ALPS-GPS QUAKENET Project booklet
Sub-contractors
Università degli Studi di Trieste
Dipartimento di Scienze della Terra
www.dst.units.it
Interreg IIIB
This project has received European Regional Development
Funding through the INTERREG IIIB Community Initiative
UNIVERSITÀ DEGLI STUDI
DI GENOVA
Università degli
Studi di Genova
Dipartimento di Macchine,
Sistemi energetici e Trasporti
(DIMSET)
Facolta' di Ingegneria
www.unige.it
Politecnico di Milano
Dipartimento di Ingegneria
Idraulica, Ambientale,
Infrastrutture Viarie,
Rilevamento (DIIAR) sezione Rilevamento
www.rilevamento.polimi.it
Politecnico di Torino
Dipartimento di Ingegneria
del Territorio, dell'Ambiente
e delle Geotecnologie
Politecnico di Torino
II Facolta' di Ingegneria
sede di Vercelli
www.vercelli.polito.it
Università degli
Studi di Milano
Dipartimento di Scienze
della Terra "Ardito Desio"
www.gp.terra.unimi.it
IREALP
Istituto di Ricerca per
l'Ecologia e l'Economia
Applicate alle Aree Alpine
www.irealp.it
LGCA
Laboratoire de
Geodynamique des Chaines
Alpines, Chambery
www.univ-savoie.fr/labos/lgca/
Agencija Republike Slovenije Za Okolje
www.arso.gov.si
ARPA
Regione Piemonte
www.arpa.piemonte.it
ARPA
Regione del Veneto
www.arpa.veneto.it
Interreg III B Project- Alpine Space
Bayerische Akademie der Wissenschaften
Bayerische Kommission für die
Internationale Erdmessung
www.bek.badw-muenchen.de
ALPS
Deutsches Geodaetisches Forschungsinstitut
www.dgfi.badw.de
GPSQUAKENET
Regione Liguria
Direzione Centrale Affari Organizzativi
Servizio Sistemi Informatici
www.regione.liguria.it
Alpine Integrated
GPS Network:
Real-Time Monitoring
and Master Model for
Continental Deformation
and Earthquake Hazard
Regione Lombardia
Direzione Generale Territorio e Urbanistica
Infrastruttura per l’Informazione Territoriale
www.regione.lombardia.it
Servizio Geologico,
Provincia Autonoma di Bolzano – Alto Adige
www.provincia.bz.it
Servizio Geologico
Provincia Autonoma di Trento
www.provincia.tn.it
Université Joseph Fourier
Laboratoire de Géophysique
Interne et Tectonophysique
UMR 5559 du CNRS
www-lgit.obs.ujf-grenoble.fr
Fondazione Montagna Sicura-Valle d’Aosta
www.fondazionemontagnasicura.org
Ecole et Observatoire des Sciences de la Terre
eost.u-strasb.fr
ALPS GPSQUAKENET
Galileian Plus S.r.l.
www.galileianplus.it
Alps-GPSquakeNet has promoted
transanational co-operation in the field of
space geodesy applied to natural hazards.
It has set a transanational network of more
than 35 continuous Global Positioning
System (GPS) stations across the Alps. It has
investigated the continental deformation
and the earthquake hazard within the Alpine
space, mountains and surrounding foothills,
where are concentrated attractive European
metropolitan areas and rapidly growing
urban centres with extensive infrastructures.
It has developed pilot projects on the use
of GPS in meteorology, landslide studies and
active faulting monitoring. It has favored
transnational know-how exchange between
regional authorities and alpine universities
and research centres.
For the first time in the Alpine geology, AlpsGPSquakeNet through its continuous GPS
network "GAIN" (Geodetic Alpine Integrated
Network) will allow the quantification of
the crustal deformation of the whole
mountain range. This will open new research
initiatives in earth and environmental
sciences, therefore rising the value of the
Alps as a natural laboratory. The direct result
will be an improvement in the knowledge
of earthquake potential and hazard, and
this will allow a better land use in terms of
a safe living space. GAIN is giving the ground
for a higher resolution space-based coverage
of urban and mountain areas in the Alpine
space offering therefore a robust tool for
future infrastructure investment, land use
harmonization and industrial planning.
Alps-GPSquakeNet started catalyzing space
geodetic applications in the Alps
(meteorology, landslide monitoring,
agriculture, navigation, transportation,
mapping, surveying…). GPS nowadays is a
must in navigation. The possibility of realtime positioning will undoubtedly play an
increasing role in the EC during the next
decades in the field of automation, traffic
guidance and real-time hazard detection.
www.alps-gps.units.it
Located in the north of the italian peninsula near the Swiss
With this background, Regione Lombardia - Direzione Generale
border, lying at a major junction of the great East-West and
Territorio e Urbanistica, in cooperation with IREALP (Research
North-South communication corridors, Lombardy is one of
Institute for Ecology and Applied Economics in Alpine Areas),
the main gateways to Italy.
Politecnico of Milano and University of Milano participates in
The intense development of virtually all activities, from tradi-
the ALPS-GPSQUAKENET Interreg IIIB Alpine Space Project,
tional sectors as agricolture to industry and service, together
led by the University of Trieste -Department of Earth Sciences
with its high density of 382 inhabitants per sqkm, make
in collaboration with the International Centre for Theoretical
Lombardy the leading region of Italy.
Physics.
Nevertheless, the region’s morphology (53% hills and mountains) and the typically continental climate with its intense and
This project has realized a high-precision Global Positioning
prolonged rainy season, make natural disasters - as landslides
System (GPS) network in the Alps. This network, besides re-
and flooding - very common (more than 130.000 landslides
presenting a monitoring tool for continental deformation, will
surveyed in Lombardy). Recent events (earthquake of Salò,
support the development of space based techniques, since it
lake of Garda, November 2004) remembered us that regional
satisfies the performance required by all the GPS applications
territory is also active from a seismic point of view.
(crustal deformation for earthquake potential, meteorology,
Within this scenery, the regional authority has developed risk
landslide monitoring, agriculture, navigation, transportation,
prevention policies and action plans to mitigate the effects of
mapping, surveying, recreation & sport).
natural disasters, that are costly both in terms of damages and
human life. Hence the need to invest in advanced tools, inclu-
In this perspective, looking for investments optimization and
ding satellite technology, to improve monitoring and rational
service improvement , the 4 ALPS-GPSQuakenet receivers lo-
land management.
cated in Lombardia are directly integrated in the regional GPS
permanent stations network GPSLombardia, the first regional
Moreover, Regione Lombardia coordinates the Region’s parti-
network, realized in Italy.
cipation in the INTERREG IIIB and IIIC, the European Funding
Programmes to support a common approach to the sustaina-
This experience also represents an example of excellence in
ble development of the territory. INTERREG programmes are a
the cooperation between regional governments, research in-
basic aspect of the Structural Funds and conform the European
stitutions and universities, sharing experience and resources
Union’s principle of common economic and social policy.
to achieve a common goal: to improve the knowledge of our
Regione Lombardia participated to the INTERREG IIC “South
land.
Zone” and has been involved in three areas within INTERREG
IIIB programme: Alpine Space, Western Mediterranean Space
and Cadses Space (Adriatic Danube). Lombardia has been
strongly committed to taking advantage of all the opportunities offered by such cooperative programmes, with over 30
projects approved in different sectors.
Davide Boni
Regione Lombardia
“Territorio e Urbanistica” District Councillor
INDEX
PREFACE
1
THE PROJECT
1.1
1.2
1.3
1.4
1.5
2
Project abstract
The partnership
Background and objectives of the project
Main activities and expected results
Coherence with European policies and
Programme objectives
THE GAIN NETWORK
2.1 Network design
2.2 Monumentation
2.3 GAIN stations
2.4 GAIN Datacentre
2.5 GAIN and referente frames
2.6 The Geodetic Alpine Integrated Network (GAIN)
and its relation to EUREF
2.7 Data analysis
2.8 Outputs and Products
2.9 GAIN and regional services
8
10
11
11
12
12
13
14
15
19
20
41
44
45
45
48
49
3
CONTINENTAL DEFORMATION IN THE ALPS
3.1 First results from the GAIN network
3.2 Active deformation in the western Alps
3.3 Active Deformation in the South-Eastern Alps
3.4 Glacier shrinkage and modeled uplift of the Alps
3.5 Instrumental earthquakes in the Alpine region:
source parameters from moment tensor inversion
4
EARTHQUAKE HAZARD IN THE ALPS
4.1 Unified Scaling Law for Earthquakes in the Alps:
a multiscale application
4.2 Deterministic earthquake hazard assessment in the Alps
5
PILOT PROJECTS
52
53
55
57
64
67
72
73
95
5.1 GPS and Meteorology
5.2 GPS and Landslides
5.3 GPS and active faults
5.4 Active tectonics and Paleoseismology
104
105
109
113
115
PARTNERS LIST
120
BIBLIOGRAPHY
122
Interreg Project ALPS-GPS Quake Net
PREFACE
More than its Roman ruins and Renaissance cities,
more than its medieval castles and meticulously
manicured agricultural landscape, more than its
matchstick forests and endlessly varied coastlines, the Alps - Europe’s majestic mountain chain
that stretches some 1100 kilometres from southern France in the west to Slovenia in the east symbolizes Europe’s permanence and solidity. In
fact, when you think of an immovable object, the
Alps would certainly qualify. Or so it seems. But
the truth is that the Alps are in constant motion shifting about few millimeters each year,
as Africa continues to creep northward towards
Europe in an endless display of nature’s raw and
relentless energy.
In spring 2004, the European Union’s Interreg
III-B Alpine Space Programme funded a 3-year €
2.424.638 grant, the ALPS-GPSQUAKENET, to study and monitor the continental deformation and
the earthquake hazard over the Alps blending seismology and space geodesy through the Global
Positioning System (GPS) technology. The Alps
being a single geological entity required an observing space geodetic network whose geometry
should be built without any cross-border relevance and the characteristics of the single observing
station identical for the whole network. The existing GPS networks were either national or regional, heterogeneously distributed, with different
characteristics and precisions accordingly with
the required singular and specific application.
The ALPS-GPSQUAKENET project aimed at the build-up of a high-performance transnational space
geodetic network of Global Positioning System
(GPS) receivers in the Alpine Space. This GPS array
denominated “GAIN” (Geodetic Alpine Integrated
Network), within the millimeter-per-year precision, represents the first ever installed transna-
tional space geodetic network in the Alps. GAIN,
through the availability of higher resolution space geodetic data will contribute in advancing natural disasters prevention in the Alps. GAIN satisfies the performance required by all the GPS
applications and further increase the precision of
existing stations.
Another important and long-term investment of
the ALPS-GPSQUAKENET project stands in partnership build-up. The partnership brought to bear on
the project objectives is of public typology. It is
represented by research institutions with powerful internationally recognized education and outreach programmes, national and governmental agencies, regional public departments. This
transnational structure with both the geoscience and end-users communities provided an excellent means for the cross training and interaction
of governmental and regional employees involved
in GPS applications and public policy, and young
research scientists within the Alpine space.
The project has contributed with new research
initiatives in earth and environmental sciences,
therefore rising the value of the Alps as a natural
laboratory. The project has favoured transnational and national know-how exchange between regional authorities and alpine universities and research centres.
This project volume is one part of the final output
of the ALPS-GPSQUAKENET project. It reports on
some of the principal activities and pilot projects
carried out by the partnership of the project.
Major issues discussed in this volume are:
the transnational space geodetic network, GAIN
and all its components;
the Alpine continental deformation over different spatial and temporal scales;
Interreg Project ALPS-GPS Quake Net
the Alps-wide multiscale and deterministic earthquake hazard assessment;
the GPS Pilot Projects spanning active fault
and transient deformations monitoring, meteorology and landslides, active tectonics and
paleoseismology
Three years of successful collaboration within
a partnership made by participants from twenty different institutions represent an experience that is difficult, if not impossible, to fully describe in words. Though we would like to report
every single activity that has been performed,
every meeting that has taken place and every
presentation that has been given, in this final
publication of the ALPS-GPSQUAKENET project
we have chosen to focus our attention to the
most relevant results that we have achieved
and tried to give a flavour of the whole project
experience. The main output is represented by
the creation of the Geodetic Alpine Integrated
Network (GAIN), which has been fully realized
within the project. One of the basic requirements to obtain reliable measurements of crustal deformation in continental areas, especially slowdeforming ones, is that the observations span
a minimum of 3-4 years, in order to account for
seasonal effects contained in the data. Since
the whole duration of the project has been of
about three years and most GPS stations have
been installed in the last two years, it is too early to present here significant deformation patterns for the whole Alpine Space. Therefore, in
chapter 2 we have focused our attention to a
detailed description of the GAIN network, from
site monumentation to data management (collection, storage, reduction and analysis). In the
years to come, the collected data will be regularly processed, leading to a continuous impro-
vement in our knowledge of the geodynamics of
the Alps and their surroundings. Nonetheless,
a large amount of scientific research on the
Alpine Space has been carried on by the universities and research centers participating to the
project. Thanks to the funds made available by
the Alpine Space programme, it has been possible to make progress into a number of multi-disciplinary research studies, making use of the
available resources in terms of existing expertise and structures. Chapters 3 and 4, with their
collection of scientific papers about continental
deformation and earthquake hazard in the Alps,
show a sample of the variety of issues that can
be addressed once different groups share their
knowledge and exploit the newly available data
provided by the GAIN network. The last chapter concerns the outcome of the four main pilot
projects, which have all obtained successful results. Therefore, they open the way to a broader
spreading of the newly applied techniques, and
to a further exploitation of GAIN, to the fields
of meteorology, landslide and active faults monitoring and their paleoseismicity.
The success of the ALPS-GPSQUAKENET project
would not have been possible without the continuous support of the different programme bodies:
the Joint Technical Secretariat, the Managing
Authority and the different National and regional Contact Points.
I wish to thank all the partners and sub-contractors of the project for their contributions
and endeavours to have made of the ALPSGPSQUAKENET project a successful contribution.
I am particularly grateful to Marco Scuratti,
Michela Fioroni and Riccardo Riva. Without their
help, this document would not exist.
Karim Aoudia
Project Coordinator
Department of Earth Sciences, University of Trieste - Trieste, Italy
&
Earth System Physics Section, the Abdus Salam International Centre for Theoretical Physics - Trieste, Italy
Interreg Project ALPS-GPS Quake Net
1 - THE PROJECT
10
Interreg Project ALPS-GPS Quake Net
1.1 - Project abstract
The use of modern space based techniques gives
us new potential to monitor and prevent natural
risk, reduce economic losses, and save lives.
The main output of the ALPS-GPSQUAKENET is the
installation of a high-performance transnational space geodetic network of Global Positioning
System (GPS) receivers in the Alpine Space. This
GPS array, called Geodetic Alpine Integrated
Network (GAIN), is capable to measure deformations within the millimetre-per-year precision, and
represents the first ever installed transnational
space geodetic network involving Italy, Austria,
France, Germany, Switzerland and Slovenia. This
will support the use of space-based techniques
since it will satisfy the performance required by
all the GPS applications (crustal deformation for
earthquake potential, meteorology, landslide monitoring, agriculture, navigation, transportation,
mapping, surveying, recreation & sports…).
The transnational structure comprising both the
geoscientists and end-users will provide an excellent means for the cross training and interaction
of regional employees and young scientists.
1.2 - The partnership
1.2.1 - Presentation of the
partnership
The ALPS-GPSQUAKENET Partnership is of public
typology, including the sub-contractors.
It is represented by research institutions with
powerful internationally recognized education
and outreach programs, national and governmental agencies, and regional public departments.
This transnational structure with both the geoscience and end-users communities provides an
excellent means for the cross training and interaction of governmental and regional employees involved in GPS applications and public policy, and
young research scientists within the Alpine space.
The partnership is made of the most outstanding
European expertise in space geodesy. The subcontracting activities are handled in close cooperation with different universities and regional or national agencies. For LGIT: Laboratoire
de Geodynamique des Chaines Alpines. For RLB:
IREALP coordinating Politecnico of Milano and
University of Milan. For RLG: University of Genova.
For ARPA-P and FondMS: Politecnico of Torino.
Former contacts already exist between the
Universities and the regional departments through regional and national projects.
All the universities and governmental agencies
co-operate in the framework of national and international projects, under the umbrella of the
EUREF (http://www.epncb.oma.be/) that represents the European Reference System (ETRS89)
that coordinates the activities related to exi-
sting local permanent GPS networks in Europe
since 1995.
1.2.2 Composition of the
partnership
Lead partner:
DST-UNITS: Università degli Studi di Trieste
- Dipartimento di Scienze della Terra, Trieste,
Italy.
Project manager: A. Aoudia.
Sub-contractor: Galileian Plus S.r.l., Roma, Italy.
Contact: A. Amodio.
Project partners:
ARPA-P: Agenzia Regionale Per la Protezione
Ambientale del Piemonte, Torino, Italy.
Manager: C. Troisi.
Sub-contractor: POLITO (Politecnico di Torino),
Dipartimento di Ingegneria del Territorio, dell’Ambiente e delle Geotecnologie - II Facolta’ di
Ingegneria, Vercelli, Italy.
Contact: A. Manzino.
ARPA-V: Agenzia Regionale per la Prevenzione e
Protezione Ambientale del Veneto, Padova, Italy.
Manager: A. Luchetta.
BEK: Bayerische Akademie der Wissenschaften
/ Bayerische Kommission für die Internationale
Erdmessung, München, Germany.
Manager: C. Völksen.
DGFI: Deutsches Geodäetisches Forschungsinstitut,
München, Germany.
Manager: H. Drewes.
EARS: Environmental Agency of the Republic of
Slovenia, Ljubljana, Slovenia.
Manager: M. Zivcic.
EOST-IPGS: Ecole et Observatoire des Sciences
de la Terre - Institut de Physique du Globe de
Strasbourg CNRS/ULP UMR7516GSB, Strasbourg,
France. Manager: J. Van der Woerd.
FondMS: Fondazione Montagna Sicura - Montagne
Sûre, Courmayeur, Italy.
Manager: J.P. Fosson.
Sub-contractor: POLITO.
GSB: Geological Survey of the Autonomous
Province of Bolzano - South Tyrol, Kardaun,
Italy.
Manager: C. Carraro.
GST: Servizio Geologico, Provincia Autonoma di
Trento, Trento, Italy.
Manager: G. Zampedri.
LGIT: Université Joseph Fourier, Laboratoire de
Géophysique Interne et Tectonophysique, UMR
5559 du CNRS, Grenoble Cedex 9, France.
Manager: A. Walpersdorf.
Sub-contractor: LGCA (Laboratoire de Geodynamique
des Chaines Alpines), Chambery, France.
RLB: Regione Lombardia, Direzione Terriotrio Urbanistica, Sistema Informativo Territoriale, Milano,
Italy.
11
Interreg Project ALPS-GPS Quake Net
Manager: R. Laffi.
Sub-contractor: IREALP (Istituto di Ricerca
per l’Ecologia e l’Economia Applicate alle Aree
Alpine), Milano, Italy.
Contact: M. Fioroni.
Sub-contractor: POLIMI (Politecnico di Milano),
Dipartimento di Ingegneria Idraulica, Ambientale,
Infrastrutture Viarie, Rilevamento - sezione
Rilevamento, Milano, Italy.
Contact: R. Barzaghi.
Sub-contractor: UNIMI (Universita’ degli studi
di Milano), Dept. of Earth Sciences “A. Desio” Geophysics section, Milano, Italy.
Contact: R. Sabadini.
RLG: Regione Liguria - Direzione Centrale AffariI
Organizativi - Servizio Sistemi Informatici, Genova, Italy.
Manager: L. Pasetti.
Sub-contractor: UNIGE (Universita’ degli Studi di
Genova), Dipartimento di Macchine, Sistemi energetici e Trasporti, Facolta’ di Ingegneria, Genova,
Italy.
Contact: D. Sguerso.
1.3 - Background and
objectives of the project
Advances in natural disasters prevention, are driven by the availability of higher-resolution space geodetic data than were previously available.
This is achieved with a GPS network observing the
entire area of interest with a homogenous distribution and identical station characteristics.
The Alps represent a single geological entity,
therefore the geometry of the observing network
should be built without any cross-border relevance and the characteristics of the single observing
GPS station should be identical for the whole
network. The existing GPS networks are either national or regional, heterogeneously distributed,
with different characteristics and precisions.
The ALPS-GPSQUAKENET unprecedented precision,
millimeter-per-year objective, satisfies the performance required by all the GPS applications and further increase the precision of existing stations.
The primary goals of ALPS-GPSQUAKENET are: earthquake hazard reduction, landslides monitoring,
and meteorology.
Solutions as overall objectives:
Install the first transnational GPS network (~29
stations) for the entire Alps;
Test innovative continental deformation models
for earthquake risk reduction;
Provide an excellent means for cross training
& interaction of regional employees in GPS applications and public policy, and young research scientists;
Catalyze multidisciplinary applications (meteorology, landslide monitoring, agriculture, navigation, transportation, mapping, surveying, recreation - sports).
12
1.4 - Main activities and
expected results
The project partnership ensures innovative
methodologies and a direct transfer of knowledge
to the local and regional authorities through annual meetings and workshops.
Four main activities:
1 Set-up of the transnational GPS network (infrastructure investment) and quality check of its
performance.
The activities of the infrastructure investment
have consisted in: network design study; evaluation of the already operational GPS receivers
(compliant or not); procurement of the new GPS
receivers; site construction; logistic and set up of
the new receivers; acceptance test and operation
start; realization of the projet Datacentre; realization of the procedure and standard for data collection, transfer and archiving.
The validation and quality check control consists
in analysis on the collected data itself in particular to evaluate: data noise; receiver clock performance; multipath or interference effects.
Output: real-time broadcasting GPS network, covering the Alps, among all the PPs, real-time data
collection, real-time monitoring of the transnational natural disasters.
Result: knowledge transfer to the regional operators, change of behaviour among the operators
since involved in a transnational network, increase in know-how exchange among PP, real-time laboratory for students and post-docs.
Impact: improve public services with a policy of
transnational commitment, creation of new jobs
through other sub-nets, improvement of modern
technologies in real-time mode, support European
Space Agency missions, sustain environmental
management and planning.
2 Continental deformation and time-variable
earthquake hazard assessment of the Alps.
The time series of the ALPS-GPSQUAKENET are
converted in strain to evaluate the full deformation pattern in the Alps.
This part consists in:
Modelling the Earth Structure of the Alps from
Geophysical data;
Real-time monitoring: seismicity – GPS – InSAR;
Geodetic strain from GPS observations;
Dynamic modelling of the Alps: present day deformation and stress pattern.
Output: structure of the earth beneath the Alps,
master model for continental deformation, recognition of zones of high seismogenic potential
and earthquake deficit in the Alps.
Result: Scientific excellence (publications), training of PhDs and post-docs, development of integrated methodologies, grants and fellowships for
young research scientists.
Impact: Change in earthquake risk policy by informing the decision makers, reduction of earthquake disasters, increase of the worldwide
Interreg Project ALPS-GPS Quake Net
excellence and opportunities of the European
geo-science community.
3 Pilot projects in test sites.
GPS network is not limited to earthquake risk reduction. Four pilot projects have been realized
to establish procedures and methodologies which can be implemented in user friendly software
packages, or standard approaches ready to be used
by Regional and Local Services for day by day monitoring of: meteo, landslides, and active faults.
Output: know-how transfer, software packages
for regional authorities, regional databases.
Result: modern technologies and real-time actions for prevention, export experiences in the
Alpine space.
Impact: reduction of natural risks, develop and
efficient emergency response.
4 Databases, website, networking, information
and publicity.
The project website represents as a major instrument of information, database archiving-handling and results distribution among the project
partners.
Output: project Web site, databases ready to use,
real-time data collection, project dissemination,
information and publicity, download open access.
Result: networking regional authorities, catalyze
other sub-networks, attract beginning students,
and promote European Space Geodesy at the worldwide level.
Impact: promote Alpine Space co-operation at
the worldwide level.
1.5 - Coherence with
European policies and
Programme objectives
ALPS-GPSQUAKENET promotes transnational cooperation in the field of Space geodesy applied
to natural hazards. It delineates the seismogenic
potential within the Alpine space, mountains and
surrounding foothills, where are concentrated the
most attractive European metropolitan areas and
rapidly growing urban centres with extensive infrastructures. It favours transnational know-how
exchange between regional authorities and alpine universities and research centres. It reinforces the European Space geodetic and geoscience
communities and support European Space Agency
missions. For the first time in the Alpine geology,
ALPS-GPSQUAKNET provides values for the crustal
shortening of the whole mountain range. This
opens new research initiatives in earth and environmental sciences, therefore rising the value
of the Alps as a natural laboratory. These results
are changing the state of the knowledge in terms
of earthquake hazard and constrain earthquake
hazard scenarios therefore better harmonise the
land use in terms of a safe living space.
ALPS-GPSQUAKENET gives for the first time the
ground for a higher resolution space based coverage of urban and mountains areas in the Alpine
space, better resolves satellite imagery, and the-
refore offers a robust tool for future infrastructure investment, land use harmonisation and industrial planning. Highly resolving remote sensing
methods (InSAR), give us new potential to monitor and prevent environmental degradation
and limit the impacts of natural disasters. ALPSGPSQUAKENET is a catalyser of space geodetic applications in the Alps (meteorology, landslides
monitoring, agriculture, navigation, transportation, mapping, surveying, recreation & sports…).
GPS nowadays is a must in navigation. The ALPSGPSQUAKENET contributes to better resolve the
simple and handy GPS used for classical routing,
thus attracting recreation & sports purposes even
in isolated areas. This possibility of regional realtime positioning will undoubtedly play an increasing role in the EC during the next decade (e.g.
GALILEO) and will dictate the forth-coming century in the field of automation, traffic guidance
and real-time hazard detection.
Contribution to the improvement of insti-
tutional setting and to the decision making
process
This project has the goal of making regional authorities talking and interacting at the national
and transnational level, therefore it gives more
weight and credibility to the institutions. The interaction of the regional authorities with universities and research centres has opened job opportunities to newly graduated students during the
project span. The GPS reference network supports
the creation of local and regional sub-networks
with direct access to ALPS-GPSQUAKENET software and databases.
This project supports transnational natural risk
prevention actions avoiding singular adapted actions and assists the decision-making authorities
in law and legislation proposal and implementation. ALPS-GPSQUAKENET contributes to delineating areas of high earthquake potential, providing
maps of maximum credible earthquake occurrence
that may change the earthquake hazard zoning at
regional, national and transnational scale.
Contribution to multisectoral integration and
co-operation
The ALPS-GPSQUAKENET Partnership with universities, research institutions and governmental
agencies with powerful, nationally and internationally recognized education and outreach programs, regional public sections and departments
directed by regional authorities, and the different hazard lines and aspects tackled, highlights
the cross-sectoral approach and favour the vertical and horizontal co-operation. The benefits
added by our cross-sectoral approach and multisectoral integration reside in the proposal of
concrete measures to reduce the natural hazards
informing straightforwardly the institution since directly involved, and driving towards transnational preventive actions and dismissing adapted
and regional actions. This contributes to reinforcing the cross-interaction and the emergence of
new ideas both in natural hazard reduction and
also in the emergency response.
13
Interreg Project ALPS-GPS Quake Net
2 - THE GAIN NETWORK
14
Interreg Project ALPS-GPS Quake Net
2.1 Network design
DST-UNITS, POLITO-ARPAP, POLIMI-RLB
The first step to realize the GAIN network consisted in gathering the partnership to discuss the
location of the GPS stations. At this level, considerations were only made on the basis of geological and tectonic aspects, and taking into account
the administrative boundaries of each project
partner.
In Figure 2.1.1, we show the map used during
the first partnership meeting in Trieste, where
black lines indicate national boundaries, red lines Italian regional boundaries and the purple
line the Transalp seismic section. Red starts represent existing CGPS stations, while yellow stars
are the preliminary locations of the GAIN sites;
note that, since EOST-IPGS and FondMS had not
yet joined the partnership, no station was foreseen in regions Alsace and Valle d’Aosta.
A major effort was put into locating the new stations within the plate boundary and at the same
time obtaining a station distribution as homogeneous as possible, for both tectonic and geodetic purposes.
The proposed sites location were afterwards explored by each project partner, in order to
find the suitable location for station installation, where a number of other technical and logistic factors had to be considered (bedrock type,
sky view, background noise, safety, accessibility,
power availability, data transmission).
sites location, besides all the previously mentioned elements, is represented by the analysis of
preliminary GPS observations. In this section, we
explain how this quality control has been performed at a number of sites.
Twenty-four hours of preliminary GPS observations have been acquired in the candidate sites,
with sampling rate of 10 sec. The cutoff angle,
or the elevation under which all the observation
are neglected, has been fixed to 0°, in order to
have e complete description of obstructions and
of signal quality at low elevations. However, because of the observation under an elevation an-
The final network design is displayed in Figure
2.1.2.
2.1.1 Quality control of
candidate sites: methodology
An important aspect when looking for candidate
Fig. 2.1.2 - the GAIN network.
Fig. 2.1.1 - Geological map of
the Alps and preliminary location
of the GAIN sites (yellow stars).
Red stars indicate existing CGPS
stations.
15
Interreg Project ALPS-GPS Quake Net
gle of 10° are neglected by the processing, we
distinguished the quality control parameters for
two observation subsets, over and under the cut
off line.
2.1.2 Error sources
In quality control we find observations characterized by a dysfunctional behaviour, caused mainly
by receiver clock jump, cycle slip, quasi-random
error or outlier. We will briefly examine the origin
of these errors, focusing on the site dependent
ones, cycle slips and quasi-random errors.
Clock jumps
The majority of the receivers maintain their internal clock synchronized to GPS time, adjusting
periodically the clock by inserting a clock jump.
These errors are completely dependent on the receiver model, in fact the synchronization procedure is proprietary, and it depends on the firmware. However the effects of clock jump on code and
carrier phase observations are well known and
can be simply recovered or removed by means of
appropriate algorithms, before looking for cycle
slips and multi-path.
Cycle slips
Tab. 2.1.1
16
The GPS receiver, after the start of the acquisition, observes the difference between the received signal and its internal duplicate, measures
the fractionary part of carrier phase and initializes an integer counter. During the observation
session the counter will be incremented of one
cycle every time that the fractionary part change from 2 to 0. So the carrier phase observation
is the sum of phase fraction j plus a counter n.
The initial number N of integer cycles between
the satellite and the receiver is unknown. This
carrier phase ambiguity N remain a constant value until a loss of signal happens. In this case
the counter n is reinitialized, causing in carrier
phase observation a jump of an integer number
of cycles.
The cycle slips have many causes. The most common is the loss of signal due to obstructions,
such as trees, buildings or other obstacles. Cycle
slips can be due also to a low signal to noise ratio, caused by bad ionospheric conditions, multipath, low satellite elevation or receiver dynamic.
The last cause can be the firmware fails, quite uncommon in modern receivers. We must underline
that in modern receivers, thanks to the good algorithms implemented in the firmware, the cycle
slip rejection and recovering is quite good. In fact
usually we found cycle slips of the first type only,
cause by obstructions; this type of cycle slip is
characterized by a zeroing of n counter, producing a large and easily detectable carrier phase
jump. The cycle slips can occur on one or both
frequencies, in particular on that one with lower
signal to noise ratio. However usually, if cycle slips are caused by loss of signal, they occur on both
frequencies.
Quasi-random errors
Multi-path, diffraction, ionospheric scintillation,
etc. are the main sources of quasi random errors,
usually neglected by the functional and stochastic models used in data processing. The LMS
adjustment leads to reduce and to distribute their
effect over the entire set of observations, so it is
preferable to treat the quasi random errors separately by cycle slips and clock jumps.
The multi-path, or the multiple reflection of the
signal, happens when the received signal is reflected by some obstacles. The terrain, buildings, or the objects can be reflecting surfaces in
the 1.6 GHz band. The multi-path is an error or disturbance that depends on the observation site,
so it must be monitored in the choice of candidate sites. Two important characteristic of multi-path are:
The multi-path signal reach the antenna always
after the direct signal, because of the longer propagation path.
The power is usually lower than the direct signal, because of the loss of power due to absorption and diffraction.
The multi-path signal distorts the correlation
function of the signal, producing measurement
errors. The effect of diffraction causes a droop
(drop) of the signal to noise ratio in the directions near to the obstacles. It can be mistaken
Receiver tracking capability
Maximum ionospheric rate (L1)
Report data gap greater than
Expected rms level of P1 multi-path
Expected rms level of P2 multi-path
Multi-path slip sigma threshold
% increase in MP rms for C/A | A/S
Points in MP moving averages
Minimum signal to noise for L1
Minimum signal to noise for L2
Elevation mask (cut-off)
Elevation comparison threshold
12 SVs
400 cm/min
10 min
50 cm
65 cm
4 cm
100 %
50
0
0
10°
25°
Orbit path spline fit sample time
SVs w/ code data for position try
Width of ASCII summary plot
Data indicators on summary plot
Do ionospheric observable
Do ionospheric derivative
Do high-pass ionosphere observable
Do multi-path observables
Do 1-ms receiver clock slips
Tolerance for 1-ms clock slips
Do receiver LLI slips
Do plot file(s)
10 min
5
72
yes
yes
yes
no
yes
yes
1.e-02 ms
yes
yes
Interreg Project ALPS-GPS Quake Net
Cut off angle % obs [expected/have] Obs / cycle slips IOD or MP slips Average MP on L1 [m] Average MP on L2 [m]
0°
69 %
35853
0
0.53
0.45
10°
88 %
35113
0
0.49
0.41
Tab. 2.1.2 - Some results from
Teqc summary file.
Fig. 2.1.3 - Pseudorange multipath at Sondrio for L1 and L2
frequencies respectively.
with other effects due to the shape of antenna gain at low elevation and to the atmosphere. The effect of quasi random errors spans over
some epochs of observation with not forecasting behaviour. They make more difficult to
fix the correct value for carrier phase integer
ambiguities.
on multi-path variance; it is used to search for
cycle slips.
2.1.3 TEQC software
Elevation comparison threshold: threshold value defined to distinguish low elevation observations. Some quality control parameters are computed separately for low elevation observations.
Orbit path spline fit sample time: the satellite
position is computed using the keplerian parameters reported in the ephemeris files; the computation is quite slow, so the satellite positions
are not computed at every observation epoch but
only at fixed intervals, then they are interpolated
by spline curves.
TEQC software (pronounced “tek”) is a simple yet
powerful and unified approach to solving many
pre-processing problems with GPS, GLONASS, and
SBAS (Satellite Based Augmentation System)
data. It includes data translation, data editing and quality control functions (Translation,
Editing, Quality Check). It is available at
http://www.unavco.org/facility/software/teqc/
teqc.html
A short TEQC tutorial for quality control procedures is available at
http://www.unavco.org/facility/software/teqc/
tutorial.html#sec_11
http://www.unavco.org/facility/software/teqc/
tutorial.html#sec_21
The quality control has been performed with the
following processing parameters for all the candidate sites:
We will explain some of these parameters.
