SPATIAL FRAMEWORK FOR STORAGE AND ANALYSES OF

SPATIAL FRAMEWORK FOR STORAGE AND ANALYSES OF
SPATIAL FRAMEWORK FOR STORAGE AND ANALYSES OF
FISH HABITAT DATA IN GREAT LAKES' AREAS OF CONCERN:
HAMILTON HARBOUR GEODATABASE CASE STUDY
A. G. Doolittle, C. N. Bakelaar, and S. E. Doka
Great Lakes Laboratory for Fisheries and Aquatic Sciences
Central & Arctic Region
Fisheries and Oceans Canada
867 Lakeshore Rd. Box 5050
Burlington, Ontario L7R 4A6
August 2010
Canadian Technical Report of Fisheries and
Aquatic Sciences 2879
Fisheries and Oceans Pêches et Océans
Canada
Canada
Canada
Canadian Technical Report of Fisheries and Aquatic Sciences
Technical reports contain scientific and technical information that contributes to
existing knowledge but which is not normally appropriate for primary literature.
Technical reports are directed primarily toward a worldwide audience and have an
international distribution. No restriction is placed on subject matter and the series reflects
the broad interests and policies of DFO Canada, namely, fisheries and aquatic sciences.
Technical reports may be cited as full publications. The correct citation appears
above the abstract of each report. Each report is abstracted in the data base Aquatic
Sciences and Fisheries Abstracts.
Technical reports are produced regionally but are numbered nationally. Requests for
individual reports will be filled by the issuing establishment listed on the front cover and
title page.
Numbers 1-456 in this series were issued as Technical Reports of the Fisheries
Research Board of Canada. Numbers 457-714 were issued as Department of the
Environment, Fisheries and Marine Service, Research and Development Directorate
Technical Reports. Numbers 715-924 were issued as Department of Fisheries and
Environment, Fisheries and Marine Service Technical Reports. The current series name
was changed with report number 925.
Rapport technique canadien des sciences halieutiques et aquatiques
Les rapports techniques contiennent des renseignements scientifiques et techniques
qui constituent une contribution aux connaissances actuelles, mais qui ne sont pas
normalement appropriés pour la publication dans un journal scientifique. Les rapports
techniques sont destinés essentiellement à un public international et ils sont distribués à
cet échelon. II n'y a aucune restriction quant au sujet; de fait, la série reflète la vaste
gamme des intérêts et des politiques de Pêches et Océans Canada, c'est-à-dire les sciences
halieutiques et aquatiques.
Les rapports techniques peuvent être cités comme des publications à part entière. Le
titre exact figure au-dessus du résumé de chaque rapport. Les rapports techniques sont
résumés dans la base de données Résumés des sciences aquatiques et halieutiques.
Les rapports techniques sont produits à l'échelon régional, mais numérotés à
l'échelon national. Les demandes de rapports seront satisfaites par l'établissement auteur
dont le nom figure sur la couverture et la page du titre.
Les numéros 1 à 456 de cette série ont été publiés à titre de Rapports techniques de
l'Office des recherches sur les pêcheries du Canada. Les numéros 457 à 714 sont parus à
titre de Rapports techniques de la Direction générale de la recherche et du développement, Service des pêches et de la mer, ministère de l'Environnement. Les numéros 715 à
924 ont été publiés à titre de Rapports techniques du Service des pêches et de la mer,
ministère des Pêches et de l'Environnement. Le nom actuel de la série a été établi lors de
la parution du numéro 925.
Canadian Technical Report
of Fisheries and Aquatic Sciences 2879
2010
SPATIAL FRAMEWORK FOR STORAGE AND ANALYSES OF FISH HABITAT
DATA IN GREAT LAKES' AREAS OF CONCERN: HAMILTON HARBOUR
GEODATABASE CASE STUDY
By
A. G. Doolittle, C. N. Bakelaar, and S. E. Doka
Great Lakes Laboratory for Fisheries and Aquatic Sciences
Central & Arctic Region
Fisheries and Oceans Canada
Burlington, Ontario L7R 4A6
© Her Majesty the Queen in Right of Canada, 2010.
Cat. No. Fs 97-6/2879E ISSN 0706-6457
Correct citation for this publication:
Doolittle, A.G., Bakelaar, C.N., and Doka, S.E. 2010. Spatial framework for storage and
analyses of fish habitat data in Great Lakes' Areas of Concern: Hamilton Harbour
geodatabase case study. Can. Tech. Rep. Fish. Aquat. Sci. 2879: xi + 68 p.
ii
ABSTRACT
Doolittle, A.G., Bakelaar, C.N., and Doka, S.E. 2010. Spatial framework for storage and
analyses of fish habitat data in Great Lakes' Areas of Concern: Hamilton Harbour
geodatabase case study. Can. Tech. Rep. Fish. Aquat. Sci. 2879: xi + 68 p.
A spatial framework approach for storage and analyses of fish habitat data has
been used to compile physical fish habitat data into a geographic information system
(GIS). This approach synthesizes data from different projects into a geographic database
(geodatabase), it can be utilized as a guide in standardizing formats, data structures and
data layers that are used in generating and mapping key habitat features (vegetation,
substrate, depth). These layers support fish habitat suitability, habitat supply, fish
population and ecosystem models. Difficulties encountered will be discussed as well as
rationale for the approach used. Construction and storage of these spatial layers within a
GIS enables quantitative measurements of habitat and analysis of trends over time and
space.
Hamilton Harbour (Lake Ontario) has been identified as a Great Lakes’ Area of
Concern (AOC) signifying that its ability to support aquatic life has been impaired.
Contributing to the effort to restore this degraded area is in concert with Fisheries &
Oceans Canada’s (DFO) commitment to healthy and productive ecosystems in Canada.
In 1989, DFO’s Great Lakes Laboratory for Fisheries and Aquatic Sciences (GLLFAS)
began a number of projects to assess the current state of Hamilton Harbour. Together,
they will assess progress toward Remedial Action Plan (RAP) targets for fish habitat and
populations of phytoplankton, zooplankton, benthos and fish, and evaluate the ability of
the ecosystem to meet all of the RAP's targets.
Parallels can be drawn from the Hamilton Harbour Area of Concern case study to
DFO’s “place-based” management goals in the Great Lakes because it uses a sciencebased approach to identify the spatial influence of factors that contribute to ecosystem
health. Incorporating scientific, biological and physical information into a
geodatabase/GIS is one method in which data can be synthesised and visualized;
thus, decisions can be based on closer integration among professionals who strive to
manage our damaged ecosystems.
iii
RÉSUMÉ
Doolittle, A.G., Bakelaar, C.N., and Doka, S.E. 2010. Spatial framework for storage and
analyses of fish habitat data in Great Lakes' Areas of Concern: Hamilton Harbour
geodatabase case study. Can. Tech. Rep. Fish. Aquat. Sci. 2879: xi + 68 p.
Nous avons utilisé un cadre spatial pour le stockage et l’analyse des données
relatives à l’habitat du poisson afin de compiler des données physiques dans un système
d’information géographique (SIG). Cette approche fait la synthèse des données de
différents projets dans une base de données géographiques. Elle peut aussi servir de guide
quant à la normalisation des formats, des structures et des couches de données qui sont
utilisés pour la création et la cartographie des caractéristiques essentielles de l’habitat
(végétation, substrat, profondeur). Ces couches de données seront utiles en ce qui a trait à
la qualité de l’habitat, les reserves de l’habitat, la population des poissons et aux modèles
d’écosystème. De plus, l’établissement et le stockage de ces couches spatiales dans un
SIG rendent possibles les mesures quantitatives de l’habitat et l’analyse des tendances
spatio-temporelles. Par ailleurs, il sera question des difficultés rencontrées ainsi que de la
justification scientifique de l’approche utilisée.
Le havre Hamilton (lac Ontario) a été désigné comme secteur préoccupant (SP)
des Grands Lacs, ce qui signifie que sa capacité de servir d’habitat aux organismes
aquatiques s’est vue réduite. L’effort en vue de restaurer cette zone détériorée s’inscrit
dans la foulée de l’engagement pris par le ministère des Pêches et des Océans (MPO)
relativement au maintien d’écosystèmes sains et productifs au Canada. En 1989, le
Laboratoire des Grands Lacs pour les pêches et les sciences aquatiques (LGLPSA) du
MPO a lancé plusieurs projets pour évaluer l’état actuel du havre Hamilton. Ensemble,
les deux entités mesureront les progrès accomplis par rapport aux objectifs du plan
d’assainissement (PA) pour l’habitat du poisson et les populations de phytoplancton, de
zooplancton, de benthos et de poissons, et elles évalueront la capacité de l’écosystème à
atteindre tous les objectifs du PA.
Des parallèles peuvent être établis entre l’étude de cas du secteur préoccupant du
havre Hamilton et les objectifs de la gestion « axée sur le milieu » du MPO pour les
Grands Lacs, puisqu’une approche scientifique est utilisée pour déterminer l’influence
spatiale des facteurs qui contribuent à la santé de l’écosystème. Incorporer de
l’information scientifique, biologique et physique dans un SIG est une méthode grâce à
laquelle les données peuvent être synthétisées et visualisées; ainsi, les décisions peuvent
s’appuyer sur une intégration plus étroite parmi les professionnels qui s’efforcent de gérer
nos écosystèmes endommagés.
iv
TABLE OF CONTENTS
ABSTRACT....................................................................................................................iii
RÉSUMÉ ........................................................................................................................iv
LIST OF TABLES .........................................................................................................vii
LIST OF FIGURES .......................................................................................................viii
LIST OF APPENDICES ...............................................................................................xi
1.0 INTRODUCTION...................................................................................................1
1.1 Spatial Data Framework for Fish Habitat Information ..........................................3
1.2 The Geodatabase....................................................................................................4
1.3 The Stages of Geodatabase Development ............................................................7
2.0 PURPOSE AND OBJECTIVES............................................................................9
3.0 METHODS ..............................................................................................................10
3.1 Hamilton Harbour Geodatabase Stage 1: Design ................................................10
3.2 Hamilton Harbour Geodatabase Stage 2: Input ....................................................12
3.2.1 Base Data (Cartographic Layers).....................................................................14
3.2.2 Sample Data
................................................................................................16
3.2.3 Modeled Data...................................................................................................19
3.2.3.1 Elevation & Bathymetry ...........................................................................20
3.2.3.2 Substrate....................................................................................................27
3.2.3.3 Fetch..........................................................................................................38
3.2.3.4 Slope .........................................................................................................40
3.2.3.5 Turbidity (Water Clarity/Secchi Depth) ...................................................42
v
3.2.3.6 Toxic Sediment .........................................................................................44
3.2.3.7 Aquatic Vegetation ...................................................................................47
4.0 RESULTS ................................................................................................................53
4.1 Hamilton Harbour Geodatabase Stage 3: Testing.................................................53
