima W., Okada M. (1999). Improvement of extraction method of coagulation active components from Moringa Oleifera seed. Water Research, Vol
33(15), pp. 3373–3378.
Okuda T., Baes A.U, Nitshijima W, Okada M. (2001). Coagulation Mechanism of Salt Solution-Extracted Active Component in Moringa oleifera Seeds. Water Research, Vol 35(3)
pp. 830–834 (a.).
Okuda T., Baes A.U, Nitshijima W, Okada M. (2001). Isolation and charcterization of coagulant extracted from Moringa Oleifera seed by salt solution. Water Research, Vol
35(2) pp. 405–410 (b.).
Sánchez, L.D., Sánchez, A., Galvis, G and Latorre, J. (2006).
Multi-stage Filtration. International Water and Sanitation
Centre, The Hague, The Netherlands, pp 11–18.
Sadgir, Parag. (2007). Traditional water purification, 5th
IWHA Conference, Tampere, Finland.
Sutherland J.P., Folkard G.K., Mtawali M.A. and Grant W.D.
(1994). Moringa Oleifera as a natural coagulant, Affordable water supply and sanitation, 20th WEDC Conference
Colombo, Sri Lanka.
Sutherland J.P., Folklard G.K., Poirier Y.L. (2001). Moringa
Oleifera. The constraints to commercialization, Development potential for Moringa products, October 29th –November 2nd 2001, Dar es Salaam, Tanzania.
WHO (2004). Guidelines for drinking water quality. Vol 1
Recommendations (3rd edn.). Geneva, Switzerland.
Göttsch E. (1992). Purification of turbid surface water by
plants in Ethiopia. Walia 14, pp 23–28 http://www.deutsch-aethiopischer-verein.de/Walia-1992-Purification.pdf
(Retrieved on 13th of November).
Schwarz Discha (2000). Water clarification using Moringa
Olefiera. Gate information Service, Germany. http://www.
deutsch-aethiopischer-verein.de/Gate_Moringa.pdf (Retrieved 13th of August 2007).
WELL, (1999). Water, Engineering and Development Centre,
Loughborough University. http://www.lboro.ac.uk/well/
resources/technical-briefs/60-water-clarificationusingmoringa-oleifera-seeds.pdf (Retrieved 13th of August
2007).
VATTEN · 2 · 08
VATTEN – tidskrift för vattenvård
Officiellt organ för FÖRENINGEN VATTEN
Redaktion:
VATTEN, Teknisk Vattenresurslära, Lunds Universitet, Box 118, S-221 00 Lund.
Ansvarig utgivare:
Lars Bengtsson. Tel. 046-222 89 80.
Redaktör:
Rolf Larsson. (Rolf.Larsson@tvrl.lth.se) Tel. 046-222 73 98, Fax 046-222 44 35.
Sekreterare:
Clary Nykvist Persson. Tel. 046-222 98 82, Fax 046-222 44 35.
Tidskriften VATTEN har som mål att publicera artiklar om teknik och forskning rörande vattenresurser, vattenbehandling och vattenvård. Vid bedömning av en artikel tas hänsyn inte bara till
dess vetenskapliga nyhetsvärde utan även till dess praktiska värde (driftfrågor). Översiktsartiklar och
debattinlägg av allmänt intresse publiceras också. Referee-system tillämpas ej i formell mening, dock
granskas artiklarna av redaktionen och tillfälliga medarbetare.
Tidskriften VATTEN ska också informera Föreningen Vattens medlemmar om föreningens aktiviteter.
Särskilt utrymme reserveras för notiser om litteratur, konferenser, kurser etc. och meddelanden av
skilda slag från föreningar, företag, institutioner, universitet och högskolor samt massmedia.
Kortfattad instruktion för författare
Artiklar, inklusive figurer, skickas elektroniskt (via e-post eller annat medium) och som papperskopia
till redaktionen. Figurer och tabeller med tillhörande text placeras sist. Ett foto på varje författare ska
också bifogas.
Artiklar kan skrivas på skandinaviskt språk eller på engelska.
Alla artiklar ska ha abstract på engelska. Artiklar på skandinaviskt språk ska dessutom ha sammanfattning på samma språk som artikeln. Abstract och sammanfattning ska vardera omfatta högst 200
ord.
En lista med högst 10 key words bifogas artikeln.
Ange dignitet på rubriker, men minimera i övrigt layout.
Referenser ordnas alfabetiskt och skrivs enligt exempel:
Lindholm, T. & Ohman, P., 1996. Some advantages of studying living phytoplancton. Vatten 52: 1,
9–14.
Författare erhåller särtryck av artikeln i form av pdf-fil utan kostnad.
Antagna artiklar får ej publiceras i annan tidskrift utan redaktionens medgivande.
Utförlig instruktion för författare
– Se www.foreningenvatten.se (under »Verksamheten/Tidskriften VATTEN»)
– Kan även erhållas från redaktionen.
152
VATTEN · 2 · 08
VATTEN · 2 · 08
151