Maximum ionospheric rate: threshold value on
ionospheric rate; it is used to search for cycle
slips.
Multipath slip sigma threshold: threshold value
Points in MP moving averages: the multi-path
is estimated as difference from a moving average value. It is necessary to set the dimension of
the time window used to compute the moving
average.
Do receiver LLI slips: it finds cycle slip previously marked by receiver with Loss of Lock Indicator,
LLI.
Do plot file: output plot files of SNR (*.sn1,
*.sn2) and multi-path (*.mp1, *.mp2) of both
frequencies, of ionospheric delay and its variation (*.ion, *.iod), of azimuth and elevation
(*.azi, *.ele). These files can be used to produce output sky-plots using the QC2SKY software or others.
2.1.4 Quality control of
candidate sites: examples
In order to clarify the above described strategy,
17
Interreg Project ALPS-GPS Quake Net
Fig. 2.1.4 - Signal to Noise
Ratio at Sondrio for L1 and L2
frequencies respectively.
Fig. 2.1.5 - Loss of Lock (yellow
line) and Visibility Obstruction at
Sondrio for L1 and L2 frequencies respectively.
we will show the results for two sites located in
Regione Lombardia, namely SOND and PORA.
SOND - Sondrio
Teqc summary file provides some preliminary results which can be divided into two categories: cut off angle fixed to 0 degree and to 10
degrees.
Looking at Figure 2.1.3 and Figure 2.1.5, it is clear
that Sondrio site is affected by a significant obstruction with an azimuth between 250 and 330
degrees, till an elevation of about 25 degrees:
this obstruction matches with the peak of the
surrounding mountains. The rural buildings, which accommodates the receiver, cover the sky with
an azimuth between 30 and 65-70 degrees, till an
elevation of about 20 degrees.
At lower elevation angle the amount of pseudorange multipath is greater than at upper eleva-
Fig. 2.1.6 - Signal to Noise
Ratio at Mt. Pora for L1 and L2
frequencies respectively.
18
Interreg Project ALPS-GPS Quake Net
Cut off angle % obs [expected/have] Obs / cycle slips IOD or MP slips Average MP on L1 [m] Average MP on L2 [m]
0°
90 %
13095
4
0.31
0.36
10°
100 %
11241
4
0.21
0.21
Tab. 2.1.3 - Some results from
Teqc summary file.
Fig. 2.1.7 - Signal to Noise
Ratio at Mt. Pora for L1 and L2
frequencies respectively.
Fig. 2.1.8 - Loss of Lock and Visibility Obstruction at Mt. Pora.
Fig. 2.2.1 - The antenna mount
device.
tion angle: it also can indicate how susceptible a
GPS antenna is to ground bounce.
As it can be seen there are no observations at
elevation less then 10 degrees, but once the signal is locked the antenna do not loose the satellite signal.
Looking at Figure 2.1.4 it is possible to see how
SNR on L2 is higher than on L1 frequency, where
the ratio is good.
PORA - Mount PORA, Bergamo
Teqc summary file provides some preliminary results which can be divided into two categories: cut off angle fixed to 0 degree and to 10
19
Interreg Project ALPS-GPS Quake Net
degrees.
Looking at Figure 2.1.6 and Figure 2.1.8, it results clear that Mount Pora site is not affected by
any significant obstructions as we can seen from
Tab.2.1.3.
At lower elevation angle the amount of pseudorange multipath is greater than at upper elevation angle, which is a quite standard effect.
Looking at Figure 2.1.7 it is possible to see how
SNR on L2 is higher than on L1 frequency, where
the ratio is good.
2.2 Monumentation
DST-UNITS, POLITO-ARPAP, POLIMI-RLB
The international standards for permanent GPS
stations involved in geodynamical studies, and
the fact that tectonic motions in the Alps are
only a few millimetres per year, require achieving
the highest possible stability of the monument.
The best choice to increase monument stability is
to tie a concrete pillar to the bedrock: this minimizes the risk for the monument to be affected
by motions not strictly connected with crustal
deformation. The very common practice of installing stations on building roofs is in this case not
advised, because the observations could be affected by oscillations or seasonal building motions related to thermal expansion.
The following guidelines to monument the permanent GPS stations have been followed whenever possible (other designs have been devised for
particular situations):
1 The pillar of the GPS stations must be well
anchored into solid bedrock. In order to define whether exposed rocks have optimal features it is mandatory to perform inspections with
a geologist.
2 The anchorage between pillar and bedrock
must be done by means of iron bars of suitable suitability section, which must be inserted
into the ground for 2-3 meters in depth. The
iron bars, which are usually fixed to the bedrock
by means of special glues, have to emerge from
the ground to allow the coupling of the pillar’s
framework.
3 The square base of the pillar must have a length of about 100 centimetres, while its height
can be 20-30 centimetres.
4
The pillar’s framework should be cylindrical and centered with respect to the base,
with a diameter of about 40 centimetres. The
cylindrical pillar should be about 150 centimetres high. In any case, the height of the pillar
should be greater than the average registered
snow height.
20
5 The concrete used to build the pillar must
have optimal properties and be able to withstand
temperature variability and weather change.
6
The GPS antenna and the concrete pillar
must be linked by means of an iron mount device (Fig. 2.2.1) inserted in the pillar by bi-component glue. This device must be levelled upon
installation.
2.2.1 Local control network
setup
Site effects and monument stability at permanent
GPS stations are controlled by monitoring a local
control network, formed by three or more GPS and
levelling points. These control points have usually been materialized on sub-horizontal rock surfaces by means of steel geodetic markers. Where
it has not been possible to find suitable surfacing
bedrock, the control points have been set up on
the foundations of massive and stable buildings.
In some cases only sub-vertical surfaces are available to set up the control points, or the only possible locations for the control points are not suitable to acquire GPS observations: in these cases
the control points can be used only for levelling.
The distance of the control points from the GPS
station depends on the wavelength of the local
phenomena to be monitored. Moreover this distance must be a compromise between different
factors as: bedrock availability, GPS satellite visibility, time necessary to reach the control point
or to perform the levelling, chance to preserve
the control point.
Local control network set up and monitoring is of
primary importance. However the monitoring is
quite expensive in terms of time and human resources in the field. The GAIN network stations
will be generally monitored by levelling and GPS
survey campaigns once per year.
An additional very important role of the control
network is represented by the possibility to determine on-site the previous location of the GPS antenna, in case the pillar is damaged or destroyed
by natural causes, accidents or acts of vandalisms. If it would ever be necessary to substitute
or relocate the monument, in fact, control points
are the only way to link the new antenna location
to the previous position and avoid wasting previous measurements. For referencing reasons, it
is very important that the control points are also
observed by GPS.
2.3 GAIN stations
2.3.1 Station monographies
Interreg Project ALPS-GPS Quake Net
1
2
4 ID name: AGNE
Site Name: Lago Agnel
4 ID name: ALPE
Site Name: Alpe d’Huez
Receiver Type: LEICA GRX1200PRO
Receiver Type: ASHTECH Z-XII3
Antenna Type: LEIAT504
Antenna Type: ASH701945C_M SCIS
Latitude: 45° 28’
Latitude: 45° 5’
Longitude: 7° 8’
Longitude: 6° 5’
Ell. height: 2354.591 meters
Ell. height: 1892.495 meters
Bedrock type: gneiss
Bedrock type: limestone bedrock of the Jura fold belt
Installed by: ARPA-P
Installed by: LGIT
21
Interreg Project ALPS-GPS Quake Net
3
4
4 ID name: AUBU
Site Name: Aubure
4 ID name: BASO
Site Name: Basoviza
Receiver Type: TRIMBLE NetRS
Receiver Type: LEICA GRX1200Pro
Antenna Type: Dorne Margolin choke ring antenna
Antenna Type: LEIAT504
Latitude: 48° 12’
Latitude: 45° 38’
Longitude: 7° 12’
Longitude: 13° 52’
Ell. height: 970.139 meters
Ell. height: 448.501 meters
Bedrock type: granitic bedrock of the Palaeozoic of the
Vosges massif
Bedrock type: limestone bedrock
Installed by: UNITS
Installed by: EOST
22
Interreg Project ALPS-GPS Quake Net
5
6
4 ID name: BOSC
Site Name: Bosco Chiesa Nuova
4 ID name: BREI
Site Name: Breitenberg
Receiver Type: LEICA GRX1200pro
Receiver Type: LEICA GRX1200PRO
Antenna Type: LEIAT504
Antenna Type: LEIAT504
Latitude: 45° 35’
Latitude: 47° 32’
Longitude: 11° 2’
Longitude: 10° 33’
Ell. height: 910.177 meters
Ell. height: 1887.912 meters
Bedrock type: sedimentary
Bedrock type: dolomite
Installed by: ARPAV-B
Installed by: DGFI
23
Interreg Project ALPS-GPS Quake Net
7
8
4 ID name: BURE
Site Name: Bure (Haute-Marne)
4 ID name: CARZ
Site Name: Monte Carza
Receiver Type: TRIMBLE NetRS
Receiver Type: LEICA GRX1200PRO
Antenna Type: Dorne Margolin choke ring antenna
Antenna Type: LEIAT504
Latitude: 48° 29’
Latitude: 46° 2’
Longitude: 5° 21’
Longitude: 8° 40’
Ell. height: 365.274 meters
Ell. height: 1164.497 meters
Bedrock type: mesozoic sedimentary limestone and marls
Bedrock type: gneiss-amphibolites
Installed by: EOST
Installed by: ARPA-P
24
Interreg Project ALPS-GPS Quake Net
9
10
4 ID name: CLTN
Site Name: Monte Coltignone
4 ID name: DEVE
Site Name: Alpe Devero
Receiver Type: TPS ODYSSEY_E
Receiver Type: LEICA GRX1200PRO
Antenna Type: TPSCR3_GGD CONE
Antenna Type: LEIAT504
Latitude: 45° 53’
Latitude: 46° 18’
Longitude: 9° 23’
Longitude: 8° 15’
Ell. height: 1440.037 meters
Ell. height: 1679.418 meters
Bedrock type: igneous
Bedrock type: calcareous schist
Installed by: IREALP - RLB
Installed by: ARPA-P
25
Interreg Project ALPS-GPS Quake Net
11
12
4 ID name: ELMO
Site Name: Monte Elmo
4 ID name: EOST
Site Name: Strasbourg
Receiver Type: LEICA GRX1200PRO
Receiver Type: TRIMBLE NetRS
Antenna Type: LEIAT504
Antenna Type: ‘Trimble Zephyr Geodetic
Latitude: 46° 42’
Latitude: 48° 34’
Longitude: 12° 23’
Longitude: 7° 45’
Ell. height: 2397.875 meters
Ell. height: 137.327 meters
Bedrock type: quartzite
Bedrock type: (building)
Installed by: GSB
Installed by: EOST
26
Interreg Project ALPS-GPS Quake Net
13
14
4 ID name: FAHR
Site Name: Fahrenberg
4 ID name: FDOS
Site Name: Fort Dossaccio
Receiver Type: LEICA GRX1200PRO
Receiver Type: LEICA GRX1200PRO
Antenna Type: LEIAT504
Antenna Type: LEIAT504
Latitude: 47° 36’
Latitude: 46° 18’
Longitude: 11° 18’
Longitude: 11° 43’
Ell. height: 1674.298 meters
Ell. height: 1888.929 meters
Bedrock type: dolomite
Bedrock type: outcropping bedrock formed by rhyolitic
Installed by: DGFI
ignimbrites of the atesina volcanic platform
Installed by: Geological Survey of Provincia Autonoma di
Trento
27
Interreg Project ALPS-GPS Quake Net
15
16
4 ID name: FERR
Site Name: Ferret valley
4 ID name: GORS
Site Name: Gorjuse
Receiver Type: LEICA GRX1200PRO
Receiver Type: LEICA GRX1200Pro
Antenna Type: LEIAT504
Antenna Type: LEIAT504
Latitude: 45° 52’
Latitude: 46° 19’ 2.604”
Longitude: 7° 1’
Longitude: 13° 59’ 59.532”
Ell. height: 2404.873 meters
Ell. height: 1048 meters
Bedrock type: granitic rocks
Bedrock type: bed rock
Installed by: FondMS
Installed by: UNITS - EARS
28
Interreg Project ALPS-GPS Quake Net
17
18
4 ID name: HGRA
Site Name: Hochgrat
4 ID name: HRIE
Site Name: Hochries
Receiver Type: LEICA SR520
Receiver Type: LEICA SR520
Antenna Type: LEIAT504
Antenna Type: LEIAT504
Latitude: 47° 29’
Latitude: 47° 44’
Longitude: 10° 4’
Longitude: 12° 14’
Ell. height: 1764.160 meters
Ell. height: 1615.181 meters
Bedrock type: conglomerate
Bedrock type: limestone
Installed by: DGFI
Installed by: DGFI
29
Interreg Project ALPS-GPS Quake Net
19
20
4 ID name: JANU
Site Name: Fort du Janus
4 ID name: JAVS
Site Name: Javornik
Receiver Type: ASHTECH µZ-CGRS
Receiver Type: LEICA GRX1200Pro
Antenna Type: ASH700936A_M SCIS
Antenna Type: LEIAT504
Latitude: 44° 54’
Latitude: 45° 53’ 36.24”
Longitude: 6° 42’
Longitude: 14° 3’ 51.48”
Ell. height: 2583.061 meters
Ell. height: 1100 meters
Bedrock type: limestone
Bedrock type: bed rock
Installed by: LGIT
Installed by: UNITS - EARS
30
Interreg Project ALPS-GPS Quake Net
21
22
4 ID name: LEBE
Site Name: Col de Lebe
4 ID name: LFAZ
Site Name: Le Faz
Receiver Type: ASHTECH Z-XII3
Receiver Type: ASHTECH micro-Z CGRS
Antenna Type: ASH710945.02B SCIS
Antenna Type: ASH701945C_M SCIS
Latitude: 45° 54’
Latitude: 45° 6’
Longitude: 5° 37’
Longitude: 5° 23’
Ell. height: 940.487 meters
Ell. height: 1071.398 meters
Bedrock type: fresh bedrock (limestone)
Bedrock type: metamorphic
Installed by: LGIT
Installed by: LGIT
31
Interreg Project ALPS-GPS Quake Net
23
24
4 ID name: LUCE
Site Name: Lucelle
4 ID name: MARK
Site Name: Le Markstein, Oderen
Receiver Type: “TRIMBLE NetRS”
Receiver Type: “TRIMBLE NetRS”
Antenna Type: Dorne Margolin choke ring antenna
Antenna Type: Dorne Margolin choke ring antenna
Latitude: 47° 25’
Latitude: 47° 55’
Longitude: 7° 14’
Longitude: 7° 1’
Ell. height: 620.100 meters
Ell. height: 1180.280 meters
Bedrock type: limestone bedrock of the Jura fold belt
Bedrock type: granitic bedrock of the Palaeozoic of the
Installed by: EOST
Vosges massif
Installed by: EOST
32
Interreg Project ALPS-GPS Quake Net
25
26
4 ID name: MATA
Site Name: Mount Matahur
4 ID name: MAVE
Site Name: Monte Avena
Receiver Type: LEICA GRX1200PRO
Receiver Type: LEICA GRX1200pro
Antenna Type: LEIAT504
Antenna Type: LEIAT504
Latitude: 46° 12’
Latitude: 46° 1’
Longitude: 13° 31’
Longitude: 11° 49’
Ell. height: 1629.062 meters
Ell. height: 1465.510 meters
Bedrock type: flysch bed rock
Bedrock type: sedimentary
Installed by: UNITS
Installed by: ARPAV-B
33
Interreg Project ALPS-GPS Quake Net
27
28
4 ID name: MBEL
Site Name: Montebelluna
4 ID name: MITT
Site Name: Kleine Mittagerspitze – Loc.
Merano 2000
Receiver Type: LEICA GRX1200pro
Antenna Type: LEIAT504
Receiver Type: LEICA GRX1200PRO
Latitude: 45° 46’
Antenna Type: LEIAT504
Longitude: 12° 2’
Latitude: 46° 41’
Ell. height: 214.453 meters
Longitude: 11° 17’
Bedrock type: sedimentary
Ell. height: 2305.675 meters
Installed by: ARPAV-B
Bedrock type: quartzite
Installed by: GSB
34
Interreg Project ALPS-GPS Quake Net
29
30
4 ID name: MOCA
Site Name: Mount Calisio
4 ID name: NIDE
Site Name: Niedersteinbach
Receiver Type: LEICA GRX1200PRO
Receiver Type: TRIMBLE NetRS
Antenna Type: LEIAT504
Antenna Type: Dorne Margolin choke ring antenna
Latitude: 46° 5’
Latitude: 49° 1’
Longitude: 11° 8’
Longitude: 7° 44’
Ell. height: 1146.372 meters
Ell. height: 445.845 meters
Bedrock type: outcropping bedrock formed by dolomite
Bedrock type: pink sandstone of the Triassic sedimentary
Installed by: GST
cover of the Vosges
Installed by: EOST
35
Interreg Project ALPS-GPS Quake Net
31
32
4 ID name: OATO
Site Name: Osservatorio Astronomico di
Torino
4 ID name: PALA
Site Name: Palazzolo
Receiver Type: LEICA GRX1200PRO
Antenna Type: LEIAT504
Antenna Type: LEIAT504
Latitude: 45° 47’
Latitude: 45° 2’
Longitude: 13° 2’
Longitude: 7° 45’
Ell. height: 4.942 meters
Ell. height: 658.815 meters
Bedrock type: alluvial sediments, test monumentation in
Bedrock type: arenite
unconsolidated material
Installed by: ARPA-P
Installed by: UNITS
Receiver Type: LEICA GRX1200PRO
36
Interreg Project ALPS-GPS Quake Net
33
34
4 ID name: PARO
Site Name: Paroldo
4 ID name: POGG
Site Name: Poggio Grande
Receiver Type: Ashtech Z – FX CORS
Receiver Type: LEICA GRX1200PRO
Antenna Type: Ashtech Dorne Margolin
Antenna Type: LEIAT504
Latitude: 44° 26’
Latitude: 44° 6’
Longitude: 8° 4’
Longitude: 8° 9’
Ell. height: 848.778 meters
Ell. height: 855.578 meters
Bedrock type: tertiary flish
Bedrock type: limestone
Installed by: ARPA-P
Installed by: RLG
37
Interreg Project ALPS-GPS Quake Net
35
36
4 ID name: PORA
Site Name: Monte Pora
4 ID name: PUYA
Site Name: Puy Aillaud
Receiver Type: TPS ODYSSEY_E
Receiver Type: ASHTECH Z12-CGRS
Antenna Type: TPSCR3_GGD CONE
Antenna Type: ASH700936A_M SCIS
Latitude: 45° 53’
Latitude: 44° 51’
Longitude: 10° 6’
Longitude: 6° 28’
Ell. height: 1926.518 meters
Ell. height: 1689.702 meters
Bedrock type:
Bedrock type: limestone
Installed by: IREALP - RLB
Installed by: LGIT
38
Interreg Project ALPS-GPS Quake Net
37
38
4 ID name: ROSD
Site Name: Roselend
4 ID name: SERL
Site Name: Serle
Receiver Type: ASHTECH Z12-CGRS
Receiver Type: TPS ODYSSEY_E
Antenna Type: LEIAT504 SCIS
Antenna Type: TPSCR3_GGD CONE
Latitude: 45° 41’
Latitude: 45° 35’
Longitude: 6° 37’
Longitude: 10° 20’
Ell. height: 1693.313 meters
Ell. height: 942.919 meters
Bedrock type: gneiss
Bedrock type: igneous
Installed by: LGIT
Installed by: IREALP - RLB
39
Interreg Project ALPS-GPS Quake Net
39
40
4 ID name: SOND
Site Name: Sondrio
4 ID name: WART
Site Name: Wartsteinkopf
Receiver Type: TPS ODYSSEY_E
Receiver Type: LEICA GRX1200PRO
Antenna Type: TPSCR3_GGD CONE
Antenna Type: LEIAT504
Latitude: 46° 10’
Latitude: 47° 39’
Longitude: 9° 51’
Longitude: 12° 48’
Ell. height: 529.015 meters
Ell. height: 1749.424 meters
Bedrock type: igneous
Bedrock type: limestone
Installed by: IREALP - RLB
Installed by: Deutsches Geodaetisches Forschungsinstitut
40
Interreg Project ALPS-GPS Quake Net
2.3.2 Recommendations
The realization of a network of 40 CGPS stations
capable of providing measurements with millimeter precision requires maximum care about every
single step involved in the process.
For this reason, we are listing here a number of
minimum technical requirements and practical
measures that we consider important and that
have allowed us to successfully realize the GAIN.
The first set of recommendations regards minimum requirements for the GPS kit, as they were
agreed upon during the project scientific meeting
held in Munch in February 2004.
Receiver:
geodetic receivers recognized by the IGS that
could handle a meteo data logger;
dual frequency ;
memory size up to needs;
embedded with a computer according to site
characteristics;
external frequency input for real time purpose (according to the needs of the of the project
partner);
possibility of remote control;
lightening protector;
UPS.
Antenna:
geodetic antenna with a ground plane;
phase center variations must be officially
certified;
spherical radome.
The second set of recommendations regards general criteria for site selection: it has been compiled by ARPA-P and is representative of the way
most project partners have handled their sites.
Technical and scientific requirements:
The site must allow the foundation of the monument in the bedrock.
The site must be optically unobstructed, over
the cut-off angle of 15°, 360° around.
In order to be significant at the alpine scale,
from the geologic point of view, the site has to
be:
- relevant, at a regional scale, from the general
geologic, tectonic and structural point of view;
- out of landslide and out of deep-seated-deformations. The point is critical; while it is quite
easy to locate and define landslides, deep-seated-deformations are often overlooked and absent on landslide maps and inventories. Since,
in the alpine area, deep-seated-deformations may
cover entire mountain flanks for extremely large
areas the geological investigation must include a
thorough analysis aimed to be sure that the selected site is located in a totally stable area.
The site has to be far from any source of electro-
magnetic disturbs (power lines, repeaters etc.).
Practical requirements:
The site must preferably be on a public property and it is advisable to provide strong and good
contacts with the local communal authorities.
The site must be easily reachable for maintenance and should also be somehow sheltered, in
order to prevent theft and vandalism. These two
requirements may be mutually exclusive.
Connection to the mains is advisable but not
indispensable. Power may easily be supplied by
means of a unit consisting of solar panels and
buffer batteries (allowing at least a one week autonomy), which is also less lightning-sensitive
than mains connection. In case of connection to
the mains, the ground-fault-device (compulsory
in most countries) has to be of the self-retriggerable type, in order to prevent power blackouts
due to lightnings.
A proper, insulated, shelter must be provided
for power supply, modems and receiver electronics. A double-case is to be preferred: an inner waterproof case and an outer ventilated one, to protect for direct sun, rain and snow.
A wire telephone connection is to be preferred
to GSM connection, which is a common source of
technical troubles. If available, a DSL connection
is the best choice.
2.4 GAIN Datacentre
GalileianPlus for DST-UNITS
The GAIN Data Centre is a host computer with
server architecture performing the following
major functionalities:
Data archiving for the whole GAIN network;
FTP data distribution to authorized users (typically to Processing Centres);
Internet server for the web site of the project;
Web front-end for the Data Quality Evaluation
(DQETM) process.
These functionalities are described in the next
sections.
2.4.1 Specifications
The physical characteristics of the GAIN Data
Centre are:
HW:
Dell PowerEdge 1800 with server architecture
Processor Intel Xeon 3.0GHz;
RAM: 1GB DDR2 Memory, (2x512MB);
3.5 inch 1.44MB Floppy Drive;
3 x 146GB SCSI Ultra320 (10,000rpm) Hard Drive
in SCSI 5 configuration, offering about 280GB of
user-available disk size;
RAID Disk Controller PERC 4/SC single channel
RAID card, 64MB cache, 1 int channels (U320)
41
Interreg Project ALPS-GPS Quake Net
20/48X IDE CD-ROM
Redundant power supply
3 years of Dell Next Business Day Premier
Enterprise support
SW environment:
Operating System: Linux Fedora Core 4 with kernel 2.6.14
Scripting Language: PHP 5.0.4
Database Management System: MySQL 4.1.14
Web server: Apache HTTP Server 2.0.54
FTP server: ProFTPd 1.210
Firewall: Iptables 1.3.0
Connection to Windows system: Samba Server
3.0142
As can be seen in figure 2.4.1, the GAIN Data
Centre has interfaces with the following items:
Processing Centres: to send GAIN GPS data via
the Internet (using ftp)
Data Receiving Centre (DST-UNITS local data
store) to retrieve DST-UNITS GPS data and to
receive DQE TM output via the LAN (using SMB
protocol)
Partners Data Centres: to receive GAIN GPS data
via the Internet (using ftp)
The World Wide Web (WWW) to host the project
web site and to distribute GPS data to authorised
users using common web browsers
2.4.2 Alps-GPS network data
archiving capabilities
Fig. 2.4.1.
Fig. 2.4.2.
Fig. 2.4.3.
The GAIN Data centre provides the interface
among Project Partners that collects GPS data
and the Project Partners in charge of GPS data
processing.
Data retrieval from the remote GPS local networks
(Partner Data Centres) is made using ftp over the
Internet, with two different modalities available:
ftp-push and ftp-get.
In the ftp-push modality, the GAIN Data centre makes available an ftp server to receive the
GPS data ftp-pushed by the Project Partners. Only
the hosts of the Project Partners are allowed to
upload data.
In the ftp-get modality, scheduled scripts connect to the Project Partners ftp servers to retrieve data.
The retrieved data are then available to
Processing Centres, that can retrieve them using
ftp. Scheduled scripts are in charge of avoiding
corrupted data to be available for the download.
2.4.3 FTP data distribution to
authorized users
The GPS data are downloadable by authorised
users also using a web interface, that allows a easier retrieving of data. An example of the query
web interface is shown in figure 2.4.2, while the
output of the query is shown in figure 2.4.3.
The organization of the archive has been designed
according to the FTP structure of the IGS data
centres, available at CDDIS-GNSS Data References
official web site: http://cddis.gsfc.nasa.gov/
gnss_datasum.html
2.4.4 Internet server for the
official web site of the project
The GAIN Data Centre acts as web server for the
official web site of the project. Goals of the web
site are:
42
Interreg Project ALPS-GPS Quake Net
to distribute to the World Wide Web users all the
information relevant to the project and its development state
to allow data browsing and their downloading to
authorised users
to publish the results of the GPS data quality
ranking for all the active stations
to allow the sharing of ideas, information and
comments using a forum
2.4.5 Web front-end for the
Data Quality Evaluation (DQETM)
process
One of the most important task problems arising
in controlling the GAIN network efficiency is to
evaluate continuously and automatically the reliability of the retrieved GPS data. The GAIN Data
Centre, through the software called Data Quality
Evaluation (DQETM), accomplishes this task.
The output of the DQE is processed by a proper
script that updates the WEB pages containing a
quality ranking of all the active GAIN stations.
The output quality parameters of the DQE module
published on the WEB site are:
Acquired data versus expected data (Acq/Exp (%)):
ratio between the number of data acquired (Acq)
and the expected data (Exp):
Acq / Exp =
Acq
Exp
* 100
Acquired data are the GPS data collection found
on RINEX file fulfilling the nominal session temporal range and within the receiver cut-off angle. Expected data are computed according to the
navigation RINEX data within the same nominal
temporal session range and the same receiver cutoff angle.
This parameter is important to check that the receiver is acquiring data properly or that there is
no obstruction over the cut-off angle.
Double frequency-acquired data versus acquired data (DF/Acq (%)): ratio between double frequency acquired data (DF) and total number of acquired data (Acq):
DF / Acq =
Acq - SF
Acq
ch. The total edited data is thus the sum of two
contributions:
1 Single-frequency acquired data editing (SF);
2 Short satellite passage data editing (Shrt).
This parameter is the ratio between the difference of double frequency-acquired data and total
edited data, and the double frequency-acquired
data:
Ed / DF =
Acq - Shrt - SF
DF
* 100
Cycle Slips (CS): this value represents the total number of loss of satellites lock followed by
a sudden relock of the same satellite. When the
loss of lock and the relock happens, a cycle slip
occurs. In terms of acquired data it means that
phase observables show a sudden jump due to the
reset of the ‘initial ambiguity’ value, since ‘initial ambiguity’ is an unknown integration integer
constant summed by the receiver to its own internal clock integration, representing the phase observable. Initial ambiguity can be estimated only
on double differenced data.
Noise on L1, L2, P1, P2 (N/L1, N/L2, N/P1, N/
P2): the parameters representing the noise on
the phases and codes observable are the residuals
standard deviation coming from an Euler-Goad algorithm based estimation process. Euler-Goad algorithm is used for obtaining, via a least square
estimation with undifferenced single satellitesingle receiver data, the following variables:
i the total non-dispersive delay at each epoch( (t));
ii the total dispersive delay at each epch (Î(t));
iii the initial ambiguity estimation for both L1
and L2 (N 1, N 2).
The observations model, according to the EulerGoad algorithm, is, for each epoch a for a given
satellite:
L1 = ~ - I + l1 N1 + e1
L 2 = ~ - aI + l2 N2 + e2
P1 = ~ + I + e3
P2 = ~ + aI + e4
* 100
DF is the difference between the acquired data
and the single frequency data (SF) acquired that
are edited.
This parameter is useful to check that both frequencies are acquiring data correctly, during the
satellite visibility period over the receiver.
Total edited data versus double frequency acquired data (Ed/DF (%)): data editing is also due
to short satellite passages, i.e. when a satellite is locked to the receiver for less then ten epo-
f1
, the frequencies ratio, the four
f1
unknowns are above displayed, and each ei represents the sum of the whole mismodeled signals
and instrumental noise. Once the estimation in
performed, the evaluation of ei is straightforward,
and it is possible to compute their distribution
around their mean value.
High value of noise parameters are the clear indication that mismodeled signals, for instance due
to multipath phenomena, have a great impact on
data reception.
Where a =
43
Interreg Project ALPS-GPS Quake Net
broadcast ephemeris provided in real time by the
system operator (National Geospatial-Intelligence
Agency, NGA, on behalf of the United States
Department of Defense, DoD). They are continuously computed from the observations of a small
network of about a dozen global tracking stations
and given in the World Geodetic System (WGS84).
Another type are the ephemeris provided by the
International GNSS Service (IGS). They are computed from observations of a large network of global
tracking stations and given in the International
Terrestrial Reference Frame (ITRF). There are different orbit products of the IGS, e.g.,
broadcast ephemeris (real time, daily updated,
~160 cm precision),
ultra-rapid ephemeris (real time, four times daily updated, ~10 cm precision),
final ephemeris (~13 days delayed, weekly updated, <5 cm precision).
Fig. 2.6.1 - The distribution of
GNNS sites of the EPN network.
Fig. 2.6.2 - The distribution of
the GAIN sites and the adjacent
EPN sites.
2.5 GAIN and reference
frames
The objectives of the ALPS-GPS Quakenet project
(see 1.1) call for highest precision in the determination of station positions and position changes of the entire station network (millimetres
and tenth of millimetres per year, respectively).
In order to achieve this extreme requirements, a
unique and consistent reference system has to be
used in all steps of the data processing for coordinate determination. Position coordinates refer
always to a coordinate system that has to be defined unequivocally, and position changes need a
stable origin to which the motions refer. If space
geodetic observations are used, like in the present project the measurements between GPS satellites and terrestrial points, it is necessary that
the reference system for the coordinates in space
(of the satellites) and on ground (terrestrial stations) must be completely identical. The reference system is realized in practice by a reference
frame, i.e., a set of stations with coordinates according to the definition of the reference system.
The global stations which are tracking the satellites and used for the orbit computations form a
fundamental part of the reference frame.
The GPS observations are measurements of travel
times of microwave signals transmitted by the satellites and received by the instruments on ground
with different clocks. As the clocks are not synchronized with an accuracy sufficient to reach for
the highest precision in point positioning (millimetres), only relative positions (baselines) can
be computed between stations in short distances,
where the clock errors are estimated in an adjustment procedure. For coordinate determination we
need thus terrestrial stations with coordinates in
the same reference system as the satellite orbits as a reference frame close to the network stations. The WGS84 has not such a dense reference
frame. The ITRF and its densification (EUREF) as
well as the IGS offer a large set of stations in the
alpine region. Due to crustal deformation (tectonic plate motions and regional deformations) the
coordinates of the stations change with time relative to the satellite orbits. Therefore, we have
to transform the coordinates of the reference frame by the station velocities to the observation
epoch in the data processing. The velocities are
shown in figure 2.5.1. The computed station positions of new stations refer then always to the observation epoch.
There are, in principle, two types of coordinates
publicly available for the orbits of all GPS satellites (satellite ephemeris). One of those are the
In summary we conclude that the GAIN reference frame has to be the same as the reference frame of the satellites, and needs a dense spacing
Hermann Drewes for DGFI
44
The global IGS network used as the reference frame for the orbit computations consisted until
end of 2006 of 98 tracking stations (IGb00) with
coordinates in the ITRF2000. Since end of 2006
it includes 132 stations with coordinates in the
ITRF2005, corrected for the absolute phase centre variations. The complete ITRF2005 as well as
the global IGS network includes about 350 stations with given position coordinates and velocities caused by crustal movements (figure
2.5.1). The global networks are densified by regional networks, e.g., in Europe by the European
Reference Frame (EUREF).
Interreg Project ALPS-GPS Quake Net
in the region of the measurements. This is only
fulfilled by the ITRF, at present in its realization
ITRF2005. For GPS users the ITRF2005 is realized by the IGS05, where the GPS receiver phase
centre variations are corrected according to the
IGS conventions of 2006. In order to get a dense
terrestrial reference frame we may also use stations of the ITRF/IGS densification in the alpine region.
2.6 The Geodetic Alpine
Integrated Network
(GAIN) and its relation
to EUREF
Christof Völksen for BEK
During the centuries each country, each federal state or even each municipality developed its
own reference system for mapping or land register. Mapping at the boundary between two systems was always and is in part still difficult since one had to work with two different coordinate
systems. EUREF is aiming at an uniform reference
systems that can be used at any place in Europe
for many different purposes.
The acronym EUREF stands for European
Reference System and is integrated in the SubCommission 1.3, Regional Reference Frames,
of the International Association for Geodesy
(IAG). It has been founded in 1987 at the
General Assembly of the International Union of
Geophysics and Geodesy (IUGG) in Vancouver. The
main objectives of EUREF are the definition, realisation and maintenance of a geodetic reference
system for Europe.