4.1.1 Fish Habitat Suitability Analysis .....................................................................53
4.2 Hamilton Harbour Geodatabase Stage 4: Implementation ...................................56
4.2.1 The Fish Habitat Data Model...........................................................................57
5.0 DISCUSSION ..........................................................................................................59
6.0 CONCLUSION .......................................................................................................63
7.0 ACKNOWLEDGEMENTS ...................................................................................64
8.0 REFERENCES........................................................................................................65
9.0 APPENDICES .........................................................................................................68
vi
LIST OF TABLES
Table 1 Use impairments ................................................................................................2
Table 2 ESRI supported data objects ..............................................................................5
Table 3 Primary data sources for the Hamilton Harbour geodatabase ..........................13
Table 4 Translation of shoreline survey type to percent substrate composition.............33
Table 5 Discrete backscatter classes identified using acoustic ranges from the
reflected multibeam ...........................................................................................36
Table 6 Backscatter class composition derived using samples from EC and NWRI .....37
vii
LIST OF FIGURES
Figure 1
Hamilton Harbour, Lake Ontario, Canada.....................................................1
Figure 2
Geodatabase stages of development ..............................................................7
Figure 3
Simple relationship between spatial and tabular information........................11
Figure 4
Conceptual grid cell attributes .......................................................................12
Figure 5
Example of base data used in the Hamilton Harbour geodatabase................15
Figure 6
Example of sample points and lines (transects) stored in the Hamilton
Harbour geodatabase .....................................................................................16
Figure 7
Example of field sample data used in the Hamilton Harbour geodatabase ...18
Figure 8
Simple modeling layers in the Hamilton Harbour geodatabase.....................20
Figure 9
Example of Canadian Hydrographic Service (CHS) bathymetric survey
data in Hamilton Harbour from 2002 and 2005.............................................22
Figure 10 Windermere Basin depth soundings from City of Hamilton Public
Works Department, 2005 ...............................................................................23
Figure 11 Example of Hamilton Harbour data provided by Ontario Ministry of
Natural Resources (OMNR) showing features delineated using
Orthoimagery in 2002 ....................................................................................24
Figure 12 Vertical datum corrections for Hamilton Harbour land elevations................25
Figure 13 Seamless Land/Water Digital Elevation Model (DEM) for the Hamilton
Harbour area...................................................................................................27
Figure 14 All substrate samples used in creating a substrate layer for habitat
assessments in Hamilton Harbour..................................................................28
Figure 15 Google Earth™ imagery of shallow shoals that were digitized and
used in creating polygon substrate data for Hamilton Harbour .....................29
Figure 16 Nearshore/offshore substrate interpolation/classification zones in
Hamilton Harbour ..........................................................................................30
viii
Figure 17 Example of shoreline survey segment with sample sheet, Bayfront Park,
Hamilton Harbour ..........................................................................................32
Figure 18 Classified substrate types based on the Shepard's Classification System
to visualize the nearshore substrate composition layer for Hamilton
Harbour ..........................................................................................................35
Figure 19 Mapping of backscatter interpreted classes in Hamilton Harbour using
acoustic ranges from the reflected multibeam ..............................................36
Figure 20 A complete substrate classification layer for Hamilton Harbour based
on the Shepard's Classification System ........................................................38
Figure 21 Average fetch for Hamilton Harbour from prevailing wind (270º) in 2007
(year macrophyte survey was completed)......................................................40
Figure 22 Slope (represented as percent change) for Hamilton Harbour derived
using ArcGIS™ Spatial Analyst showing high to low values .......................41
Figure 23 Example of turbid input from Grindstone Creek in Hamilton Harbour,
delineated using Google Earth™ imagery and used to model final
turbidity layer.................................................................................................42
Figure 24 Final Secchi depth layer for Hamilton Harbour generated using a spline
method to model submergent vegetation .......................................................44
Figure 25 Hamilton Harbour sediment toxicity classes from laboratory toxicity
assays conducted from part sampled sediments.............................................45
Figure 26 Draft sediment toxicity levels in Hamilton Harbour as determined by
interpolating classified point sample sediments using a spline method ........46
Figure 27 Example of an emergent aquatic vegetation polygon from Hamilton
Harbour ..........................................................................................................48
Figure 28 Submergent aquatic vegetation model layers ................................................49
Figure 29 Predicted SAV (% cover) for Hamilton Harbour based on current
condition ........................................................................................................51
Figure 30 Draft aquatic vegetation layer for Hamilton Harbour, highlighting areas
where validation is required and where toxic sediment and water quality
layers may overrule........................................................................................53
Figure 31 Fish Habitat Suitability model requirements .................................................54
ix
Figure 32 Draft output of high, medium and low suitability ranges for Hamilton
Harbour fish communities using the HSA model ..........................................56
Figure 33 Fish Habitat Data Model................................................................................58
Figure 34 Status of spatial data compiled and added to the geodatabase.......................59
x
LIST OF APPENDICES
Appendix 1 Sample records from a Defensible Methods input data file for
Hamilton Harbour. .....................................................................................68
xi
1.0 INTRODUCTION
Located on the west end of Lake Ontario, Hamilton Harbour is situated between
the city of Hamilton (primarily to the south) and the city of Burlington (north) (Figure 1).
Figure 1. Hamilton Harbour, Lake Ontario, Canada.
The harbour is an important hub for shipping activities through the Burlington canal,
primarily for industries along the southeast shoreline in Hamilton. The north shore in
Burlington is mainly residential with mixed recreational use to the west. Recreational
boating is a popular activity in the harbour, supported by various marinas along the north
shore and south west in and around the Bayfront area. Residents in the area can enjoy
1
hiking trails, powered and non-powered watersports, fishing, beaches and cultural
activities, including restaurants, museums, parks and festivals.
Over 200 years of history, the harbour encountered a significant amount of habitat
destruction and pollution. According to the Bay Area Restoration Council (BARC), “By
1926, canals and infill eliminated more than two-thirds of the original wetlands, protected
inlets and shallow areas. By the early 1900s, the harbour ecosystem was severely
degraded as a result of direct sewage discharges, habitat loss, toxic spills and sediment
contamination.” (BARC 2008). Recognizing the need for change, in 1987 the
International Joint Commission (IJC) designated Hamilton Harbour as an “Area of
Concern” (AOC), one of 43 areas identified in the Great Lakes Basin. As such, the
harbour is recognized as having at least one or more of the 14 use impairments identified
for AOCs (Ontario Ministry of Environment and Environment Canada 1992).
Table 1. Use impairments.
I.
II.
III.
IV.
V.
VI.
VII.
VIII.
IX.
X.
XI.
XII.
XIII.
XIV.
Restrictions on fish and wildlife consumption
Tainting of fish and wildlife flavour
Degraded fish and wildlife populations
Fish tumours or other deformities
Bird or animal deformities or reproductive problems
Degradation of benthos
Restrictions on dredging activities
Eutrophication or undesirable algae
Restrictions on drinking water consumption or taste or odor problems
Beach closings
Degradation of aesthetics
Added costs to agriculture or industry
Degradation of phytoplankton and zooplankton populations
Loss of fish and wildlife habitat
In 1992, a Remedial Action Plan (RAP) for Hamilton Harbour was established.
This plan offered various recommendations to improve water quality and environmental
2
conditions for humans and other biota. As a member of the Bay Area Implementation
Team (BAIT), Fisheries and Oceans Canada (DFO) has assisted in the coordination and
implementation of remedial actions required in an effort to delist Hamilton Harbour as an
AOC. This support includes research and expert advice, and addresses a number of use
impairments listed above.
As indicated in a report by Minns et al. (2006), “The health of fish communities,
and their dependent fisheries, is a key indicator of ecosystem health.” The goal of this
document is to explain how spatial data was collected and compiled in a Geographic
Information System (GIS) to support research related to habitat requirements of healthy
fish populations. The development of a data framework for spatial information serves as
a geodatabase in which data is stored, processed and retrieved, and as a key metadata
tool.
1.1 SPATIAL DATA FRAMEWORK FOR FISH HABITAT INFORMATION
The primary driver behind a spatial data framework for fish habitat information is
the ability to store, integrate and process habitat data required and collected by DFO
researchers for further analysis and modeling in the harbour. Key habitat features are
mapped such as macrophyte distributions, substrate composition and depth, and then
applied in fish habitat supply analysis (HSA), population and ecosystem modeling. A
number of factors driving the development of the spatial data framework include:
* Organization (of compiled, collected and partner data)
* Coordination (of data that is supplied by and used by more than one project)
3
* Filling data gaps (identification of spatial data gaps)
* Integration and standardization (between projects and across AOCs)
* Ease of use (facilitate the process of querying, extracting and sharing data).
This spatial data framework, structured within a geodatabase, will serve as a common
storage and management framework for geographic information and spatial data for DFO
science. It includes geographic features, associated attribute tables, point and transect
field information as well as remotely sensed acoustic, satellite and aerial imagery.
1.2 THE GEODATABASE
A “geodatabase” is a geographic database that has the capability of storing spatial
layers and attribute data in a relational database management system (RDBMS). A
geodatabase can be simple or complex, but largely depends on the nature of its
application. There are many advantages of utilizing a geodatabase for storing spatial
data, a few are listed below:

Store a rich collection of spatial data in a centralized location (facilitates data
access and maintenance).

Apply sophisticated rules and relationships to the data (e.g. sample locations must
fall within shorelines).

Define advanced geospatial relational models (e.g. topologies, networks).

Maintain integrity of spatial data with a consistent, accurate database.

Integrate spatial data with other IT databases.

Support custom features and behaviour through programming, scripts and data
models (ESRI 2010a).
4
The geodatabase stores a variety of data “objects”, in many different formats. The
following table represents a sample of supported objects (ESRI 2010b).
Table 2. ESRI supported data objects.
Object
Tables
Feature
Classes
Feature
Datasets
Relationship
Classes
Geometric
Networks
Topologies
Raster
Datasets
Raster
Catalogs
Survey
Datasets
Toolboxes
Behaviour
Rules
Description
non-spatial objects or descriptive information
spatial features, such as points, lines or polygons
containers for feature classes that share the same spatial
reference
manage thematic relationships between tables/feature
classes/both
used in flow network analysis – relationships between point
and line features
spatial relationships within and between feature classes –
used to find and fix spatial errors
gridded data derived from various formats (.img, .jpg,
interpolations, etc.)
tables that reference a collection of raster image files
store survey information and can group survey data into
projects, and multiple projects into a project folder
geoprocessing tools used in the ArcGIS geoprocessing
framework
define legal or permitted attribute values, relationships
between classes, topological relationships between features,
and connections between network features
In addition to the above, a geodatabase supports a variety of spatial modeling, data
management and analysis functions. One of the key benefits is that the implementation
of data structures and validation rules allows users to model reality more closely than was
possible with other data models (ESRI 2004). This approach to modeling benefits the
analysts who manage and manipulate the data directly, and also the scientists who
develop fish habitat and population models. Certain approaches and data structures were
adapted from the Marine Data Model, developed by ESRI and the marine GIS
community. This model assists in the organization and maintenance of data within DFO,
5
and supports modeling and management decisions related to habitat and biota including
water quality.
Commonly known as “Arc Marine”, users in the marine GIS community have
worked to establish a model that supports applications in oceans and coastal areas. This
community, including academic, government and non-government organizations and
researchers, required a database design that facilitated the collection of dynamic and
multidimensional data from the oceans, seas and coasts. It also strived to provide a more
logical way to represent these data in the object-oriented world of the geodatabase
(Wright 2007).
As a result, the Marine Data Model or framework assists users with data input,
storage, and dissemination of data using a pre-configured geodatabase template. It can
also assist in improving performance with data processing (particularly with larger
datasets, such as bathymetry) and analysis of this data (e.g. time series, coastal processes,
etc.). Wright (2007) indicates that the data model improves the ability to manage and
exchange large marine datasets using a framework that can be shared and implemented
across many platforms and applications.
The same benefits have driven the development of a DFO fish habitat data model
that could be applied to AOCs. DFO Science has been working with other organizations
to improve conditions in areas such as Hamilton Harbour, Bay of Quinte, and the Detroit
and St. Clair Rivers in the Huron–Erie Corridor (HEC), in an effort to delist them as
Areas of Concern. A common data model was a logical step in the planning process,
particularly because each AOC has common delisting targets, sampling protocols, and
data needs, such as fish habitat information gathering. Much of this document will focus
6
on the Hamilton Harbour fish habitat data model and resulting geodatabase. The
concepts, methods, and applications can be applied to other areas providing the
appropriate data exists and that the relationships and data structure provide the desired
results.
1.3 THE STAGES OF GEODATABASE DEVELOPMENT
The development of a geodatabase can be conceptualized as a four-step process:
(1) the design stage, (2) the input stage, (3) the test stage, and (4) the implementation
stage (Figure 2).
Figure 2. Geodatabase stages of development.
7
This is an iterative process that may require adjustments at each stage to ensure all
project requirements are met.
1. THE DESIGN STAGE
The design stage is important because it identifies data types collected and
required by researchers. It allows the data manager and scientist to scope additional data
needs and manage expectations based on resources. Often adaptation of the design is
necessary to achieve the desired end-user application, as in our study (e.g. web mapping).
Establishing a flow diagram or conceptual model of the design can be valuable in
documenting needs and examining relationships between the spatial layers and their
attributes. Documenting how information will be used is a prerequisite, particularly with
complex modeling. It is important to structure data in a way that facilitates the
mathematical modeling process, recognizing that information may be processed directly
within a GIS or summarized for use in an external model. Having the ability to link
information back to features is important, and must be addressed in the design stage.
2. THE INPUT STAGE
The input stage often represents the bulk effort of a project as it requires a
significant amount of time with the input and management of data. Data collected or
acquired from different sources may not fit into the standardized data model framework.
Depending upon the location of a project, it may be necessary to establish a common
spatial coordinate system in order to align data. Identification of data gaps may require
further research into different data sources or methods that could be used to generalize
and represent the dataset as a whole (through interpolation), or may even require a revisit
to the design stage to address these concerns.
8
3. THE TESTING STAGE
The testing stage provides an opportunity to query, extract, and report on the data,
identifying whether or not all user requirements have been met. A revisit to the design
stage will be required if they are not.
4. THE IMPLEMENTATION STAGE
The implementation stage represents the final goal of the project. Data is
consolidated, modeling (both mathematical and spatial) is complete, and results have
been generated through mapping and/or reporting on methods used in the spatial analysis.
The foundation for the fish habitat geodatabase model lies within the design or
conceptual model stage, and are examined in greater detail below.
2.0 PURPOSE AND OBJECTIVES
The main objective of this project was to create and store spatial data layers that
will be used by DFO researchers in support of fish habitat suitability, supply and
population modeling (Doka et al., unpublished data). The goal of this report is to
document the stages of development used in the Hamilton Harbour geodatabase. It
explores, in detail, the methods required to create fish habitat layers, the fish habitat
geodatabase model, and the range of data required by DFO researchers to make decisions
and recommendations on remedial actions within the Hamilton Harbour AOC.
9
3.0 METHODS
3.1 HAMILTON HARBOUR GEODATABASE STAGE 1: DESIGN
The design stage is important because it identifies information required to
accomplish certain tasks or achieve certain goals. Visually, this can be achieved with a
conceptual model diagram or a data flow diagram. A conceptual model is a tool that
bridges the gap between a graphical representation of a process and a computer data
processing model. A GIS has been used to model and manage both tabular and spatial
data in a geographic database (geodatabase). Each object in the diagram represents an
object in the geodatabase that is required or necessary to create the final spatial layers for
visualization and future simulation modeling. Each object may have a relationship to an
attribute table (providing more detailed information about the object) or may have a
relationship to another object. For example, survey transect data may be represented as a
sample point (Object A). A sample point has a defined relationship to the transect layer
(Object B) and to an attribute table. A relationship between the point layer and the
survey transect layer provides the user with information about the transect via a data
model (Figure 3).
10
Figure 3. Simple relationship between spatial and tabular information.
Tabular data can be extracted through queries based on key relationship identifiers
(primary key), or unique IDs. Therefore, a sample point sharing a common ID with a
sample transect allows a user to draw the same information from the tabular data.
Relationships can be simple or complex in nature. Complexity reflects the nature
of the project, the available data that is used in the analysis, and the number of variables
needed to identify patterns and relationships in the spatial data for the area of interest.
The Hamilton Harbour data model relies heavily upon a grid format to store information
and perform various spatial analyses. To simplify this concept, the harbour is represented
on a map as evenly spaced grid cells, and for the purpose of this case study, uses a 5 m x
5 m cell size. Unlike traditional cartographic symbols such as points, lines and polygons,
habitat features are related or spatially associated in a series of overlapping grids. If each
grid was merged together, and each cell was converted into a point (at the centroid of
11
each cell), each point would have a number of habitat variables associated to it that would
feed into a model that would help classify each individual grid cell or point as low,
medium or high suitability (Figure 4).
5 m
5 m
Figure 4. Conceptual grid cell attributes.
There are a number of advantages of storing information as a grid:





A simple data structure—A matrix of cells with values representing a coordinate
and sometimes linked to an attribute table
A powerful format for advanced spatial and statistical analysis
Has the ability to represent continuous surfaces and to perform surface analysis
Has the ability to uniformly store points, lines, polygons, and surfaces
Has the ability to perform fast overlays with complex datasets (ESRI 2010c).
The design stage is fundamental in identifying layers and information needed to
accomplish the spatial analysis objectives. Completion of this stage facilitates the
transition into the second stage, data input.
3.2 HAMILTON HARBOUR GEODATABASE STAGE 2: INPUT
DFO researchers required a number of key datasets in order to complete the
classification and analysis of fish habitat in Hamilton Harbour. These datasets were
physical, chemical or biological in nature. Each is regarded as a key component of fish
habitat, indicating the overall health of aquatic habitats (Minns et al. 2006). Data input
for these geodatabase layers required 1) assembly and synthesis of data (e.g. emergent
and submergent vegetation, depth, substrate, and toxicity) and 2) tabulating the layers for
12
use in habitat suitability modeling used to classify the habitat of the study area depending
on fish usage. Data sources for the Hamilton Harbour geodatabase can be found in Table
3.
Table 3. Primary data sources for the Hamilton Harbour geodatabase.
In reference to the Hamilton Harbour geodatabase model, this section will
examine three types of data:
1. Base Data
2. Sample Data
3. Modeled Data
 data that can be used for general mapping, spatial
reference
 data that represents field collected sample information
used to support spatial analysis
 base and sample data compiled and spatially modeled
for fish habitat applications and analysis
Each component contributes to the input stage of geodatabase development. Information
and data is compiled and layers are created that reflect the nature of the spatial
phenomena or the format of the data structure needed for further modeling. A number of
13
challenges can arise in this stage particularly when compiling a single modeled GIS layer
from multiple data sources (e.g. substrate).
3.2.1
Base Data (Cartographic Layers)
Base data features represent layers in the data model that can be used as inputs for
the simple layer generation, mapping, context, and even further analysis. Point layers in
the base data include geographic features (for labeling maps or locating named features),
and elevation points (including spot elevations, bathymetry points). Line features
represent road networks, stream networks, contours and shorelines. Polygon features
represent geographic boundaries (including townships, municipalities, etc.), watersheds,
and waterbody polygons (e.g. permanent and intermittent). Orthoimagery and satellite
imagery is valuable as a cartographic layer as it provides a snapshot picture of the study
area – useful to compare temporally with other imagery, or to be used in interpretation or
map display (Figure 5).
14
Figure 5. Example of base data used in the Hamilton Harbour geodatabase.
15
3.2.2 Sample Data
Sample data within the geodatabase supports layers used for modeling habitat,
often represented as point or line (transect) locations (Figure 6).
Figure 6. Example of sample points and lines (transects) stored in the Hamilton
Harbour geodatabase.
Point sample data has been provided by various organizations including DFO,
Environment Canada (EC), and other agencies that have sampled in the harbour. While
not all data collected and stored is relevant or necessary to the habitat modeling, often
information could be extracted or derived from the sample data to add value to one or
many habitat model input layers (e.g. fish community interaction or Secchi values for
16
submergent vegetation modeling). Examples of sample data types include fish species,
zebra mussel abundance, zooplankton and phytoplankton density, benthic samples,
emergent or submergent aquatic vegetation (SAV), substrate type, temperature profiles,
dissolved oxygen concentration, toxicity sediment classes, substrate classes (using
acoustics) and Secchi depth. This list is not static and additional variables or sample data
could be incorporated into the dataset, if necessary. Temperature and dissolved oxygen
profiles or time series are also collected by field crews in an effort to capture as much
information about a sample location as needed to accurately reflect water quality and
habitat conditions. Several in situ monitoring stations were set up across the harbour for
extended periods to capture much needed temporal data or key limnological processes
that may impact the fish habitat availability. All the series data have not been
incorporated into the current geodatabase, but are being analysed for future incorporation
of key time series elements (i.e. seasonal patterns).
Sample transect data, represented as linear geographic features (point A to point
B), have also been incorporated into the geodatabase. DFO researchers have collected
information including fish community data, macrophyte densities, temperatures taken at
start/middle/end location, dissolved oxygen (same as temperature) and substrate (using
acoustics). Much of this work is on-going, adding temporal information to the transect
data (Figure 7).
17
Figure 7. Example of field sample data used in the Hamilton Harbour geodatabase.
18
3.2.3 Modeled Data
Modeled data represents base and sample data that has been combined and
manipulated to create layers required for fish habitat analysis and modeling. In reference
to the fish habitat data framework, these modeling layers have been classified as either
“simple” or “complex”:
Simple:
 a single thematic layer or feature (e.g. depth)
 Can be complicated to assemble because it requires data and locationspecific interpretation
Complex:
 a single thematic layer created by combining simple layers to produce a
unique output (e.g. submergent aquatic vegetation)
Or
 Integrated multiple spatial data layers with statistical models (e.g.
Habitat Suitability Analysis)
Simple modeling layers include bathymetry (elevation), substrate, fetch, slope, turbidity,
aquatic vegetation and toxic sediments (Figure 8).
19
Figure 8. Simple modeling layers in the Hamilton Harbour geodatabase
3.2.3.1 Elevation and Bathymetry: “Historically, mean sea level (MSL) has been
used as the zero of elevation.”(Mahoney 2010). Conceptually, this fixed reference point
is used to derive the elevation of a geographic location. This elevation reflects a vertical
datum or reference point against which measurements are made. This vertical datum
differs significantly from nautical chart or bathymetric data (or chart datum), which, for
safety reasons, identifies the minimum depth of water that could occur at any point
20
(Canadian Hydrographic Service 2010). Therefore, synthesizing land and water
elevations requires an adjustment so that all values reflect a single datum, whether land or
chart-based, directed by the nature or goal of a project. For this project, all elevation
values have been adjusted to chart datum (International Great Lakes Datum 1985, or
IGLD85), maintaining a high level of accuracy in the bathymetric survey data, and
facilitating the generation of a bathymetry layer using elevation values. Bathymetry data
is one of the most important components to aquatic habitat modeling as it defines and
describes the topology (shape) of the underwater space and its features that broadly
define habitat for fishes.
Highly detailed bathymetric point data was assembled from single and multibeam
surveys completed by the Canadian Hydrographic Service (CHS) in 2002 and 2005
(Leyzack, CHS, Burlington, pers. comm.). This data (n > 5 000 000), represented as
points with associated depth values, were corrected to IGLD85, which is 74.2 m in
Hamilton Harbour (Figure 9).
21
Figure 9. Example of Canadian Hydrographic Service (CHS) bathymetric survey
Data in Hamilton Harbour from 2002 and 2005.
Computer-Aided Design data (CAD) from Windermere Basin was provided by the City
of Hamilton, Public Works Department (Helka, Public Works Dept., Hamilton, pers.
comm.). Soundings were extracted from CAD drawings and used in generating
bathymetry points to align with CHS data and elevation values for the harbour (Figure
10).
22
Figure 10. Windermere Basin depth soundings from City of Hamilton Public Works
Department, 2005.
The Ontario Ministry of Natural Resources (OMNR) provided a number of terrain
datasets (derived from the Greater Toronto Area Orthophoto Project in 2002) for the
Hamilton Harbour study area. The orthophoto project used soft-copy photogrammetric
techniques to produce a highly accurate and precise elevation dataset. This was used to
generate a digital elevation model (DEM) at a resolution of 5 m (+/- 0.5-1 m horizontal
and vertical accuracy) along with other data products, like linear features such as
shorelines, islands, breakwalls and waterbodies (Figure 11). Cartographically, these
features are used in map production, to reflect a current picture of the harbour. They are
also used to define extents of land and water features, and also to select data from the
23
DEM that could be used for interpolating elevations in the nearshore area where it was
too shallow for bathymetric survey equipment.
Figure 11. Example of Hamilton Harbour data provided by Ontario Ministry of
Natural Resources (OMNR) showing features delineated using
Orthoimagery in 2002.
Methods: Interpolation of elevations in the nearshore area was necessary to bridge
the gap between adjacent land and seafloor elevations. Shoreline features were extracted
from the CAD data (based on air photo interpretation done in 2002) to acquire an
accurate and recent representation of the harbour. These shoreline features were used to
identify the average extent of the water, or to be used as a “mask” for interpolation
purposes. A 50 m buffer of the water features was created and used to extract elevation
data from the detailed land DEM. Centroids of the grid cells were converted to points,
24
and land elevation values were adjusted from the current height reference system
(Canadian Geodetic Vertical Datum or CGVD28) to IGLD85 (i.e. subtracting 0.102 m
from land elevations as a correction factor) (Herron, CHS, Burlington; pers. comm) using
vertical benchmark data from the area (Sauvé, NRCAN, Ottawa, unpublished data
(Figure 12).
Figure 12. Vertical datum corrections for Hamilton Harbour land elevations.
Features that were not included in the bathymetric survey (man-made islands, docks,
breakwalls, etc.) were added after the interpolation to ensure they were captured as part
of the assessment.
Survey information from CHS was converted into an elevation value based on the
IGLD85 datum of 74.2 m above sea level (5.0 m depth = 69.2 m elevation, or 74.2 m
minus 5 m). These points were merged with the land elevation data into one layer.
25
Using ArcGIS™, a spline interpolation method was used to generate an elevation
grid. This technique “minimizes the overall surface curvature, resulting in a smooth
surface that passes exactly through the input points…and is best for generating gently
varying surfaces such as elevation.” (ESRI 2010d). The “tension” option was chosen in
an effort to constrain the results based on the character of the data being modeled – in
other words, to reflect the original sample data as closely as possible. Where land
elevation values did not exist (particularly with small restoration islands), a value of 74.2
m (0 m depth) was assigned to ensure these features were not lost. These areas were
spatially merged with the spline elevation grid. The final elevation grid facilitates the
calculation of depth values using a standard calculation (Datum or water level elevation –
interpolated elevation), and can be applied and/or modified to address different water
level scenarios.
Final Elevation Layer: The final elevation layer (map) is a seamless coverage
from land to water and can be used to represent both elevation and bathymetry (Figure
13). Since most of the work related to this project requires depth information, grid cell
elevation values less than 74.2m (cut off for dry land) were extracted and a new
bathymetry grid layer was created for modeling under a low water scenario
(standardized).
26
Figure 13. Seamless Land/Water Digital Elevation Model (DEM) for the Hamilton
Harbour area.
3.2.3.2 Substrate: Identifying and classifying substrate for Hamilton Harbour
was a prerequisite for classifying fish habitat. A number of organizations have looked at
classifying substrate type based on sample data collected at point locations (Rukavina and
Versteeg 1995). For this analysis, data was assembled from various sources to compile
and develop a comprehensive spatial layer of substrate compositions based on both
qualitative assessments and quantitative grain size analysis.
Based on the assumption that the bottom composition of the harbour has not
changed dramatically over the time of the surveys, point sample data has been obtained
from a variety of sources and temporally spanned several decades. Core sample data
collected by the National Water Research Institute (NWRI) from the 1980s and 1990s
27
(NWRI 1995) provided a solid foundation for the substrate layer. Recent point sample
data was also used from DFO to fill gaps. Projects related to zebra mussels, electrofishing
transect habitat surveys, and targeted habitat sampling (that included ponar, shoreline and
acoustic surveys) provided quantitative and qualitative substrate data. Sample data has
also been provided by EC (Milani 2010, unpublished data), CHS, and the Hamilton
Harbour RAP (through detailed designs of restoration projects) (Hall, Hamilton Harbour
RAP, Burlington, pers. comm.). Figure 14 represents all point samples used in creating a
substrate layer for habitat assessments.
Figure 14. All substrate samples used in creating a substrate layer for habitat
assessments in Hamilton Harbour.
28
In addition to point samples, a shoreline survey (with photos and samples) was
conducted in 2006-2007 (Doka, in prep) was also used to attribute shoreline segments
(shoreline features provided by the OMNR) with a general substrate composition.
Polygon data contributing to the substrate layer was largely based on a visual assessment
of detailed orthoimagery (Google 2007). The harbour high resolution photos provided
data for features not captured in sample data, such as shallow shoals (Figure 15).
Figure 15. Google Earth™ imagery of shallow shoals that were digitized and
used in creating polygon substrate data for Hamilton Harbour.
A final detailed source of substrate was multibeam backscatter data provided by
CHS, used in conjunction with sample points in the “offshore” area of the harbour to
29
classify substrate types into 4 discrete categories based on smoothness, size and
composition.
Methods: Two areas (or zones) were used in the creation and interpretation of the
final substrate layer – the nearshore zone and the offshore zone. Each zone represented a
different spatial challenge for interpolation/classification methods based on available data
(quantity and quality) and its interpretation (Figure 16).
Figure 16. Nearshore/offshore substrate interpolation/classification zones in
Hamilton Harbour.
Nearshore: Data compiled from a number of sources were assigned to specific
classes (bedrock, boulder, cobble, rubble, gravel, sand, silt, clay, hardpan, pelagic)
described in Minns et al. (2006) and based on a modified Wentworth scale (Bain and
30
Stevenson 1999). Each substrate sample point was attributed with the percent
composition of each class (e.g. 50% gravel, 50% sand) so that the total percent summed
to 100%. Quantitative grain size samples used these classes listed above but could only
capture grain size small enough to field sample (but percent composition are more
reliable). Qualitative field measurements/visual assessments of substrate types classified
as dominant, subdominant and trace were allocated percent values (post survey).
Dominant and subdominant values were allocated percent compositions based on a ratio
of to ⅔ to ⅓ (66% - 33%). If a visual sample had dominant, subdominant and trace
values assigned, it was allocated percent compositions based on a 60:30:10 split (i.e.
60%-30%-10%).
It was necessary to use shoreline characteristics to fill gaps in the coastal area
where substrate point data was sparse. A number of field methods were used to describe
and assign the shoreline types and then to use those types to assign the substrate
composition needed for modeling (e.g. bedrock, boulder, cobble, rubble, etc.) These
methods included data collection with Global Positioning Units (GPS) to identify unique
reaches, site photos for verification, and orthoimagery (Figure 17). See Doka et al. (in
prep) for detail regarding the shoreline survey.
31
Figure 17. Example of shoreline survey segment with sample sheet, Bayfront Park,
Hamilton Harbour.
Coordinates were taken with a GPS at each change in shoreline composition, and
observations such as shoreline type, land-use and surficial nearshore geology were
documented. Each segment represents a shoreline type different from its neighbour.
Substrate samples were also taken at midpoints of segments to quantify and determine if
shoreline substrate classifications could predict nearshore substrate type.
32
The following table outlines changes made to the original data based on how the
original characteristic might emulate a type of substrate:
Table 3. Translation of shoreline survey type to percent substrate composition.
Original Description
Sand
Clay/Silt
Gravel
Cobble
Gabian cribs
Crib Dock
Rubble
Boulder
Armour Stone
Artificial Fill
Bedrock
Steel Wall
Wooden Wall
Sand Barrels
Zebra Mussels/Shells
% Functional Composition
100% Sand
50% Clay, 50% Silt
100% Gravel
100% Cobble
50% Cobble, 50% Boulder
33% Rubble, 64% Cobble
100% Rubble
100% Boulder
100% Boulder
100% Cobble
100% Bedrock
100% Bedrock
100% Bedrock
100% Boulder
100% Gravel
Shoals and man-made habitat features visible in the orthoimagery were digitized
into polygons. Habitat structures used in restoration projects (materials ranging from
cobble to armour stone blocks) were classified as either 100% cobble or armour stone,
and subsequently converted to 5 m x 5 m grid cells.
A mask of the nearshore zone (areas <7 m) was created to facilitate the
interpolation of sample points found within this zone, which is much different from the
offshore substrates as it offers a more realistic representation and method for
interpolation. Based on this notion, all point data were classified into two categories:

Fine Substrate (gravel, sand, silt, clay)

Coarse Substrate (bedrock, boulder, cobble, rubble)
33
The fine substrate point data were interpolated using a spline function to create a
“smooth” surface within the nearshore zone. In an effort to closely reflect original
sample data values, the “tension” option was chosen (for details see spline interpolation
description in Elevation/Bathymetry methods section).
Coarse substrate points were buffered by 10 m and converted to grid cells, then
superimposed on the soft sediment grid. This approach is based on the transitional nature
of sand and silt areas transitioning to clay substrates in deeper waters throughout the
harbour; much of the coarse substrate materials sampled (e.g. restoration structures,
scattered boulders, shoals, etc.) are either rare in the offshore, or associated with mainly
man-made shoreline features in the nearshore. A final grid representing all substrate (fine
and coarse) in the nearshore was created; each cell’s composition summing to 100%
(Figure 18). Shoreline segments (line features) were converted into a 5 m grid cell raster
to align with the original digital elevation model. This would eventually be used to
supercede grid cell values extrapolated to the edge of the harbour.
34
Figure 18. Classified substrate types based on the Shepard's Classification System
(Poppe et al. 2003) to visualize the nearshore substrate composition layer
for Hamilton Harbour.
Offshore: CHS collected multibeam bathymetric data for the harbour (Leyzack,
CHS, Burlington, pers. comm.). This survey provided depth information (>5 m depth),
and also contributed to the modeling of substrate data through expert interpreting and
classifying backscatter data (Tekmap, unpublished data) collected from the multibeam
Simrad EM3000 system.
The data stream from the Simrad system includes both depth and amplitude data.
The amplitude data (or backscatter) data are a function of the angle at which the sonar
beam reflects off the seafloor (grazing angle), the smoothness of seafloor, and the
seafloor composition. After applying a series of backscatter correction functions, a
35
simple reclassed map was created identifying four discrete classes (Table 5) representing
backscatter acoustic ranges from the reflected multibeam: (Figure 19) (Tekmap,
unpublished data).
Table 5. Discrete backscatter classes identified using acoustic ranges from the
reflected multibeam.
Class
1
2
3
4
Backscatter start (dB)
0.0
30.7
33.7
39.1
Backscatter end (dB)
30.7
33.7
39.1
50.0
Figure 19. Mapping of backscatter interpreted classes in Hamilton Harbour using
acoustic ranges from the reflected multibeam.
36
Classified backscatter information was averaged in the offshore zone. Offshore
substrate samples were used to allocate a percent composition to each backscatter
category. Sample points (processed samples) were spatially joined to grid cells at the
same location. All values from the same backscatter class were averaged. The
composition results are shown in Table 6.
Table 6. Backscatter class composition derived using samples from EC and NWRI.
Class
1
2
3
4
% Composition
0% gravel, 8% sand, 49% silt, 43% clay
0% gravel, 19% sand, 44% silt, 37% clay
1% gravel, 22% sand, 44% silt, 33% clay
0% gravel, 36% sand, 39% silt, 25% clay
As with the nearshore zone, points with larger substrata (> sand) were buffered and
superimposed on the final grid layer to ensure this information was not lost due to the
interpolation of predominantly fine substrate.
Final Substrate Layer: Interpolated and classified values within the nearshore
zone, combined with the offshore classification, created a complete substrate layer with
no data gaps (Figure 20). Represented as a 5 m cell size grid, this layer is used as a stand
alone product, a component in predictive modeling of macrophytes (SAV), and as an
input into fish habitat suitability supply or population models.
37
Figure 20. A complete substrate classification layer for Hamilton Harbour based on
the Shepard's Classification System (Poppe et al. 2003).
3.2.3.3 Fetch: Fetch can be defined as “…the unobstructed distance that wind
can travel over water in a constant direction” (USGS 2008). As a component to modeling
submerged aquatic vegetation, fetch plays a key role in determining whether or not
vegetation is able to colonize. Using a wind fetch and wave model created by the United
States Geological Survey (USGS 2008), fetch was calculated for the harbour. This model
generates fetch data at user-specified wind direction angles using a grid and specified
shoreline. The extent of bathymetry grid for the harbour was used as the primary input
for the fetch model. The model assumes that the input raster is properly projected in
38
Methods: The fetch model requires a number of inputs:



a land raster dataset – values > 0 indicate land, <= 0 or NODATA indicate water
a wind direction list – text file containing values of wind directions (angles
from which fetch data is needed)
a calculation method – three different calculation methods are available, 1) SPM,
2) SPM restricted, and 3) Single.
A land raster dataset was created from the DEM. Land grid values were re-classed
to 99 (i.e. Elevation > 74.2 m) and water grid values were set to 0. In 2007, there was a
dominant westerly wind (or 270º) in Hamilton Harbour, and a text file was created to
reflect these average conditions. The “SPM” method was chosen as it uses a
recommended procedure from the Shore Protection Manual (U.S. Army Corps of
Engineers 1984), spreading 9 radials around the wind direction in 3-degree increments
and averaging the values.
Final Fetch Layer: The result of the “SPM” model is a grid of values which
identifies the distance to shore based on the directions identified in the text file. With a
dominant westerly wind in 2007, the year that the submergent vegetation was sampled, it
is apparent that higher fetch values are found in the east end of the harbour. This grid
was used in the generation of a submergent vegetation model for the harbour (Leisti,
Bouvier and Doka, pers. comm.) (Figure 21) as wind driven forces determine vegetation
presence (Baird 1996). However, fetch could also be useful for hydrodynamic and other
biotic models.
39
Figure 21. Average fetch for Hamilton Harbour from prevailing wind (270º) in 2007
(year macrophyte survey was completed).
3.2.3.4 Slope: In previous studies, the relationship between slope and
macrophyte growth has been examined. According to Duarte and Kalff (1986), “there is
a great influence of the slope of the littoral on the biomass of submerged macrophyte
communities.” This conclusion based largely upon the physical stability of the sediment,
and impacts of erosion. The bathymetry grid was used to generate slope values for the
harbour based on elevation changes.
Methods: Slope can be calculated in degrees or as a percentage, and the lower the
calculated slope value, the flatter the surface. Percent slope was obtained using ArcGIS
™ Spatial Analyst, derived by calculating the maximum rate of change between each grid
40
cell and its neighbours (rise/run * 100). For example, the steepest downhill descent for
the cell (i.e. the maximum change in elevation over the distance between the cell and its
eight neighbors) (ESRI 2010e).
Final Slope Layer: The final slope layer was a 5 m x 5 m grid resolution where
each cell represented a percent change in slope value (Figure 22). On average, the
steepest slopes were found in the 2-5m depth range, along breakwalls, and on the
southeastern shore where excavation/dredging deposits had occurred.
Figure 22. Slope (represented as percent change) for Hamilton Harbour derived using
ArcGIS™ Spatial Analyst showing high to low values.
41
3.2.3.5 Turbidity (Water Clarity/Secchi Depth): Turbidity refers to how clear the
water column is. High concentrations of particulate matter can modify light
penetration….reduced significantly, macrophyte growth may be decreased.”(NRRI
2010). Particularly within Hamilton Harbour, water clarity impacts aquatic vegetation
growth by restricting light from penetrating into the water column. Apart from sediment
re-suspension from within the harbour, there are turbid inputs such as the outflows of
Grindstone Creek in the west (Figure 23) and Indian Creek in the east.
Figure 23. Example of turbid input from Grindstone Creek in Hamilton Harbour,
delineated using Google Earth™ imagery and used to model final
turbidity layer.
42
Point data from DFO (Doka et al. unpublished data) and EC (Hiriart-Baer et al.,
unpublished data) was compiled to create a layer for spatial water clarity. This layer
would be used as an input to the SAV model. Secchi depth values were collected at a
number of locations across the harbour (although spatial coverage was poor), often
seasonally and occurring at different depths (maximum depth of 3 m).
Methods: Point data for Secchi depths were used to interpolate a turbidity layer
for Hamilton Harbour. A Spline method (with a smoothing factor of 0) was used in an
effort to spatially represent Secchi depths in the harbour as a whole. Grindstone Creek
and Indian Creek are known to have high turbidity values, resulting in “plumes” was
extend out into the harbour. Based on orthoimagery, the average extent of these plumes
were captured in a GIS and merged with the Secchi grid. Values in these areas are
known to have limited macrophyte growth, and in an effort to model macrophyte
coverage, lower Secchi values have been attributed to these plume areas.
Final Secchi Depth Layer: The Secchi layer represents a preliminary assessment
of collected samples and interpreted turbid input in the harbour (Figure 24). Further
analysis is needed to ensure that the interpolated layer accurately represents the
characteristics of the phenomena being represented. This includes an investigation into
turbid inputs such as Grindstone Creek, Indian Creek and Cootes Paradise, and their
temporal or permanent influences on water clarity that may restrict macrophyte growth in
the harbour.
43
Figure 24. Final Secchi depth layer for Hamilton Harbour generated using a Spline
method to model submergent vegetation.
3.2.3.6 Toxic Sediment: A sediment toxicity layer was created to define areas
that might be toxic to aquatic vegetation and biota, including fishes or their habitat. This
layer serves as a “mask” to spatially identify highly toxic areas that should be remediated
or avoided for restoration initiatives.
Sample point data provided by EC (Milani and Grapentine 2006b) has been
classified into distinct levels of toxicity based on lab assays: Non-toxic, potentially toxic,
toxic, and severely toxic (Milani and Grapentine 2006a). In total, 177 sediment samples
collected between 2000 and 2006 were provided. The spatial distribution of samples are
directly correlated to reference sites throughout the harbour, with significant clustering of
44
samples in known, highly-toxic areas especially the Randle Reef area to the south and the
Windermere Arm to the east (Figure 25).
Figure 25. Hamilton Harbour sediment toxicity classes from laboratory toxicity
assays conducted from part sampled sediments (Milani and Grapentine
2006b).
Methods: Sample points were interpolated using a Spline method (with a
smoothing factor of 0) in an effort to spatially represent toxicity in the harbour as a
whole. This is currently a draft output. Additional statistics will be conducted to ensure
that areas are not under/over represented, and that the interpolated results represent an
accurate picture of toxic sediment in the harbour.
45
Final Toxicity Layer: The final toxicity layer represents a generalized map of
toxicity in Hamilton Harbour. The interpolation method makes obvious assumptions
about distributions of contaminants and toxic zones (evident in some of the larger areas
represented by one point). Likely, further research is needed to investigate spatial
patterns, including potentially modeling sediment transport, as well as current and wave
impacts on nearshore sediments. A draft interpolated result can be seen in Figure 26
highlighting some of the issues raised and zones potentially needing further investigation
(Marvin, pers. comm.).
Figure 26. Draft sediment toxicity levels in Hamilton Harbour as determined by
interpolating classified point sample sediments using a Spline method
(Milani and Grapentine 2006a).
46
3.2.3.7 Aquatic Vegetation: Generation of a final aquatic vegetation layer
requires two vegetation input layers, including (1) emergent vegetation, and (2)
submergent vegetation. Each will be discussed in detail below.
1. Emergent Vegetation
“Marshes are typically characterized by emergent vegetation and relatively high
oxygen levels in the rooting zone. The vegetation often shows distinct zonation with
changes in water depth and exposure to wave action.” (Newmaster et al.1997). Emergent
plants provide valuable cover and habitat for aquatic species, including nursery habitat
for young fish and adults, and spawning habitat for some species.
Emergent vegetation for Hamilton Harbour was provided by the OMNR.
Methods: Emergent vegetation or wetland areas that were visible or classified
from the 2002 orthoimagery were converted into a 5 m resolution wetland grid for the
Hamilton Region. When converting vegetation layers into cover, an assumption was
made that emergent wetlands represent high density cover (100% cover). Another
assumption would be that emergent vegetation extents changed or new wetlands have not
appeared since that time and this generally represents current conditions.
Final Emergent Aquatic Vegetation Layer: The final layer represents emergent
vegetation in Hamilton Harbour, which is represented as a 5 m x 5 m grid from 2002
(Figure 27).
47
Figure 27. Example of an emergent aquatic vegetation polygon from Hamilton Harbour.
2. Submergent Vegetation
Submergent aquatic vegetation is used by all trophic levels of the ecosystem,
providing life-cycle necessities including nutrients and shelter. Mapping SAV requires a
significant amount of field time to effectively capture the spatial distribution within a
given area because remote aerial sensing may not work and ground truthing is necessary.
With a surface area of approximately 200 km2, an exhaustive spatial survey of Hamilton
Harbour was not feasible. Therefore, the development of an SAV model was necessary
to predict SAV presence and percent cover in Hamilton Harbour from empirical
48
relationships and various datasets including field transect data (Leisti, Bouvier and Doka,
pers. comm.).
Primary inputs to the percent cover model included elevations, slope, and
effective fetch. Depth (derived from the elevation layer) and Secchi were also used to
predict SAV presence based on light penetration to support plant growth in different
turbidity zones (Figure 28).
Figure 28. Submergent aquatic vegetation model layers.
Methods: Using a multiple linear regression equation for percent cover, input
grids were combined into a percent cover value and then classified. In the establishment
or growth of vegetation, certain variables are weighted higher in the predictive regression
model than others. The following formula was applied:
49
Percent Cover SAV = 86.3783 + (-0.7201 * [ps]) + (-10.4607 * [d]) + (-0.0099 * [ef]) +
([d] - 2.3082) * ([d] -2.3082) * -3.3981 + ([d] - 2.3082) * ([ef] –
1299.6220) * 0.0026
ps =
d=
ef =
percent slope
depth
effective fetch at @270º (Leisti, Bouvier and Doka. pers. comm.)
The final classification assigns a percent value to each cell based on every value
or variable input into the model. Grid cells with a depth > 5.75 m (maximum depth of
colonization), as well as those that have a Secchi value of < 0.6 m (insufficient light
penetration to support plant growth), were removed from the analysis (set to 0% cover).
Submergent Aquatic Vegetation Layer: The preliminary SAV layer represents a
synthesis of habitat features or characteristics that are required or limit macrophyte
growth (Figure 29). This layer will likely be further modified for Hamilton Harbour
specific conditions that could further limit this predicted coverage of SAV, especially by
toxic sediments in slips and the Randle Reef area where vegetation growth has yet to be
verified.
50
Figure 29. Predicted SAV (% cover) for Hamilton Harbour based on current conditions.
Aquatic Vegetation: Aquatic plants, both submergent and emergent, provide a
foundation and support for local food webs and are indicators of healthy systems
(Jeppesen et al1998). In Hamilton Harbour, a GIS layer was created to capture the spatial
extent of all aquatic vegetation, both emergent and submergent.
Methods: The HSA model requires vegetation to be classified into categories of
emergent, submergent, or no cover. Combining results from the submergent aquatic
vegetation layer (predicted % cover) and the wetlands layer (100% cover) provided a
means of generating the layer needed. While there are limited emergent wetlands within
Hamilton Harbour, these two layers could potentially overlap. In these situations, a
51
hierarchical method was used and tested to assign a value. If emergent vegetation exists,
and there is no predicted submergent cover, then the emergent value is used (e.g. 100%
emergent). If there is overlap of emergent and submergent data, the predicted SAV value
takes precedence and the remainder is classified as percent emergent (e.g. 50% SAV,
50% emergent). If the predicted cover does not equal to 100% and there is no emergent,
the remainder is % no cover (e.g. 60% SAV, 40% no cover). Where there is no vegetated
cover, it is classified as 100% no cover.
Final Aquatic Vegetation: The final aquatic vegetation layer is represented as a 5 m grid,
containing submergent, emergent and no cover % values (Figure 30).
Figure 30. Draft aquatic vegetation layer for Hamilton Harbour, highlighting areas
where validation is required and where toxic sediment and water quality
layers may overrule.
52
4.0 RESULTS
4.1 HAMILTON HARBOUR GEODATABASE STAGE 3: TESTING
The testing stage provides an opportunity to query, extract, and report on the data,
and evaluate if all project requirements have been met. For this case study, the testing
stage reflects various steps used in generating input and receiving output from the habitat
models, as well as any input layers used directly in the models.
4.1.1 Fish Habitat Suitability Analysis
Fish habitat suitability analysis for Hamilton Harbour uses a number of habitat
layers. At minimum, these layers can be used:

Vegetation: Submergent vegetation and emergent vegetation are combined to
represent an overall percent cover (percent no cover, percent
emergents, percent submergents).

Depth:
Categorized into classes including 0-1 m, 1-2 m, 2-5 m, 5-10 m,
>10 m.

Substrate:
Percent composition categorized into bedrock, boulder, cobble,
rubble, gravel, sand, silt, clay, hardpan, pelagic (Figure 31).
53
Figure 31. Fish Habitat Suitability model requirements.
Methods: A spatial overlay of three habitat layers produced a new layer that
combined all habitat characteristics and attributes together. A layer with unique
combinations of habitat characteristics was generated, which lumped areas with identical
habitats while ensuring a unique identifier could be used to remap the output results back
and summarized into a table that was used as input to the HSA. Specifically,
requirements of the model include:

A unique ID (assigned based on unique combinations of all variables)

Total area (of each unique combination)

Area type (changed or unchanged) for pre- and post-scenario assessment

Depth (classified into categories: 0-1 m, 1-2 m, 2-5 m, 5-10 m, 10 m+)

Substrate (% of bedrock, boulder, cobble, rubble, gravel, sand, silt, clay, hardpan
or pelagic)

Vegetation (classified into categories: % no cover, % emergent, % submergent).
54
For a detailed look at the input table, see the example in Appendix 1.
Fish Habitat Suitability Layer - Example: Once these variables were processed by
the model (e.g. Minns et al. 1997) the output from the model was remapped and classified
by habitat suitability for different life stages (adult/spawning/young of the year), and
thermal preference (cold, cool, warm). Combining all species together for a composite
suitability index reflects the fish community’s habitat needs as a whole, ranked from 0
(low suitability) to 1 (high suitability). For a detailed description into habitat suitability
matrix (HSM) input requirements and output results, see the report from Minns et al.
(2006). Initial results from the HSA model can be seen below, and is used to illustrate
the type of model output. When testing the output, it is possible to identify where
adjustments are required (e.g. Randle Reef) in either the data or the modeling (Figure 32).
55
Figure 32. Draft output of high, medium and low suitability ranges for Hamilton
Harbour fish communities using the HSA model.
4.2 HAMILTON HARBOUR GEODATABASE STAGE 4: IMPLEMENTATION
The implementation stage is the desired goal or “end-result” of the geodatabase.
The framework used in generating the output for the HSA is one of the final products of
the geodatabase, although the results are still preliminary. Implementing the fish habitat
data model (structure, format and concepts) will facilitate work in other areas, especially
AOCs, where an area-based or an ecosystem approach is required.
56
4.2.1 The Fish Habitat Data Model
The Fish Habitat Data Model integrates information from a variety of sources, in
a variety of ways. Consolidated “views” of habitat information from different agencies
have contributed to the generation of thematic spatial data (represented as points, lines
and polygons). These core spatial datasets provide the foundation to the development of
other complex or modeled fish habitat layers, and provides components to other modeling
efforts, such as HSA models.
For modeling fish habitat, the value and importance of incorporating raster data
becomes apparent. Orthoimagery, one form of raster data, can be used as a mapping
reference, but also as an indicator of change in a study area over time (with a series of
images). Using a raster to map data can also be an effective method to present
information (such as elevation, temperature, etc.) as a continuous surface. Much of the
modeled data is continuous, and opportunities exist to include other valuable information,
such as temperature. Categorizing raster data into discrete classes provides a means for
grouping and mapping thematic data. This method was used in a number of grid layers
including the toxicity layer (grouped by level of toxicity) and the final habitat suitability
layer.
Development of each spatial layer or grid can be completed separately, however,
spatially it is important to ensure that these grid cells align spatially. In doing so, one can
visually recognize spatial patterns, further analyse and assess this information, and
demonstrate to others how the spatial influence of factors (e.g. aquatic vegetation) can
contribute to the overall health of ecosystems. Figure 33 represents a “simplified
framework” of the Hamilton Harbour geodatabase.
57
Figure 33. Fish Habitat Data Model.
58
5.0 DISCUSSION
Hamilton Harbour has been used as a case study to demonstrate a method of
integrating spatial information into a data model framework that will assist habitat
managers in making informed decisions about fish habitat requirements and delisting
targets. This type of data model may be applied to other AOCs, such as the St. Clair
River and the Detroit River in the Huron-Erie Corridor. Work has already been
completed in the Bay of Quinte (Minns 2008), with the exception of new data to add.
The following chart (Figure 34) identifies the status of the work to date:
Figure 34. Status of spatial data compiled and added to the geodatabase.
Much of the data has been collected for the Detroit and St. Clair Rivers. The “Inprogress” status is reflected by some of the issues or challenges listed below.
As with any given project, outlining the successes and challenges is beneficial in
both the documentation process and also in applying knowledge to new projects as
“lessons learned”. One of the key successes of the project was the development and
compilation of GIS layers needed by DFO researchers that are suitable for use in
59
scientific models, including fish habitat models. Assembled from different sources, these
layers represent a synthesis of data that has been collected, and reflects the expertise of
the individuals that collect it.
Another key success is the ability to apply the fish habitat framework (concepts
and design) for Hamilton Harbour to other areas. This portable approach will streamline
the decision-making process (with regards to data requirements) and also facilitate data
processing. It will also assist in identifying and anticipating challenges for other AOCs
where fish habitat and fish populations are considered to be impaired, particularly with
recognizing data gaps.
Applying the concepts and design from this case study will help implement an
“area-based” approach to managing resources. This type of comprehensive approach
looks at the “management of human activities in a place rather than dividing management
according to individual sectoral activities wherever they occur” (Young et al. 2007). In
addition, Young states “The boundaries of ecosystems are difficult to define….. the
boundaries of governance systems can be distinct (as lines on maps) but often have little
to do with the spatial structure of either the biophysical or human dimensions of marine
ecosystems.” From a Great Lakes’ perspective, using a GIS as a tool can assist with
integration and spatial understanding of factors influencing ecosystems. The output helps
with collaborative planning and adaptive management of resources locally, regionally,
and internationally.
A side benefit of this project is the communication between data managers, GIS
specialists and research partners. Recalling the stages of data model development,
researchers are involved in the design, input and testing of spatial data. With the
60
improvement of technology and the growing need and case for spatial analysis using GIS,
this tool is invaluable, especially for area-based management. GIS provides new ways of
storing, presenting, analyzing and modeling simple and complex spatial data. The
development of a “fish habitat framework” data model allows DFO researchers to share
their knowledge, expertise and advice with other stakeholders and managers, particularly,
in identifying fish habitat supply (using depth, substrate, vegetation and other data used
to develop layers needed for habitat modeling), but also by learning from the various data
management and spatial layer creation challenges in the process.
A number of challenges, including project and data-related issues, were noted
throughout this document. One in particular was the scope of the project and of the
geodatabase. Completion of the geodatabase work was limited to Hamilton Harbour
AOC initially, and while significant progress has been made on the other AOCs in terms
of data collection and processing, there is still a considerable amount of work to be
completed for spatial input and analysis. In Hamilton Harbour, historic and future habitat
conditions have yet to be analysed.
Another challenge encountered relates to the data model and representing the
complex relationships between the GIS layers in a web environment. Specifically,
presenting this information in a web mapping application required a different
geodatabase model that is compatible and “web” friendly. The original layers had to be
scaled down into a simplified format that could easily be accessed and queried. This
approach is much different than structuring the data in a traditional, normalized RDBMS.
Recognizing the nature of the data, and the type of information collected and
compiled into the geodatabase, it is often difficult to share with partners due to data
61
sharing constraints and agreements. This does not necessarily restrict access to the
interpreted results, but the raw information used to generate some of the GIS layers.
While still a challenge, it can often be rectified easily and, depending on the nature of the
application or organization, agreements can be made on an ad-hoc basis for sharing.
While data accessibility can often pose a challenge, there are a number of inherent
data challenges when using information from external organizations or third parties. It is
common that the condition of the data needs to be addressed prior to implementing in an
analysis or map display. Specifically:

formatting: spreadsheets or tables need to be compatible with GIS input format

georeferencing: identifying the projection and datum to be used/that were used,
and aligning with those of the project

GIS feature type: translating data features into a useful data type (e.g. transect
lines into points)
Other notable spatial or data challenges include:
1. Scale – all data is not available at the same scale, which dictates the level of detail
possible in analysis and layer generation (e.g. national vs. provincial data).
2. Accuracy – when more than one temporal dataset is available, one must choose
the most appropriate representation of the time period being studied (e.g. historic,
current, future shoreline features) or make assumptions.
3. “Alignment” of related layers – implications often arise from choices made in (2)
with subsequent data collected to fill gaps or new layers created from assumptions
based on expert opinion (e.g. bathymetry).
4. Consistency within a dataset – lack of standardized classifications and qualitative
versus quantitative observations often require interpretation (e.g. substrate).
5. Data gaps – where no data exists, proper interpolation methods need to be
evaluated and implemented (e.g. SAV).
62
6. Cumulative error – within a dataset/layer – compounded by all the challenges
mentioned above it is difficult to quantify uncertainty but this must be considered
when evaluating results and using layers for further modeling.
7. Application – particularly with other systems and geographic areas, data model
and spatial methods may need to be adjusted (e.g. riverine vs. lake including flow
modeling/currents).
Identifying and recognizing these challenges is valuable for the GIS analyst and
researchers, and will assist in the communication of impacts, gaps and uncertainties and
future decisions throughout the course of the project. As with all assessments and
models, caveats are necessary to their application and will be stated up front.
6.0 CONCLUSION
The fish habitat data model for Hamilton Harbour is useful for making decisions
regarding fish habitat targets. Recognizing some of the challenges can aid in identifying
future effort (i.e. gap filling), particularly with data collection in the nearshore zone (1-5
m depth) in the case of Hamilton Harbour. With the standard base layers completed,
other potential impacts to fish habitat supply can be examined, including temperature,
oxygen and contaminant loading issues in the harbour. Much of this work is on-going,
particularly within Hamilton Harbour and Bay of Quinte.
Applying this framework to other AOCs or degraded areas is certainly not trivial,
particularly with some of the challenges outlined in this document. Identifying the
habitat layers needed is the first goal and certainly not the most difficult step. Under
different circumstances, data gaps may exist in each area, and subsequently models must
adapt to reflect these changes. On the other hand, new information may exist in these
63
areas, offering new elements to be included in the modeling process. Recognizing the
value of a GIS and its ability to present and synthesize information will assist in the
management of resources and also the recognition and understanding of ecosystem
interactions. A spatial framework for fish habitat information is a building block for
defining these relationships.
7.0 ACKNOWLEDGEMENTS
This study was funded from the DFO Great Lakes Action Plan 2006-2010. A
number of agencies have supported this work with spatial data, including Environment
Canada, Ontario Ministry of Natural Resources, City of Hamilton, Hamilton Harbour
RAP, Natural Resources Canada, and DFO, including Fish Habitat Science and Canadian
Hydrographic Service. A complete list of all data contributors can be found in Table 3.
The authors would also like to acknowledge Lynn Bouvier and Kathy Leisti
(DFO) for their contributions, particularly with the submergent macrophyte model, as
well as Terese Herron (CHS) for her assistance with vertical benchmarks and datums.
Special thanks to Robert Randall (DFO) and John Hall (EC) for their thorough
review and comments on this technical report.
64
8.0 REFERENCES
Bain, M.B. and N.J. Stevenson, eds..1999. Aquatic habitat assessment: common methods.
American Fisheries Society. Bethesda, MD. 216 p.
Baird, W.F. and Associates. 1996. Defensible methods of assessing fish habitat: physical
habitat assessment and modelling of the coastal areas of the lower Great Lakes. Can.
Manuscr. Rep. Fish. Aquat. Sci. 2370: 38+ pp.
Bay Area Restoration Council. Bay History. [Online],
http://www.hamiltonharbour.ca/whysave-about.htm (accessed 12 July, 2008).
Canadian Hydrographic Service. Vertical Datums. [Online],
http://www.lau.chs-shc.gc.ca/english/VerticalDatums.shtml (accessed 18 January, 2010).
Doka, S.E. Fisheries and Oceans Canada. 867 Lakeshore Rd., Burlington, ON L7R 4A6.
Unpublished Data.
Doka, S.E., Leisti K., Doolittle, A.E., Avlijas, S., Bouvier, L.. In Prep. Shoreline survey
of Hamilton Harbour, 2007. Can. Data Rep. Fish. Aquat. Sci..
Duarte, M. and J. Kalff. 1986. Littoral slope as a predictor of the maximum biomass of
submerged macrophyte communities. Limnol. Oceanogr. 31(5), 1986, 1072-1080
© 1986, by the American Society of Limnology and Oceanography, Inc.
ESRI. Geodatabase: Data Storage and Management for GIS. [Online],
http://www.esri.com/software/arcgis/geodatabase/index.html (accessed 1 March, 2010a).
ESRI. Geodatabase: Data Storage. [Online],
http://www.esri.com/software/arcgis/geodatabase/data-storage.html (accessed 1 March,
2010b).
ESRI. What is Raster Data? [Online],
http://webhelp.esri.com/arcgisdesktop/9.2/index.cfm?TopicName=What_is_raster_data%
3F (accessed 25 February, 2010c).
ESRI. Applying a Spline Interpolation [Online],
http://webhelp.esri.com/arcgisdesktop/9.2/index.cfm?TopicName=Applying_a_spline_int
erpolation (accessed 1 April, 2010d).
ESRI. Calculating Slope. [Online],
http://webhelp.esri.com/arcgisdesktop/9.2/index.cfm?TopicName=Calculating_slope
(accessed 6 April, 2010e).
ESRI. 2004. Building Geodatabases I (Course Material).
65
Google Earth. 2005.
Hiriart-Baer, V., J. Milne, and M. Charlton. Environment Canada. 867 Lakeshore Rd.
Burlington, ON L7R 4A6. Unpublished Data.
Jeppesen, E., M. Søndergaard, M. Søndergaard and K. Christoffersen (eds.).1988. The
structuring role of submerged macrophytes in lakes. New York, Springer-Verlag Inc.
ISBN 0-387-98284-1. 423 p.
Mahoney, M.J. A discussion of various measures of altitude. [Online],
http://mtp.jpl.nasa.gov/notes/altitude/altitude.html (accessed 31 March, 2010).
Milani, D. and L.C. Grapentine. 2006a. The application of BEAST sediment quality
guidelines to sediment in Hamilton Harbour, an Area of Concern. NWRI Contribution
No. 06-407. National Water Research Institute, EC, Burlington, Ontario.
Milani, D. and L.C. Grapentine. 2006b. Identification of toxic sites in Hamilton Harbour.
NWRI Contribution No. 06-408. National Water Research Institute, EC, Burlington,
Ontario.
Milani, D. 2010. Indicators of Randle Reef, Hamilton Harbour recovery: monitoring of
benthic conditions 2005 to 2007. Interim Report on Sediment Toxicity and
Bioaccumulation. EC, Burlington, Ontario. March 2010.
Minns, C.K., J.E. Moore, M. Stoneman, and B. Cudmore-Vokey. 2001. Defensible
methods of assessing fish habitat: lacustrine habitats in the Great Lakes basin –
conceptual basis and approach using a habitat suitability matrix (HSM) Method. Can.
Manuscr. Rep. Fish. Aquat. Sci. 2559: viii + 70p.
Minns, C.K., A. Bernard, C.N. Bakelaar, and M. Ewaschuk. 2006. A fish habitat
classification model for the upper and middle sections of the Bay of Quinte, Lake
Ontario. Can. Manuscr. Rep. Fish. Aquat. Sci. 2748: vii + 61p.
Newmaster, S.G., A.G. Harris, and L.J. Kershaw. 1997. Wetland plants of Ontario. Lone
Pine Publishing, Edmonton, AB, Canada. 240 p.
NRRI (Natural Resources Research Institute). Turbidity in Lakes. [Online].
http://lakeaccess.org/russ/turbidity.htm (accessed 21 April, 2010).
Ontario Ministry of Environment and EC. 1992. Remedial Action Plan for Hamilton
Harbour: RAP Stage 2. Hamilton, Ontario.
Poppe, L.J., A.H. Eliason, , and M.E. Hastings. 2003, A Visual Basic program to classify
sediments based on gravel-sand-silt-clay ratios. Computers and Geosciences, v. 29, no. 6,
p. 805-809.
66
Rukavina, N.A. and J.K. Versteeg. 1995. The physical properties of the surficial
sediments of Hamilton Harbour. NWRI Lakes Research Branch Technical Note
No. 95-150.
Sauvé, P. Natural Resources Canada. 615 Booth Street, Ottawa, ON K1A 0E9.
Unpublished Data.
U.S. Army Corps of Engineers. 1984. Shore Protection Manual. Fort Belvoir, VA,
Coastal Engineering Research Center.
USGS. Application of wind fetch and wave models for habitat rehabilitation and
enhancement Projects. [Online],
http://www.umesc.usgs.gov/management/dss/wind_fetch_wave_models.html (accessed 5
July, 2008).
Wright, D.J., M.J. Blongewicz, P.N. Halpin, and J.Breman. 2007. Arc Marine: GIS for a
Blue Planet. ESRI Press, California, USA. 202 p.
Young, O.R., G. Osherenko, J. Ekstrom, L.B. Crowder, J. Ogden, J.A. Wilson, J.C. Day,
F. Douvere, C.N. Ehler, K.L. McLeod, B.S. Halpern, R. Peach. 2007. Solving the crisis in
ocean governance: place based management of marine ecosystems. Environment, Vol.
49, No. 4, p. 20-32. Heldref Publications, Philadelphia, PA.
67
9.0 APPENDICES
Appendix 1. Sample records from a Defensible Methods input data file for Hamilton
Harbour.
;HAMILTON HARBOUR DEFENSIBLE METHODS
*UnitType=Area
*Units=m2
*Order=ID,Area,AreaType,Depth,Substrate,Vegetation
*Proportions=Depth:Z0_1,Z1_2,Z2_5,Z5_10,Z10+
*Proportions=Substrate:Bedrock,Boulder,Cobble,Rubble,Gravel,Sand,Silt,Clay,Hardpan,Pelagic
*Proportions=Vegetation:NoCover,Emergent,Submergent
…
10000,25,UNCH,"0,0,100,0,0","0,0,0,0,2,51,35,12,0,0","62,0,38"
10001,25,UNCH,"0,0,100,0,0","0,0,0,0,2,51,35,12,0,0","79,0,21"
10002,25,UNCH,"0,0,100,0,0","0,0,0,0,2,51,36,11,0,0","85,0,15"
10003,25,UNCH,"0,0,100,0,0","0,0,0,0,2,51,40,7,0,0","44,0,56"
10004,25,UNCH,"0,0,100,0,0","0,0,0,0,2,51,42,5,0,0","56,0,44"
10005,25,UNCH,"0,0,100,0,0","0,0,0,0,2,51,42,5,0,0","77,0,23"
10006,25,UNCH,"0,0,100,0,0","0,0,0,0,2,51,47,0,0,0","45,0,55"
10007,25,UNCH,"0,0,100,0,0","0,0,0,0,2,52,23,23,0,0","66,0,34"
10008,25,UNCH,"0,0,100,0,0","0,0,0,0,2,52,24,22,0,0","49,0,51"
10009,25,UNCH,"0,0,100,0,0","0,0,0,0,2,52,25,21,0,0","61,0,39"
10010,25,UNCH,"0,0,100,0,0","0,0,0,0,2,52,25,21,0,0","67,0,33"
10011,25,UNCH,"0,0,100,0,0","0,0,0,0,2,52,25,21,0,0","68,0,32"
10012,25,UNCH,"0,0,100,0,0","0,0,0,0,2,52,25,21,0,0","70,0,30"
10013,25,UNCH,"0,0,100,0,0","0,0,0,0,2,52,26,20,0,0","44,0,56"
10014,25,UNCH,"0,0,100,0,0","0,0,0,0,2,52,26,20,0,0","57,0,43"
10015,50,UNCH,"0,0,100,0,0","0,0,0,0,2,52,26,20,0,0","59,0,41"
10016,25,UNCH,"0,0,100,0,0","0,0,0,0,2,52,26,20,0,0","72,0,28"
10017,25,UNCH,"0,0,100,0,0","0,0,0,0,2,52,26,20,0,0","78,0,22"
10018,25,UNCH,"0,0,100,0,0","0,0,0,0,2,52,27,19,0,0","56,0,44"
10019,25,UNCH,"0,0,100,0,0","0,0,0,0,2,52,27,19,0,0","57,0,43"
10020,25,UNCH,"0,0,100,0,0","0,0,0,0,2,52,27,19,0,0","59,0,41"
…
68
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