In close cooperation with the International GNSS
Service (IGS) a network of permanently installed GNSS receivers has been established on voluntary basis by universities, research institutes
and other groups. This network, which is known
as the EUREF Permanent Network (EPN), observes continuously the signals of the GNSS satellites. The EPN consists currently of more than
180 stations and covers Europe almost completely (see Figure 2.6.1). It serves as a densification of the IGS network in Europe and also plays
a major role in the realisation of the European
Terrestrial Reference System 89 (ETRS89). This
system is defined as connected to the stable
part of the Eurasian plate and as identical to
the International Terrestrial Reference System
(ITRS) at epoch 1989.0. The ETRS89 is the EUrecommended frame of reference for geodata
in Europe. Therefore it is used as the standard
for precise positioning, surveying and geodynamic studies throughout Europe. It is supported
by EuroGeographics, which represents nearly all
European national mapping and cadastral agencies (NMCA), and is therefore the dominating reference system in Europe.
The main objective of the GAIN network is the
monitoring of crustal deformation caused by plate tectonics or by earthquakes, land slide monitoring and meteorology. These studies request a
solid and well defined reference system such as
the ITRF2000 (International Terrestrial Reference
Frame 2000) or its subset IGb00. The network
can also be used for some applications in navigation, agriculture, mapping and surveying, which also need a unified reference system. It is therefore obvious to integrate stations of the EUREF
network into the gain network and consequently
in the data analysis. The selection of suitable reference station from the EUREF network is governed by the following arguments
covering the area of interest (the ALPS)
well known station behaviour
coordinates should be available in the reference
frame IGb00 and ETRS89
data availability / data access.
A selection of EPN sites are shown in Figure 2.6.2.
Almost 15 sites cover the ALPS with a concentration in Italy and Austria. This number of sites is
more than sufficient for the integration of the
GAIN network into the ETRS.
Although the GAIN network is processed in the
IGb00 the integration of the EUREF sites with
their well known coordinates in the ETRS89 allows
the transformation between the two systems. An
additional advantage of the integration of EPN
sites is the study of continental deformation outside the Alpine region. This study is currently
under investigation by the Bavarian Comittee for
International Geodesy (BEK), but is not directly
linked to this project.
2.7 Data analysis
POLIMI-RLB, BEK, LGIT, POLITO-ARPAP
The monumentation of most part of the GAIN stations and the availability of continuous GPS observations in the data centre, have been the starting point for the data analysis.
Different aspects have been considered by the
three analysis centres of the project.
The Bavarian Comittee for International Geodesy
(BEK) and Polytechnic of Milano (POLIMI) analysis differ from each other for the GPS stations included in their analysis and for the reference stations taken into account, whereas the Université
Joseph Fourier, Laboratoire de Géophysique
Interne et Tectonophysique (LGIT) analysis also
differs for the GPS Software used.
Briefly, the purpose of each analysis centre could
be summarized as follows:
1 The aim of the BEK has been the analysis
and the integration of the GAIN Network in the
EUREF Permanent Network (fig.2.6.2). GPS data
have been processed by means of the scientific
Bernese Software v.5.0 [Beutler et al., 2007].
45
Interreg Project ALPS-GPS Quake Net
Fig. 2.7.1 - LGIT processed
network over the Alps.
Fig. 2.7.2 - The POLIMI processed Network over the Alps. The
EUREF stations MATE, NOT1 and
CAGL, used with the other EUREF stations as reference frame,
are not plotted.
46
Interreg Project ALPS-GPS Quake Net
2 The aim of the LGIT has been the analysis
and the integration of the GAIN Network with 17
European IGS sites and with other stations belonging to the RENAG/OSUG Network, taking care of
the Western part of the Alps (fig.2.7.1). GPS data
have been processed by means of the scientific
program GAMIT/GLOBK [Herring et al, 2006].
3 The aim of POLIMI was the analysis and the integration of the GAIN Network with the FReDNet
Network [Zuliani et al., 2002] and some RING stations [Selvaggi et al., 2006] taking care of the
Eastern part of the Alps (fig. 2.7.2). GPS data
have been processed by means of the scientific
Bernese Software v. 5.0.
The GPS data processing strategy selected in the
Bernese GPS Software v.5.0 by BEK and POLIMI
analysis centres is based on the following
assumptions:
Precise satellite orbits and Earth Rotation
Parameter provided by IGS – International GPS
Service [http://igscb.jpl.nasa.gov/]
Absolute Phase Centre Variation (PCV) both of
satellites and antennas, provided by IGS.
Global ionospheric modelling provided by CODE
[ftp://ftp.unibe.ch/aiub/CODE]
Ocean tide loading provided by OSO – Onsala
Space Observatory [http://www.oso.chalmers.se/
~loading/] – was applied.
Processing of the phase double difference observations .
Tropospheric modelling was realised applying
Niell mapping function and estimating tropo-
spheric parameters every hour with one gradient
per station.
QIF (Quasi Ionosphere Free). ambiguity resolution strategy is applied.
Cut-off angle has been fixed to 10 degrees.
According to the processed network, different
IGS stations have been used as reference stations, constraining the latest coordinate realization of the IGS Reference Frame (IGb00/IGS05)
The GPS data processing strategy selected in the
GAMIT/GLOBK by LGIT is based on the following
choices:
IGS final orbits adjusted
Cut-off angle fixed to 10º
Fig. 2.7.3 - The POLITO processed Network.
Fig. 2.7.4 - TORI coordinate
time series. The red dots are the
adjusted coordinates obtained
removing less reliable stations.
Note that the jump in week
1412 have been removed.
47
Interreg Project ALPS-GPS Quake Net
A priori meteorological values GPT (Boehm et
al., 2007)
Mapping function VMF1 (Boehm et al., 2006)
Ocean loading model FES2004 (Letellier, 2004)
Absolute Phase Centre Variation (PCV) both of satellites and antennas, provided by IGS (IGS_05).
Daily solutions have been computed obtaining
coordinates and variance-covariance matrix of
the GPS stations
Polytechnic of Torino (POLITO) was also interested in elaborating GAIN Network. Its purpose was the analysis and the integration of the
GAIN Network with other Italian GPS Permanent
Network (fig. 2.7.3) for mapping and surveying
and the implementation of an automatic procedure for data quality control, based on Bernese
Software v.5.0. GPS data have been processed
by means of the scientific Bernese Software v.
5.0 following the same standards quoted above.
Moreover POLITO has realised a software called
NetDownload [[email protected]], which retrieves and prepares data for the Bernese
Processing Engine (BPE) The output of this procedure is the improvement of time series stability, as shown for the EUREF TORI station in fig.
2.7.4
Fig. 2.8.1 - WART coordinate
time series for the year 2006.
Fig. 2.8.2 - CANV coordinate
time series for the year 2006.
2.8 Outputs and Products
POLIMI-RLB, BEK, LGIT, POLITO-ARPAP
The most important results of the performed data
analysis described in Paragraph 2.7 are the daily
coordinates of the GPS Permanent Stations.
The time series of daily solutions are then analysed to get estimates of station velocities, as explained in Paragraph 3.1.
In this paragraph some coordinate time series of
GAIN stations are shown.
In Fig. 2.8.1 and Fig. 2.8.2 the computed daily
coordinates (X,Y,Z) of the WART and CANV stations are respectively plotted: these time series
are quite long as compared to those of other stations (about one years).
Green points are the coordinate in the IGS05
Reference Frame, introduced at day 309 of 2006,
whereas purple points are the coordinate in the
IGb00 Reference Frame. From DoY 309 of 2006
the points are double (green and purple points)
because IGS05 coordinates are transformed into
IGb00 coordinates applying the IGS transformation parameters (for further explanations see
Paragraph 3.1).
A problem in the time series analysis can occur
in case of Reference Frame change. Even though coordinate transformation is applied with proper parameters, residual discontinuities can be
present after datum shift compensation (see Fig.
2.8.1, Fig. 2.8.2).
The residual discontinuities are further estimated
48
Interreg Project ALPS-GPS Quake Net
and removed from the time series. In this way refined time series are obtained suitable for reliable
velocity estimates.
As an example of this procedure the coordinate
time series (North, East, Up) of WART station are
shown in Fig. 2.8.3 after the recovering of the residual jumps.
Fig. 2.8.3 - WART coordinate
time series after jump recovering.
2.9 GAIN and regional
services
IREALP-RLB, ARPAP, FondMS, UNIGE-RLG
2.9.1 Lombardia region
In Italy, the first example of GNSS regional service is given by GPSLombardia, the GNSS positioning service of Regione Lombardia.
The network project started in 2003 and involved
Regione Lombardia, Politecnico di Milano, and
IREALP, the regional research institute that finally created GPSLombardia (http://www.gpslombardia.it)
The network now consists of 18 permanent stations and covers 75% of the population of the region; it will cover all the territory after the forthcoming new installation of a permanent station
in the city of Chiavenna.
Starting from July 2005, GPSLombardia offers the
users both RINEX data for post-processing positioning and real time corrections. RINEX data are
available 24/365 from all the stations: through
a web interface, the user has the possibility to
retrieve the data by selecting the station, the
time period and the data decimation of interest.
Real time differential corrections are distributed
in different ways (also known as “mountpoint”)
through a NTRIP server (http://131.175.88.151:2
101), allowing a real time accuracy of few decimetres if using a GIS rover and few centimetre accuracy if using a geodetic double frequency receiver.
The GPS hardware of permanent stations consists
of TPSCR3_GGD + CONE antenna and TPS Odissey
receiver, from Topcon, and tracks data of both
GPS and GLONASS constellations. All stations are
provided with long duration backup battery.
Most of the stations are installed on the roofs of
civil buildings while the 4 permanent stations installed in the framework of ALPS – GPSQuakenet
project were installed on concrete pillar on rock.
Some more details are given below:
PORA: installed on the top of Monte Pora,
between the provinces of Brescia and Bergamo.
Data are retrieved using a wireless link, hiperlan standard;
SOND: installed in the vicinities of the city of
Sondrio. Data are retrieved using a wireless link
with HI-FI standard. Solar panels were used;
CLTN: installed on Monte Coltignone, in the vicinities of the city of Lecco. Data are retrieved
using a wireless link and hiperlan standard. Solar
panels were used.
SERL: installed in the “Parco Naturale di
Cariadeghe” (province of Brescia). Data will be
retrieved using a wireless link, hiperlan standard.
Solar panels are used.
All permanent stations are daily monitored by several automatic procedures:
Daily and weekly adjustment using RINEX
data, performed by Politecnico di Milano, Polo
Regionale di Como;
Real time control over communication lines;
Real time control over GPS data availability;
Controls over percentage of lost GPS data vs.
expected data;
Comparison of active GPS and GLONASS satellites vs expected ones;
Controls over “low level” state of software (caster server, RINEX server, etc);
Coordinates of the stations are distributed in
IGB00 reference frame; moreover, 7 parameters
for the transformation in the Italian local ETRF89IGM95 system are available on the website.
The website (http://www.gpslombardia.it) offers
a wide range of interesting pages for GPS users
working in Lombardia region. First of all, there is
a list of most-used GPRS contracts, allowing the
49
Interreg Project ALPS-GPS Quake Net
user to select his best one choice in order to use
GPSLombardia real time corrections.
Moreover, some interesting planning tools are
available: the GPS user can see, for example, how
many satellites are expected to be available in
the next hours and their elevation, for both GPS
and GLONASS constellation.
An expected value of GDOP and PDOP is also given. All these information are available for all
the permanent stations of GPSLombardia and are
updated 4 times a day, using ultra rapid ephemeris from CODE. The website also offers the real
time status of the software that distributes differential corrections (GNSMART from GEO++): it is
therefore possible to see how many satellites are
seen and if any station is temporarily unavailable
for TLC problems. A password limited section is
also available: GPSLombardia users can see their
total traffic, for both RINEX data and connections
to NTRIP server.
Regione Lombardia will soon contribute to the
creation of an Italian permanent GPS network,
created and managed by IGM (Istituto Geografico
Militare), whose aim will be the update of the
Italian ETRF89-IGM95 reference system.
2.9.2 Liguria Region
Services distributed on the whole territory allow
integrating data with spatially correlated effects
especially for Real Time applications. Obviously
they required a strong synergy with neighbouring
networks for the congruency of data at the border, and the continuity of the service. A national
coordination about Reference Frame and management aspects is recommended.
In particular for Liguria Region, the POGG
Permanent Station will be the first of a planned
GNSS regional network.
The institution of a Survey Regional Service concerns the realization, management and analysis
of a GNSS permanent stations network on the regional territory. Foreseen activities regard the
realisation of a data collection centre, elaboration and analysis of the regional data. Primary
importance is give not only to data quality, but
also to continuity and guaranty of the services
offered to the territory in the different institutional activities.
A brief review on possible applications, for different level of their precisions, is:
a
High precision applications (sub-centimetre precision level), to monitor continental deformation and landslides, and for meteorological
modelling.
b Medium precision applications (sub-decimetre precision level) for Real Time positioning: real
time tracking and searching targets for technical applications (technologic network, roads, …),
updating cartography, cadastral applications and
many others.
50
c Low precision applications (sub-meter precision level) in particular for navigation purposes (e.g. aircraft positioning), and for specialized
users (ambulances, firemen, etc.).
Several categories of people are potential users of
a regional Positioning Service:
1 Professional categories: Surveyors, Geologist,
Engineers, Architects.
2
Cartographic agencies: IGMI, Agenzia del
Territorio.
3 Local territorial administrative agencies, in
different sectors like civil protection (prevention
and management of risk and emergency), urban
and transport planning, tourism, cadastral, environment and energy services.
4 Association of management of territorial districts (Autorità di bacino, Comunità Montane).
5 Public and private company of transport management (terrestrial, sea, aerial)
6 Private users.
Several people like scientist researcher, administrative and technical staff of Regional Public
Administration, University, local Public structures
are involved. Thus, a dedicated staff for the GNSS
Services is very important to guarantee optimal
results and this requires to create new expertise
for young scientists.
In synthesis, the Services has to provide to the
local users the necessary tools to make the use of
satellite survey techniques easier and more economic, assist the final users, study and develop
new applications.
2.9.3 Piemonte Region
Arpa Piemonte has created a unit, which is part of
the Centro regionale per le ricerche territoriali e geologiche in order to manage and maintain the five
permanent GAIN receivers installed in Piemonte
in the framework of the AlpsGPSQuakenet project.
The related costs will be covered by ARPA annual
budget. Arpa Piemonte is currently preparing an
internet site in order to distribute the data acquired by the receivers. The data will be available
to everybody, free of charge, and may be useful
for surveyors, engineers, geologists and to the local administration for several applications (topographic surveys; surveys related to the register of
landed property; local networks etc.).
2.9.4 Valle d’Aosta Region
The permanent GPS station and the GAIN network
supply an additional instrument to the glaciologists in their monitoring activities:
glaciers tongue variation and ice-flow velocity measurements with differential GPS techniques
can be lead with high accuracy; during summer
activities the tongue position is measured and
compared with the previous one in order to reveal
annual displacement. Superficial sensors on the
Interreg Project ALPS-GPS Quake Net
glacier measure the rate displacement of the ice.
hanging glaciers displacement rate; hanging
glacier dynamic is a matter of study, in some case
the fall down of enormous ice blocks can be dangerous for downhill settlements or alpine path.
An experimental study can be implemented positioning a chain of economical remote GPS receivers along the front and recording the displacement until the falling.
The Civil Protection of the Aosta Valley
Autonomous Region investigates some alpine rescue applications of the GPS station. An example is the precise positioning of the rescue teams
over the territory, their portable GPS receivers
transmit via radio their position to the central
station that elaborates the correction and locates them. Another application under study is the
alpine path mapping with the help of the local alpine guides.
51
Interreg Project ALPS-GPS Quake Net
3 - CONTINENTAL
DEFORMATION
IN THE ALPS
52
Interreg Project ALPS-GPS Quake Net
3.1 First results from the
GAIN network
BEK, POLIMI-RLB, LGIT
The main product of the EU co-funded project
ALPS-GPSQUAKENET is the permanently installed
GAIN network in the Alps that tracks continuously
the signals of GNSS satellites. Apart from collecting
the data it is of course necessary to process the observations and estimate precise coordinates. Only
the monitoring of the coordinates allows drawing
conclusions on the tectonics of the Alps.
The collected GNSS data of the whole GAIN network
are processed by three different analysis centres:
BEK: Bayerische Kommission für die Internationale
Erdmessung (Munich, Germany)
LGIT: Université Joseph Fourier, Laboratoire de
Géophysique Interne et Tectonophysique (Grenoble,
France)
POLIMI-RLP: Politecnico di Milano (Milan, Italy).
While BEK and POLIMI-RLB are using the GNSS
analysis package BERNESE 5.0, LGIT exerts GAMIT
for the data processing. The data of the network
are evaluated independently applying identical
standards like correction models for the antenna (absolute), realising the same reference frame
(ITRF) and applying similar models for the correction of the troposphere and for example the effects
of ocean loading. These standards have been defined by the three groups beforehand to ensure that
comparable results will be achieved. The networks
of the different analysis centres are slightly different due to the selection of different GNSS sites
that are added to the network (e.g. stations of the
EUREF Permanent Network [EPN]). Nevertheless,
the independently estimated solutions of the three
analysis centres agree quite well.
Data are available for each day and therefore daily solutions for the coordinates are computed for
the whole network. The coordinates for each station change with time due to plate tectonics, earthquakes and possibly due to local movements of
the monument. The strongest signal seen in the position changes is of course caused by plate tectonics. But also snow on top of the antenna, which
usually stays only for a couple of days, can have
an impact on the coordinate solution. These undesirable effects like local movements of the monument carrying the antenna or snow on top of the
antenna compromise the signal, which is of course
the change of position due to tectonics. One has
to be extremely careful with the interpretation of
position changes. It is therefore very important to
exclude all these undesirable effects. This ensures
that the detected signals are truly dependent on
tectonics only.
North
In order to estimate the precision of the coordinate components it is necessary to remove a linear
trend from the time series of the position changes,
which is mainly caused by plate tectonics. And in
some cases it is also necessary to remove jumps
(e.g. snow on the antenna etc.). The mean root
mean square of these residuals is given in table
3.1.1 for the three coordinate components in the
GAIN network. The achieved precision is very satisfying. Already daily position changes in the order
of a few millimetres are detectable. This indicates
that the impact of rather small earthquakes in the
vicinity of the station should be visible.
The GAIN network has just recently been completed. Some GNSS sites are installed for more than
two years. Others are operational only for a few
months. It is quite obvious that the precision of
the estimated velocities for each site depend very
much on the length of the observation period.
Sites like ACOM, BASO, HRIE and ZOUF are operational for the entire observation period of almost
two years, while sites like SOND, POGG, MOCA and
STATION
ACOM
AFAL
AGNE
BASO
BOSC
BREI
CANV
CARZ
DEVE
FAHR
FDOS
GROG
HGRA
HRIE
JANU
LEBE
LFAZ
MAVE
MBEL
MDEA
MOCA
MONC
MPRA
OATO
POGG
PUYA
ROSD
SOND
WART
ZOUF
East
Vj
1.43
1.44
sVj
Vh
sVl
sVh
0.5
0.6
0.8
0.5
0.7
0.6
0.5
0.6
2.5
0.5
0.6
0.9
0.6
0.6
1.7
0.9
0.6
0.7
0.7
0.5
1.0
0.6
0.5
0.5
1.3
0.6
0.7
0.9
0.5
0.5
0.7
0.7
1.1
0.7
1.0
0.7
0.7
1.0
2.5
0.7
0.9
1.3
0.8
0.7
3.0
1.4
0.8
1.0
1.1
0.7
1.6
0.9
0.6
0.7
1.9
0.8
0.9
1.4
0.7
0.7
mm/a
14.7
14.5
10.7
18.5
17.5
15.6
15.9
14.1
18.8
12.7
12.9
21.8
18.2
20.1
15.3
16.3
16.1
19.0
19.3
16.3
13.8
14.0
15.9
14.2
11.6
14.2
15.2
18.7
16.3
12.8
22.2
16.2
22.1
23.1
21.1
23.0
21.3
23.9
19.1
25.0
23.8
18.4
12.5
22.1
27.4
21.7
19.3
22.3
16.8
21.2
15.2
23.4
22.5
20.6
27.3
23.3
15.6
14.6
20.4
23.5
Height
RMS [mm]
GAIN Network
Vl
3.74
2.8
2.2
2.1
-1.7
3.8
1.3
2.2
-5.3
-6.9
4.3
4.8
-0.1
7.1
-1.3
-0.2
-6.2
-0.7
3.7
-6.1
1.6
-2.7
--1.2
3.2
0.8
4.5
2.8
0.7
0.7
1.1
0.7
1.0
0.7
0.7
1.0
2.5
0.7
0.9
1.3
0.8
0.7
3.0
1.4
0.8
1.0
1.1
0.7
1.6
0.9
0.6
0.7
1.9
0.8
0.9
1.4
0.7
0.7
Tab. 3.1.2 - Estimated velocities
in the GAIN network.
Tab. 3.1.1 - Average precision
achieved in the GAIN network.
53
Interreg Project ALPS-GPS Quake Net
Experience has shown that it is in general not possible to estimate significant vertical motion within
a time span of less then 3 to 4 years. Heights estimated with GPS are the weakest of all three coordinate components as table 3.1.1 shows. Therefore
the vertical motion of the sites GROG, JANU, MBEL
and POGG are not presented in Table 3.1.2.
Fig. 3.1.1 - The horizontal velocities of the GAIN network in the
ITRF2005.
Fig. 3.1.2 - The velocities of a
part of the Bavarian Real-time
network SAPOS. The network
is fixed to the Eurasian plate
(based on the NUVEL-NNR1A
model).
Table 3.1.2 shows also that the horizontal velocities of most stations are quite similar. This is of
course no surprise since all the stations are affected
by the plate motion of the Eurasian plate. This becomes also visible by looking at figure 3.1.1, which shows the horizontal velocities only. The general trend of the motion points into east-north-east
with a rate of 20 to 30 mm.
others have observation periods as little as half a
year. Significant estimates for the site movement
cannot be expected as long as the observation period is rather short. Table 3.1.2 shows the estimates for the horizontal and vertical velocities for a
number of sites from the GAIN network.
Table 3.1.2 shows clearly that the accuracy of the
estimated components increases with time. The
best results were achieved with stations, which are
operational close to two years [BASO, HRIE etc.].
The opposite is visible for the above mentioned
stations with rather short observation periods. But
other aspects are proven by this table as well. JANU
has been covered by snow for some weeks. The receiver was also not always operational. The analysis of the GPS data estimates a subsidence of more
than 50 mm per year. It is quite clear that this cannot be the realistic vertical motion of the station.
54
Figure 3.1.2 shows the Bavarian part of a network
that belongs to the Satellite Positioning Service
(SAPOS) of the German State Survey. This network
has been analysed by the BEK for the last 2.5 years.
Compared to figure 3.1.1 the horizontal residual velocities are much smaller because the network is
fixed to the Eurasias plate. Most stations show the
same residual velocity apart from some stations like
the WANK, which is caused by the instability of the
local marker. This shows again the importance of
stable markers. But this figure gives us a good indication of what to expect in one or two years time.
The analysis of the gain network should produce
velocities for the horizontal components with accuracy better than 1 mm/a. Sudden position changes,
as small as a few millimetres, should be detectable.
The correlation with seismic events might give an
indication for an earthquake.
Interreg Project ALPS-GPS Quake Net
Fig. 3.2.1 - Global tectonic
settings, showing the AfricaEurasia collision with velocities
increasing from west to east and
the counter-clockwise rotation of
the Adriatic micro-plate.
Fig. 3.2.2 - Seismicity along the
Eurasian-African plate boundary.
Fig. 3.2.3 - Historical and Instrumental seismicity in the western
Alps.
3.2 Active deformation
in the western Alps
Andrea Walpersdorf for LGIT
3.2.1 Introduction
The western Alps result from the Europe-Africa collision which leads to the indentation of Europe by
the Adriatic promontory (Tapponier, 1977). A probable counterclockwise rotation of the indenter
has been suggested (Anderson and Jackson, 1987;
Ménard, 1988; Vialon et al., 1989) (Figure 3.2.1).
The present day convergence velocity between
Africa and Europe is less than 1 cm/yr and increases from the western to the eastern Mediterranean
(Figure 3.2.1).
Although the countries bordering the western
Alps have low to moderate seismotectonic activity (Figure 3.2.2), several earthquakes of Ml > 5
are recorded in historical catalogues (Beauval and
Scotti, 2003) and by paleoearthquake evidence
(e.g. Sébrier et al., 1997) (Figure 3.2.3). A mean
of one magnitude 6 earthquake occurs in the western Alps per century.
The Alpine region is the most deforming part of
western and central Europe. A synthesis of focal mechanisms by Sue et al. (1999) and Delacou
et al. (2004) shows that the kinematics are characterized by a continuous area of orogen-perpendicular extension following closely the large-sclae topographic crest line of the Alpine arc
(corresponding to E-W extension in the western
Alps), while thrusting is observed locally, limited to areas near the border of the Alpine chain
(Figure 3.2.4). The complex current tectonics of
the Western Alps would result from the interfe-
rence of the Europe-Adria collision and relative
rotation at the limits of the belt, and buoyancy
forces within the western Alpine lithosphere (Sue
et al., 1999).
55
Interreg Project ALPS-GPS Quake Net
S compression (which could have been associated
to the motion of the Corsica-Sardenia-block), but
is rather characterized by a dominating E-W extension (Figure 3.2.5).
a
The third measurement campaign in 2004 has identified measurement outliers for a certain number of
sites. These errors could be due to blunders in the
tripod installation, unidentified tribrach offsets,
unidentified antenna phase center offsets, etc. and
cannot be evidenced by only two campaign measurements. However, even after the third measurement, it is difficult to identify the epoch bearing
the outlier. Therefore a forth campaign will be necessary to confirm the velocity field of this dense
GPS network.
Permanent measurements
b
Fig. 3.2.4 - a) Map of the strain/
stress states in the Western Alps
(stress tensors represented by
black and open arrows, P and
T axes represented by red and
blue lines) and b) 3D view identifying regions of dominating
extension (blue), compression
(red) and strike-slip mechanisms
(green) (from Delacou et al.,
2004).
56
3.2.2 Existing GPS measurements
covering the western Alps
Campaign measurements
To measure the present day deformation in the western Alps, a 60 stations campaign GPS network
has been established in 1993 covering the French,
Italian and Swiss part of the western Alps (Figure
3.2.5, from Vigny et al., 2001). Three measurement campaigns have been done in 1993, 1998
and 2004. The two first campaigns have shown relatively slow velocities (1-3 mm/yr) on all of the
sites with respect to stable Europe. The uncertainties have been evaluated by the respective campaign repeatabilities to 1.3 mm/yr, which means that
most of the velocities are close to or inside of the
measurement error. The deformation pattern evidenced by these measurements has no visible N-
Velocity estimates with an accuracy of less than
1 mm/yr have been obtained by measurements of a network of permanent GPS stations from
1996 to 2001 covering central and western Europe
(Nocquet and Calais, 2003, Fig. 3.2.6). The authors have shown that central Europe behaves
as a rigid block with internal deformation of no
more than 0.4 mm/yr. There is almost no motion
west of the Rhine Graben and on the Iberian peninsula, and less than 0.6 mm/yr across the Rhine
Graben and the Pyrenees. The current strain pattern in the western Alps combines E-W extension
and right-lateral shear. There is some evidence for
a counter-clockwise rotation of the Adriatic microplate which appears to control the strain pattern
along its boundaries in the Friuli area, the Alps and
the Apennines (Nocquet and Calais, 2003) (Fig.
3.2.6).
The first permanent GPS network dedicated to the
observation of the deformation of the western Alps
has been installed in the French Alps and its foreland from 1997 on (the REGAL network, now alpine part of the French RENAG consortium, http://
renag.unice.fr). By now, about 30 stations installed directly on the bedrock are operational (Figure
3.2.7).
First results of this permanent network have been
published by the initiators (Calais et al., 2002,
Figure 3.2.8), based on data from 1997 to 2001.
The authors find velocities in the French Alps of
less than 2 mm/yr with respect to stable Europe,
with uncertainties from 0.3 to 1.4 mm/yr depending on the age of the station. The central part of
the mountain belt is dominated by E-W extension.
Compressional strain oriented N-S to NW-SE is noted in the southern Alps. These geodetic data as
well as seismotectonic observations (Figure 3.2.8)
are coherent with a model where the deformation
of the western Alps is mainly controlled by the
anti-clockwise rotation of the Adriatic micro-plate
with respect to stable Europe.
The comparison of these first velocity estimates based on the REGAL data up to 2001 with a more
recent solution constrained by REGAL data up to
Interreg Project ALPS-GPS Quake Net
2004 presented in Walpersdorf et al. (2006) (Figure
3.2.9) shows a general decrease of site velocities
(from typically velocities of 2 mm/yr to 1 mm/yr).
The compression in the southern Alps between stations GRAS and MICH has not been confirmed, but
seems to persist further west between sites MARS
and MICH, and MTPL and SAUV.
Semi-permanent networks
One of the regions in the western Alps where the
highest deformation rates have been expected is
the Jura area, located between the alpine orogen
and its foreland. The Jura is known to have been
an active area during the Neogene. Some evidence suggests that this is still the case, but precise knowledge of deformation and slip rates is still
unavailable. A local semi-permanent GPS network
was installed in 2000 to address this issue and to
improve the seismic hazard assessment of the region. The 6 sites are measured at least twice a year
for about 10 days, to obtain position time series
approaching the quality of permanent stations.
This semi-permanent approach requests the use of
a single GPS receiver for the 6 sites. The semi-permanent sites have lower constraints for the site selection as permanent sites as they need no electricity and telephone. Therefore, semi-permanent
observations can be an efficient means of densifying permanent GPS networks for geodynamic purposes. Figure 3.2.10 shows the Jura velocity field
obtained in Walpersdorf et al. (2006). The uncertainties are evaluated to about 0.3 mm/yr. Most of
the Jura velocities with respect to stable Europe
are lower than this error limit. The major feature
in this velocity field is a 1 mm/yr relative velocity
between sites JU02 and JU04, giving a hint of the
present day activity of the Vuache fault, situated
between the two sites.
3.2.3 Contribution of the
ALPS-GPSQUAKENET project:
the GAIN network
In the framework of the Interreg IIIB project ALPSGPSQUAKENET, 8 new permanent GPS stations have
been installed in the western Alps, 6 in France
and 2 in Italy (Figure 3.2.11), as part of the GAIN
network covering the entire Alpine arc. These 8 stations will contribute to quantify precisely the slow
but complex deformation in the western part of
the Alps. The resolution of displacement rates expected to be inferior to 1 mm/yr requires however
continuous observations during at least 5 years.
For most of the GAIN stations a 1 to 2 years data
span has been available at the end of the project
(spring 2007). The velocities evaluated from these
data have typical formal errors of the order of 2 to
1 mm/yr as seen on Figure 3.2.11. We need to pursue the measurements until 2010 at least to obtain
significant displacement rates characterizing the
slow deforming western part of the Alps.
3.3 Active Deformation
in the South-Eastern Alps
Fig. 3.2.5 - GPS velocities from
2 measurement campaigns in
1993 and 1998, and the dominating extensive strain pattern
(from Vigny et al., 2001).
Riccardo Riva and Alessandra Borghi for DST-UNITS,
POLIMI-RLB and UNIMI-RLB
3.3.1 Introduction
The South-Eastern Alps represent one of the most
seismically active regions in Europe and an outstanding natural laboratory for the study of active deformation.
Collision between the African and Eurasian plates,
in fact, leads to deformation of the plate boundary,
with consequent stress localization on active faults that leads to a very high seismic activity.
The largest reported earthquake in the region took
place in Western Slovenia in 1511, with an estimated magnitude of 6.9 [Fitzko et al., 2005], while in
the last three decades we have observed the 1976
Friuli sequence, where the largest shock had a magnitude of 6.5 [Aoudia et al., 2000], and more recently the 1998 and 2004 earthquakes in Western
Slovenia, with magnitude 5.7 and 5.3 respectively
[Bajc et al, 2001].
In order to study the complex geodynamics of the
region, a dense GPS network has been established
in the last few years, as it is possible to see in
Figure 3.3.1, where blue dots represent the continuous stations of the FreDNet, the red dots the
newly installed GAIN stations and the black dots
the campaign GPS sites initially installed in collaboration by University of Trieste, Politecnico di
Milano and University of Milano and later incorporated in the ALPS-GPSQuakenet project.
In Figure 3.3.1 we show a topographic map of the
region, where the main mapped active faults are
identified by blue lines and each dot represents a
Fig. 3.2.6 - From Nocquet and
Calais, 2003. The crustal velocity
field of western Europe from
permanent GPS array solutions,
1996-2001.
Fig. 3.2.7 - The French REGAL
network (REseau Gps permanent
dans les ALpes) in 2006.
Fig. 3.2.8 - From Calais et al.
(2002). GPS velocity field and
strain tensors in the western Alps
compared to focal mechanisms.
Fig. 3.2.9 - REGAL velocity fields
based on data from 1997 to
2001 (Calais et al., 2002) and
on data from 1997 to 2004
(Walpersdorf et al., 2006).
57
Interreg Project ALPS-GPS Quake Net
Figure 3.2.10. From Walpersdorf
et al. (2006). The velocity field
in the semi-permanent Jura
network with respect to stable
Europe, based on measurements
from 2000 to 2004.
GPS station: it is evident how the regional geodynamics is extremely complex, so that a high density distribution of GPS stations is absolutely necessary to understand how internal deformation is
allocated on the different faults.
Fig. 3.2.11 - First velocity estimates with respect to stable Europe,
including 1 to 2 years of data
on most of the GAIN stations
installed in the framework of the
ALPS-GPSQUAKENET project.
3.3.2 GPS data processing
58
In 2002 a GPS non-permanent network has been
set up in Friuli Region and in 2004 it has been enlarged with six sites in western Slovenia (Figure
3.3.2).
The FreDNet GPS continuous stations have also
been integrated in our network [http://www.
crs.inogs.it/frednet/ItalianSite/XFReDNetHome.
htm] to increase its spatial resolution and improve the geodetic description of the area. Moreover,
the use of continuous GPS stations allows a better link between our local network and a global reference frame (IGS05 for the late 2006, IGb00 for
2006, 2005 and 2004; ITRF00 for 2003 and 2002)
and helps in connecting campaigns performed in
Friuli Region with those in Slovenia, which typically take place on different days. Since the new GAIN
station BASO has been operating from July 2005,
it has been added to our network. The Slovenian
EUREF station GSR1 has been also integrated into
the network.
For each campaign site, the monumentation was
carefully performed to ensure a sub-millimetre centring at every occupation. At each station point, a
25 cm long still rod was fixed in solid rock and a
properly designed steel pillar holding the antenna was centred on this ground part, using toroidal
levels calibrated before every campaign. For each
GPS site, we installed networks of control points and performed spirit levelling measurements to
check for spurious local site effects.
During all campaigns, data were collected for four
consecutive days, with daily sessions of eight
hours, and at some safer sites we have measured
continuously for four days. The sampling rate was
15” for non-permanent stations and 30” for CGPS.
Data analysis was performed with the Bernese
Software v.5.0 [Hugentobler et al., 2004], where
the Quasi Iono Free (QIF) strategy for ambiguity
fixing was selected. Tropospheric parameters were
estimated on one hourly basis and wet delays were
modelled as stochastic parameters, using the mapping function Dry_Niell. The ionospheric disturbance has been treated using global ionospheric models estimated by CODE [Hugentobler et al., 2000]
in the L1&L2 ambiguity estimation step and using
the iono-free observations (L3) in the co-ordinate
computation.
Observations were first analysed on a daily basis with multi-base approach, without constraining any stations and saving the normal equations. Subsequently, the daily normal equations
were combined into multi-day solutions for each
year with the ADDNEQ program of the Bernese software, fixing the co-ordinates of the five reference
stations used in this network (GRAS, GRAZ, WTZR,
ZIMM and MATE), which were propagated to the
current epoch. These stations were selected because they are the closest to the Friuli-Slovenia
Region. In this way we have obtained a set of coordinates for each campaign station, which allows us
Interreg Project ALPS-GPS Quake Net
Fig. 3.3.1 - topographic map of
the South-Western Alps. Blue lines indicate mapped active faults (strike-slip with red arrows,
dip-slip with blue triangles). Red
dots are GAIN stations, black
dots campaign stations and light
blue dots FreDNet stations.
Fig. 3.3.2 - map of the GPS sites
of the Friuli-Slovenia non-permanent network (blue points), the
FreDNet continuous stations (red
points), the EUREF GSR1 station
(green point) and the GAIN
station BASO (pink point).
59
Interreg Project ALPS-GPS Quake Net
Y(ti) is the daily coordinate at time ti,
a is the site velocity in the coordinate component taken into account
b the interceptor with the ordinate axis
w is the annual frequency
c and d are the magnitude of the periodic signal
H is the Heaviside step function,
ng is the number of jump i.e. due to change in
reference frames, changes in antenna/receivers or
co-seismic displacement etc..,
gj are the amplitude of the jump accursed at time
Tgi.
The unknown parameters (a, b, c, d and gj) have
been computed by Least Squares Estimator using
a proper stochastic model that takes into account
of the time correlations existing in GPS time series
[Barzaghi et al, 2004], applied to a dataset cleaned from outliers.
3.3.3 GPS results for continuous
stations
Fig. 3.3.3 - continuous GPS
stations, horizontal displacement
rates for years 2003-2006 with
respect to the Eurasian plate
velocity as defined by model
NNR-NUVEL1A.
Fig. 3.3.4 - sketch of the geodynamics of the Western Alps,
explaining the formation of
right-lateral and thrust faults as
a consequence of the regional
tectonics.
to estimate the displacement vectors between the
different campaigns.
For the CGPS stations we have elaborated the whole available database, from installation to the end
of 2006, with the use of Bernese Software v.5.0,
selecting the same procedure used for the GPS
campaign analysis up to the daily solution computation. The daily coordinate solutions constitute a time series for each station and allow a deeper
analysis of the GPS site velocities. All the coordinates have been reported in the same reference frame (IGb00) using the official transformation
parameters.
Each coordinate component (North, East, Up) has
been fitted using the following deterministic model: a linear trend plus a periodic signal with annual frequency. Due to the shortness of the time
series, about three years, any estimate of the frequency of the periodic signal will not be reliable,
and it has been fixed in accordance to standard
GPS practice.
Y(t i) = a·t i + b + c·sinwt + d·coswt +
where
60
ng
j=1
gjH(ti - Tgj)
The first results obtained by processing three years
of data for the FredNet network give a picture of
the large scale deformation of the whole area and
are presented in Figure 3.3.3.
In order to provide a clear representation of the
motion of the GPS stations, we have taken as reference the Eurasian velocity as provided by the geological model NNR-NUVEL1A.
In this way, any displacement different from zero
means that the GPS site is actually moving with respect to the velocity field predicted by the geological model.
Considering the error ellipses, we see how the three
northernmost sites AFAL, ZOUF and ACOM, and the
south-western site CANV, do not show significant
motions. On the contrary, we can clearly observe
compression in north-west direction along the line
TRIE, MDEA and MPRA.
The magnitude and direction of the observed motion is consistent with the counter-clockwise rotation of the Italian peninsula, as an effect of the
African plate moving northwards: the whole process is sketched in Figure 3.3.4.
In the left cartoon, Plate A represents Eurasia,
Plate B Africa, and the central part, where the faults are located, the Alps. The combination of thrust
faults and strike-slip faults arises from the presence of a pole of rotation in the western part of the
domain, as further clarified represented in the right cartoon of Figure 3.3.4.
3.3.4 GPS results for campaign
measurements
The analysis of the motion of the continuous stations might be enough to depict the large-scale
deformation of the whole region. However, considering the density of mapped fault already described in Figure 3.3.1, a densification of the network
Interreg Project ALPS-GPS Quake Net
is necessary in order to understand if and where
the regional deformation is going to be localised,
leading to potential seismic risk.
Preliminary results, after the occupation of the
GPS campaign sites in 2002, 2004 and 2005, are
displayed in Figure 3.3.5. We can observe as large differences in displacement magnitude and direction arise between sites close to each other,
meaning that deformation is not homogeneously
spread over the whole region. Interpretation of
those motions requires a further effort of campaign geology and modelling work, in order to understand which geological features are responsible for
the local geodynamics. More GPS campaigns and
the addition of other space-geodetic techniques,
such as Interferometric Synthetic Aperture Radar
(InSAR), will also help to increase the measurements accuracy.
gins to relax, with a process that can last for several centuries, depending on the intensity of the
earthquake and the properties of the Earth crust.
In the right panel of Figure 3.3.6, we show the accumulated postseismic deformation between 1511
and the present: we see how the region affected by
significant deformation is much wider than in the
coseismic case, and that displacement amounts to
Fig. 3.3.5 - campaign GPS
displacements accumulated
between 2002 and 2005.
Fig. 3.3.6 - map view of horizontal displacement following the
1511 earthquake. Left panel:
coseismic; right panel: postseismic relaxation between 1511
and 1976.
3.3.5 Co- and post-seismic
deformation of past
earthquakes
In this section we will study the effect of the largest reported earthquakes in the area, where coseismic deformation means the elastic (instantaneous) deformation that follows the rupture, while
postseismic deformation is the effect of stress relaxation in the lower crust and upper mantle. The
moment an earthquake takes place, in fact, a permanent deformation occurs in the area around
it, leading to a situation of non-equilibrium: the
excess stress causes material to flow in deeper crustal layers, which results in a long-lasting deformation process.
Starting from the largest reported seismic event,
the 1511 Western Slovenia earthquake, we show
coseismic deformation in the left panel of Figure
3.3.6, where the scale is in millimetres: we can
see how some areas where displaced by more than
one meter.
After the earthquake, the accumulated stress be-
61
Interreg Project ALPS-GPS Quake Net
Fig. 3.3.7 - map view of horizontal displacement following the
1976 earthquake. Left panel:
coseismic; right panel: postseismic relaxation between 1976
and 1998.
Fig. 3.3.8 - map view of horizontal coseismic displacement
following the 1998 (left panel)
and the 2004 (right panel)
earthquakes.
62
about 20 cm in the part where our GPS sites are located, i.e. around the yellow lobe at the centre of
the picture. Most of this deformation has occurred
in the first few centuries after the earthquake, but
a small residual is still present today, as we will
show later.
Analogously, we can study co- and post-seismic deformation for the main shock of the 1976 Friuli
earthquake, displayed in Figure 3.3.7. Due to the
smaller size of the earthquake, deformation is concentrated in a much smaller area; nonetheless, we
obtain up to 30 cm of maximum coseismic deformation and more than 4 cm of postseismic relaxation between 1976 and today.
In Figure 3.3.8 we show coseismic deformation for
the last two relevant earthquakes, occurred near
Bovec in western Slovenia in 1998 and 2004 (left
panel 1998, right panel 2004): we neglect the postseismic signal, which is below the millimetre level. Since the 2004 event took place after the first
GPS campaign in western Slovenia, it might have
affected later measurements: model results, however, indicate that only the site TRIG could have
experienced a noticeable displacement.
Modelling postseismic relaxation from past earthquakes gives us the opportunity to evaluate their
effect on the GPS measurements that we have been
performing in the last few years. In Figure 3.3.9
we show model postseismic relaxation rates due
to the cumulative effect of all the above discussed earthquakes. The small blue arrows show the
expected motion of our GPS sites: we can see how
all of them are at the sub-millimetre level, suggesting that we can consider postseismic deformation as a second order effect when we interpret
the GPS measurements, as those of Figures 3.3.3
and 3.3.5.
3.3.6 Stress transfer
In the previous sections we have analysed what is
the impact of past earthquakes on surface displacement. Here we want to discuss another important aspect of the interaction between earthquakes
in the same region, namely Coulomb Failure Stress
(CFS).
Interreg Project ALPS-GPS Quake Net
Fig. 3.3.9 - map view of presentday postseismic relaxation rates.
Fig. 3.3.10 - map view of CFS
due to the 1511 earthquake
and projected on the 1976 fault,
computed at a depth of 5 km
with μ=0.4. The three black solid
segments indicate the surface
traces of the 1976, 1998 and
1511 earthquakes, respectively,
starting from the West.
The mathematical definition of CFS is the
following:
CFS = t + m·sn
where t is tangential stress, sn normal stress and μ
the friction coefficient: tangential stress is directly
related to slip along the fault plane, while normal
stress influences the degree of locking of the fault.
A positive value means that an earthquake is facilitated, while a negative value means that an earthquake is prevented. When an earthquake occurs,
in fact, the stress field around it is modified and
this can affect the behaviour of active faults located in the neighbourhoods. Modelling CFS does not
tell where and when another earthquake will take
place, but gives an indication about which regions
will see an increase or a reduction of seismic risk.
In order to see whether there has been a positive interaction between past earthquakes in the
South-Eastern Alps, we have modelled CFS starting
from 1511 and accounting for both coseismic and
postseismic effects.
In Figure 3.3.10 we show the effect of the 1511
earthquake on the 1976 earthquake, identified by
the westernmost black segment and focal mechanism. The colour scale portrays the logarithm of
the Coulomb stress in Pascal: generally, values are
considered significant when they are above 1 bar,
equal to 105 Pa, here represented in orange and
purple.
Since the 1976 earthquake is located in the yellow
region, where CFS has a value of about half a bar,
we cannot say that this seismic event has really
been triggered by the 1511 earthquake; however,
it looks like there has been a positive interaction
between the 1511 and 1976 events.
Next, we add the contribution of the 1976 Friuli
earthquake and its relaxation up to 1998. In Figure
3.3.11 we plot CFS on the fault that ruptured during the Bovec earthquake: the earthquake falls in
an area where there has been definite interaction
with the previous earthquakes, in particular with
the 1511 event. The distance between the 1976
and the 1998 earthquakes, however, appears to be
too large to implicate a noticeable stress transfer
between the two events. The reason why Figure
3.3.10 and 3.3.11 are so different, even if both
are dominated by the stress field induced by the
1511 earthquake, is in the fact that the 1976 and
1998 earthquakes have very different focal mechanisms, so that the result of the projection of shear
63
Interreg Project ALPS-GPS Quake Net
The mechanism is similar to the1998 event, i.e. a
right lateral strike-slip, with the addition of a dipslip component. If the location is very close to the
1998 event, than the similarity between the two
earthquakes leads to a negative CFS value, consistent with the fact that the first earthquake has
released most of the tectonically accumulated
stress. However, the is an uncertainty about the
exact location of the 2004 event, that could put
it outside the 1998 stress shadow, in an area where the positive influence of the 1511 earthquake is
still dominant.
In conclusion, as far as Coulomb stress transfer is
concerned, all major seismic activity in the SouthEastern Alps seems to be dominated by the long lasting effect of the 1511 earthquake. The 1976 seismic event, though rather large in terms of surface
motions and damage, is probably too far from the
area of the 1998 and 2004 earthquakes to be able
to influence them. The similarity between the 1998
and 2004 earthquakes, in turn, suggests that either the location of the 2004 event is not accurately determined, or that not all of the accumulated
tectonic stress had been released during the 1998
seismic sequence.
3.4 Glacier shrinkage and
modeled uplift of the Alps
Valentina Barletta for UNIMI-RLB
3.4.1 Introduction
Region Area
Yearly Reduction, % Mass Balance m/yr, w.e.a
Austria
0.63
- 0.64
France
0.47
- 0.97
Italy
0.47
- 1.19
Switzerland
0.84
- 0.40
Germany
0.63
- 0.64
Fig. 3.3.11 - map view of CFS
due to the 1511 and 1976 earthquakes and projected on the
1998 fault, computed at a depth
of 5 km with μ=0.4.
64
and normal stress on the fault plane is also very
different.
Last, we can study the combined effect of the
1511, 1976 and 1998 earthquakes on the most recent 2004 earthquake in western Slovenia; results
are shown in Figure 3.3.12.
Alpine glaciers are subject to rapid decline starting from 1850 A.D., as inferred from analyses of
regional and national glacier inventories, as done
in the present study, or from comparison of glacier data obtained from Landsat Thematic Mapper
(TM) data with previous glacier areas [Paul et al.,
2004]. Alpine glacier shrinkage is consistent with
worldwide large ice mass reductions [Paul et al.,
2004; Haeberli et al., 1999] and can be considered an indicator of global climate change [IPCC,
2001]. Changes in volume of glacier masses in the
Alps are expected to induce vertical uplift due to
the Earth’s elastic and viscoelastic response to surface load redistribution, as modeled in this study. Uplift rates from Alpine glacier wasting can be
compared with those predicted for larger ice complexes, for example the Patagonian ones treated by
Ivins and James [2004], thus providing testable
tools for modeling lithosphere-cryosphere interaction, within two different environments, the Alps
and the Andes.
3.4.2 Alpine Glacier Mass Loss
World Glacier Inventory (WGI) data have been used
to evaluate the mass loss affecting the Alpine glaciers on a known time interval. WGI reports glacier data (surface area, lenght and main dimensional parameters) collected between the 50’s and
Interreg Project ALPS-GPS Quake Net
the 80’s of the XX century: it has been necessary to
apply an area reduction factor in order to estimate
more recent glacier surface areas. Updated glacier
areas are evaluated for the years 1996, 1997, 1998
and 1999 by applying a surface reduction factor
specific for each glacier, considering the Alpine region where it is located. This factor is obtained
from the available literature dealing with glacier
shrinkage rate in the different Alpine regions in
Table 3.4.1. Main sources for calculating the reduction factor are multitemporal national and regional glacier inventories, whose comparison permitted the evaluation of reliable glacier reduction
over a time frame of one or two decades [Paul et
al, 2004; Käab et al., 2002; Biancotti and Motta,
2001]. Besides, the numerical variation affecting
the whole glacier sample is considered: glaciers
indicated as ``Glacierets’’ in the WGI are removed
from the data-set. This choice is due to the authors’ findings in the Italian Alps where all the glacierets reported in the WGI data base at the end
of the nineties disappeared. Thus the glaciers considered in this study covered in 1999 an area of
2215 km2. Moreover the choice of evaluating the
glacier mass loss for the years 1996-1999 is also
supported by the availability, for that time interval, of specific mass balance data of a representative sample of Alpine glaciers. The mass balance data published on FOG [IAHS-UNESCO, 2005]
can be considered representative of the mass variations which affected Alpine Glaciers on the studied time frame. In fact, according to the “glacier
regionalism’’ introduced by Reynaud et al., [1984],
glaciers located in the same region show stronger
correlation among their mass balances. Since more
than one mass balance for the period 1996-1999
has been published for each region, a mean mass
balance value per year has been obtained by averaging the mass balances for each region. These yearly values have then been averaged over the considered four years to obtain a unique value for each
Alpine region (second column, Table 3.4.1), to be
used within a high spatial resolution normal mode
scheme. The average of the values reported in Table
3.4.1 equals -0.71 m/yr water equivalent: this value represents an estimate of the mean mass loss
affecting the whole sample of Alpine glaciers and
is used within the low spatial resolution calculation. In this scheme it is necessary to include the
water feeding the Mediterranean sea: by applying
the above reported glacier loss value, the sea-level rise (limited to the Mediterranean sea) of 0.46
mm/yr is obtained.
Fig. 3.3.12 - map view of CFS
due to the 1511, 1976 and 1998
earthquakes and projected on
the 2004 fault, computed at a
depth of 5 km with μ=0.4.
Tab. 3.4.1 - Coefficient of Glacier
Area Reduction Per Year and the
Mass Balance, Averaged Over
1996-1999.
Fig. 3.4.1 - Caption: Low
resolution map of the uplift rate
caused by glacier mass balance
of -0.71 m/yr.
Fig. 3.4.2 - Computed high
resolution (elastic response) to
ice mass loss. In the inset (corresponding to the the dashed box
in main figure), our white contour lines are superimposed to
the vertical rates obtained from
the new national height system
(LHN95) of Switzerland [Schlatter
et al., 1999], for comparison.
65
Interreg Project ALPS-GPS Quake Net
Layer
r km
r kg/m3
u Pa s
Pa
1
6371.0
2650.0
2.97*10
1.00*1035
2
6352.5
2750.0
5.58*1010
2.15*1019
3
6341.0
2900.0
6.81*1010
5.00*1021
4
6331.0
3439.3
7.27*1010
4.64*1020
Tab. 3.4.2 - Rheologic Structure.
5
5951.0
3882.3
1.09*1011
4.64*1020
Fig. 3.4.3 - High resolution
viscous contribution, for 155
Km^3 of ice volume loss 1850
A.D.
6
5701.0
4890.6
2.21*1011
1.00*1021
7
348.0
1093.2
0.00
0.00
3.4.3 Traditional Normal Mode
Approach
Uplift rates are based on a spherical and radially
stratified, viscoelastic Earth model (linear Maxwell
rheology, 7 layers, Table 3.4.2), similar to that used
for post-seismic calculations, but tuned to the viscosity profile used by Burov et al. [1999], inferred from ECORS profile, and with volume averaged
rigidities from PREM [Dziewonski and Anderson,
1981]. In normal mode theory for a viscoelastic
Earth model, the vertical displacement is expanded
in spherical harmonics. Truncation of the expansion results into information loss, determining the
spatial resolution. The Alpine glaciers have characteristic lengths of about 1 Km, thus a harmonic decomposition up to 40,000 degrees should be
exploited. A low resolution calculation can be carried out considering only an equivalent load spread
over the area of the geographical region occupied
by all the glaciers; we can thus limit the harmonic
expansion to about 500 degrees.
The effective mass balance (s_E) is smaller with respect to the real one (s_R) by a factor chosen to
respect the total mass balance, s_E=(A_R/A_E)*s_
R, for a real area A_R of 2215 Km2 and a mean mass
balance s_R of -0.71 m/yr.
Results of this low resolution calculation are shown
in Figure 3.4.1, where the largest elastic uplift rate
is about 0.1 mm/yr and the position of the maximum is roughly located in the centre of the area,
as expected. Besides the uplift rate localized in
the surrounding of the alpine chain, we can also
66
10
take into account the effect of Mediteranean basin
sinking caused by the melting water. In this calculation we used about 1.151 Gt/yr of water feeding the Mediterranean sea, obtaining a small subsidence down to -0.007 mm/yr in the southern
Mediterranean basin. The value of 0.1 mm/yr
obtained from this low resolution calculation, can
be compared with the present-day uplift rate of
2 mm/yr obtained by Ivins and James [2004] for
the Patgonian ice fields and a sub-cratonic mantle,
prediction very similar to that for an elastic model,
not shown in that article. The difference can be
understood by considering that in Patagonia the
mass loss considered by Ivins and James [2004],
38.4 Gt/yr for years 1990-2000 A.D, is 24 times
larges than the Alpine 1.54 Gt/yr.
The low resolution calculation of this section is
appropriate for estimating long wavelength features of the vertical displacement, but the need for
more detailed uplift rates in correspondence with
the largest ice complexes, requires accurate high
resolution calculations.
3.4.4 High Resolution Approach
Differently from the case of global problems, such
as sea level computations where the load function
is extended over the whole globe, loads now occupy a very limited part of the sphere, and thus it
is convenient to integrate straightforwardly over
the load. Furthermore, the glaciers can be easily
treated as a discrete ensemble of point-like sources. A similar use of point-like source distribution
can be found for co- and post-seismic deformation problems as recentely done by Dalla Via et
al. [2005].
Le Meur and Hindmarsh [1999] and Barletta and
Sabadini, [2006] have shown that for a surface
load the elastic part of the solution tends asymptotically to a non-zero value for increasingly high
harmonic degrees. Le Meur and Hindmarsh [1999]
have shown how to overcome the Gibbs phenomenon due to truncation by making use of the sum of
Legendre polynomials series.
The Green function thus can be rewritten in a more
suitable form and so we can compute the approximate solution for one disk and compare with the
highly accurate solution where the sum of the series is carried out at 40,000 harmonic degrees to
resolve 1 Km loads.
Interreg Project ALPS-GPS Quake Net
The comparison shows that for distances from the
load larger than three times the radius of the load
our approximation overlaps the highly accurate solution. This indicate that use of the approximated relation for uplift rates at points far from every glacier is correct, at least for distances larger
than three times the radius of the load. When the
distance of the observation point from a glacier is
smaller than three times the radius of the load, the
contribution of that specific glacier is treated with
the highly accurate decomposition. Small size loads can thus be accurately treated by means of the
above computationally efficient technique, optimizing the traditional normal mode approach.
To show the results over the whole Alpine chain
of Figure 3.4.2, a regular 5 Km spaced grid resolution has been used. Nonetheless, uplift rate can be
evaluated correctly using a restricted set of points
only, as benchmarks (leveling or GPS) in the inverse problem scheme.
3.4.5 Results
Figure 3.4.2 shows the modeled uplift rates for
present-day mass balance of Table 3.4.1 and for
the Earth model of Table 3.4.2, while the inset
shows the measured uplift rate, from leveling and
GPS, limited to Switzerland after Schlatter et al.
[1999]. Uplift rates, in the range 0.1-0.2 mm/
yr, characterize the response of the whole Alpine
belt to present-day glacier reduction, in agreement
with the findings of Figure 3.4.1, encircling patches of high uplift rates, localized over the major
glacial complexes.
The largest uplift rate spot, of 0.9 mm/yr, indicated
by the arrow, occurs in the French Alps, in proximity of Mount Blanc Group, where the largest mass
loss is located. Uplift values of about 0.4 mm/yr,
a factor two lower than the maxima in the west,
are located in Austria. In the inset the (white contours) modeled and measured uplift rate pattern,
closely resemble in shape, indicating that present day glacier shrinkage contributes a substantial fraction of observed uplift rate. The 0.1 mm/yr
white contour in the north almost overprints the
0.4 mm/yr measured, indicating a 20% contribution from present day glacier shrinkage in the periphery of the uplifting region. The largest patches
of modeled uplift rate overlap the regions of largest observed ones, although the highest modeled
uplift rates of about 0.4-0.5 mm/yr, dark blue patches in the main panel of Figure 3.4.2, are about
a factor two lower than measured ones.
Modeling fails to reproduce the northernmost part
of the large measured uplift patch, right in the inset. These findings suggest on one side that glacier shrinkage sensibly contributes to the Alpine
chain uplift, on the other side that there are other
phenomena, such as active tectonics, drainage and
erosion [Schlunegger et al., 2001a, 2001b], affecting both the values of the largest uplift rates and
their geographical distribution.
Figure 3.4.3 portrays the effects of viscous relaxation in the lower crust and in the asthenosphere on
uplift rates, assuming a constant volume loss since 1850 A.D. for a total of 155 Km3, equally distributed over the representative glacier distribution
of the 1973, and rheological parameters in Table
3.4.2. According to Haeberli and Beniston [1998]
the estimated total glacier volume in European
Alps was about 130 Km3 for mid-1970s and since
the end of the Little Ice Age the glacierization has
lost around half of its original volume. This means
130 Km3 since 1850, plus 25 Km3 since 1973, according to Paul et al [2004].
The viscous uplift pattern is smoother than the elastic one. In the western Alps, where also the elastic
part is most important, the largest viscous uplift
rates are concentrated, hitting 0.32 mm/yr. The viscous part can be more than 50% of the elastic
part, but it is generally smaller.
Summing up the largest elastic contribution and
the geographically corresponding viscous one, an
uplift of 0.7-0.8 mm/yr is obtained in the western
Alps, a substantial fraction (half) of the largest
observed one of 1.5 mm/yr. The estimate of the
viscous contribution, on the other hand, can be
affected by uncertainties in the rheological parameters, and must thus be taken with caution.
The dimensions of the patches where uplift is concentrated, of tens of kilometers, is comparable with
the thickness of the crust, of 40 km, as in Table
3.4.2, indicating that contributions to uplift originates from stress relaxation in the soft lower crust
of 2.15* 1019 Pa s, as in Table 3.4.2.
3.4.6 Conclusions
According to our model, rapid glacier shrinkage
gives a substantial contribution to Alpine chain
uplift. The specialized normal mode technique we
introduced here provides an high spatial resolution
scheme, thus becoming a necessary tool within the
forward problem, which will be the kernel of the inverse problem for future work. Our results are the
key for a correct interpretation of uplift data in the
Alpine chain and for quantifying the contributions
from different drivers of uplift, present-day glacier
instability, active tectonics and drainage.
3.5 Instrumental
earthquakes in the
Alpine region: source
parameters from
moment tensor
inversion
Mariangela Guidarelli for DST-UNITS
We consider moment tensor inversion solutions
for a set of earthquakes, with moment magnitude 4.8, that occurred in the Alpine region during
67
Interreg Project ALPS-GPS Quake Net
Fig. 3.5.1 - Map with fault plane
solutions determined for the
events analysed in this study.
the period 1998-2004. The source parameters have
been retrieved using a robust methodology (INPAR
method), that performs a dynamic relocation of
the hypocentre (latitude, longitude and depth) simultaneously with the determination of the focal
mechanism (Šílený et al., 1992; Šílený, 1998).
To help constrain stress conditions and tectonic features in the Italian region it is necessary
to supplement the information provided by global seismology (e.g. CMT-Harvard solutions, USGS)
with regional and local broad-band studies. In
fact the Italian region is seismically very active
but most of the relevant seismicity is concentrated
in the magnitude range 5-6, with a little number
of events with MW>6.0. The study of earthquakes
with MW<6.0 is necessary to obtain information
about the tectonic structures and it is important in
the framework of seismic risk assessment, because
even moderate magnitude events can contribute to
the seismic risk due to the uniqueness of the cultural heritage of Italy.
In this study we analyse a set of earthquakes,
with magnitude between 4.8 and 5.7 that occurred in the Alpine region during the period 19982004. To determine the source parameters, we apply the methodology called INPAR (Šílený et al.,
1992; Šílený, 1998) that performs a full waveform
inversion to obtain the source moment tensor.
The INPAR method had successfully been applied
over a quite broad range of magnitudes (1.5-6.0)
and within a wide variety of tectonic (Campus
et al., 1996; Radulian et al., 1996; Vuan et al.,
2001; Chimera et al., 2003; Guidarelli et al., 2003;
Guidarelli et al. 2006a), volcanic, and geothermal
environments (Campus et al., 1993; Cespuglio et
al., 1996; Guidarelli et al., 2000; Kravanja et al.,
2000; Saraò et al., 2001; Guidarelli et al., 2002;
Guidarelli et al., 2006b) and it has been shown
68
that it can handle earthquakes in a wide range of
magnitudes, from MW=1.5-2.0 for events in volcanic environments (Guidarelli et al., 2002), up to
MW=6.0 in tectonic settings (Vuan et al., 2001;
Guidarelli, 2004; Guidarelli et al., 2006a). A significant feature of the INPAR method is the possibility
of retrieving useful information about the seismic
source even in case of a limited number of records
available (Vuan et al., 2001). Therefore the INPAR
method is particularly useful for regional and local
studies, especially when few stations are available
because of logistic problems or sparse seismometric
networks. It follows that such methodology can be
complementary to global scale methodologies (e.g.
CMT), mainly when it permits a reliable estimation
of focal depth, fixed a priori in CMT inversion, and
can confidentially be extended to the analysis of
smaller events.
INPAR method (Šílený et al., 1992; Šílený, 1998)
uses the point source approximation and consists
of two main steps. The first step is a linear inversion and the six moment tensor rate functions
(MTRF) are retrieved. They are obtained extracting
from the data, with a damped least squares algorithm, the Green functions, in this case with a time
dependence given by a Heaviside function, computed by the modal summation method (e.g. Panza,
1985; Florsch et al., 1991; Panza et al., 2000).
Therefore the procedure does not require the a priori assumption of an initial source model.
The broad-band modelling of local and regional seismic waveforms needs a quite precise location of
the hypocentre: any mislocation affects the confidence of the source mechanism determination. For
this reason, the INPAR method performs, whenever
necessary, dynamic relocation of the hypocentre simultaneously with the determination of the mechanism (Šílený et al., 1992). The base functions are
Interreg Project ALPS-GPS Quake Net
Event number
Date
Time
Lat
Lon
Depth
Mw
1
12.4.1998
10:55:32.50
46.27±0.01
13.55±0.01
10±4
5.7
2
10.5.2000
15:09:00.81
44.33±0.03
11.90±0.16
8±7
4.9
3
21.8.2000
17:14:31.10
44.90±0.01
8.25±0.04
7±1
5.0
4
17.7.2001
15:06:15.24
46.67±0.04
11.10±0.04
14±8
5.0
5
14.2.2002
03:18:01.39
46.38±0.08
13.11±0.12
3±3
4.8
6
14.9.2003
21:42:51.86
44.20±0.17
11.55±0.02
32±1
5.4
7
12.7.2004
13:04:07.16
46.27±0.01
13.64±0.01
7±1
5.3
8
24.11.2004
22:59:40.05
45.70±0.18
10.72±0.07
10±3
5.2
Event number
Nodal planes
DC (%)
CLVD (%)
N° of waveforms
1
216 67 - 20/314 72 - 156
31
69
14
2
106 82 75/350 17 153
36
64
13
3
343 50 - 105/187 42 - 73
68
32
8
4
207 64 - 5/299 85 - 154
74
26
11
5
50 62 74/262 32 118
71
29
15
6
227 56 89/49 34 91
44
56
8
7
101 36 122/243 60 69
89
11
8
8
231 47 64/87 49 115
82
18
6
thus computed for a set of values of the hypocentral coordinates lying between two extremes, defined on the basis of hypocentral estimates. We
indicate the different values of the source coordinates with the variables (X1, X2, X3). In the course
of the inversion intermediate values of the parameters (X1, X2, X3) are computed, incrementing the
initial values with steps chosen a priori. The base
functions corresponding to intermediate values of
the source coordinate are computed with a linear
interpolation of the base functions evaluated at
the grid defined by the assumed set of coordinates. The difference between the observed records
and the synthetic seismograms, corresponding to
a given source coordinate set is computed using
a L2 -norm. The norm can be considered as a function of the parameters (X1, X2, X3) and its minimum
is searched. The second step is non-linear, and the
six MTRF, obtained after the first step of the inversion, are reduced to a constant moment tensor and
the corresponding source time function taking only
the correlated part from each MTRF. This is a basic feature of the INPAR method since, when taking
only the coherent part at different stations, the influence on the solution of non-modelled structural
details and of scattering by non-modelled heterogeneities is reduced (Kravanja et al., 1999). The
problem is non-linear and it is solved iteratively by
imposing constraints such as positivity of the source time function and, when clear readings of first
arrivals are available, consistency with polarities.
A genetic algorithm is used in the search of solutions and in the estimate of the error areas for the
different source parameters (Šílený, 1998).
We report here the results of the inversion for a
set of damaging events recorded within the Alpine
region in the period 1998-2004 with magnitude
Mw 4.8: we include also two events recorded in
Slovenia, near the border with Italy, due to their
relevance for the Italian tectonic setting. We used
waveform data from IRIS consortium, ORFEUS, and
MedNet network recorded at local and regional distance. The structural models used for the inversion are taken from the EUR-ID data set (Du et
al., 1998) updated with recent surface wave tomographic studies (Raykova et al., 2004; Pontevivo
and Panza, 2006; Panza et al., 2006).
We inverted a set of 8 events and the results of the
inversions are reported in Table 3.5.1 and Table
3.5.2, while the fault plane solutions are plotted
in Fig. 3.5.1.
We performed the inversions using a maximum of
15 signals (vertical, NS and EW components). To
retrieve information about the error of the solution we use the posterior probability density function to mark confidence zones of the model parameters (Šílený, 1998). From the size and shape of
the confidence areas we can decide about the reliability level of the solution. The MTRFs retrieved
from the waveform inversion, and then the average
mechanism and source time function, are considered to be affected by three types of errors, generated respectively by:
1 the noise present in the data;
2 the horizontal mislocation of the hypocenter
adopted to compute the base functions in the dep-
Tab. 3.5.1 - Source parameters
for the events analysed in this
study. The values for latitude,
longitude and depth are those
retrieved after the inversion with
dynamic relocation of the hypocentre, as provided by the INPAR
method. Mw is the magnitude
from scalar seismic moment
(Kanamori, 1977).
Tab. 3.5.2 - Nodal planes and
moment tensor component percentages for the studied events
in the Italian Peninsula; DC(%):
percentage of the double couple
component; CLVD(%) percentage of the CLVD component. In
the last column is reported the
number of records used in the
inversion for each event.
69
Interreg Project ALPS-GPS Quake Net
Fig. 3.5.2 - Fault plane solutions
of the studied earthquakes with
their confidence error areas.
70
th grid used in the inversion;
3 the improper structural models used to compute the base functions (Šílený et al., 1996). The
variance is turned into confidence regions of the
eigenvalues and eigenvectors of the moment tensor. From the confidence areas for each mechanism, shown in Fig. 3.5.2, we can see that most of
the fault mechanisms are well resolved.
Table 3.5.2 lists the nodal planes and the percentages of the moment tensor components, Double
Couple (DC) and Compensated Linear Vector Dipole
(CLVD), corresponding to each event. We observe
in several cases a departure from a pure double
couple mechanism and the presence of a significant CLVD component.
The active tectonic region localized at the border
between Italy and Slovenia was affected by two
earthquakes, recorded in 1998 and 2004, with moment magnitude 5.7 and 5.3 (events 1 and 7) respectively. The mechanism for the event of 1998 is
a strike slip and seems to correlate with the NW-SE
trend of the Dinaric structures (Bajc et al., 2001);
the fault plane solution of the 2004 event shows
a dominant thrust component, that matches the
mean focal mechanism for the area (ZS9 seismic
zonation, ZS9 Working Group, 2004) that is characterized by the convergence of the Adriatic and
European plates. Both of the events are shallow,
with a considerable CLVD component in case of
event 1. Quite large confidence regions characterize the mechanism of event 7; nevertheless the dominant thrust component is confirmed.
Two of the analysed earthquakes are located in
north western Italy. Event 3, in the Monferrato
area, is characterized by a normal source mechanism, but large confidence areas indicate uncertainty in the solution. Event 8 occurred near the city
of Brescia and the Garda Lake with a dominantly
thrust mechanism.
Event 6 presents a thrust mechanism while event
2 a thrust mechanism with a small strike slip component. They evidence the compressional environment of the Northern Apenninic chain that can originate thrust faulting earthquakes, in agreement
with the indications contained in ZS9 seismic zonation (ZS9 Working Group, 2004).
The 17 July 2001 event 4 fault plane solution is
a well resolved strike-slip mechanism. For the 14
February 2002 event 5 is characterized by a thrust
mechanism, in agreement with the thrust mechanisms obtained for the 1976-1977 Northern Italy seismic sequence (Aoudia et al., 2000; Pondrelli et al.,
2001) and the average focal mechanism proposed
by the ZS9 seismogenetic zonation (ZS9 Working
Group, 2004).
Interreg Project ALPS-GPS Quake Net
71
Interreg Project ALPS-GPS Quake Net
4 - EARTHQUAKE
HAZARD IN THE ALPS
72
Interreg Project ALPS-GPS Quake Net
4.1 Unified Scaling Law for
Earthquakes in the Alps:
a multiscale application
Anastasia Nekrasova for DST-UNITS
4.1.1 Introduction
The evident heterogeneity of seismicity distribution and dynamics (Mandelbrot, 1982; Turcotte,
1997, 1999) are apparently scalable according to
the generalized Gutenberg-Richter recurrence law
that accounts for the fractal nature of faulting
(Kossobokov & Mazhkenov, 1988). The results of the
global and regional analyses (Keilis-Borok et al.,
1989; Kossobokov & Mazhkenov, 1994; Nekrasova &
Kossobokov, 2002, 2003; Kossobokov & Nekrasova,
2003, 2004, 2005) imply that the recurrence of
earthquakes in a seismically prone site, for a wide
range of magnitudes M and sizes L, can be described by the following recurrence law
log10 N (M, L) = A + B · (5 – M) + C · log10 L
where N(M, L) is the expected annual number
of earthquakes at a seismically active site of linear dimension L. Following Ba et al.(2002) and
Christiansen et al. (2002) who gave an alternative
formulation using the inter-event time distribution
we denote the relationship as the Unified Scaling
Law for Earthquakes or USLE.
The characterization of the geometrical nature of
the earthquake epicenters distribution is not an
obvious question. When epicenters are uniformly
distributed over the surface, their number will be
proportional to the area, whereas when they are
distributed along a narrow strip, e.g., along a liner fault zone, the number will be proportional to
the length of a segment. There are other possibilities that cannot be excluded a priori. Seismic activity is by no means a uniform process. Therefore,
the question of spatial and temporal scaling arises
with the necessity in seismic hazard and risk assessment, as well as in the studies of earthquake sequences prior to the largest earthquakes aimed at
earthquake prediction, which require a transfer of
estimates and/or criteria from one area to another.
When the data available permits the evaluation of
the USLE coefficients, a reliable answer to the question posed above could be given.
Here we apply the methodology to a three different length scales defined as S1, S2 and S3. S1 covers the whole Central Mediterranean area including Alps, Dinarides, and Apennines. S2 covers the
Alps and surroundings. S3 covers the Friuli-Venezia
Giulia region and Western Slovenia. We analyze
two earthquake catalogues. The first one is the
UCI2001 (Peresan et al., 2005) covering S1 and
S2 scales. The second one, the Friuli-Venezia Giulia
Seismometric Network bulletins (Dipartimento
Centro di Ricerche Sismologiche) covering S3 scale. These catalogues are enough homogeneous and
do not require any kind of stability tests like in
Molchan et al., (1997). We provide maps of the
A, B, C coefficients, for the different scales and
discuss their usage in seismic hazard assessment.
We analyze the seismicity flow before major events in the region. Furthermore, we provide seismic
risk estimates for selected cities in the Alps and
surroundings.
4.1.2 Motivations of the method
(Kossobokov & Mazhkenov, 1988)
The Gutenberg-Richter recurrence law (Gutenberg &
Richter, 1954; 1956), the most reliable and generally accepted law of similarity in seismology, establishes the relation for a given space-time volume
between the annual number of earthquakes, N, and
magnitude, M,
log10 N(M) = a-b(M-5), M
_ M
M*
(1),
where M
_ and M* are lower and higher magnitude cutoffs. The coefficient a characterizes the expected level of seismic activity in the area, and b
reflects changes in the number of earthquakes in
successive magnitude ranges.
Evidently (1) does not provide any information on
the size of the region considered. Let us assume
that a sequence of earthquakes is self-similar in
space, that is, there are no principal differences in
the geometry of a set of epicenters, when considered at different scales. Let N(M, L) be the expected
annual number of earthquakes in a seismic area
of linear dimension L (note that spatial averaging
is performed over the areas supporting earthquake
epicenters; non-seismic areas, those where no earthquakes have been recorded in the period of investigation, are excluded). In the case of similarity, the correspondence between N(M, L) and N(M)
from the Gutenberg-Richter law can be presented
as N(M, L) = N(M)(L/l)c where N(M) = N(M, l) and l
is the characteristic length of the region. Thus, the
recurrence law (1) can be rewritten as
log10 N(M, L) = A + B (5 – M) + C log10 L (2).
The objective of the study is to estimate the parameters of this law over the Alps, assuming site
specific homogeneity and similarity for the set
supporting earthquake epicenters. The coefficients
A and B are similar to a and b from equation (1).
According to the concept of hierarchy and self-similarity in Earth dynamics, the coefficient C is the
fractal dimension (Mandelbrot, 1982) of the set of
epicenters which shows how the number of earthquakes, N, is changing with linear dimension, L,
of an area.
Generally speaking, the fractal dimension is a notion that can be defined only locally (unless, a special case of mono-fractal geometry is considered),
that is, at a given point of a set. It is the same for
all points in case of a homogeneous and self-similar set. It is rather difficult to estimate the fractal
73
Interreg Project ALPS-GPS Quake Net
dimension of a set using a sample of points, because, first of all, the fractal dimension of any finite
(even very large, as well as of any infinite discrete) set equals 0, and, second, the estimate can be
biased by the inadequate choice of areas used for
the computation.
Usually, in the case where the fractal dimension of
a set of epicenters is estimated from earthquake
catalogs, statistics are only representative for
areas with a linear size of about several tens of kilometers. A set of epicenters is approximated by a
finite number of events from a catalog where the
coordinates are subject to some error. Furthermore,
it is obvious that the spatial distribution of earthquakes is not uniform, and it is clear that this distribution is regionally inhomogeneous.
4.1.3 Algorithm
Notwithstanding the abovementioned difficulties, here we accept the hypothesis that the seismic process is self-similar, at least locally, and
estimate the coefficients of equation (2) using
the following algorithm for accounting Scaling
Coefficients Estimation, hence its name SCE:
A catalogue of earthquakes is used as initial data
source. A space-time-magnitude volume, S ×T × M,
is considered. Here S is the territory, T is time interval from T0 to T1, and M is the magnitude range
M M0 ; all events with M M0 are reported since
time T0. The data are processed as follows:
1 The magnitude range M is subdivided into
m adjacent intervals of length DM without
overlapping
M0 + (j - 1) DM
Mj <M0 + j DM, j = l,2,...,m.
2 The entire area S is subdivided into a hierarchy
of h levels. The 0-level corresponds to the entire S
included in a square with side L0. (To specify the
boundaries, a square here is a set {(x, y) : a x <
a + L0, b y < b+L0 }). In the two successive levels
(i and i+1) of hierarchy Li+1 = 0.5Li, for i = 0, l,...,
h-1. A square at the level h of this hierarchy will
be denoted by wh.
3 Using the earthquake catalog, for each one out
of the m magnitude ranges and for each one out
of the h levels of hierarchy, the following number
Nji is computed
Nji = [ ( nj (Qi))2 ]/ Nj
(3),
where summation extends over all areas {Q i } at
the i-th level of hierarchy; nj (Q i) is the number of
events from a magnitude range Mj in an area Q i of
linear size Li ; Nj is the total number of events from
a magnitude range Mj .
The number Nj i can be considered as a mean of the
number of events in the magnitude range Mj in an
area at the i-th level, where the mean is calculated
over the Mj epicenter set. To show this, let us call
a “telescope” a set of h+1 embedded squares, W =
74
{wo, wl,..., wh}, so that each wi belongs to the i-th
level of hierarchy. Note that for the hierarchy defined above, each “telescope” corresponds unambiguously to a single square cell wh from the lowest
level. Let us assume that the Mj epicenters’ set is
defined by a sample of events from the catalog, Xj
= {x1 xNj}. Each xk defines a “telescope” W(xk) such
that wh (xk) contains the epicenter xk. Let {W(xk)}
be a set of “telescopes” defined by Xj. and nj (wi)
be the number of events from Xj that fall within wi.
Then, the mean number of events in an area of the
i-th level of hierarchy over the sample Xj is
Nji =
{k=1,…,Nj}
nj (wi (xk)) / Nj .
Substituting the summation over the events with
the summation over the areas wi (xk) for the i-th level, we obtain equation (3).
Thus, a set of numbers nj (wi (xk)), i = 0, 1,..., h
provides data for an estimate of fractal dimension
for an epicenter locus set at the point xk, and the
numbers Nji are averages of these data.
4 Estimates of A, B, and C in (2) are derived from
the set of linear algebraic equations log10 Nji = AB(Mj - 5) + Clog Li by the least squares method.
Unlike many other recent applications (e.g., Bak
et al., 2002) the method makes heuristic adjustments for heterogeneity of seismic distribution, as
well as for consistency of the real data statistics in
different magnitude ranges: specifically, the equations that correspond to the evidently incomplete
samples are excluded from computations. For this
purpose a heuristic limitation requiring log10 (Nj,i /
Nj+1,i) > const on transfer from the magnitude range Mj to Mj+1 (where const is a free parameter of the
SCE algorithm, usually set to 2) is used. Similar limitation - log10 (Nj,i / Nj+1,i) > const - is introduced
for the transfer from (i-1)-th to i-th level of spatial hierarchy.
5 In addition to the original prototype algorithm (Kossobokov & Mazhkenov, 1988), at each seismically active location, the steps 1-4 are applied
many (usually 100) times with randomized box
counting settings (Nekrasova & Kossobokov, 2002).
The resulting series of estimates are used to determine the final A, B, and C averages and their standard errors sA, sB, and sC.
Nji are normalized in time and space: they are computed for unit of time of 1 year and unit of length
of 1 degree of the Earth meridian.
The estimate of C, as suggested here, is very close to the definition of the correlation dimension
D2 (Atmanspacher & Scheingraber, 1988) although our reasoning aims originally at estimation of
the Hausdorff capacity dimension D0 (Mandelbrot,
1982).
4.1.4 Implications for seismic
hazard and seismic risk estimates
Any kind of risk estimates results from a convolution of the hazard with the exposed object under
Interreg Project ALPS-GPS Quake Net
consideration along with its vulnerability –
R(g)=H(g) · O(g) · V(O(g)),
where H(g) is natural hazard at point g, O(g) is the
exposure of objects of risk at point g, and V(O) is
the vulnerability of objects of risk. Note that distribution of risks, as well as objects of concern
and their vulnerability could be time-dependent.
In the case of seismic phenomenon the key role
in the risk assessment is related to the choice of
a probability model describing the occurrence of
earthquakes in a specified space-magnitude-time
volume V = {g, M, t}. A rough description of the
leading features for long-term seismic activity is
usually provided by assuming the flow of events
(g, M, t) V to be a stationary point Poisson process with annual rate of N(M), which according to
the well-established Gutenberg-Richter law (1) is
parameterized in a log-linear form (Molchan et.
al.,1996). Seismic reality evidences many contradictions to this assumption that have led to complications of the existing hypotheses by introducing declustered sequences of main events and
their associates (fore- and after-shocks) superimposed with hypothetical distributions of the associate size, time and location. In any case, the estimation of N(M) at a given site of interest remains
the basic source of seismic risk assessment, as well
as the basic source of inadequate seismic engineering decisions.
Following the Unified Scaling Law for Earthquakes,
USLE, that generalizes Gutenberg-Richter recurrence relation, one can demonstrate that the traditional estimations of seismic risk for cities and urban agglomerations are usually underestimated.
In fact, any estimate of seismic hazard rate (e.g.,
N(M)) depends on the size of territory that is used
for averaging and, therefore, may differ dramatically when scaled down in the proportion to the
area of interest.
Let us consider a city located at g which area equals
S and supports seismic locations. In this case, in
accordance with USLE, the rate N(M) equals:
N(M)=10A ×10B×(5-M) SC/2
(3)
If the USLE coefficients are estimated from scaling
that starts with seismic region of linear dimension
L0 >> S0.5, then underestimation of N(M) by traditional proportion S/L02 will account to the factor of
(L02 /S)C/2, implying “surprises” of inadequate scaling from large to small size area.
The following two estimates of seismic risk are
among the most simplest and natural. Of course,
they do not use complicated procedures that might be more realistic and adequate convolutions
of hazard, objects and their vulnerability and are
used here to illustrate the general approach. The
first one, rc, estimates for a given city the rate of
hits by earthquakes of magnitude M0 and, therefore, equals to N(M0) at the city location g. In this
case, the city is the object of concern and the convolutions are trivial unity factors. The second one,
rp, considers the city inhabitants as objects of concern. In a city, where citizens are equal in vulnerability, with the population of Pc the suggested risk
estimates the annual number of city inhabitants
affected by earthquakes of magnitude M0, which
makes rp = N(M0) × Pc.
We shall use both rc and rp estimates for selected
cities in the Alps. Evidently, more appropriate estimates of seismic risks of different nature would require involvement of the specialists in social sciences and economics.
4.1.5 Data
We have performed the analysis of basic parameters of seismic activity in the Alps at the three levels of geographical scale (Figure 4.1.1) – (i) a
scale, S1, covering: Alps, Apennines and Central
Mediterranean (ii) a scale, S2, covering Alps and
surroundings, and (iii) a regional scale, S3, covering the Friuli-Venezia-Giulia and western Slovenia
regions. Specifically, we were trying to get an insight into the hierarchical scaling of sizes and earthquake magnitudes on transition from 4° - 1/4°
to 1° - 1/16° and magnitude range from 3 down
to the to-date level of the regional earthquake catalog completeness of 2.2. As concerning the time
scale we have investigated the entire periods of
the complete data coverage, i.e., 136, 32, and 29
years as well as sliding time windows down to 6
years.
Note that all catalogues considered are not homogeneous either in space or in time. Therefore the
reliability of the results we present below varies
in space and time, which fact complicates straightforward interpretations and conclusions. We try
avoiding this difficulty by setting formal a priory limitations on an acceptable estimation of the USLE
parameters as well as by additional joint analysis
of these parameters computed at locations from
extended uniform seismogenic zones.
The ultimate purpose of our study is related to
seismic risk estimates for large cities and urban
agglomerations. Therefore, we (i) apply the obtained local estimates of seismic hazard for assessments of seismic risk at selected cities in Italy
and surrounding countries and (ii) demonstrate the importance of specificity of geometrical
structure and scaling properties of seismic activity distribution in this case. Such an application requires specifying data on urban population
and city areas.
Joint data for the S1 scale (Alps + Apennines +
Dinarides + Central Mediterranean), 1870-2005
We consider the territory of Italy and adjacent area
as a whole within the boundaries from 36°N to
50°N and from 2°E to 20°E (Figure 4.1.1). To characterize the USLE parameters we apply the method
described in section 3 to the Italian seismic activity data from 1870 to the end of 2005 as repor-
75
Interreg Project ALPS-GPS Quake Net
Fig. 4.1.1 - Map of epicenters
of earthquakes with M=3 or
larger in 1870-2005. Light curve
outlines the region of S2 scale,
while the heavy one encloses
the S3 scale region.
Fig. 4.1.2 - Map of epicenters
of earthquakes with M=2.6 or
larger in 1974-2005 in the Alps
and surroundings.
76
Interreg Project ALPS-GPS Quake Net
Fig. 4.1.3 - Annual number
of earthquakes of different
magnitudes in the DCRS/OGS
catalogue, 1900-2005.
Fig. 4.1.4 - Map of epicenters
of earthquakes with M=2.2 or
larger reported by DCRS/OGS in
1977-2005 in Friuli-Venezia Giulia and western Slovenia regions.
ted in the Updated Catalogue of Italy (UCI2001).
This is the same UCI2001 catalogue as in the regional studies aimed at real-time prediction of strong
earthquakes (Peresan et al., 2005) restricted to
the end of 2005. The UCI2001 catalog is a compilation composed from the CCI1996 (Peresan et
al., 1997) for the period 1900-1985 and updated
form the USGS/NEIC Preliminary Determinations
of Epicenters (PDE data) since 1986. According to
Peresan & Panza (2002), since 1950 the composite catalogue provides a rather complete coverage
of the territory under study. The magnitudes of reported events in this period are usually above magnitude 3 determined with the accurately to the
first decimal digit. In addition to it we have used
the historical part of the Current Catalogue of Italy
(CCI1996), which compliments information on earthquakes of practically the same completeness
from 1870 to 1900. This part of the CCI1996 data
is based on “Catalogo dei Forti Terremoti in Italia
dal 461 a. C. al 1980” (Boschi et al., 1995). As a
result, we have obtained the join catalogue that
covers the entire period from 1870 to the end of
2005. We have used the whole time span of the catalogue except for the data affected by the Second
World War.
The epicenters of earthquakes with magnitude 3
or larger for the entire period are plotted in Figure
1. The joint catalog data for magnitudes 3.0 or
larger permit calculation of the USLE parameters
that uses 5 bisecting steps of the spatial hierarchy
from the linear size L0 =4° down to L4 =1/4°. The
territory of the catalogue completeness differs in
time due to the evident re-arrangements of the seismographic networks and will be specified in the
analysis of the results.
77
Interreg Project ALPS-GPS Quake Net
According to our observation that confirms the results of a previous study (Romashkova, 2006), the
completeness of the OGS data set depends strongly on the territory and appears the best in the
Tolmezzo-Gemona-Bovec region within much wider
limits from 45.5°N to 46.7°N and from 11.7°E to
14°E (the rectangle outlined with a heavy line in
Figure 4.1.1). Figure 4.1.3 displays annual number of earthquakes of different magnitude ranges
and demonstrates that from the mid-nineties the
OGS catalog reports presumably most of the earthquakes about magnitude 2 or even smaller. Figure
4.1.4 shows the map of epicenters of magnitude
2.2 or larger earthquakes from the middle of 1977
to the end of 2005 in Friuli-Venezia-Giulia region.
We accepted magnitude 2.2 cutoff as the level of
the catalogue completeness since 1977. For the
fine scale analysis the OGS catalogue permits calculation of the USLE parameters making use of the
5 bisecting steps of the spatial hierarchy from the
linear size L0 =1° down to L4 =1/16°.
Variability of the USLE parameters in the
eight major seismogenic zones
Fig. 4.1.5 - The eight major
seismogenic zones combined by
Molchan et. al. (1996) from the
79 smaller ones (GNDT, 1994).
Yellow dots mark the grid points
inside the regions.
78
Scale S2 (Alps and surroundings), 1974-2005
Figure 4.1.2 displays the epicenters of magnitude
2.6 or larger earthquakes reported in the UCI2001
data set from 1974 through 2005 within the boundaries from 43.5°N to 48.5°N and from 5°E to 16°E
(outlined rectangle in Figure 4.1.1). According to
Peresan et al. (2005) this set of earthquakes appear to be reasonably complete. The completeness
of data permits us to use 5 bisecting steps of the
spatial hierarchy from the linear size L0 =2° down
to L4 =1/8° for characterizing the USLE parameters in Northern Italy and surrounding territories
in another ranges of scales than those specified in
section 5.1.
As it was mentioned above the catalogues are not
homogeneous either in space or in time. Therefore
the reliability of the results we present below varies
in space and time, which fact complicates straight forward interpretations and conclusions. We try
avoiding this difficulty by additional joint analysis of the basic parameters of seismic activity computed at many locations from extended seismogenic zones, for example, the same large unions
of seismogenic zones (GNDT, 1992), as defined in
(Molchan et. al., 1996): Eastern Alps, Western Alps,
Northern, Central and Southern Apennines, Calabria,
Sicilia and Etna Volcano zones (Figure 4.1.5).
Seismic risk estimates for selected cities
The seismic risk for the largest cities of the territories under investigation was calculated. The
Italian cities population and areas data were retrieved from the web site http://www.comuniditalia.biz/dati/ (Comuni d’Italia). For other cities we
used the data from the Internet free encyclopedia
(Wikipedia).
Scale S3 (Friuli-Venezia Giulia region and
Western Slovenia), 1994-2005
4.1.6 The results of analysis cities
The earthquake database of the National Institute
of Oceanography and Experimental Geophysics,
Centre of Seismological Research, consists of
network bulletins, 1977-1999 (published on CDROM), and preliminary bulletins, 2000-2001 (both
available on web-site http://www.crs.inogs.it/).
To characterize the USLE parameters for the FriuliVenezia-Giulia region we have used the corrected
version of the OGS catalogue kindly provided by
Dipartimento Centro di Ricerche Sismologiche,
Istituto Nazionale di Oceanografia e di Geofisica
Sperimentale updated to the end of 2005. The magnitudes in the revised catalogue have been estimated all with the same most recent relation though the entire time span and updated through 2005.
Long-term (1870-2005) estimates of A, B,
and C for S1 scale
In this section we present the results of our investigation of the USLE parameters at the three scales of spatial accuracy, from large to small. We also
try to search for patterns of temporal variability, as
well as to understand the importance and limitations of brining the data from lower magnitude levels for obtaining finer resolution.
We start from the analysis the entire territory of
Italy and surroundings and the largest available
time span of 136 years (from 1870 through 2005).
Each plate of the three ones in Figure 4.1.6 consists of the map (on the left), the probability density distribution functions of the USLE coefficient
Interreg Project ALPS-GPS Quake Net
Fig. 4.1.6 - The USLE coefficients of S1 scale,
1870-2005: (A) Logarithm of the annual
number of magnitude 5 earthquakes in
1°×1°, A; (B) magnitude balance, B; (C)
fractal dimension of the epicenter locus, C.
79
Interreg Project ALPS-GPS Quake Net
Fig. 4.1.7 - 2-D projections of
the 858 combinations of A,
B, C coefficients for S1 scale,
1870-2005.
Fig. 4.1.8 - 2-D projections (a) of
the 147 combinations of A, B, C
coefficients in the eight seismogenic zones, 1870-2005. Larger
symbols on 2-D projections mark
the center of gravity of the grid
points from a single seismogenic
zone.
(on the right, top, vertical abscissa) and its standard error (on the right, bottom, horizontal abscissa) and corresponds to the one of A, B, and C coefficients, respectively.
The reliable estimations of the USLE parameters
were possible in 2352 grid points of the 1/4° mesh
80
covering the entire region considered. The standard
errors of the coefficients do not exceed 0.08 and
confirm claimed accuracy of the values plotted on
the maps.
The density bulk of the logarithm of seismic activity (Figure4.1.6, A) normalized to recurrence of a
Interreg Project ALPS-GPS Quake Net
Fig. 4.1.9 - Intermediate-term variations of the USLE coefficients
for S1 scale, 1870-2005.
moderate magnitude 5.0 earthquake, at a unit area
of 1°×1° in a unit time of one year is distributed
from –0.7 to –1.8, which values correspond to the
recurrence rates from one event in five years to one
event in 60 years. The highest values are being observed in Central Apennines, Calabria and Sicily.
The recurrence of moderate earthquakes is much
lower in the Western Alps.
As can be judged from Figure 4.1.6, B, the density
bulk of the coefficient of magnitude balance B concentrates mainly around 0.8. Higher values of B are
observed in northern Sicily and Western Alps where
they reach 1.2 or more.
One can find from Figure 4.1.6, C that the estimate of fractal dimension of the earthquake epicenters’ locus C has a smooth density distribution
widespread from 0.9 to 1.5. The highest values of
C are located mainly in the Swiss Alps and Northern
Apennines, while the lowest ones are found in
Sicily.
Figure 4.1.7 displays the observed combinations of
the USLE parameters in the three projections on
A - B, A - C, and C - B planes and in 3D, respectively. The USLE parameters in Italy and adjacent
areas do not display any evident correlation although their distribution is far from being random.
Let us follow the ABC values when their location
points move along the eight seismogenic zones
from Eastern Alps down to Sicily (Figure 4.1.8).
When the points move from more seismically active
Eastern though less active Western Alps they enter
and remain in the highly fractured regions (indi-
81
Interreg Project ALPS-GPS Quake Net
Tab. 4.1.1 - Strong magnitude 6
or larger earthquakes inside or
next to the three regions.
Fig. 4.1.10 - Intermediate-term
variations of the USLE coefficients in the three seismogenic
zones, 1870-2005: (a) Eastern
Alps; (b) Western Alps; (c)
Northern Apennines; (d) map of
epicenters of strong, magnitude
6 or larger, earthquakes.
Region
Magnitude
1873-06-29
Location
Depth, km
°N
°E
6.3
46.15
12.38
25
1891-06-07
6.0
45.55
11.17
40
1936-10-18
6.2
46.05
12.42
18
1976-05-06
6.5
46.23
13.13
12
1976-06-17
6.1
46.08
12.93
14
1976-06-17
6.1
46.45
13.50
35
1976-09-15
6.0
46.25
13.13
12
1976-09-15
6.0
46.30
13.18
2
1998-04-12
6.0
46.24
13.65
10
1961-01-17
6.3
46.50
7.40
0
3. Northern Apennines 1920-09-07
6.7
44.20
10.25
18
1. Eastern Alps
2. Western Alps
a
82
Date
Interreg Project ALPS-GPS Quake Net
b
cated by high values of C about 1.4) while making
a loop to join via Northern Apennines the domain
of the most active areas of Central Apennines. This
move is better observed on A - C plane. The balance of magnitudes B on this move is mostly wondering around 0.77 except an excursion to higher values about 0.92 on the Western Alps loop (see C
– B plane). The points enter the domain of moderate activity when passing Southern Apennines, and
then migrate to the linear fracture zone of Calabria
and Sicily. At the same time the values of magnitude balance demonstrate a transfer from the domain of stable values within 0.72-0.82 to an unstable one where the values of B alternate from
0.65 in Calabria and from 0.78 in Sicily to just below 1. As could be judged from this example the
values of the USLE parameters are hardly related to
an arbitrary seismotectonic zonation and, therefore, contribute yet unexplored information to the
description of seismic activity at regional and local scales.
Intermediate-term variability of the A,
B, and C estimates in the three seismogenic zones of S1 and magnitude 6 or larger
earthquakes
The joint data from CCI1996 and UCI2001 catalogues permits to investigate the temporal variability of the USLE parameters at the intermediateterm scale from 1870. For reliable estimations the
SCE algorithm requires substantial number of earthquakes, for example, in the long-term estimates this number is larger than 128 in all cases and
more than 256 in 95% cases. Of course, it is not
possible to investigate temporal variability of the
83
Interreg Project ALPS-GPS Quake Net
c
USLE parameters with the same level of reliability. To regularize the temporal analysis and stabilize the results of the A, B, C estimates we accepted
the level of 32 events, which seems a reasonable
retreat from the standard, and considered the averaging, applied to the points from a single seismogenic zone, which presumably improves the accuracy. The accepted limitations permitted us to plot
the sexennial maps of USLE coefficients, which, according to graph at the bottom of Figure 4.1.9,
seems to be about the lowest admissible limit. In
fact, the numbers of the grid points that allow application of the SCE algorithm in the time span of
the analysis is usually less than one half of the grid
points of the long-term estimation (i.e., 2352).
This is the reason for an additional averaging in
a systematic study of intermediate-term variability of the USLE parameters. Since 1970 the numbers
84
are above one thousand possibly due to the better
seismic survey of the territory in the recent years.
Same as in Figure 4.1.9, each of the plates in Figure
4.1.10 displays the trailing sexennial estimates of
the A, B, C medians (top three graphs) along with
the trailing number of points that contribute to the
average in the region considered (bottom). The medians are embraced by 75 and 25 percentiles, which
span characterizes the variability of the coefficient
estimate in the region.
The red vertical lines in Figure 4.1.10 mark the times of strong magnitude 6 or larger earthquakes
that happen in or next to the region and, therefore, might be associated with the observed changes of the USLE coefficients. Their magnitudes and,
if possible, names are provided on the top of the
plate. The green dotted lines indicate the times
of the smaller moderate earthquakes (magnitudes
Interreg Project ALPS-GPS Quake Net
d
between 5.5 and 5.9). The parameters of the earthquakes from each of the eight regions are listed
in Table 4.1.1.
Figure 4.1.10a characterizes temporal changes of
the USLE parameters in the Eastern Alps seismogenic zone. The region trajectory wanders most of
the time within the following ranges: A between
–1.0 and -0.5, B between 0.6 and 0.8, C between
0.7 and 1.2. In the last four decades the sexennial averages of A raised above -0.5 in 1972-1974,
i.e., two years in advance of the 1976 Friuli earthquakes, and right after it, then in 1996-1997,
i.e., a year in advance the 1998 Bovec earthquake
and right after it. At the same times except for right after the Friuli events, B was in its low values
about 0.6 or less, while C remained about 1.2. The
occurrence of the 1976 Friuli earthquake resulted
in a dramatic drop of C from its maximum about 1.3
down to about 0.6. In the years between the two
strong earthquakes A and B graphs were anti-correlated demonstrating a substantial steady growth
of activity from -1.5 to -0.5 and decline of the magnitude balance from 1 to 0.6 in the fifteen years
before the 1998 Bovec earthquake.
The temporal changes of the USLE coefficients in the Western Alps seismogenic zone (Figure
4.1.10b) do not compare with the occurrence of
strong earthquakes here due to transient abundance of data. On the other hand, the behavior of the
USLE coefficients in the last four decades are apparently similar to the one between the two strong
earthquakes in the Eastern Alps seismogenic zone,
i.e., anti-correlated changes of A and B accompanied with a small variation of C.
In the Northern Apennines (Figure 4.1.10c) A rose
above -0.3 in 1917-1918, i.e., two years in advance of the 1920 magnitude 6.7 strong earthquake.
At the same time B was in its low values of 0.5-
0.6, while the sexennial C wandered just above
0.9. Similar to the two regions described earlier,
in the absence of strong earthquakes we observe
once again anti-correlated changes of A and B accompanied with a small variation of C. The rise of
A from -0.5 to about 0 and the drop of C from 1.2
to under 0.9 in 1998 is hardly inflicted from a magnitude 5.9 earthquake in 1996, but might be related to extended activity associated with the 1997
Umbria, magnitude 6.4 earthquake in the adjacent
Central Apennines. It should be noted that the dramatic difference of the sexennial average (C = 1)
and the long-term one (C = 1.44; see Figure 4.1.8
above) is not a contradiction, but indicates partial
involvement of the complex multitude of fractures
of different ranks and orientations into seismic activity at the intermediate-term scale, which eventually combines into overall involvement at the
long-term scales.
The long-term (1974-2005) estimates of
A, B, and C coefficients for the Alps and
surroundings
We continue our analysis of the USLE in lower scales. Specifically, we consider about one third of the
S1, seismic events from lower magnitude range,
smaller time span of 32 years and, finally, smaller linear size of the spatial hierarchy from L0 =2°
down to L4 =1/8°. Similarly to Figure 4.1.6, Figure
4.1.11 displays the maps and distributions of the
USLE coefficients for the Alps and surroundings
being computed by the same SCE algorithm though
applied in the lower ranges of scales in 2533 grid
points of the 1/8° mesh covering the region. The
standard errors of the coefficients shown on the right bottom of each plate confirm high accuracy of
the values plotted on the maps.
Figure 4.1.11, A evidences the four modes of seismic activity distribution characterized with the
85
Interreg Project ALPS-GPS Quake Net
Fig. 4.1.11 - The
USLE coefficients in Alps and
surroundings,
1974-2005: (A)
Logarithm of the
annual number
of magnitude
5 earthquakes
in 1°×1°, A;
(B) magnitude
balance, B; (C)
fractal dimension
of the epicenter
locus, C.
a
b
c
86
Interreg Project ALPS-GPS Quake Net
Region
1. Eastern Alps
Data
Magnitude
1976-05-06
Locus
Depth
°N
°E
6.5
46.23
13.13
12
1976-06-17
6.1
46.08
12.93
14
1976-06-17
6.1
46.45
13.50
35
1976-09-11
5.8
46.28
13.20
24
1976-09-15
6.0
46.30
13.18
2
1976-09-15
6.0
46.25
13.13
12
1998-04-12
6.0
46.24
13.65
10
2004-07-12
5.7
46.30
13.64
7
2004-11-24
5.5
45.63
10.56
17
1996-10-15
5.8
44.79
10.78
10
2003-09-14
5.6
44.33
11.45
10
2. Western Alps
3. Northern Apennines
Tab. 4.1.2 - Moderate and strong
earthquakes of magnitude 5.5 or
larger in the three regions of S2.
Fig. 4.1.12 - Intermediate-term
variations of the USLE coefficients in the three seismogenic zones, 1974-2005: (a) Eastern Alps;
(b) Western Alps; (c) Northern
Apennines; (d) map of epicenters
of moderate, magnitude 5.5 or
larger, earthquakes.
a
87
Interreg Project ALPS-GPS Quake Net
b
d
88
Interreg Project ALPS-GPS Quake Net
c
values of A about -0.2, -0.7, -1.1, and -1.8. These
levels of activity are observed in Central Apennines
at the southern limits of the region, at the northeastern borders of Italy and Northern Apennines,
in between these regions and the north-western
borders of Italy, and at the north-western borders
of Italy, respectively. The density distribution of
the B values has also the four modes about 1.3,
1.1, 0.95, and 0.8, however, these modes are related to different locations except for the areas
of the highest B values that coincides with areas
of the lowest estimates of seismic activity situated in the north-eastern Switzerland and at the
French-Italian border. The density distribution of C
has one skewed mode with the maximum 4% above 1.4 and the “long tail” about 0.5% ranging from
1.1 to 0.8 and below. This tail is evidently inflicted
by the 1976 Friuli sequence of strong earthquakes,
which is the dominant feature of seismic activity
in the entire 32-year period considered and could
also explain the location of the high values of A at
the north-eastern borders of Italy. The highest values of C are observed in the most fractured mountain territories of the region, i.e., in Northern
Apennines and Swiss Alps.
Intermediate-term variability of the A, B,
and C estimates in the three seismogenic
zones of S2 and magnitude 5.5 or larger
earthquakes
The level completeness of the UCI2001 data from
1974 through 2005 is substantially better than
for the other times covered by this catalogue.
Following the same rules and limitations that we
accepted in section 6.2 it is possible to apply the
SCE algorithm systematically in the three seismogenic regions of the Alps and surroundings. Similar
to Figure 4.1.10, the three plates of Figure 4.1.12
display temporal variability of the A, B, C values in
seismogenic zones of Eastern and Western Alps and
Northern Apennines along with the dates of the
89
Interreg Project ALPS-GPS Quake Net
Fig. 4.1.13 - The USLE coefficients in Friuli-Venezia Giulia
and western Slovenia regoin,
1977-2005: (A) Logarithm of the
annual number of magnitude
5 earthquakes in 1°×1°, A; (B)
magnitude balance, B; (C) fractal
dimension of the epicenter
locus, C.
a
b
c
90
Interreg Project ALPS-GPS Quake Net
City
Country
Long-term estimates over 1870-2005
Area Population
Milano
182
Torino
rp
A
B
C
1,256,,211
-1.08
0.76
1.45
0.00388
0.00068
4,868
856
130
865,263
-1.29
0.84
1.37
0.00225
0.00032
1,948
280
Genova
159
601,388
-1.07
0.77
1.45
0.00365
0.00062
2,194
371
Bologna
244
373,539
-0.82
0.78
1.41
0.00949
0.00159
3,545
595
Firenze
102
356,118
-0.76
0.78
1.40
0.00601
0.00099
2,139
352
Venezia
413
271,073
-0.89
0.76
1.35
0.01300
0.00227
3,525
615
Verona
207
253,208
-0.92
0.73
1.42
0.00656
0.00122
1,660
309
Trieste
84
211,184
-0.97
0.81
1.24
0.00486
0.00076
1,026
160
Padova
93
204,870
-0.89
0.76
1.37
0.00457
0.00080
937
164
Italy
Name
rc
moderate strong moderate strong
Friuli
Venezia
Giulia
region
Udine
57
95,030
-0.98
0.81
1.23
0.00387
0.00060
367
57
Pordenone
38
49,122
-0.97
0.80
1.28
0.00267
0.00043
131
21
Gorizia
41
35,667
-0.98
0.81
1.23
0.00316
0.00049
113
18
Slovenia
Lubiana
275
265,881
-1.01
0.82
1.22
0.00969
0.00148
2,576
394
Switzerland
Zurich
92
366,145
-1.51
0.73
1.37
0.00107
0.00020
393
74
B-value
Region
SCE algorithm
Long-term estimates 1870-2005
Molchan, Kronrod, 2005
Eastern Alps
0.77
0.70
Western Alps
0.84
1.00
Northern Apennines
0.76
0.75
Central Apennines
0.79
0.75
Southern Apennines
0.77
0.70
Calabria
0.78
0.65
Sicily
0.87
0.60
Etna
0.90
0.80
moderate and strong earthquakes of magnitude 5.5
or larger. The parameters of the earthquakes from
each of the three regions are listed in Table 4.1.2.
Figure 4.1.12 characterizes temporal changes of the
USLE coefficients in Eastern Alps seismogenic zone.
Although A and B are apparently anti-correlated as
in the previously considered range of scales, the
values of activity decline steadily at the approach
of the strong 1998 Bovec earthquake from above 0.5 in 1991 to below -1.4 in 1997. A justified conclusion on whether it is a reliable indication of the
integral activity difference at different scales or it
results from data deficiency in the western half of
the seismogenic zone is hard to make. We cannot
reject any of the two possibilities.
In the Western Alps seismogenic zone, which did
not experience any magnitude 5.5 or larger earthquake in the last two decades, there are enough data for application of the SCE algorithm in the
majority of the grid points since 1986. The inter-
Tab. 4.1.3 - Seismic risk estimates
for selected cities of Italy and
surrounding countries.
Tab. 4.1.4 - Balance of earthquake magnitudes in the eight
seismogenic regions: Estimates
of coefficient B (this study) and
b-value (Molchan and Kronrod,
2004).
mediate-term estimates of the USLE parameters in
this zone (Figure 4.1.12b) also show anti-correlated behavior of A and B along with notable stability of C.
In the Northern Apennines seismogenic zone systematic intermediate-term estimation of the USLE
parameters in the majority of the grid points is
possible in 1985. There were two earthquakes of
magnitudes 5.8 and 5.6 in the region. In both cases A values were at the high levels of -0.5 and 0.2, while B drop to below 0.65 and to 0.6 three
years in advance each of the two earthquakes.
After 1984 the values of C were wandering around
1.2. C was decreasing from above 1.3 at the approach of the first earthquake and was rising from
1.1 during three years prior to the second one.
The estimates of A, B, and C coefficients for Friuli-Venezia Giulia region and
Western Slovenia
Finally, we consider the smallest scales available for
91
Interreg Project ALPS-GPS Quake Net
Fig. 4.1.14 - The USLE
B coefficient for S1
scale: (a) 1870-2005;
(b) 1870-1976.11.1;
(c) 1870-1980.5.1.
a
b
the Friuli-Venezia Giulia and Western Slovenia (S3)
region in 1977-2005 within the area of about 120
by 200 km, spatial resolution of 1/16°, and magnitudes down to 2.2. Figures 4.1.13a and 4.1.13b demonstrate predominance of the characteristic values of A about -1.5 and B about 1.15, each of
which appears in 12% of the obtained estimates.
These values are localized mainly in the TolmezzoGemona-Bovec area of linear dimension about 40
km where the completeness of the OGS data set appears to be the best as mentioned above in section 5.3. The values of C here are evidently biased
due to the 1998 Bovec sequence. The seismic activity associated with the 1976 Friuli sequence of
earthquakes might be still in effect in 1977 and,
therefore, could have affected the estimates in the
western areas of the region.
In comparison with the long-term estimates given
in Figure 4.1.6 the fine scale maps of the USLE parameters plotted in Figure 4.1.13 do not match,
confirming a banality that the data sets covering
hundreds of years are essential for a reliable adequate estimation of seismic hazard and seismic risk.
The high quality data of a few recent decades is not
enough and its use could be misleading without
bringing in historical and/or even paleoseismological evidence. On the other hand the variability of
the USLE parameters computed at different scales
appears to be rich enough for an objective description of seismic activity and its scaling properties in
a given region.
Seismic risk estimates for selected cities
c
We have used the long-term estimates of the USLE
coefficients determined by using the SCE algorithm
to characterize seismic risk for the major Alpine cities (more than 200,000 inhabitants), for selected
cities in surrounding countries and for the three
principal cities of the Friuli-Venezia-Giulia region.
Specifically, we have considered the values of A, B,
and C at the grid points inside the 1/4° vicinity of
the cities and have calculated the values of rc and
rp twice, for “moderate” (magnitude 5.0 or larger)
and “strong” (magnitude 6.0 or larger) earthquakes.
The results are given in Table 4.1.3 below.
It is interesting to compare the obtained estimates
of risk with those achieved previously when using
the NEIC Global Hypocenters Database, 1964-2002
(Nekrasova, A., V. Kossobokov, 2005). The global
estimates for Milan (mega-citie of Italy) are as follows: Milan, A=-0.76, B=1.24, C=1.22. These values
imply overestimation of seismic risk (either rc or
rp) for moderate earthquakes about magnitude 5.5
in the citie by the factors of 3.4. For strong earthquakes about magnitude 6, the risk is slightly
overestimated (by factor 1.1). The example shows
explicitly the importance of bringing into consideration prolonged regional data on earthquakes for
accurate and reliable estimates of seismic risks.
4.1.7 Discussion and conclusions
Thus, the parameters of USLE for the Alps and surroundings were evaluated making use of the diffe-
92
Interreg Project ALPS-GPS Quake Net
Fig. 4.1.15 - Intermediate-term
variations of the USLE coefficients in the CN regions, (S1 scale,
1870-2005): (a) Northern; (b)
Central; (c) Southern.
a
rent earthquake data and different ranges of magnitude and scaling dimensions. Of course, this is
just the beginning of understanding the complexity of earthquake recurrence and many other computations could be suggested for clarification of
different aspects of seismic energy release at different scales of space, time and energy. Nevertheless
some conclusions are already evident from the present analysis.
The results confirm high accuracy of the USLE coefficient estimates obtained with the help of the SCE
algorithm at different scales from regional to local
ones, as well as from long-term to intermediateterm ones. The accuracy along with overall concordance of the USLE parameters in the calculations
at different spatial, temporal and magnitude scales permits a few comparisons with the estimates
of the basic parameters of seismic activity in Italy
published earlier. For example, Table 4.1.4 compares the values of B coefficient obtained as an
overall average computed for the points inside the
eight seismogenic regions with the b-value computed earlier by Molchan and Kronrod (2004) for
the same regions. Except for the Western Alps, B
is slightly larger than b. The average difference B
and b is about 0.07 with the standard deviation
0.12, which values imply insignificant differences
for each of the eight regions but Sicily (at the confidence level about 95%). The difference observed
for Sicily might be due to the above mentioned
instability of the B determinations in the region
(6.1, Fig. 4.1.10), which could not be observed
in a single determination of b-value (Molchan and
Kronrod, 2004).
For analyzing the spatial distribution of B coefficient values one may find useful to redraw Figure
93
Interreg Project ALPS-GPS Quake Net
b
4.1.6-B in a form of Figure 4.1.14, which outlines more clearly the areas of the magnitude balance extremes, i.e., Po Valley and Switzerland for
low and Western Alps and Tyrrhenian Sea for high
values. It should be noted, however, that such a
sharp differentiation carry a certain degree of arbitrariness and might lead to illusory perception of
the anomaly extent in space. Therefore, it should
not substitute, but rather supplement the continuous mapping of the USLE coefficients. As shown
in 6.1 the USLE parameters are contributing to an
arbitrary seismotectonic zonation yet unexplored
information on seismic activity at regional and local scales.
The monitoring of the long- and intermediate-term
trailing estimates of the A, B, C coefficients may
94
contribute to the analysis aimed at earthquake forecast and or prediction. Essentially, the trailing A
rescaled to a target magnitude of interest deliver
the observed expectation of the rate for such events, which could be used by itself as a forecast tool.
On the other hand our analysis of the A, B, C variability suggests a finer candidate for being used as a
precursory signature: rise of A to its high values accompanied with low values of B appears to indicate at intermediate-term scale of years the approach of strong earthquakes. Figures 4.1.15 and 4.1.16
show a possibility to supplement the on-going test
of earthquake prediction algorithms with information either on the overall average of the observed
values of the USLE coefficients inside the CN regions of investigation (Figure 4.1.15) and/or with
Interreg Project ALPS-GPS Quake Net
Fig. 4.1.16 - Intermediate-term
variations of the USLE coefficients for the set of neighbors on
a regular grid, (S2 scale, 19742005): (a) A, B, C coefficients;
(b) map of point set.
those derived locally, e.g., from a set of neighbors
on a regular grid (Figure 4.1.16). As could be judged from the figures the observed variability suggests further investigations into the problem of the
A, B, C predictive power and efficiency.
The analysis of the USLE in Italy and surrounding
regions are to be continued in the future.
4.2 Deterministic
earthquake hazard
assessment in the Alps
Franco Vaccari for DST-UNITS
A deterministic procedure (Panza et al., 2001) for
seismic hazard assessment has been applied to the
Alpine region. This approach addresses some issues
largely neglected in probabilistic hazard analysis,
namely how crustal properties affect attenuation:
ground motion parameters are not derived from
overly simplified attenuation functions, but rather
from synthetic time histories.
Starting from the available information on the
Earth’s structure, seismic sources, and the level of
seismicity of the investigated area, it is possible
to estimate maximum ground displacement, velocity, and acceleration or any other parameter relevant to seismic engineering, which can be extracted from the computed theoretical signals. This
procedure allows us to obtain a realistic estimate of the seismic hazard in those areas for which
scarce (or no) historical or instrumental information is available and to perform the relevant para-
95
Interreg Project ALPS-GPS Quake Net
metric analyses.
Synthetic seismograms can be constructed to model ground motion at sites of interest, using knowledge of the physical process of earthquake generation and wave propagation in realistic media.
The signals are efficiently generated by the mo-
Fig. 4.2.1 - Flow-chart of the deterministic procedure for seismic
hazard assessment at regional
scale. The vertical component is
routinely not used.
Figure 4.2.4. Flowchart of the
algorithm adopted to determine
the structural model of the
Alpine region (modified from
Farina, 2006).
Fig. 4.2.2 . Regional polygons associated with average
structural models. GNDT:
model previously defined in the
framework of GNDT activities.
ALPS: structural boundaries
defined in the framework of the
ALPS-GPSQUAKENET project.
On the left, a sample of the
digital data extracted from the
database is shown.
96
Interreg Project ALPS-GPS Quake Net
Fig. 4.2.3 - An example of
structural model obtained after
the application of the algorithm
shown in Figure 4.2.4. A sample
of the digital data extracted from
the database is shown.
Fig. 4.2.5 - Discretized seismicity
(0.2° x 0.2° cells) obtained from
Italian earthquake catalogue CPTI
(2004), merged with the Slovenian catalogue used by Zivcic
et al. (2000) and the Croatian
catalogue used by Markusic et
al. (2000). The yellow polygons
show the areas where the Italian
(CPTI), Slovenian (S) and Croatian
(C) catalogues have been used.
dal summation technique (Panza, 1985; Florsch et
al., 1991), so it becomes possible to perform detailed parametric analyses at reasonable costs. For
instance, different source and structural models
can be taken into account to create a wide range
of possible scenarios from which to extract essential information for decision making. The flowchart
of the deterministic procedure is shown in Figure
4.2.1.
4.2.1 Structural models
In first-order zoning, average structural models are
defined, representing the lithospheric properties at
a regional scale. Synthetic seismograms are computed, taking into account the effects of lateral
heterogeneities in a rough way: if the source-receiver path crosses one or more boundaries between
adjacent structural models, the signal is computed
assuming the model of the receiver as representa-
97
Interreg Project ALPS-GPS Quake Net
tive of the whole path.
In Figure 4.2.2 the regional polygons associated
with average structural models are shown. The map
marked GNDT (top) shows the boundaries between
different structures as they were defined in previous studies carried on mainly in the framework of
the Italian GNDT group activities. The map marked
ALPS shows the new boundaries corresponding to
the structural models defined in the framework of
the ALPS-GPSQUAKENET project. On the left, a sample of the digital data extracted from the database, defined during the project activities, is shown.
An example of structural model is given in Figure
Fig. 4.2.6 - Seismogenic zones
das defined in ZS9 (Meletti and
Valensise, 2004). On the left,
a sample of the digital data
extracted from the database is
shown.
Fig. 4.2.7 - Focal mechanisms
associated with the seismogenic
zones, as defined in ZS9 (Meletti
and Valensise, 2004). On the
left, a sample of the digital data
extracted from the database is
shown.
98
4.2.3. The physical properties of the models have
been obtained according to the scheme shown in
Figure 4.2.4.
4.2.2 Seismic sources
The first problem to tackle in the definition of seismic sources is the handling of seismicity data.
Basically, what is needed is an evenly spaced distribution of the maximum magnitude over the
territory, but the data available from earthquake
catalogues are widely scattered. Furthermore, earthquake catalogues are both incomplete and af-
Interreg Project ALPS-GPS Quake Net
fected by errors, so a smoothed distribution is preferable (Panza et al., 1990).
For the Alpine region, three earthquake catalogues
have been considered: the Slovenian catalogue
used by Zivcic et al. (2000), the Croatian catalogue
used by Markusic et al (2000) and the Italian catalogue CPTI (2004). The image of seismicity obtained after discretization into 0.2° x 0.2° cells is
shown in Figure 4.2.5.
The punctual distribution of epicenters is discretized into cells and the maximum magnitude of the
events pertinent to each cell is retained. In the case
where the earthquake catalogue contains different
estimates of magnitude (e.g., magnitude computed
from body waves, from surface waves, or from macroseismic intensity), the maximum between them
is considered. In most cases, the smoothing obtained by considering just the discretized cells is not
enough. To account for errors in the location of the
source and for its extension in space a centered
smoothing window (radius = 3 cells) is then considered, so that earthquake magnitudes are analyzed
not only in the central cell but also in the neighboring ones. The smoothed image of seismicity is finally intersected with the seismogenic zones and
only the cells located within a seismogenic zone
are retained and constitute the set of sources used
for the generation of the synthetic seismograms.
ZS9 seismic zones have been considered (Meletti
and Valensise, 2004), shown in Figure 4.2.6. The
focal mechanism is equal for all the sources belonging to the same seismogenic zone, as defined in
ZS9, and are shown in Figure 4.2.7. They compare
rather well with the source parameters of the in-
strumental earthquakes studied in the framework
of the ALPS-GPSQUAKENET and reported in chapter
3.5. The sources identified with the above mentioned algorithm are shown in Figure 4.2.8.
Fig. 4.2.8 - Sources identified
in the Alpine region, within
the seismogenic zones defined
in ZS9 (Meletti and Valensise,
2004) that have been used for
the generation of synthetic
seismograms. The magnitude
is obtained by discretizing
and smoothing the seismicity
reported in the CPTI (2004)
earthquake catalogue.
Fig. 4.2.9 - Map showing the
peak displacements obtained
from the synthetic seismograms
(top) and the period of the
peak of the Fourier spectrum
(bottom). Structural model ALPS,
earthquake catalogue CPTI. A
set of synthetic seismograms (80
s duration, peak value 4.1 cm),
extracted from the database at
a single site and generated by
different sources is shown as
well (right).
99
Interreg Project ALPS-GPS Quake Net
Fig. 4.2.10 - Same as Figure
4.2.9, considering velocities
(peak value 8 cm/s).
Fig. 4.2.11 - Same as Figure
4.2.9, considering accelerations
(peak value 25 cm/s2). Clearly
the period saturation at 1Hz
indicates that acceleration values
are underestimated. This limit is
overcome by estimating the Design Ground Acceleration (DGA),
as shown in Figure 4.2.12.
Fig. 4.2.12 - Distribution of
DGA, obtained as a result of
deterministic zonation extended
to high frequencies using the
design spectra of EC8 for soil
A. Structural model ALPS,
earthquake catalogue CPTI. The
increase of values when compared with the 1 Hz accelerations
of Figure 4.2.11 (top) is evident.
100
4.2.3 Computations
Once the structures and the sources have been defined, sites are considered on a grid (0.2° X 0.2°)
that covers the whole territory. The synthetic signals are computed for an upper frequency content
of 1 Hz, and the scaled point-source approximation
(Gusev, 1983) is still acceptable. This is fully justified by practical considerations, as several-story
buildings have a peak response in the frequency range below 1 Hz (Manos and Demosthenous,
Interreg Project ALPS-GPS Quake Net
1992), and by the fact that modern seismic design
approaches and technologies, such as seismic isolation, tend to lower the free oscillation frequencies of buildings. As a rule of thumb (in Italian legislation) the resonance period (in seconds) of a
building can be expressed as 0.1H/B0.5, where H
is the height and B is the maximum lateral extension, in meters. When shorter periods are considered, it is no longer possible to neglect the finite
dimensions of the faults and the rupturing process
at the source.
To reduce the number of computed seismograms,
the source-receiver distance is kept below an upper threshold, which is taken to be a function of
the magnitude associated with the source. The maximum source-receiver distance has been set equal
to 25, 50, and 90 km, respectively, for M < 6, 6
M < 7 and M 7. All seismograms are computed for
a hypocentral depth which is a function of magnitude (10 km for M < 7, 15 km for M 7). Keeping
the hypocentral depth fixed (for classes of magnitude) and shallow is important due to the large errors generally affecting hypocentral depth reported
in the earthquake catalogues and due to the fact
that strong ground motion is mainly controlled by
shallow sources (Vaccari et al., 1990).
Synthetic seismograms for P-SV-waves (radial and
vertical components) and SH-waves (transverse component) are originally computed for a seismic moment of 10-7 Nm. The amplitudes are then
properly scaled according to the smoothed magnitude associated with the cell of the source.
For the moment-magnitude relation, the one given by Kanamori (1977) has been used. The finiteness of the source is accounted for by scaling the
spectrum using the spectral scaling law proposed
by Gusev (1983) as reported in Aki (1987). At each
site, the horizontal components are first rotated to
a reference system common to the whole territory
(North-South and East-West directions) and then
the vector sum is computed.
Among the parameters representative of strong
ground motion we have focused our attention
on maximum ground acceleration, velocity, and
displacement.
The Fourier spectra of displacements and velocities show that an upper frequency limit of 1 Hz is
sufficient to take into account the dominant part
of seismic waves, while this is definitely not true
for accelerations (Panza et al., 1999). On the other
hand, the required knowledge about seismic sources and lateral heterogeneities, which might justify the choice of a higher frequency limit in the
computations, is not available at the scale of the
entire Alpine region.
The maps of peak displacement, velocity and acceleration, and the period associated with the
peak in the Fourier spectrum, are shown in Figures
4.2.9, 4.2.10 and 4.2.11 respectively. Wherever
long periods in the range between 20 and 30 s dominate in the displacements, it means that are related to signals generated by strong earthquakes
Fig. 4.2.13 - Map showing the
peak velocities obtained for
the structural model ALPS (top)
and the structural model GNDT
(bottom). The map of the differences is shown on the right,
where triangles pointing up indicate values bigger for the ALPS
model, while triangles pointing
down indicate values bigger for
the GNDT model.
Fig. 4.2.14 - Map showing the
magnitude assigned to the seismic sources for the earthquake
catalogue CPTI (top) and for
the earthquake catalogue
UCI (bottom). The map of the
differences is shown on the
right, where triangles pointing
up indicate values bigger for the
CPTI catalogue, while triangles
pointing down indicate values
bigger for the UCI catalogue.
101
Interreg Project ALPS-GPS Quake Net
4.2.4 Parametric tests and
conclusions
Fig. 4.2.15 - Map showing the
peak velocities obtained for the
earthquake catalogue CPTI (top)
and for the earthquake catalogue UCI (bottom). The map of
the differences is shown on the
right, where triangles pointing
up indicate values bigger for the
run made using the CPTI catalogue, while triangles pointing
down indicate values bigger
for the run made using the UCI
catalogue.
occurring at large distances from the site (about
90 km), while the magnitude of the closer events, which are responsible for the higher frequencies
(between 2 and 5 s in our computations), is not
big enough to let these frequencies dominate the
ground motion scenario.
For accelerations, deterministic results has been
extended to frequencies higher than 1 Hz by using
the Eurocode 8 design response spectrum for soil
A (EC8, 1993), which defines the normalized elastic acceleration response spectrum of the ground
motion, for 5% critical damping. The procedure to
obtain the Design Ground Acceleration (DGA) is described in detail in Panza et al. (1996).
The results of the deterministic procedure are particularly suitable for civil engineers as seismic input
for the design of special buildings. In fact, the relevance of the displacements at periods on the order of 10 s or so is a key issue for seismic isolation
and in general for lifelines with large linear dimensions, such as bridges and pipelines, where differential motion plays a relevant role in their stability (Monti et al., 1996).
102
A couple of parametric tests have been performed
to check the stability of the results to variations of
the input data.
One test has been done on the influence of the
structural model properties, adopting the GNDT
model instead of the ALPS model, both shown in
Figure 4.2.2. Not only the regional polygons differ,
but also the properties of the layered models associated with them. The comparison of the results in
terms of peak velocities is shown in Figure 4.2.13.
In Figure 4.2.14 two maps with the sources used to
generate the synthetic seismograms are shown. The
map on the top has been obtained using the CPTI
catalogue, and corresponds to the map of Figure
4.2.8. The map on the bottom is the corresponding
map of sources that is the result of the discretization and smoothing of the UCI earthquake catalogue (Peresan and Panza, 2002). The map of the
magnitude difference is shown on the right. The
maps of peak velocities obtained for the sources of
Figure 4.2.14 are shown in Figure 4.2.15.
The results of the parametric tests show how important the definition of the input data is. Wherever
the differences in the ground shacking scenarios
are relevant, a better understanding of the structural models, and of the properties of the seismic
sources, is desirable.
Further investigations and developments in the
methodology might be considered for a possible
continuation of the project, like an improved definition of the seismogenic zones based on the
analysis of the GPS data that will be made available by the network installed during the ALPSGPSQUAKENET project.
Another area where improvements might be achieved is in the definition of the structural model.
The LSO smoothing algorithm (Farina, 2006) might be applied to smaller cells, wherever possible,
so that the boundaries between the Alps and the
surrounding plains could be better respected in the
modeling.
Finally, the possibility of applying a more realistic
scaling of the seismic source for the magnitude
should be considered. If the peak values obtained
with the deterministic approach compare satisfactorily with observations (Panza et al., 2001),
the signal duration in the synthetic seismograms
may underestimate the duration of the recordings,
if the rupturing process at the source is a complicated one, like the Irpinia 1980 earthquake (e.g.
Vaccari et al., 1990).
Interreg Project ALPS-GPS Quake Net
103
Interreg Project ALPS-GPS Quake Net
5 - PILOT PROJECTS
104
Interreg Project ALPS-GPS Quake Net
5.1 GPS and Meteorology
Andrea Walpersdorf for LGIT
5.1.1 Introduction
Predictability of the atmosphere and particularly
of precipitation is of extraordinary societal, economic, and social significance. Its improvement represents a task of provident character for our future existence. Agriculture and water resources
management, air and shipping traffic, road transport and energy economy directly depend on the
state of the atmosphere. Damage caused by extreme precipitation events heavily burdens the budgets of industry, national governments and international organizations. People affected by extreme
precipitation events often face economic ruin.
Susceptibility to extreme events, e.g. strong precipitations, hailstorms or storms, will further increase
in the industrialized nations due to the increasing
accumulation of material assets and the optimization of economic processes. In Europe, this became obvious in 2002 again during the catastrophic
flash flood event in Saxonia, which caused an economic loss of 10B US$ (Munich Re Group 2002).
The devastating hurricane season 2005 demonstrated that even the most developed countries such
as the US can hardly handle these events.
Quantitative forecast of non-extreme precipitation events is of comparable value, although the
avoidable losses mostly do not appear to be that
spectacular. Complemented by estimates of their
potential uncertainties, such forecasts are of inestimable value as input for hydrological applications or for consulting in agriculture and the construction sector.
In spite of this strong societal demand for quantitative precipitation forecasting (QPF), progress
during the last two decades has been very slow.
While many aspects of numerical weather forecasting have made great advances, model skills for
precipitation remain much lower than for other
atmospheric parameters. Among the reasons explaining QPF failures, the following arguments are
usually mentioned:
Inaccuracy of the model initial states due to a
bad quality or lack of observations – this is particularly true for moisture which has a strong spatial variability
Sub-optimal use of the observations or a lack of
assimilated systems suited for the existing but non
standard observations
Deficiencies in the parameterization of the physical processes
Too coarse resolution of the models and also remaining problems in their dynamical cores
Nowadays, to measure water vapour in the atmosphere, meteorological services rely on standard synoptic radiosoundings. However, even though such
devices provide reasonably accurate measurements
of water vapour profiles, they are far too sparsely
distributed in space and time to support reliable
forecasting of such rain events. However, Global
Positioning System (GPS) data, from stations with
well known coordinates, permit to estimate tropospheric delays which can be transformed into precipitable water vapour (PWV). Methodological studies of meteorological GPS applications have been
carried out since more than 10 years (e.g. Bevis
et al., 1992, Businger et al., 1996, Tregoning et
al., 1998, Bock and Doerflinger, 2001) and enable
us now to infer PWV from GPS observations with
the same precision as conventional meteorological
measurements, such as radiosondes and microwave
radiometers (WVR), to about 1-2 mm PWV.
GPS tropospheric delay or PWV estimation has several advantages over traditional meteorological measurements of water vapour: It can be done
at low cost (either by using already existing GPS
stations or by installing new GPS stations which
are less expensive than other instruments), it is
performing under all weather conditions and the
method is intrinsically stable. Effectively, GPS
PWV measurements are based on the exploitation
of propagation delays excluding any instrumental
drifts. Nevertheless, modifications of the analysis strategy or the change of instruments (receiver and/or antenna) can induce instantaneous offsets in the coordinate time series, an effect which
is however reduced on the tropospheric parameter
estimates. The GPS performances have been tested mainly for mid-latitude networks where they
have shown high efficiency. As more and more GPS
are being deployed and operated in a continuous
mode for geodetic purposes, they offer the potential for a dense and reliable water vapour measurement network. Presently, applications in meteorological analysis and weather forecast are widely
spread in e.g. European, US and Japanese weather
services (Gendt et al., 2004, Guerova et al., 2006,
Gutman et al., 2004, Nakamura et al., 2004).
Examining potential meteorological applications of
the permanent GPS stations installed in the framework of ALPS-GPSQUAKENET is part of the Work
Package “Pilot Projects” proposed in this study.
To illustrate the capacities of our network, we recall the principles of tropospheric measurements
by GPS, show existing applications in comparable
networks and present the results of a test period
which has been scheduled in November 2006.
5.1.2 Principles of tropospheric
measurements by GPS
The GPS signals at radio frequencies (1575 and
1228 MHz) are delayed when travelling through the Earth’s atmosphere, in a dispersive way by
the charged part (the ionosphere) and in a nondispersive way by the neutral part of the atmosphere (the troposphere). While the ionospheric
delay can be identified by comparing the different
influence on the two wavelengths of the GPS signals, the effect of the troposphere on the two
GPS signals is the same and can only be determined by independent measurements or by indi-
105
Interreg Project ALPS-GPS Quake Net
rect methods. The tropospheric delay has a typical value of 2.30 m at zenith, which can vary by
several cm in a few hours, mainly due to the rapidly varying water vapour along the travel path.
This delay must be precisely known to achieve high
precision positioning (sub-centimetric). Model predictions or estimations from meteorological ground
measurements have shown to be not precise enough to correct efficiently for the tropospheric delay.
Therefore, an additional parameter has been introduced in the GPS data inversion representing the
temporally and spatially varying troposphere, the
tropospheric delay. To distinguish the tropospheric delay from the geometric distance between satellite and antenna, a simple model is applied to
characterize the troposphere: a mapping function
describing the increase of the tropospheric delay
for decreasing satellite elevations. The tropospheric parameter represents the average delay over all
available satellites in a defined time interval, attributed to the zenith direction: the zenith total
delay (ZTD) (Figure 5.1.1). A typical time resolution of ZTD evaluations is 15 min. Permanent GPS
stations can therefore provide ZTD evaluations every 15 min continuously, independently of daytime,
season or weather conditions.
The primary GPS observable ZTD can be converted into an observation of water vapour via simple
ground measurements of pressure and temperature. ZTD is the sum of two components, the Zenith
Hydrostatic Delay (ZHD) and the Zenith Wet Delay
(ZWD). In fact, ZHD can be determined from ground
pressure measurements. Then, ZWD can be isolated
by subtracting ZHD from GPS observed ZTD. The
ZWD is almost proportional to Precipitable Water
Fig. 5.1.1 - Schematic representation of the Zenith Total
Delay (ZTD), estimated from all
simultaneously available satellite
observations.
Fig. 5.1.2 - GAIN ZTD time series
(units of the vertical scale are
metres).
Fig. 5.1.3 - GAIN ZTD time
series. Zoom on 15 days in
2006 (day of year 165 to 179).
The ZTD on the vertical axis is
indicated in metres.
106
Vapour (PWV). The relation between ZWD and PWV
can be approximated by a function of ground temperature. Typically, PWV = 0.15 ZWD with ZWD and
PWV in mm. The precision which can be reached
for the evaluation of ZTD is 6 - 12 mm. This means
that PWV estimations by GPS are precise to 1 - 2
mm, reaching the precisions of standard meteorological measurements (radio soundings, water vapour radiometers). Radio sound and water vapour
radiometer measurements are the main sources of
water vapour observations over the continents, but
they are expensive and therefore sparse in time and
space. The persistent lack of water vapour observations for the characterization of the water cycle in
the atmosphere could be filled by the exploitation
of data from GPS permanent stations.
5.1.3 Applications of GPS tropospheric parameter monitoring
GPS meteorological applications can be divided in
three classes:
1
Climatology (long stable time series of
PWV/ZTD)
2 Meteorological analysis (detailed PWV field for
the study of precipitation events, 3D PWV fields
from tropospheric tomography)
3 Numerical weather prediction (continuous PWV
observations in near real time for assimilation in
numerical weather prediction models)
For the first two applications, GPS ZTD is provided by the most precise, post processing solution
using precise satellite orbits published 2 weeks
after the measurement date. The time delay chosen for a post processing also generally permits to
wait until the data transmission of all stations included in the network is completed. ZTD observations from the GAIN network obtained in a post
processing analysis are shown in Figure 5.1.2. The
upper, middle and lower boxes represent the western, central and eastern stations. The dominating feature is the annual signal with high ZTD values in summer and low values in winter, related
to the amount of tropospheric humidity. Not only
the annual trend but also variations over shorter
time scales seem to be correlated between stations, which however show a constant offset, due
to the different station heights resulting in different hydrostatic delays (the higher the station the
lower the hydrostatic delay). The zoom on a period of 15 days presented in Figure 5.1.3 (western,
central and eastern stations again in upper, middle and lower box) shows in fact time offsets and
differences in amplitude in the time series even
between close-by stations. Also complete de-correlation can be observed between more distant stations. Most of the variability in ZTD is related to
the tropospheric water vapour. The differences and
the de-correlation of the time series are due to the
limited spatial and temporal scale of the water vapour distribution.
The spatial and temporal resolution of ZTD obser-
Interreg Project ALPS-GPS Quake Net
vations in GPS networks can be used to establish
precise instantaneous PWV maps and their temporal evolution. Figure 5.1.4 shows the example of
PWV monitored by GPS during the flash-flood event
in SE France on 8-9 September 2002 (Champollion
et al., 2004)
The precise knowledge of the amount and variation of tropospheric water vapour is substantial for
the understanding of the water cycle in the atmosphere as well as for the prediction of precipitations. Important advances could be made in both
fields by exploiting the water vapour observations
from the still increasing GPS permanent networks.
However, the relation between Precipitable Water
Vapour and precipitations is not direct. Figure
5.1.5 presents PWV time series for a precipitation
event in SE France in 2004. The onset of heavy rain
is marked by the arrow and corresponds to the end
of a 2 days interval with high PWV. The rain dries
out the troposphere as attested by a rapid drop of
PWV. This shows that the presence of high PWV
does not lead instantaneously to precipitations.
Figure 5.1.6 illustrates the spatial offset between
regions of high PWV and zones of rain, by comparing maps of PWV and rain for the same epoch, the
9th September 2002 at 6 h (during the flash-flood
event of 2002). While the zone of high humidity is
spread in a V shape from the Mediterranean coast
to the first reliefs of the Cevennes mountains, the
strongest precipitation cells are seen at the northern extremity of the humidity zone.
Accounting for the complex relationship between
tropospheric humidity and precipitations, the assimilation of GPS ZTD or PWV in Numerical Weather
Prediction (NWP) is probably the most efficient
way of exploiting the tropospheric information of
GPS measurements. NWP models can be implemented with the physical rules describing the water
cycle in the atmosphere, and in particular with the
detailed topography constraining upward motion
of the air which can lead to condensation. Also,
all necessary information is contained in NWP models to extract the water vapour observation directly from ZTDs without converting them first into
PWV passing by empirical formulations increasing
(however slightly) the uncertainties.
The technical challenge to provide significant observations from GPS measurements for assimilation
is the short life time of water vapour in the troposphere (some hours). The European project COST
716 has studied the GPS capacities for NWP and
emits the recommendation to provide GPS ZTD 1
h 45 after the measurement. This implies a data
analysis done in “near real time”, through hourly data download and by a rapid analysis adapted
to the short delay between the acquirement and
the analysis of the data. The rapid analysis has to
deal with ultra rapid (and therefore less precise)
orbit solutions instead of final solutions published
2 weeks after the measurement date. Moreover, a
precise ZTD value needs to be constrained by a few
hours of data before and after the time tag of the
parameter. However, the near real time ZTD is evaluated for the most recent (1 hour) observation
session (Figure 5.1.7). This means the data after
the time tag of the ZTD evaluation are missing. For
both reasons (orbits and sessions) the near real
time ZTD is less precise than the post processed
ZTD, but some methodological developments can
help limiting the errors. Our tests have shown that
most precise solutions are obtained from an analysis calculating every hour a session of 7 hours ending with the most recent just downloaded hour of
data. ZTDs are estimated every 30 min, so in particular at the beginning, in the middle and at the
end of the last hour of data. The middle value corresponds to the time tag of the hourly data file and
is the one representing best the ZTD of this data
span. This hourly analysis has to be rapid enough
to leave some time (e.g. 20 min) for the stations
to upload their data and to be terminated before 1
h 15 after the end of the hourly measurement session (Figure 5.1.7) so that the results are available
for assimilation within 1 h 45 after the time tag
of the session.
5.1.4 Near real time test period
in the GAIN network
The hourly data download is an operational challenge and needs stable data transfer lines. The performance of the GAIN network to this account has
been tested during a near real time test period
scheduled in November 2006. To have time enough
for the hourly data analysis, a data upload within
20 min after the end of the hourly session has been
required. 12 stations from the GAIN network have
Fig. 5.1.4 - From Champollion et
al., 2004. 2D PWV fields inferred from GPS ZTD observations
in the French REGAL/RENAG
network during the flash-flood
event of September 2002.
Fig. 5.1.5 - GPS PWV time series
during a precipitation event in SE
France in 2004.
Fig. 5.1.6 - Precipitations and
rain in SE France during the
flash-flood event of 2002. Both
maps are established by meteorological modelling.
Fig. 5.1.7 - Near real time data
analysis strategy.
107
Interreg Project ALPS-GPS Quake Net
Fig. 5.1.8 - Examples of near real
time data upload performances
in the GAIN network.
Fig. 5.1.9 - ZTD estimation with
final IGS orbits compared to
ultra rapid orbits.
Fig. 5.1.10 - Comparison
between ZTD evaluation in post
processing sessions (triangles,
from 12 hour sessions shifted
by 4 hours with only the middle
4 hours kept) and in near real
time sessions (dots, from 7 hour
sessions shifted by 1 hour with
only the last hour kept).
Fig. 5.1.11 - Map of specific
humidity (kg/kg, see color code)
for the 4th of November 2004.
Superposed dark lines delimit
zones of increased specific
humidity due to assimilation
of one GPS observation (ZTD).
Increment between successive
lines is 0.1 g/kg. The maximum
increase is located at the GPS
site (AXPV) and is 0.8 g/kg.
Fig. 5.2.1 - Distribution of
monitored landslides (red dots)
in Piemonte.
108
participated in this test period, 8 of them providing more than 90 % of data in due time (Figure
5.1.8). The participation rate and the rate of “success” shows that operational hourly data download
of geodynamic GPS sites (more often on bedrock
than on internet) is non trivial.
The data of this test period have been analyzed to
test the influence of the near real time mode compared with a post processing strategy. Figure 5.1.9
shows ZTD differences between a solution with ultra rapid orbits and a solution with final IGS orbits. The differences are less than 10 mm (corresponding to 1.5 mm PWV).
A second test aimed at evaluating the loss of precision due to ZTD evaluation at the end of the session. Figure 5.1.10 displays ZTD calculated with a
post processing strategy shifting 12 h sessions by
4 hours and keeping only the middle 4 hours in
the ZTD solution (graph “post processing” in the
bottom of Figure 5.1.10). This post processed ZTD
is compared to near real time ZTD, evaluated in the
last hour of 7 h sessions shifted by 1 hour (graph
“near real time processing” in the bottom of Figure
5.1.10). Momentary offsets of up to 20 mm (3 mm
PWV) are observed between the two time series
which seem to be related to the lack of constraints on the ZTD evaluation at the end of the observation session.
In collaboration with Météo France, assimilation
tests are done to prepare the use of GPS ZTD in
the new French operational weather forecast model (AROME). The influence of the assimilation
of a single ZTD observation (GPS site AXPV of the
French RGP network) is shown in Figure 5.1.11
(V. Ducrocq, personal communication, 2006).
Assimilation of GPS ZTD is already done operatio-
nally in Germany and Switzerland. For the GAIN
stations, providing hourly data would be a (relatively) simple means to valorise the data by making
them available for operational meteorological applications. A dedicated water vapour application is
scheduled for summer 2007, where the GAIN GPS
stations located in Alsace will be included into the
COPS experiment (Convective and Orographicallyinduced Precipitation Study).
5.2 GPS and Landslides
Carlo Troisi for ARPA-P, Giorgio Zampedri for GST
Landslide monitoring by means of GPS system developed rapidly in the past few years. GPS measuring proved to be an effective and reliable tool
especially for measuring large and slow-moving
landslides; application for small and/or fast moving
landslides is not so widespread for, in such cases,
other monitoring systems are to be preferred.
Each GPS monitoring system consists of a number
of benchmarks on the landslides the positions of
which is measured vs. the position of some reference benchmarks located off-slide. The ideal number of reference points should be four, geometrically disposed around the slide, but the usual number
is two (which is also the minimum) or three. Since
locating off-slide reference point is no easy matter,
the presence of permanent GPS stations (within 710 km from the landslide to me monitored) can greatly help.
The following paragraphs shortly report the GPS
landslide monitoring experience of two partners of
the Alps Gps Quakenet project: Arpa Piemonte and
Servizio Geologico della Provincia di Trento.
Interreg Project ALPS-GPS Quake Net
5.2.1 GPS for landslide
monitoring in Piemonte
The Centro Regionale per le ricerche territoriali e
geologiche of Arpa Piemonte (The Regional agency
for environmental protection) manages a landslidemonitoring network including about 300 landslides (fig. 5.2.1). Most of the monitored landslides
threaten built-up areas or important structures.
The network includes about 680 inclinometers; 400
piezometers; 120 automated data recording units;
conventional tacheometric surveys on monumented benchmarks; GPS systems; extensometers and
joint-meters.
Large landslides and deep-seated-deformations are
very common along the Alps; about 600 of them
are mapped in Piemonte. Monitoring this kind of
slope movements may be extremely expensive and
poses several technical problems such as:
conventional tacheometric surveys are often
inapplicable for, given the wide area, monitoring
points are not normally intervisible;
conventional local monitoring devices, such as
inclinometers, may be representative of a minimum
portion of the deformation only.
between each reference station and each on-slide
benchmark. The measuring network works in “local” mode, i.e. is not connected to topographic
vertex related permanent geodetic networks; only
relative positions of points, on and out of the landslide, are evaluated. Overall precision is in subcentimetric order; resulting displacement vectors,
in order to be validated, are to be consistent with
slope attitude and with the general conceptual model of the slide.
A good example of GPS monitoring is the “Cima
Given this framework GPS is, up to now, the best
technique (and, by far, the cheapest one also) for
monitoring very large landslides and deep-seateddeformations. Several benchmarks, scattered over
a wide area, may be effectively measured in one
workday. Moreover, in areas with many rock outcrops (which is common on large alpine landslides and deformations) installation of GPS benchmarks is easy and can be made directly from the
surveying Arpa personnel, without need to have recourse to any building enterprises.
A GPS monitoring system is presently active on fifteen landslides in Piemonte. Apart from one site,
monitored since 2000, GPS monitoring is active
since 2005. Definite displacements, up to 8 cm/
yr, were measured in seven cases. Each system
consists of a number (three to twenty) of on-slide benchmarks and two or three off-slide reference points.
GPS data are acquired on the field by personnel
of Arpa Piemonte, with double-frequency receivers in static mode, acquisition times of 1-1.5 h
and 15 s epoch. Data post-processing is made with
Leica™ LGO software. The displacement , between
the zero-reference survey and each displacement
survey, is calculated along for all the baselines
Fig. 5.2.2 - Map of the Cima
Brenvetto Landslide. Displacement vectors on GPS benchmarks (red arrows) are shown;
blue figures (mm) refers to the
2004-2006 period.
Fig. 5.2.3 - The map shows four
large landslides in the area of
Ceresole Reale (TO). The landslides are monitored by means of
GPS benchmarks (blue dots); the
AGNE GAIN receiver is used as
reference station.
Permanent receiver
CODE
N. of gps monitored landslides using the
permanent receiver as reference station
Ceresole Reale Agnel
AGNE
4
Baceno Alpe Devero
DEVE
2
Trarego Viggiona Carza
CARZ
1
Torino Osservatorio
OATO
0
Paroldo
PARO
7
Fig. 5.2.4 - The map shows a
large deep-seated-deformation
in the area of Baceno (VB). The
deformation is monitored by
means of GPS benchmarks (blue
dots); the DEVE GAIN receiver is
used as reference station.
Fig. 5.2.5 - Benchmark bolts for
GPS monitoring.
109
Interreg Project ALPS-GPS Quake Net
Fig. 5.2.6 - A benchmark bolted
in place.
Fig. 5.2.7 - On rock outcrops
the benchmarks can be easily
and cheaply installed by simple
hand-drilling.
Brenvetto Landslide” (Valprato Soana, Provincia di
Torino) which is a large deep-seated-deformation
extending for about 1 km2 from Cima Brenvetto to
the Soana stream (fig. 5.2.2).
The crown is clearly defined by prominent upwardfacing slopes a feature which, in the alpine area, is
typical of deep-seated-deformations. Toe erosion
by the stream often triggers large debris slides.
110
GPS techniques are the best way to monitor such
a landslide, for steep slope angles and terrain roughness makes a hard field for conventional instruments. On the other hand, it was easy and cheap
to install a monitoring GPS network consisting of
eight benchmarks: three off-slide grouted-in-place reference monuments and five on-landslide benchmarks, simply bolted in place in hand-drill-made
holes on large boulders or exposed rock. An Arpa
field team measures the benchmarks once a year.
Results of the measurements shows displacements
between 3 and 8 cm/yr; displacement vectors are
consistent with slope attitude.
One critical feature of GPS monitoring systems is
the position and the characteristics of reference
off-slide benchmarks; having permanent ones strongly helps so that, as part of the WP6 of the ALPS
Quakenet project, Arpa Piemonte identified several landslides in the range of 7 km about from the
five permanent receivers installed for the project,
which can thus be used as reference stations (see
table). A benchmark for a second non-permanent
reference station was also installed for each landslide. Several GPS benchmarks were installed on
each landslide and zero-reference measurements
were made on all of the landslides. Displacement
measurements will start in 2007; the foreseen frequency is 1 or 2 measurements per year.
Figure 5.2.3 and 5.2.4 show large GPS monitored
landslides close to the Ceresole Reale and Baceno
receivers respectively.
Figure 5.2.5 shows the benchmarks bolts custommade for GPS landslide monitoring. They can easily bolted in place on hard rocks or boulders (fig.
5.2.6) in hand-drill-made holes (fig. 5.2.7) or
placed on a monument (fig. 5.2.8). Around the
Paroldo GAIN permanent receiver there are several
large permanent, soft-rock translational landslides.
Interreg Project ALPS-GPS Quake Net
The slides are currently monitored by means of several inclinometers. Since, on these sites, it’s both
unadvisable and technically difficult to install permanent monuments, we devised and made a simple
device in order to use the inclinometers wellheads
as reference benchmarks (fig. 5.2.9).
5.2.2 GPS for landslide monitoring in the Province of Trento
The Geological Survey of Provincia Autonoma di
Trento manages a landslide-monitoring network including about 15 landslides. The network includes
about 174 inclinometers; 125 piezometers; 10 automated data recording units; conventional topographic benchmarks; GPS systems, extensometers
and joint-meters. Remote data acquisition systems, optical systems and GPS are also used to
monitor landslides’ dynamics.
Use of GPS for landslide monitoring proved to be
extremely useful when used in association with
Fig. 5.2.8 - GPS receiver on
monumented benchmark.
Fig. 5.2.9 - Brass device (right)
made in order to use the inclinometers wellheads as reference
benchmarks. The lower, large
diameter, part exactly enters the
pipe of a conventional aluminium inclinometer casing (diam.
76 mm, shown on the left), thus
providing a stable and reliable
stand for a GPS receiver.
Fig. 5.2.10 - Campodenno
landslide; GPS network and
displacement vectors.
111
Interreg Project ALPS-GPS Quake Net
Fig. 5.2.11 - Peio landslide;
GPS network and displacement
vectors.
Fig. 5.2.12 - Peio landslide; planimetric displacements of GPS
benchmarks.
other monitoring techniques. This integrated approach allowed optimal use of the advantages offered by GPS, such as measure repeatability, comparability and use of long baselines (> 1.5 km) .
Some of the results from GPS monitored landslides
in Trentino are hereinafter described.
Campodenno landslide
The landslide develops in Campodenno, Val di Non,
at an elevation of about 550 m a.m.s.l.; it’s about
350 m wide and about 300 m long along a gently dipping slope. Most of the landslide surface
is built-up. At the moment about 15 buildings
are severely fissured (three of which evacuated by
mayors’ ordinance); minor fissuring affects other
25 buildings about. The landslide was classified
as “area a rischio idrogeologico elevato” (elevated landslide risk area) by force of a national law
(D.P.C.M. 29/9/998). The landslide affects a thick
cover deposit consisting of glacial clays and overlying detritic coarse carbonatic debris 2 to 20 m
112
thick.
The sliding surface is located around the top of the
glacial deposits, at an average depth of 8 m, and
the movement is mainly translational. Recorded
surface displacements are about 1-2 cm/yr . The
landslide is monitored by means of three inclinometers, six piezometers, three GPS benchmarks and
about sixty levelling benchmarks. In recent times
the GPS system was partly replaced by a conventional topographic survey system; this was made in
order to increase the monitoring points, which are
now about 35. Fig. 5.2.10 shows the GPS network
and the related displacement vectors.
Peio landslide
The landslide develops around the Peio village, and
affects a steep slope below the built-up area. The
elements at risk are the Peio village itself, a camping site and some houses on the valley bottom.
The landslide was classified as “area a rischio idrogeologico elevato” (elevated landslide risk area)
by force of a national law (D.P.C.M. 29/9/998).
The active landslide occupies about 500000 m2 ,
maximum depth of the sliding surface is about 45
m . Landslide materials are mainly heterogeneous
loose glacial deposits, mainly consisting of limy
sands with decimetric to metric size clasts.
The slope is constantly monitored by means of 10
inclinometers, 4 piezometers, a conventional topographic survey network, a GPS network (fig. 5.2.11;
fig. 5.2.12) and 3 automated extensometers.
Recorded displacements are in the order of 10-15
Interreg Project ALPS-GPS Quake Net
Fig. 5.2.13 - Prezzo landslide;
GPS network and displacement
vectors.
cm/yr, with clear acceleration in rainy periods. The
GPS recorded displacements are strongly consistent
both with the results from the others monitoring
techniques and with the general conceptual landslide model.
Fig. 5.2.14 - Prezzo landslide;
planimetric displacements of the
GPS benchmarks.
Prezzo landslide
The landslide develops along the slope of Mt.
Melino and affects:
the whole village of Prezzo (Val Giudicarie), where several buildings are damaged;
the provincial road S.P. n. 122 and some local
roads.
The landslide could also dam the underling Chiese
torrent; damming could cause an overflow which
could threaten an industrial area and the Pieve di
Bono village.
The landslide was classified as “area a rischio idrogeologico elevato” (elevated landslide risk area)
by force of a national law (D.P.C.M. 29/9/998).
The landslide is about 340 m wide and about 1200
m long; the thickness varies between 50 and 80 m,
the average being around 68-70 m .
It’s an old reactivated landslide, mainly consisting
of clayey deposits with fragments, blocks and lenticular rock bodies consisting of marly limestones
and dolomites (Wengen Formation).
The landslide is monitored by means of inclinometers, GPS benchmarks, conventional topographic
benchmarks and piezometers. Recorded displacements are in the order of 3-6 cm/yr . Foreseen remedial works include a deep drainage tunnel, at the
base of the slope, whit branching micro-drains.
GPS recorded displacements (fig. 5.2.13; fig.
5.2.14) perfectly dovetail with the results from the
other monitoring systems and allow a proper definition of landslide dynamics and extension.
5.3 GPS and active faults
LGIT, EOST
One particular application of permanent GPS measurements for geodynamics is the monitoring of
active faults. In the complex Alpine tectonics,
the observation of low magnitude earthquakes
by dedicated seismological networks like Sismalp
(Thouvenot, 2002) helps a lot to identify which of
the numerous faults in the mountain belt could be
presently active. In the French Alps and their foreland, three fault zones with characteristic seismicity are instrumented by permanent GPS, the
Durance fault, the Belledonne fault and a zone
of more diffuse seismicity in the Briançon region
(Figure 5.3.1). The GAIN network established in
this project contributes with 3 GPS stations to
the monitoring of the Belledonne fault and the
Briançon zone.
The oldest GPS monitoring site is set up on the
Durance fault, to complement a dense seismological network and seismotectonic studies initiated by IRSN (Institut de Radioprotection et Sureté
Nucléaire, Fontenay-aux-Roses) (Figure 5.3.2). The
two GPS stations GINA and MICH have been instal-
113
Interreg Project ALPS-GPS Quake Net
Fig. 5.3.1 - Seismicity in the
western Alps registered by the
Sismalp network. The green frames indicate the location of the
3 active fault zones monitored
by permanent GPS.
Fig. 5.3.2 - Location of the GINA
and MICH permanent GPS stations (red triangles) to each side
of the Durance left lateral strikeslip fault (red line). Seismological
stations are indicated by yellow
and blue squares.
Fig. 5.3.3 - The left graph shows
the de-trended time series of
baseline components of the
GINA-MICH baseline of 28 km
length, from mid 1998 to end
of 2003. The north, east and up
components are in the upper
3 boxes and the total baseline
length in the lower box. The
average baseline rates are
indicated for each component
on the top of the corresponding
box. The right graph represents
the convergence test of the
velocity estimates based on
increasing amounts of data.
The final values are the average
rates of the complete time series
shown in the left graph.
Fig. 5.3.4 - The 3 GPS sites
around Grenoble on the three
mountain belts bordering the
town. The alignment of seismicity along the Belledonne massif
indicates the present day activity
of this fault.
114
led in 1998, as some of the first stations of the
REGAL/RENAG network (http://renag.unice.fr).
The results of the permanent GPS measurements
across the Durance fault are shown in Figure 5.3.3,
in form of time series of the components of the
GINA-MICH baseline. Note the jump in the horizontal components at the end of each year, probably
due to hydrological phenomena, while the vertical
component shows an increase of noise each summer, probably due to higher water vapour content
of the atmosphere. These features contribute to the
noise of the positioning and increase the observation duration needed to obtain significant velocity
measurements across the Durance fault. A numerical test of the degree of convergence of the baseline rate has been performed and is shown on the right graph in Figure 3. The relative velocity between
the two GPS sites has been calculated over increasing observation lengths, from 1 year to 5.5 years
after the start of the measurements, 4 times per
year. The graphs show dispersion of the velocity
results during the first 4 years of data, and a final
convergence in the 5th year to a value of 0.2 mm/
yr of shortening on the north component, while the
eastern velocity component is insignificant. The NS shortening corresponds to transpressive left-lateral strike-slip motion across the Durance fault and
is compatible with the tectonic observations of the
Durance fault.
The first active fault which has been instrumented
with GPS permanent stations by the LGIT Grenoble
is the Belledonne fault, a dextral strike-slip fault
producing about one earthquake of magnitude 2-4
per year (Figure 5.3.4). Two GPS stations have been
installed on each side of the fault, STEY on the
Chartreuse mountain belt (in May 2003) and CHAM
on the Belledonne massif (in December 2003). A
third station installed in the Vercors mountain
belt completes the monitoring of slow deformation
around Grenoble since December 2004. This third
station LFAZ is part of the GAIN network.
Today, 3 years of data are available on the baseline
between STEY and CHAM, crossing the Belledonne
fault. The time evolution of the baseline components of this 17 km baseline is shown in Figure
5.3.5. As for the Durance fault monitoring, a convergence test has been done, estimating the baseline rates after one year of data and then every 3
months up to the end of 2006 (vertical lines in the
left graph of Figure 5.3.5). The right graph shows
the variations of the inferred velocities and the final values after 3 years in red. These baseline rates
(-0.20 mm/yr on the north component, 0.74 mm/yr
on the east component and 0.73 mm/yr on the up
component) are calculated considering the dispersion of the time series as white noise. Another calculation has been done to infer linear station velocities using a noise model of coloured noise, which
is supposed to better take into account the correlated noise in the positioning time series (green
numbers in the right graph). Both results indicate a fault velocity below 1 mm/yr but still differ
Interreg Project ALPS-GPS Quake Net
Fig. 5.3.5 - Left graph: De-trended time series and average rates
of the baseline components of
the 17 km STEY-CHAM baseline
across the Belledonne fault. Right graph: Results of the velocity
convergence test and comparison with velocities established by
a coloured noise model.
Fig. 5.3.6 - Comparison of the
velocity convergences on the Belledonne fault (left graph) and the
Durance fault (right graph). The
vertical line in the Durance plot
indicates the presently available
observation span of 3 years on
the Belledonne fault.
on the sub-millimeter level. Figure 5.3.6 compares the convergence tests done for the Belledonne
fault with the Durance fault monitoring. After a
three years time span (the available observation
span on the Belledonne fault), the Durance fault
velocities were still at about 1 mm/yr from their final values on the horizontal baseline components,
and at 3 mm/yr on the vertical component. One or
two years of additional observations will be necessary on the Belledonne fault to obtain significant
displacement rates at the level of a tenth of a millimetre per year.
In the framework of the ALPS-GPSQUAKENET program, LGIT had the opportunity to instrument the
Briançon zone where an alignment of extensive
seismic activity is localized (Figure 5.3.7, from Sue
et al., 2000). These extensional earthquake mechanisms are coherent with the prevailing deformation pattern in the western Alps, the E-W extension localized in the centre of the mountain belt.
The comparison of GPS measurements in 1996 with
classical triangulation observations yielded a first
velocity field, indicating an E-W extension of about
10 mm/yr with uncertainties of the same order of
magnitude (Sue et al., 2000).
Two permanent GPS stations (PUYA, JANU, right
graph in Figure 5.3.7) installed in 2005 on each
side of the seismically active belt could quantify
precisely in a few years which part of the global
Alpine extension is localized in the Briançon extensive zone. Moreover, the local temporary GPS
network of 30 sites measured first in 1996 has
been reoccupied in July 2006 (right graph in Figure
5.3.7). For this new measurement campaign, the
classical geodetic bold markers have been doubled
by new screw markers, increasing the measurement
precision by forced antenna centring. This implied
the measurement of local ties between the old and
the new marker to establish the link to the 1996
measurements. A first analysis of the 2006 data reveals repeatabilities (a measure of positioning precision) of 2-3 mm on the horizontal baseline com-
Fig. 5.3.7 - From Sue et al.,
2000. Left graph: Fault traces
and earthquake focal mechanisms in the Briançon zone. Right
graph: Temporary (red triangles)
and permanent (blue triangles)
GPS networks. The velocity
vectors have been established by
comparing triangulation and GPS
measurements and have maximum amplitudes of 10 mm/yr
(Sue et al.).
ponents between the stations. Supposing the local
ties have a precision of 3 mm and the 1996 campaign has a precision of 6 mm, we obtain a formal uncertainty for the velocity estimates inferred from the 1996 and the 2006 campaigns of 1.2
mm/yr. This is probably a conservative value which implies that displacements of the order of 1
mm/yr (the maximum displacement rate expected
in the Briançon zone) will be at the limit of resolution. However, after 5 years of measurement, the
two permanent GPS stations will quantify the local velocity field with a resolution of probably better than 0.2 mm/yr, and a second measurement
campaign on the screw markers at that time (in
2011) will hopefully provide some significant information about the local distribution of the deformation field.
5.4 Active tectonics and
Paleoseismology
Jerome van der Woerd for EOST
5.4.1 Introduction
The region of high relief in western Europe are
115
Interreg Project ALPS-GPS Quake Net
clearly areas that accommodate most of the strain
due to the collision between Africa and Europe.
These ranges are mainly the Alps, the Pyrenees,
and the Apennines. However, the regions that surround these ranges, while far less deformed and
with less relief, are also characterized by a certain level of seismicity that reflect slow strain rates (Figure 5.4.1). These regions are in part the
Tertiary grabens that follow the south western
Figure 5.4.1 : seismicity across
western Europe (after Nocquet
et al., 2003).
Figure 5.4.2 : Main structural
units in western Europe after
Ziegler and Dezes (2005).
Figure 5.4.3 : same as figure
5.4.2 with blocks represented
in different colours and faults
at their boundaries (convergent
boundaries or reverse faults in
blue, strike-slip faults in red,
divergent or normal faults in
green, purple when combined
strike-slip and normal) (modified
from Ziegler and Dezes, 2005).
116
Alpine arc and continue through the Upper and
Lower Rhine Graben to the North Sea. Other regions
include reactivated structures from the Paleozoic
orogenies (Figure 5.4.1).
The intracontinental regions or large collision zones like western Europe, which is the place where Africa collides with Eurasia, do accommodate
slow rates of strain over large areas. At the longitude range of the Alps, the collision zone extends
from the High Atlas in Morocco and Maghrebides
in Algeria to the North Sea in The Netherlands
and Germany, over a zone of 2000-2500 km. Plate
tectonics model (Demets et al., 1994) as well as
recent geodetic models from GPS (Nocquet et al.,
2003) confirm that Africa and Eurasia collide at a
rate of several mm/yr. How and where this convergence is accommodated remains the main issues
in the description and understanding of the strain
field in Europe. It is important to note that these
questions are debated also for the India-Asia collision zone, which is, nevertheless, characterized
by higher rates of deformation (Tapponnier et al.,
2001; England and Molnar, 2005). In Europe, the
tectonic strain rates may equal or be slower than
other processes, like post-glacial isostatic rebound
(in the Pyrenees, Alps and northern Europe)(Beck
et al., 1996; Sue et al., 2000) or mantel upwelling (in the Eifel massif)(Walker et al., 2005), for
example.
The main geological structures that characterize the present geomorphology and landscapes of
Europe, as well as the reactivated structures during
the Tertiary may be used to distinguish regions that
deform more than others (Figure 5.4.2). Seismicity
and focal mechanisms can also be used to discuss
coherent pattern of strain within slow deforming
regions. In Figure 5.4.3 we suggest a large scale frame with blocks that may be more rigid and
with strain mostly occurring on their boundaries.
For such models, where the block sizes are almost
as large as the block boundaries, the definition of
region with coherent strain is rather subjective.
However, we think that this kind of first order approach allows proposing a pattern of deformation
that can be tested against other models.
To test this model GPS data can be compared to
the modelled strain field of the blocks. For example, a four block model (Tesauro et al., 2003) with
a simplified block geometry indicates that coherent
strain is accommodated within the blocks and may
explain the abrupt changes in GPS vectors direction
close to their boundaries (Figure 5.4.4).
To better understand the deformation across western Europe it is therefore important to combine different methods that link local observations
of deformation to the regional movement of blocks
across the collision zone. For this purpose, the permanent monitoring of strain with GPS stations is a
step towards documenting the strain field across
wide areas. In the Rhine Graben area (Figure5.4.5),
north of the Alps and the Jura fold belt, the nature and amount of strain accommodated is still de-
Interreg Project ALPS-GPS Quake Net
bated. In particular, the recent instrumental seismicity may not reflect the long-term strain of the
region. The historical earthquakes enhances the
knowledge of zones exposed to seismic risk but covers still a short period compared to the characteristic time periods of geological processes (Figure
5.4.6).
To further enlarge our knowledge of past earthquakes and strain accommodation, the geological study of past events or paleoseismology is
necessary (Meghraoui et al., 2001; Sébrier et al.,
1997; Peters et al., 2005, 2007). However, this
method is limited to the largest earthquakes as
most of the seismic events do not have primary
surface breaks, but only secondary indirect consequences of ground shaking that may cause certain
damages to man-made constructions or geological
strata (Beck et al., 1996).
river beds have partially or completely remodeled
the geomorphic expression of the faults. Depicting
possible displacements caused by these faults have
been possible by careful mapping of the terrace
treads and by local paleoseismological trenching
(Peters et al., 2005, 2007).
The terraces treads can be followed over large di-
5.4.2 Example of active fault
monitoring
In the southern Upper Rhine Graben : the
Basel earthquake of 1356
The historical earthquake of 1356 occurred in the
region of Basel (Figure 5.4.6). Despite historical
records of damages and several recent investigations the possible source of this large earthquake
of magnitude over six remains controversial (Meyer
et al., 1994; Meghraoui et al., 2001; Nivière et al.,
2002; Ferry et al., 2005; Lambert et al., 2005). It
is located at the intersection of two main structures, namely, the north-northeast trending Upper
Rhine Graben, and the east-west trending Jura fold
belt (Figure 5.4.7).
Trenching south of the city of Basel have lead to
propose that the fault responsible of the 1356
earthquake is a north-south trending normal
fault (Meghraoui et al., 2001; Ferry et al., 2005).
Radiocarbon and thermoluminescence dating of
successively buried colluvial wedges have been interpreted as the traces of the recurrence time for
Mw 6.5 earthquakes every 2500 years. The average rate of displacement is on the order of 0.1-0.3
mm/yr.
Historical records of damages caused by the 1356
events to the castle of the region are numerous
but sometimes ambiguous. Recent re-interpretations of the historical records have lead to a new
evaluation of the damaged area and therefore to
the possible source of the earthquake (Lambert et
al., 2005) (Figure 5.4.8). The macroseismic data
suggests an east-west elongated seismic source
(Meyer et al., 1994; Lambert et al., 1997) under
the folded Jura front.
Figure 5.4.4 : A simplified four
block model is consistent with
GPS vectors from the blocks
(after Tesauro et al., 2003).
Figure 5.4.5 : instrumental
(yellow) and historical (red) earthquake in the Upper and Lower
Rhine Graben north of the Alps.
Black lines are mostly normal
faults bouding the Upper and
Lower Rhine Graben (after Ferry
et al., 2005).
Figure 5.4.6 : Historical
earthquakes across the alpine
domain with intensity I0 7.
Source : Sirène, ECOS and Grünthal (2003).
In the northern Upper Rhine Graben :
uplifted Rhine terraces
The northern Upper Rhine Graben is characterized
to the west by a set of normal faults that separate
it from the Mainz basin. These western border faults are at present sub-parallel to the main strike of
the Rhine River and recent as well as past Rhine
117
Interreg Project ALPS-GPS Quake Net
Fig. 5.4.7 - main structural units
of the Rhine Graben (after Peters
et al., 2007). Red square is Basel
area.
Fig. 5.4.8 - Trench wall interpretation across the western border
of the Birs valley south of Basel
proposed as the locus of the
Basel-Reinach Fault responsible
of the 1356 Basel earthquake
(Ferry et al., 2005; Meghraoui et
al., 2001).
Fig. 5.4.9 - re-evaluation of the
damaged zone due to the 1356
event in the Basel area (after
Lambert et al., 2005).
Fig. 5.4.10 - The northern Upper
Rhine Graben is characterized by
sup-parallel mostly normal faults.
These fault cut and offset abandoned terrace levels of the Rhine.
To the right, example of trench
wall interpretations dug across
one of these normal fault reveals
repeated normal displacements.
Fig. 5.4.11 - Terrace levels
across the different faults of the
western border of the northern
Upper Rhine Graben south of
Mainz. Clearly, the terraces have
been displaced relatively by the
fault activities.
stances and clearly show that the normal faults
were active during the Quaternary. The several tens
of meters of displacement in the Quaternary allow
to put bounds on the slip-rates of these faults on
the order of 0.01-0.1 mm/yr. While terraces can be
indicators of recent fault activity and can be used
to quantify the amounts of displacements, elsewhere, the absence of terraces does not mean absence
of tectonic activity. In particular, the eastern side
of the Upper Rhine Graben, which is geomorphologically one of the most spectacular stretch along
the Rhine Graben follows one of the deepest trough of Quaternary sediments in the graben. Both the
geomorphic expression of the Rhine Graben shoulder and the depth of the sediments are indicator of
long-term tectonic activity that most probably occurs along the East Border Fault. Quantifying the
amount and rate of displacements in this area reveals to be difficult in the absence of clear geomorphological markers.
The Remua fault near Chamonix
In 1905 two earthquakes have struck the region around the Mont-Blanc massif in France and
Switzerland (Figure 5.4.12). As of today, the possible sources of these earthquakes are unknown
(Alasset, 2005; Cara et al., 2006; van der Woerd et
al., 2006). A combined approach of seismological
modeling and geological investigation has lead to
suggest different potential scenarios for these earthquakes. Recent magnitude 5 earthquakes in the
area indicate a consistent pattern of deformation
characterized by left-lateral movements along NW
trending faults, right-lateral movements along NE
trending faults and thrusting along NNE trending
folds. Synthetic modeling of some of the rare records of the 1905 events are not constrained enough to decide the type of faulting involved in the
events.
In addition to the possible mechanisms expressed
by the recent seismicity, a prominent fault scarp of
the Remua fault has been explored (Figure 5.4.13).
This several tens of meters high scarp is very steep
and exhibits in its central part recently exhumed
rocks. Weather this exhumation is due to tectonic
displacements or local slope instability is debated. If tectonic, then this fault is a left-lateral northeast-trending normal faults as attested by striations on the fault scarp. To constrain the timing of
possible displacement we have conducted cosmogenic Be10 dating of rock outcrops on both side of
the scarp. The dating of glacially polished quartz
veins indicate that this part of the valley was free
of ice around 17 ka, shortly after the Last Glacial
Maximum (Coutterand and Nicoud, 2005). If the 5m
of exhumed bedrock at the base of the scarp are
of tectonic origin, then the ages provide an upper
bound of tectonic movement and allow to determine a rate of 0.3 mm/yr.
5.4.3 Future work
Other regions like the transition zone between the
Alps and the Dinarides show clear evidence of active
faulting. The Cividale and Idria faults are right-lateral strike-slip faults that accommodate northward
displacement of the Adriatic towards the Alps. Both
faults were the site of large historical earthquakes
that may have produced surface breaks.
Remote sensing techniques and high accuracy digital elevation models obtained from radar measurements (SRTM or LiDAR) help in the identifica-
118
Interreg Project ALPS-GPS Quake Net
tion of fault scarps. The analysis of topography and
high accuracy images have then to be combined
with geological field work to constrain the timing
of movements and the possible size of events.
Among the techniques used in the active fault characterization, continuous GPS measurements allow
to determine both the type and rate of displacements in active tectonic regions. The network installed during the Alps-GPSQuakenet project in
the Rhine Graben area will allow to better understand the type of deformation occurring in this region and precise the locations of the most active faults.
Fig. 5.4.12 - regional seismotectonic map of the northwestern Alps. Intensity map of the
1905 events are indicated as well
as the recent focal mechanisms
of magnitude 5 events.
Fig. 5.4.13 - View to northwest in the Chamonix-Mont
Blanc valley towards the Aiguilles Rouges massif. Below the
upper level reached by the glaciers during the Last Glacial
Maximum, a clear linear and steep cliff marks the trace of the
Remua fault (after Alasset, 2005).
Fig. 5.4.15 - Remote sensing images allow to map active fault
traces. A) active fault traces revealed by high precision digital
elevation model in Buttrio along the Cividale Fault. B) Active
fault traces mapping by high precision radar topography
(LiDAR) reveals as a useful tools in forested area of western
Slovenia along the Idria fault (Cunningham et al., 2006).
Fig. 5.4.16 - Map of permanent
GPS stations installed by the
ALPS-GPSQuakenet project in the
upper Rhine Graben to monitor
present strain across the Vosges
and Black Forest massifs.
Fig. 5.4.14 - Active fault map of western Slovenia. Idria and
Cividale faults are clean cuts through the landscape indicating
recent activity.
119
Interreg Project ALPS-GPS Quake Net
Partners list
LEAD PARTNER
Acronym
UNITS
Institution
Università degli Studi di Trieste - Dipartimento di Scienze della Terra
People involved
Karim Aoudia
Riccardo Riva
Franco Vaccari
Nastia Nekrasova
Mariangela Guidarelli
Giuliano Panza
PROJECT PARTNERS
Acronym
Institution
People involved
Agenzia Regionale Per la Protezione Ambientale del Piemonte
Carlo Troisi
Ernesto Benazzo
Michele Morelli
Giacomo Re Fiorentin
Agenzia Regionale Per la Protezione Ambientale del Veneto
Alberto Luchetta
Rodolfo Bassan
Antonio Cavinato
Mirco Pollet
BEK
Bayerische Akademie der Wissenschaften / Bayerische Kommission für die Internationale Erdmessung
Christof Völksen
DGFI
Deutsches Geodaetisches Forschungs Institut
Hermann Drewes
EARS
Environmental Agency of the Republic of Slovenia
Mladen Zivcic
Ecole et Observatoire des Sciences de la Terre, Institut de Physique du Globe de Strasbourg, UMR CNRS/ULP 7516, 5, Rue Descartes 67084 Strasbourg Cédex, France.
Jérôme van der Woerd
Gilbert Ferhat
Matthieu Ferry
Mustapha Meghraoui
Jean-Paul Boy
Frédéric Masson
Pascal Gégout
Jacques Hinderer
Patrice Ulrich
Philippe Kah
Fondazione Montagna Sicura
Fabrizio Diotri
Claudio Lucianaz
Jean Pierre Fosson
Iris Voyat
Geological Survey of the Autonomous Province of Bolzano South Tyrol
Claudio Carraro
Servizio Geologico Provincia Autonoma di Trento
Giorgio Zampedri
Saverio Cocco
Francesca De Francesch
Fulvia Demattè
Mauro Degasperi
Oscar Groaz
Giovanni Merler
Franco Rippa
ARPA-P
ARPA-V
EOST-IPGS
FondMS
GSB
GST
120
Interreg Project ALPS-GPS Quake Net
LGIT
Université Joseph Fourier, Laboratoire de Géophysique Interne et
Tectonophysique, UMR 5559 du CNRS
Andrea Walpersdorf
Nathalie Cotte
RLB
Regione Lombardia, Direzione Territorio Urbanistica, Sistema Informativo Territoriale
Stefania Crotta
Roberto Laffi
Andrea Piccin
RLG
Regione Liguria – Direzione Centrale Affari Organizativi – Servizio Sistemi Informatici
Pier Giorgio Gerbino
Lucia Pasetti
Enrica De Micheri
SUB-CONTRACTORS
Acronym
Institution
People involved
Galileian plus srl
Angelo Amodio
Massimiliano Chersich
Andrea Francia
UNIGE
Dipartimento di Macchine, Sistemi energetici e Trasporti (DIMSET) Facolta’ di Ingegneria - Universita’ degli Studi di Genova
Bianca Federici
Domenico Sguerso
POLIMI
Dipartimento di Ingegneria Idraulica, Ambientale, Infrastrutture Viarie, Rilevamento (DIIAR) - sezione Rilevamento Politecnico
di Milano
Riccardo Barzaghi
Alessandra Borghi
Letizia Cannizzaro
POLITO
Dipartimento di Ingegneria del Territorio, dell’Ambiente e delle Geotecnologie Politecnico di Torino - II Facolta’ di Ingegneria sede di Vercelli
Ambrogio Manzino
Manuele Pesenti
Marco Roggero
UNIMI
Dipartimento di Scienze della Terra “Ardito Desio” Universita’
degli Studi di Milano
Valentina Barletta
Roberto Sabadini
IREALP
IREALP (Istituto di Ricerca per l’Ecologia e l’Economia Applicate
alle Aree Alpine)
Paolo Belluomini
Luca Grimaldi
Marco Scuratti
Michela Fioroni
LGCA (Laboratoire de Geodynamique des Chaines Alpines),
Chambery
Francois Jouanne
GALILEIAN
LGCA
121
Interreg Project ALPS-GPS Quake Net
Bibliography
Cespuglio, G., Campus, P. and Šílený, J., 1996. Seismic moment tensor resolution by
waveform inversion of a few local noisy records. II. Application to Phlegraean Fields
(southern Italy) volcanic tremors. Geophys. J. Int., 126: 620-634.
Aki, K, 1987, Strong motion seismology, in M. Erdik and M. Toksöz (eds) Strong
ground motion seismology, NATO ASI Series, Series C: Mathematical and Physical
Sciences, D. Reidel Publishing Company, Dordrecht, Vol. 204, pp. 3-39.
Chimera, G., Aoudia, A., Saraò, A., and Panza, G.F., 2003. Active tectonics in Central Italy: constraints from surface wave tomography and source moment tensor inversion. Phys. Earth Planet. Int., 138: 241-262.
Anderson, H. and J. Jackson, Active tectonics in the Adriatic region, Geophys. J.
R. astr. Soc., 91, 937-983, 1987.
Christensen, K., L. Danon, T. Scanlon, & P. Bak, 2002. Unified Scaling Law for Earthquakes. Proc. National Acad. Sci. 99, suppl. 1: 2509-2513.
Comuni d’Italia, 2005, Istituto Enciclopedico Italiano, http//www.comuniditalia.biz/
dati.
Aoudia A., Saraó A., Bukchin B., and Suhadolc P., 2000. The 1976 Friuli (NE Italy)
Thrust Faulting Earthquake: A Reappraisal 23 Years Later, Geophys. Res. Lett., 27
(4), 573-576.
Aoudia, A., Saraò, A., Bukchin. B., and Suhadolc, P., 2000. The 1976 Friuli (NE
Italy) thrust faulting earthquake: a reappraisal 23 years later. Geophys. Res. Lett.,
27(4): 577-580.
Atmanspacher, H., & H. Scheingraber, W. Voges, 1988. Global scaling properties
of a chaotic attractor reconstructed from experimental data, Phys. Rev. A., 37(4),
1314-1322.
Bajc J., Aoudia A., Saraó A., and Suhadolc P., 2001. The 1998 Bovec-Krn mountain
(Slovenia) earthquake sequence, Geophys. Res. Lett., 28 (9), 1839-1842.
Bajc, J., Aoudia, A., Saraò, A., and Suhadolc, P., 2001. The 1998 Bovec-Krn mountain (Slovenia) earthquake sequence. Geophys. Res. Lett., 28: 1839-1842.
Bak, P., K. Christensen, L. Danon, & T. Scanlon, 2002. Unified Scaling Law for Earthquakes. Phys. Rev. Lett. 88: 178501-178504.
Barletta V. R. and R. Sabadini, (2006), Investigating superswells and sea level
changes caused by superplumes via normal mode relaxation theory, J. Geophys. Res,
111, B04404, doi:10.1029/2005JB003926
Barzaghi R., Borghi A., Crespi M., Pietrantonio G., Riguzzi F. (2004). GPS permanent
network solution: the impact of temporal correlations, Proceedings of the IV HotineMarussi Symposium on Mathematical Geodesy, Matera, Italy, 17-21 June 2002, F. Sansò editor, pp 179-183, Springer-Verlag, 2004.
Beauval, C., and O. Scotti, Mapping b-values in France using two different magnitude ranges: possible non power-law model ? Geophys. Res. Lett., 30 (17), 1892,
doi:10.1029/2003GL017576, 2003.
Beutler G., Bock H., Dach R., Fridez P., Gäde A., Hugentobler U., Jäggi A., Meindl
M., Mervart L., Prange L., Schaer S., Springer T., Urschl C., Walser P., Bernese GPS Software Version 5.0 (2007), Astronomical Institute, University of Bern.
Bevis, M., S. Businger, T.A. Herring, C. Rocken, R.A. Anthes and R.H. Ware, GPS Meteorology: Remote Sensing of Atmospheric Water Vapor Using the Global Positioning
System, J. Geophys. Res., 97, 15,787-15,801, 1992.
Biancotti A. and M. Motta, (2001), L’evoluzione recente ed attuale dei ghiacciai italiani, Bollettino geofisico 3-4 , 27 – 36;
Bock, O., and E. Doerflinger, Atmospheric modeling in GPS data analysis for high accuracy positioning, Phys. Chem. Earth, Vol. 26, Iss. 6-8, 373-383, 2001.
Boehm J., Heinkelmann R. and Schuh H., Short Note: A global model of pressure
and temperature for geodetic applications, Journal of Geodesy, DOI 10.1007/s00190007-0135-3 (2007)
Boehm J., Niell A., Tregoning P., Schuh H., Global Mapping Function (GMF): A
new empirical mapping function based on data from numerical weather model data,
Geophysical Research Letters, Vol. 33, L07304, doi:10.1029/2005GL025546, 2006.
Boschi, E., Ferrari, G., Gasperini, P., Guidoboni, E., Smiriglio, G. ans Valensise, G.,
1995, Catalogo dei Forti Terremoti in Italis dal 461 a C. al 1980. Istituto Nazionale di
Geogisica SGA storia geofisica ambiente.
Burov, E., Podladchkov, Y., Grandjean, G. and Burg, J.P., (1999), Thermo-mechanical approach to validation of deep crustal and lithospheric structures inferred from
multidisciplinary data: application to the Western and Northern Alps, Terra Nova 11
(2-3) , 124-131.doi: 10.1046/j.1365-3121.1999.00236.x
Businger, S., S.R. Chiswell, M. Bevis, J. Duan, R.A. Anthes, C. Rocken, R.H. Ware,
M. Exner, T. VanHove and F. Solheim, The Promise of GPS in Atmospheric Monitoring,
Bull. Amer. Meteor. Soc., Vol. 77, No. 1, 5-18, 1996.
C. Vigny, J. Chéry, T. Duquesnoy, F. Jouanne, J. Ammann, M. Anzidei, J.-P. Avouac,
F. Barlier, R. Bayer, P. Briole, E. Calais, F. Cotton, F. Duquenne, K. L. Feigl, G. Ferhat,
M. Flouzat, J.-F. Gamond, A. Geiger, A. Harmel, M. Kasser, M. Laplanche, M. Le Pape,
J. Martinod, G. Ménard, B. Meyer, J.-C. Ruegg, J.-M. Scheubel, O. Scotti, and G. Vidal,
GPS network monitors the Western Alps’ deformation over a five-year period: 19931998, J. Geodesy, 76, no. 2, 63-76, 2002.
Calais, E., J.-M. Nocquet, F. Jouanne, and M. Tardy, Current strain regime in the Western Alps from continuous Global Positioning Sytem measurements, 1996-2001, Geology, 30-7, 651-654, 2002.
CPTI, 2004. Catalogo Parametrico dei Terremoti Italiani, versione 2004 (CPTI04),
INGV, Bologna.
Dalla Via, G., R. Sabadini, G. De Natale and F. Pingue, (2005), Lithospheric rheology
in southern Italy inferred from postseismic viscoelastic relaxation following the 1980
Irpinia earthquake, J. Geophys. Res. , 110 , B06311, doi:10.1029/2004JB003539
Delacou, B., C. Sue, J.-D. Champagnac and M. Burkhard, Present-day geodynamics
in the bend of the western and central Alps as constrained by earthquake analysis,
Geophys. J. Int., 158, 753-774, 2004.
Du, Z.J., Michelini, A., and Panza, G.F., 1998. EurID: a regionalized 3-D seismological model of Europe. Phys. Earth Planet. Int., 105. 31-62.
Dziewonski, A.M., and D.L. Anderson, (1981), Preliminary reference Earth model,
Phys. Earth Planet Inter. , 25 , 297-356, doi: 10.1016/0031-9201(81)90046-7.
EC 8, 1993. Eurocode 8 structures in seismic regions - design - part 1 general and building, Doc TC250/SC8/N57A.
Farina, B., 2006, Lithosphere-asthenosphere system in Italy and surrounding areas:
optimized non-linear inversion of surface-wave dispersion curves and modelling of
gravity Bouguer anomalies. PhD thesis, Università degli studi di Trieste.
Fitzko F., Suhadolc P., Aoudia A., Panza G.F., 2005. Constraints on the location and
mechanism of the 1511 Western-Slovenia earthquake from active tectonics and modeling of macroseismic data, Tectonophysics, 404, 77-90.
Florsch, N., Fäh, D., Suhadolc, P. and Panza, G. F., 1991. Complete synthetic seismograms for high-frequency multimode SH-waves, PAGEOPH, 136, 529-560.
Florsch, N., Fäh, D., Suhadolc, P., and Panza, G.F., 1991. Complete synthetic seismograms for high-frequency multimode SH-waves. Pageoph, 136: 529-560.
Gendt, G., G. Dick, C. Reigber, M. Tomassini, Y. Liu and M. Ramatschi, Near Real
Time GPS Water Vapour Monitoring for Numerical Weather Prediction in Germany, J.
Meteorol. Soc. Japan, Vol. 82, No. 1b, 361-370, 2004.
GNDT, Gruppo Nazionale per la Difesa dai terremoti, 1992, GL Seismotettonica (zone
sismogenetiche), Modello seismotettonico del terremoto italino, versione aggiornata al Maggio 1992.
Guerova, G., J.-M. Bettems, E. Brockmann and C. Matzler, Assimilation of COST 716
Near-real Time GPS data in the nonhydrostatic limited area model used at MeteoSwiss,
Meteorology and Atmospheric Physics, Vol. 91 Nr. 1-4, 149-164, 2006.
Guidarelli, M., 2004. Models of lithosphere and seismic sources in the Scotia sea region. Ph.D. Thesis, University of Trieste, Trieste, Italy.
Guidarelli, M., and Panza, G.F., 2006a. Determination of the seismic moment tensor for local events in the South Shetland Islands and Bransfield Strait. Geophys. J.
Int., 167, 684-692.
Guidarelli, M., Russi, M., Plasencia Linares, M.P., and Panza, G.F., 2003. The Antarctic Seismographic Argentinean-Italian Network and the progress in the study of
structural properties and stress conditions in the Scotia Sea region. Terra Antartica
Reports, 9: 25-34.
Guidarelli, M., Saraò, A, and Panza, G.F., 2000. Analysis of seismicity by seismic
moment tensor: the Phlegraean Fields (Southern Italy) case. Internal Report IC/
IR/2000/5, The Abdus Salam International Centre for Theoretical Physics, 69 pp.
Guidarelli, M., Saraò, A., and Panza, G.F., 2002. Surface wave tomography and seismic source studies at Campi Flegrei (Italy). Phys. Earth Planet. Int., 134: 157-173.
Guidarelli, M., Zille, A., Saraò, A., Natale, M., Nunziata, C., and Panza, G.F., 2006b.
Shear-wave velocity models and seismic sources in Campanian volcanic areas: Vesuvio and Campi Flegrei. VESUVIUS: Education, Security and Prosperity. Editor Flavio
Dobran, Elsevier, Amsterdam.
Gusev, A. A., 1983, Descriptive statistical model of earthquake source radiation and
its application to an estimation of short period strong motion, Geophys. J. R. Astron.
Soc. 74, 787-800.
Gutman, S. I., S. R. Sahm, S. G. Benjamin, B. E. Schwarz, K. L. Holub, J. Q., Stewart,
and T. L. Smith, Rapid Retrieval and Assimilation of Ground Based GPS Precipitable
Water Observations at the NOAA Forecast Systems Laboratory: Impact on Weather Forecasts, J. Meteorol. Soc. Japan, Vol. 82, No. 1b, 351-360, 2004.
Campus, P., Cespuglio G., and Panza, G.F., 1993. Full moment tensor retrieval and
fluid dynamics in volcanic areas: the case of Phlegraean Fields (south Italy). Atti dell’Accademia dei Lincei of the International Conference Large Explosive Eruptions (The
problem of eruptions, forecasting and warning; limits and possibilities), pp. 81-101,
Accademia dei Lincei, Rome.
Haeberli W., R. FraunFelder, M. Hoelzle and M. Maish, (1999), Rates and acceleration trend of global mass changes, Geografiska Annaler 810 , 585-591.
Campus, P., Suhadolc, P., Panza, G.F., and Šílený, J., 1996. Complete moment tensor retrieval for weak events: application to orogenic and volcanic areas. In Seismic source parameters; from microearthquakes to large events. Tectonophysics, 261:
147-163.
Herring T.A., King R.W., McClusky S.C., Introduction to GAMIT/GLOBK Release 10.3,
Department of Earth, Atmospheric and Planetary Sciences, Massachussetts Institute
of Technology, 28 September 2006
Haeberli, W. and M. Beniston (1998), Climate change and its impacts on glaciers
and permafrost in the Alps, AMBIO 27 (4): 258-265
122
Interreg Project ALPS-GPS Quake Net
Hugentobler U., Dach R., Fridez P., 2004. Bernese GPS Software, Version 5.0, University of Bern, Draft.
Hugentobler U., Shaer S., Springer T., Beutler G., Bock H., Dach R., Ineichen D.,
Mervart L., Rothacher M., Wild U., Wiget A., Brockmann E., Weber G., Habrich H., Boucher C., 2000. CODE IGS Analysis Center Technical Report 2000.
Nakamura, H., K. Koizumi, and N. Mannoji, Data assimilation of GPS precipitable
water vapor into the JMA mesoscale numerical weather prediction model and its impact on rainfall forecasts, J. Meteorol. Soc. Japan, Vol. 82, No. 1b, 441-452, 2004.
Nekrasova, A., V. Kossobokov, 2002. Generalizing the Gutenberg-Richter scaling law.
EOS Trans. AGU, 83 (47), Fall Meet. Suppl., Abstract NG62B-0958.
IPCC, 2001: Climate Change 2001: Synthesis Report, A Contribution of Working
Groups I, II, and III to the Third Assessment Report of the Integovernmental Panel on
Climate Change [Watson, R.T. and the Core Writing Team (eds.)]. Cambridge University Press, Cambridge, United Kingdom, and New York, NY, USA, 398 pp.
Nekrasova, A., V. Kossobokov, 2003. Generalized Gutenberg-Richter recurrence law:
Global map of parameters. Geophysical Research Abstracts, 5, 2003. Abstracts of the
Contributions of the EGS-AGU-EGU Joint Assembly, Nice, France, 06-11 April, 2003
(CD-ROM): EAE03-A-03801.
IUGG (CCS)-UNEP-UNESCO, (2005), Fluctuations of Glaciers 1995-2000, World Glacier
Nekrasova, A., V. Kossobokov, Unified Scaling Law for Earthquakes: Mega-cities
and urban agglomerations, Eos Trans. AGU, 86(52), Fall Meet. Suppl., Abstract S23A0229, 2005.
Ivins E.R. and T.S. James, (2004), Bedrock response to Llanquihue Holocene and
present-day glaciation in southernmost South America, Geophys. Res. Lett. , 31 ,
L24613, doi:10.1029/2004GL021500.
J.-M. Nocquet and E. Calais, Crustal velocity field of western Europe from permanent
GPS array solutions, 1996-2001, Geophysical Journal International, 154 (1), 72–88,
doi:10.1046/j.1365-246X.2003.01935.x, 2003.
Kääb, A., F. Paul, M. Maisch, M. Hoelzle, and W. Haeberli (2002), The new remote-sensing-derived Swiss glacier inventory: II. First results, Ann. Glaciol. , 34 , 362–
366.
Kanamori, H., 1977. The energy release in great earthquakes, J. Geophys. Res. 82,
2981-2987.
Kanamori, H., 1977. The energy release in great earthquakes. J. Geophys. Res.,
82: 2981-2987.
Keilis-Borok, V.I., V.G. Kossobokov & S.A. Mazhkenov, 1989. On similarity of spatial distribution of seismic activity. In: Theory and algorithms of interpretation of
geophysical data, Moscow: Nauka, p. 28‑40. (Computational Seismology 22, in Russian).
Kossobokov V.G. & Mazhkenov S.A., 1988. Spatial characteristics of similarity for
earthquake sequences: Fractality of seismicity. Lecture Notes of the Workshop on Global Geophysical Informatics with Applications to Research in Earthquake Prediction
and Reduction of Seismic Risk (15 Nov.‑16 Dec., 1988), ICTP, Trieste, 15 p.
Kossobokov, V., & A. Nekrasova, Generalized Gutenberg-Richter recurrence law.
Geophysical Research Abstracts, 5, 2003. Abstracts of the Contributions of the
EGS-AGU-EGU Joint Assembly, Nice, France, 06-11 April, 2003 (CD-ROM): EAE03-A06597.
Kossobokov, V.G., & A.K. Nekrasova, 2004. Unified scaling law for earthquakes: global map of parameters. In: Analysis of geodynamical and seismic processes. Moscow:
Geos, p. 160-175 (Computational Seismology 35, in Russian).
Kossobokov, V.G., V.I. Keilis-Borok, D.L. Turcotte, and B.D. Malamud, 2000. Implications of a statistical physics approach for earthquake hazard assessment and forecasting. Pure Appl. Geophys., 157: 2323-2349
Kravanja, S., Batini, F., Fiordelisi, A., and Panza, G.F., 2000. Full moment tensor
retrieval from waveform inversion in the Larderello geothermal area. Pure and Appl.
Geophys., 157: 1379-1392.
Kravanja, S., Panza, G.F., and Šílený, J., 1999. Robust retrieval of seismic point
source time function. Geophys. J. Int., 136: 385-394.
Le Meur, E., and R.C.A Hindmarsh (2000), A comparison of tow spectral approaches
for computing the Earth response to surface loads, Geophys. J. Int. , 141 , 282-298,
doi:10.1046/j.1365-246x.2000.00068.x.
Panza G.F., Romanelli F. and Vaccari F., 2001. Seismic wave propagation in laterally heterogeneus anelastic media: theory and applications to seismic zonation. Advances in Geophysics, Academic Press, 43, 1-95.
Panza, G.F., 1985. Synthetic seismograms: the Rayleigh waves modal summation,
J. Geophys., 58, 125-145.
Panza, G.F., 1985. Synthetic Seismograms: the Rayleigh Waves Modal Summation.
J. of Geophys., 58: 125-145.
Panza, G.F., and Saraò, A., 2000. Monitoring volcanic and geothermal areas by full
seismic moment tensor inversion: are non-double-couple components always artefacts of modelling? Geophys. J. Int., 143: 353-364.
Panza, G.F., Peccerillo, A., Aoudia, A., and Farina, B., 2006. Geophysical and petrological modelling of the structure and composition of the crust and upper mantle: the
case of the Tyrrhenian Sea and surroundings. Earth Science Review, in press.
Panza, G.F., Prozorov, A. and Suhadolc, P., 1990. Is there a correlation between lithosphere structure and statistical properties of seismicity? In: R. Cassinis and G. F.
Panza (Editors), The structure of the Alpine - Mediterranean area: contribution of
geophysical methods. Terra nova, 2, 585-595.
Panza, G.F., Romanelli F., and Vaccari, F., 2000, Seismic wave propagation in laterally heterogeneous anelastic media: theory and applications to seismic zonation.
Adv. Geophys., 43: 1-95.
Panza, G.F., Vaccari, F. and Cazzaro, R., 1999. Deterministic seismic hazard assessment. In F. Wenzel et al. (Eds), Vrancea Earthquakes: Tectonics, Hazard and Risk Mitigation, 269-286. Kluwer Academic Publishers, The Netherlands.
Panza, G.F., Vaccari, F., Costa, G., Suhadolc, P. and Fäh, D., 1996. Seismic input modelling for zoning and microzoning, Earthquake Spectra 12, 529-566.
Paul F., A. Kääb, M. Maisch, T. Kellenberger and W. Haeberli, (2002), The new remote-sensingderived Swiss glacier Inventory:I.Methods, Annals of Glaciology , 34 ,
355-361
Paul F., A. Kääb, M. Maisch, T. Kellenberger, and W. Haeberli, (2004), Rapid disintegration of Alpine glaciers observed with satellite data, Geophys. Res. Lett. , 31 ,
L21402, doi:10.1029/2004GL020816.
Peresan, A. and Panza, G.F., 2002. UCI2001: The updated catalogue of Italy, ICTP,
Trieste, Italy, Internal report, IC/IR/2002/3.
Peresan, A., Costa, G., Vaccari, F., 1997. CCI1996: the current catalog of Italy, Internal report IC/IR/97/9. International Centre for Theoretical Physics, Trieste, Italy.
Peresan, A., Panza, G.F., 2002. UCI2001: The updated catalogue of Italy, Internal
report IC/IR/2002/3, ICTP, Trieste, Italy.
Letellier T., Etude des ondes de marée sur les plateux continentaux. Thèse doctorale, Université de Toulouse III, Ecole Doctorale des Sciences de l’Univers, de l’Environnement et de l’Espace, 237pp, 2004.
Peresan, A., V. Kossobokov, L. Romashkova & G.F. Panza, 2005. Intermediate-term
middle range earthquake predictions in Italy: a review. Earth-Science Reviews 69, (12), 97-132.
Mandelbrot, B.B., 1982. The Fractal Geometry of Nature. Freeman, New York, 488 p.
Pondrelli, S., Ekström, G., and Morelli, A., 2001. Seismotectonic re-evaluation of
the 1976 Friuli, Italy, seismic sequence. Journal of Seismology, 5: 73-83.
Manos, G.C. and Demosthenous, M., 1992. Design of R.C. structures according to the
Greek Seismic Code Provisions. Bull. of IISEE, 26, 559-578.
Markusic, S., Suhadolc, P., Herak, M. and Vaccari, F., 2000. A contribution to seismic hazard assessment in Croatia from deterministic modelling, PAGEOPH., 157,
185-204.
Pontevivo, A., and Panza, G.F., 2006. The lithosphere-asthenosphere system in the
Calabrian Arc and surrounding seas – Southern Italy. Pageoph, in press.
Postpischl, D. (1985), Catalogo dei terremoti italiani dall’anno 1000 al 1980.
C.N.R.-Progetto Finalizzato Geodinamica.
Meletti, C. and Valensise, G., 2004. Zonazione sismogenetica ZS9 – App.2 al Rapporto Conclusivo. In: Gruppo di Lavoro MPS (2004). Redazione della mappa di pericolosità sismica prevista dall’Ordinanza PCM 3274 del 20 marzo 2003. Rapporto Conclusivo per il Dipartimento della Protezione Civile, INGV, Milano-Roma, aprile 2004,
65 pp. + 5 enclosures.
Radulian, M., Ardeleanu, L., Campus, P., Šílený, J., and Panza, G.F., 1996. Waveform
inversion of weak Vrancea (Romania) earthquakes. Stud. Geoph. Geod., 40: 367-380.
Ménard, G., Structure et cinématique d’une chaîne de collision: les Alpes occidentales et centrales, Thèse de doctorat d’état, Université Joseph Fourier, Grenoble, 1988.
Reynaud L., M. Vallon, S. Martin, and A. Letreguilly, (1984), Spaziotemporal distribution of the glacial mass balance in the Alpine, Scandinavian and Tien Shan areas,
Geografiska Annaler 66A , 3 , pp. 239-47.
Molchan, G.; Kronrod, 2004, Frequency–magnitude relation for Italy. SAND group
internal report, ICTP, Trieste. Italy.
Romashkova L., 2006, Analysis of the Friuli-Venezia Giulia regional earthquake database, SAND group internal report, ICTP, Trieste, Italy.
Molchan, G.; Kronrod, T.L.; Panza, G.F. 1996 Hazard oriented multiscale seismicity
model: Italy. International Centre for Theoretical Physics. Internal report IC/96/23,
ICTP, Trieste, Italy.
Saraò, A., Panza, G.F., Privitera, E., and Cocina, O., 2001. Non double couple
mechanisms in the seismicity preceding 1991-1993 Etna volcano eruption. Geophys.
J. Int., 145: 319-335.
Monitoring Service, Zurich, http://www.geo.unizh.ch/wgms/
Monti, G., Nuti, C. and Pinto, P. E., 1996. Non-linear response of bridges under multi-support excitation, J. of Structural Engineering, ASCE, 122, 10.
Raykova, R., Chimera, G., Farina, B., and Panza, G.F., 2004. S-wave velocity structure of the lithosphere-asthenosphere system in Mediterranean region, 32nd Int. Cong.,
Abs. Vol., Pt. 2, abs 208-5, 970 pp.
Schlatter A., U. Marti and D. Schneider, (1999), The new national height system
(LHN95) of Switzerland, EUCOR-URGENT , Annual Reports 1999, Subproject 1.1 ,
http://comp1.geol.unibas.ch/report99/reportindex.htm
123
Interreg Project ALPS-GPS Quake Net
Schlunegger F., J. Melzer and G.E. Tucker, (2001b), Climate, exposed source-rock
lithologies, crustal uplift and surface erosion: a theoretical analysis calibrated with
data from the Alps/North Alpine Foreland Basin system, Int J Earth Sci , 90 484–
499, doi:10.1007/s005310100174
Schlunegger, F. and H. Matthias (2001a), Crustal uplift in the Alps: why the drainage
pattern matters, Terra Nova 13 (6) , 425-432. doi: 10.1046/j.1365-3121.2001.00374.x
Sébrier, M., A. Ghafiri, and J.L. Blès, Paleoseismicity in France : fault trench studies
in a region of moderate seismicity, J. Geodyn., 24, 207-217, 1997.
Selvaggi G. et al., La “Rete Integrata Nazionale GPS” (RING) dell’INGV:
Šílený, J., 1998. Earthquake source parameters and their confidence regions by a
genetic algorithm with a “memory”. Geophys. J. Int., 134: 228-242.
Šílený, J., Campus, P., and Panza, G.F., 1996. Seismic moment tensor resolution by
waveform inversion of a few local noisy records-I. Synthetic tests. Geophys. J. Int.,
126: 605-619.
Šílený, J., Panza, G.F., and Campus, P., 1992. Waveform inversion for point source
moment tensor retrieval with optimization of hypocentral depth and structural model. Geophys. J. Int., 109: 259 274.
Sue, C., Martinod, J., Tricart, P., Thouvenot, F., Gamond, J.-F., Fréchet, J., Marinier, D., Glot, J.-P. & Grasso, J.-R. 2000. Active deformation in the inner western
Alps inferred from comparison between 1972-classical and 1996-GPS geodetic surveys, Tectonophysics 320, 17-29.
Sue, C., Thouvenot, F., Frechet, J. & Tricart, P., Widespread extension in the core
of the western Alps revealed by earthquake analysis, J. Geophys. Res., 104 (B11), 25
611-25 622, 1999.
Tapponnier, P., Evolution tectonique du système alpin en Méditerranée: poinçonnement et écrasement rigide-plastique, Bull. Soc. Géol. Fr., 7, 437-460, 1977.
Thouvenot, F., La surveillance sismique du Sud-Est de la France. Risques Infos 13,
5-7, 2002.
Tregoning, P., R. Boers, D. O’Brien and M. Hendy, Accuracy of Absolute Precipitable Water Vapor Estimates from GPS Observations, J. Geophys. Res., 103, 28,70128,710, 1998.
Turcotte, D. L., 1997. Fractals and Chaos in Geology and Geophysics, 2nd edition,
Cambridge University Press, Cambridge
Turcotte, D. L., 1999. Seismicity and self-organized criticality. Phys. Earth Planet.
Int., 111: 275–294.
una infrastruttura aperta per la ricerca scientifica. Atti 10a Conferenza Nazionale
ASITA, 2006, 1749-1754
Vaccari, F., Suhadolc, P. and Panza, G.F., 1990. Irpinia, Italy, 1980 earthquake: waveform modelling of strong motion data, Geophys. J. Int. 101, 631-647.
Vialon, P., P. Rochette et G. Ménard, Indentation and rotation in the Alpine arc, in
Alpine Tectonics, Geological Society of London Special Publication 45, pp. 329-339,
eds. M. Coward, D. Dietrich and R. Park, 1989.
Vuan A., Russi, M., Costa, G., and Panza, G.F., 2001. Moment tensor waveform inversion in the sub-Antarctic Scotia Sea region: feasibility tests and preliminary results. Terra Antartica, 8(2): 55-62.
Walpersdorf, A., S. Baize, P. Tregoning, E. Calais and J.-M. Nocquet, Deformation in
the Jura Mountains (France): First Results from Semi-Permanent GPS Measurements,
Earth Planet. Sci. Lett., Vol. 245, 365-372, 2006.
Wikipedia, 2006, Wikipedia, L’enciclopedia libera. http://it.wikipedia.org/.
Zivcic, M., Suhadolc, P. and Vaccari, F., 2000. Seismic zoning of Slovenia based on
deterministic hazard computations, PAGEOPH, 157, 171-184.
ZS9 working group (2004). Redazione della mappa di pericolosità sismica prevista
dall’Ordinanza PCM del 20 marzo 2003. Rapporto conclusivo per il Dipartimento della
Protezione Civile, INGV, Milano-Roma, aprile 2004, appendice 2.
Zuliani D., Battaglia M., Murray M.H., Michelini A., Burgmann R., Marson I., FREDNet:A continuous GPS geodetic network monitoring crustal deformation in NE Italy.
AGU Fall Meeting, poster presentation, 2002
124
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement