A DECADE IN THE LIFE OF A COASTAL INFORMATION SYSTEM

(1) A.G. Sherin , (1) P. Fraser, S. Solomon, D.L. Forbes, K.A. Jenner, S. Hynes, T. Lynds, (2) P. Gareau

(1) Natural Resources Canada, Geological Survey of Canada, Dartmouth, Nova Scotia, (CA)
(2) Helical Systems, Dartmouth, Nova Scotia, (CA)

Abstract
In 1994, the Geological Survey of Canada (GSC) developed a prototype system for coastal information using a commercial geographic information system (GIS) and its dynamic segmentation feature.
The system consists of the following components:

1. A standard definition of coastal features and a hierarchical coastal feature classification system.
2. An extensible conceptual and physical data model for coastal information.
3. Customized software for data capture, mapping, and analysis.

The system has been used for several purposes including:

1. Provision of digital data to other parts of the Canadian government as a contribution to a national database for oil spill sensitivity.
2. Provision of base information for the creation of a new national park.
3. Provision of information to support land-use and environmental planning for rural communities.
4. Provision of information to support research on climate-change impacts on vulnerable coasts and the development of erosion hazard maps for these coasts.
5. Provision of information to support coastal resource inventories by communities participating in stewardship education and diversification of local economies.

The new data model for the CIS, how it integrates with other parts of the information infrastructure for the GSC, examples of applications and a vision for future enhancements are presented.

Introduction
In 1994, the Geological Survey of Canada (GSC) developed a prototype system for coastal information using a commercial geographic information system (GIS) and its dynamic segmentation feature (Sherin and Edwardson 1994 & 1997, Jenner et al. 2003). The Coastal Information System (CIS) has been modified on several occasions and as of fall 2003 uses ArcGIS 8.2 and ArcInfo 7.2. This work built upon earlier efforts at the GSC, predating the easy availability within the GSC of GIS systems, that used a commercial data base management system (Fricker and Forbes 1988) and later investigations into appropriate data structures for georeferenced coastal information (Sherin et al. 1992).

Development of the CIS had two major objectives: 1) to enhance the coastal mapping work of the GSC by taking advantage of extensive coastal aerial video surveys conducted since 1983 (Frobel and Taylor 2002). Video imagery has been obtained for the entire coast of three Canadian Atlantic Provinces (Nova Scotia, New Brunswick, and Prince Edward Island), the island portion and parts of the Labrador portion of the Province of Newfoundland and Labrador, all of Lake Winnipeg (Manitoba), and several areas in the Canadian Arctic (in the Yukon, Northwest Territories and Nunavut); and 2) to support research on physical coastal processes and hazards by providing digital maps and a foundation for analysis.

Since 1994, interpretation of physical coastal features from much of the airborne video has been captured into the CIS, accomplishing the detailed mapping of substantial parts of the coast line in southeastern Canada and the western Canadian Arctic in 13 discrete data sets. The locations of the 13 CIS data sets are shown in Figure 1.

System components
The system consists of the following components:

1. A standard definition of coastal features and a hierarchical coastal feature classification system.
The definition of coastal features used is a modified version of the one published by the Resource Inventory Committee for the Province of British Columbia (Howes et al. 1994). The coastal feature classification system classifies coastal features on two themes (i.e. form and material) and at four levels of hierarchy (i.e. super-type, type, sub-type, and feature). Examples of the classification system are presented in Table 1. Since 1994 changes have occurred to the classification system with additions such as a boulders, regressive thaw failure, and ice.

2. An extensible conceptual and physical data model for coastal information.
The CIS models the coast as a line and uses linear reference systems and dynamic segmentation. A discussion on the suitability of modeling the coast as a line and the use of linear reference systems and dynamic segmentation is found in Sherin (2000). The CIS maps the extent of individual coastal features and stores their attributes rather then an a priori division of the coastline into segments. This approach was chosen to ensure more objective interpretation of the physical attributes of the coast and enabled the mapping and analysis of actual coastal features. Aggregated classifications of larger coastal segments can be constructed from the detailed data in the CIS.
An entity-relationship diagram of the conceptual data model is presented in Figure 2. The database structure implemented in 2003 has only two event tables one for coastal units at one location (i.e. at a point) and one for coastal units that extend along the coast (i.e. linear). This is a simplification of earlier data models used by the CIS where form and material information were stored in separate tables to allow for one to many relationships. A coastal unit could have none, one or more forms and be composed of none, one or more materials. In practice, this flexibility was not used and complicated queries. The data dictionary for the LINE_UNIT table now used in the CIS is presented in Table 2. Two other aspects of the data model of note are the items for features for both form and material and change mapping (i.e. FORM_FEATURES, MAT_FEATURES and CHANGE_MAPPING). These items could have been implemented as separate tables but for easy implementation and queries have been included as delimited text strings. The data model is now implemented in ArcGIS 8.2 geodatabase.

3. Customized software for data capture, mapping, and analysis.
Data capture software has not been upgraded to ArcGIS 8.2 but remains operational in Arc Marco Language (AML) in ArcInfo 7.2. There are many features of ArcGIS that will allow the easy conversion of the present custom capabilities to the new environment. The digitize events tool of ArcGIS 8.2 will be customized as follows to match the existing capabilities of the ArcInfo 7.2 data capture tool: a) control of the language of form and material attributes; b) permit choosing entries from drop down menus; c) save information for attributes that don't change very often (e.g. MAPPER); and d) use the end measure of a unit as the start measure of the next to ensure events are continuous.

Special tools have been developed in AML to produce maps and support analysis. Tools to study coastal form adjacency relationships along the coast of the Avalon Penninsula of Newfoundland were developed to explore along coast analysis functionality (Jenner et al. 2003). A tool for the development of erosion hazard indices using data from the CIS was also developed. The results of use of this tool for the Beaufort Sea coast of Canada's Arctic will be discussed under applications (Solomon and Gareau 2003).

Applications
The CIS has been used for several purposes. The geographic locations for the individual applications are shown on Figure 1.

National Sensitivity Mapping Program
Environment Canada has prepared for most coasts of Canada, databases and digital atlases of environmental sensitivities to support spill responses (Laflamme et al. 2001). The CIS was used to provide the physical description of the coast for several Priority Areas of Response (PAR).

Meally Mountain National Park, Newfoundland and Labrador
Parks Canada asked the GSC to prepare a database of coastal physical features in support of the development of the Meally Mountain National Park in southern Labrador. Unlike most of the CIS data sets, the Meally Mountain National Park coastal interpretations were made from vertical aerial photos, the only comprehensive coastal imagery available. The scale of these photos limited the mapped resolution of features and severely inhibited the interpretation of materials.

Land-use and environmental planning

Voiseys Bay, Newfoundland and Labrador
To support environmental planning and assessment for the marine infrastructure for the proposed Voiseys Bay nickel mine, the Department of Fisheries and Oceans (DFO) asked the GSC to map the physical features of Kangeklualuk Bay, the proposed site for nickel ore loading facilities. A new type, named boulder, was added to accommodate coastal nearshore and foreshore features composed of boulders which are common in Labrador. Sub-types included scattered boulders, boulder flats and boulder barricade, a concentration of boulders in the nearshore caused by ice push. DFO also required the mapping of seaweed distribution in the nearshore. The CIS was easily modified to accommodate this requirement.

Pictou County, Nova Scotia
In collaboration with the Province of Nova Scotia and local community organizations in Pictou County (northern Nova Scotia), the GSC used the CIS to document the physical characteristics for the Pictou County's coast in the southern Gulf of St. Lawrence. Two major products were delivered to the community organizations to support their business objectives of land-use and environmental planning. A CD-based product developed using MapInfo, contained a complete database of the physical coastline features converted from ArcInfo 7.2 and a digital copy of the aerial videos selectable from a GIS based map. MapInfo was chosen for delivery of this product because one of the community organizations with GIS technical expertise used MapInfo and agreed to support its use within the community. A second product (Taylor et al. 2002) was based upon a second aerial video survey conducted in 2000. Anthropogenic features (e.g. wharves, coastal protection installations) were mapped and entered into the CIS and compared with the CIS data interpreted from an aerial video survey in 1988. More abundant, continuous and larger shore protection structures were found in 2000. The increase in backshore structures was just over 1 % between 1988 and 2000 but the increase was concentrated along a few short segments of shoreline. This application demonstrates one of two techniques for the use of the CIS to investigate coastal change. Another method is discussed in the next application. The work with land-use planning organizations in Pictou County lead to constructive consultations with the Nova Scotia Association of Land-use Planning Executives to guide the GSC's development of coastal information resources for land-use planners.

Climate-change impacts on vulnerable coasts and the development of erosion hazard indices
Shaw et al. (1998) identified sections of Canada's coastline most sensitive to the effects of sea-level rise. The GSC and partners have conducted extensive research in two such areas, viz. parts of the southern Gulf of St. Lawrence coast in Prince Edward Island (southeastern Canada) and most of the Beaufort Sea coast in the Yukon and Northwest Territories (western Canadian Arctic). The CIS played a valuable role in both projects.

Southern Gulf of St. Lawrence, Prince Edward Island
In Prince Edward Island, a multi-faceted study was undertaken (McCulloch et al. 2002) to determine potential impacts of sea-level rise and climate warming in the coastal zone and to identify possible adaptation options. The project encompassed extensive coastal seabed mapping in addition to shore-zone studies, and included detailed climatological and oceanographic studies of relative sea-level change, storm surges and coastal flood modelling, high-resolution topographic mapping of flood zones using airborne laser altimetry, coastal process studies, socio-economic impacts analysis, and adaptation and planning policy. Data from the CIS were incorporated into map products integrating these data sets and a prototype erosion hazard index was developed using CIS data (Forbes and Manson 2002).

Beaufort Sea, Yukon and Northwest Territories
On the Beaufort Sea coast, ice-bonded sediment and permafrost are important factors in determining coastal stability. The CIS classification system was extended to include a number of ice-related forms and ice was added as a material. Coastal change was mapped using an approach different from that used in the Pictou County project described above. In the Beaufort Sea project, sections of coastline were attributed according to the extent and nature of change (e.g. aggradation, retreat, re-orientation, detached barrier, retrogressive thaw failure stability). Expanding on the prototype erosion hazard index developed in PEI, an erosion hazard index (EHI) was developed for the Beaufort Sea coast using the CIS data (Solomon and Gareau, 2003). Additional morphological attributes had to be added including exposure, storm-surge water level elevations, and nearshore slope. These additional morphological attributes along with various types of backshore and foreshore forms and attributes, were assigned scores (see Tables 3 & 4). The EHI was then calculated for each segment of coast by summing all of the scores. The distribution of scores and class rank for the EHI is shown in Table 5. About half of the coastline is characterized by high to very high erosion hazard. The score was compared to historical point measurements of coastal erosion. There was good agreement between the index and the retreat rate calculated from measurements for segments of coast with high retreat rates but correlation at lower measured retreat rates was poor. Since the hazard index is based on aerial video-derived coastal mapping and a conceptual model of high latitude coastal processes, it is a guide to locations that exhibit geological and morphological susceptibility to coastal erosion and not a substitute for site-specific engineering design studies. Incorporation of sea ice, especially the duration of the open water season, and oceanographic data and/or more extensive coastal change measurements may improve the correlation between the hazard index and measured erosion rates.

Coastal resource inventories
DFO in collaboration with local communities conducted coastal resource inventories for the entire Atlantic coast of the Province of Nova Scotia. The inventories were part of DFO's coastal resource stewardship education program and contributed to defining opportunities for local economic diversification. CIS mapping was used in several of these projects. Two other CIS projects demonstrated the use of the CIS in support of the diversification of local economies. In collaboration with an environmental consulting firm, a CIS data set was developed to support the environmental assessment of a marine gas pipeline landing location near Shelburne, Nova Scotia. A demonstration CIS set was developed near Lunenburg, Nova Scotia to show the utility of CIS data for the sea kayaking tour industry.

Integration with the GSC Information Infrastructure

Metadata
The CIS data sets are documented in comprehensive product level metadata that meets current Federal Geospatial Data Committee (US) standards. The metadata can be discovered on the Internet through Canada's national metadata discovery portal (www.geoconnections.org) and the Canadian Geoscience Knowledge Network (www.cgkn.net). Both web locations are part of Canada's initiative to develop a national geospatial information infrastructure.

Web publication
The GSC's coastal research activities and information about the CIS are accessible on the web (www.gsca.nrcan.gc.ca/coastweb). This site includes a web-based mapping application that shows the geographic extent of aerial video surveys and CIS data sets.

Future Enhancements

Conclusions
The Coastal Information System has moved from a concept to an operational system supporting the research, partnership and outreach activities of the Geological Survey of Canada. The system has demonstrated its utility in a number of coastal sectors, including coastal environmental and land-use planning, coastal process and climate-change research, and contributions to local economic diversification. It has shown some promise, through the development of erosion hazard indices and analysis of coastal change to assist in the prediction of the future coastal change. This application of the system is the one of the most exciting opportunities, but not yet fully explored. The inclusion of relevant environmental and forcing information within or accessible to the system is necessary to progress in this area. In 10 years, the system has captured only a relatively small segment of Canada's long coastline, primarily because it is only populated on an opportunity basis almost always with collaborators and often funded by external partners. The transition of the CIS from a project-based information resource to a corporate resource requires ongoing support in terms of coastal science and information technology expertise, operating dollars, and a robust hardware and software infrastructure. This is a continuing challenge, which the GSC is meeting through new program and corporate initiatives in information management over the next two to four years.

Figure 1: Location and purpose for CIS data sets

 

Figure 2: Entity-Relationship diagram for the CIS. Entities (boxes) and relationships (connecting lines) which are drawn dotted are not implemented. 1-N indicates a one to many relationship between the entity COAST and the entity UNIT.

Form Supertype Form Type Form Subtype Form Features
solid cliff vertical
steep
overhanging		 terraced, unconsolidated over solid, discontinuous, arch, cave, smooth, ramping, sea stacks / pinnacles, talus cone
outcrop (1) (5)  
platform (3)  
anthropogenic (13)  
ice (7)  
unconsolidated cliff vertical
steep
overhanging		 unconsolidated over solid, discontinuous, talus cone, slump block, gullied, not stabilized, partially stabilized, stabilized
slope (2) (3)  
beach fringing
wide fringing
pocket
barrier
tombolo
attached spit
detached barrier		 cusp, berm, wave cut, washover channel, washover fan, storm ridge, relict ridge, rip channel, discontinuous
dune ephemeral
foredune
primary ridge
multiple ridge
parabolic
transverse
dome
sheet
relict
cliff-top		 blowout, wave cut, not stabilized, partially stabilized, stabilized
watercourse (10)  
delta (5)  
bar (6)  
waterbody (2)  
flat (2)  
anthropogenic (13)  
wetland (5)  
boulder (3) (3)  

Table 1: Examples of the CIS classification system for form. A complete subtype and features list is presented only for cliff, beach and dune. Number in brackets for other types represents the number of subtypes for that type.

Definitions for selected morphological forms used in the CIS:

1. outcrop: a low to moderately sloping (less than 40°) surface extending seaward from the backshore, composed of bedrock.

2. slope: horizontal to gently sloping (less than 20°) surface extending seaward from the backshore, composed of unconsolidated material and smooth, undulating or irregular in shape

3. boulder: boulder was chosen as the name for a morphological form type composed of boulders that is common in Labrador. Examples of form sub-types are scattered boulders, boulder barricade, boulder flat.

FIELD NAME
W
TYPE
DESCRIPTION
IMPLEMENTATION NOTES
ROUTE_   Long Int The number of the route from the linear reference system (route system) Present uses are 1 for the mainland route and 2 for the island route
UNIT_NO   Long Int The system defined unique number for a coastal unit unit-no is unqiue no matter what type of unit it is
FROM_   Float The distance along the route where the linear unit starts units are usually in meters
TO_   Float The distance along the route where the linear unit ends units are usually in meters
NTS_SHEET 20 Text The reference number for the topographic map the unit falls in  
ZONE_ 20 Text The name of the cross shore zone the unit is relevant to Present values are nearshore, foreshore and backshore
MAPPER 20 Text The name of the person who conducted the interpretation  
OPERATOR 20 Text The name of the person that entered the data into the CIS  
QUAL_CONTROL 20 Text The name of the person that performed quality control on the interpretation and data entry  
SOURCE 20 Text The type of information used as the source of the interpretation values include video and aerial photographs
SOURCE_DATE   Date The date the source information was collected  
SOURCE_COMMENTS 60 Text comments  
SOURCE_QUALITY 20 Text The class of quality of the source information excellent, good, poor
FORM_SUPERTYPE 20 Text The form supertype class solid, unconsolidated
FORM_TYPE 20 Text The form type class controlled vocabulary for type
FORM_SUBTYPE 45 Text The form subtype class controlled vocabulary for subtype
FORM_HEIGHT 12 Text The form height class controlled vocabulary for height
FORM_COMMENTS 120 Text Comments on the form controlled vocabulary for comments for some units, operator has a choice of taking standard comments or free form
FORM_FEATURES 100 Text The class(es) of features seen in the form a repeating group of features separated by commas, controlled vocabulary;
MAT_SUPERTYPE 20 Text The material supertype class solid, unconsolidated
MAT_TYPE 20 Text The material type class controlled vocabulary for type
MAT_SUBTYPE 45 Text The material subtype class controlled vocabulary for subtype
MAT_COMMENTS 50 Text Comments on the material Same as FORM_COMMENTS
MAT_FEATURES 50 Text The class(es) of features seen in the material essentially a repeating group of features separated by commas, controlled vocabulary;
CHANGE_MAPPING 120 Text Change mapping features comma delimited list of change features, controlled vocabulary; used in the Beaufort CIS only

Table 2. Data dictionary for the LINE_UNIT table.

For ZONE_ = Backshore      
Item Item Qualifier Item Value Score
FORM_HEIGHT For FORM_TYPE = Cliff High 1
Moderate 2
Low 3
Very Low 4
For FORM_TYPE = Slope High 1
Moderate 2
Low 3
FORM_FEATURES For stability features Unstable 3
Partially stable 2
Stable 1
For ice content features Wedge, massive or ping ice 3
polygons 3
Retrogressive thaw failure (RTF) 3
RFT stable or partially stable 2
No ice indicators 1
FORM_TYPE   wetlands 3
MAT_TYPE   organic 3
clastic 0
For ZONE_ = Foreshore      
FORM_SUBTYPE FORM_TYPE = beach Detached barrier 5
Fringing or pocket 2
Wide fringing 1
tombolo 1
FORM_TYPE = beach and
FORM_FEATURES = washover		 barrier 4
FORM_TYPE = beach and FORM_FEATURES ne washover Attached spit or barrier 3
FORM_TYPE = delta Inlet or ebb 4
flood 1
FORM_TYPE = flat Intertidal or supratidal 2
   
FORM_TYPE = bar transverse 3
FORM_TYPE   wetlands 3
FORM_FEATURES   Drowned tundra 1
MAT_TYPE   Silt, peat or mud 3
sand 2
Pebbles and cobbles with sand 1

Table 3: CIS items and assigned scores for the calculation of the Beaufort Sea erosion hazard index. Bold and capitalized text indicates an item name from the CIS LINE_UNIT Table.

 

Additional Attribute
Value
Score
Exposure Well-sheltered -2
Sheltered -1
Exposed 0
Well-exposed -1
Storm water level 0.0 - 0.5 m -1
0.5 - 1.0 m 0
1.0 - 1.5 m 1
1.5 - 2.5 m 2
Nearshore slope 0.00 - 0.01 ° 0
0.01 - 0.10 ° 1
0.10 - 1.06 ° 2

Table 4: Additional morphological attributes (i.e. not presently in the CIS), their values and scores used for the Beaufort Sea erosion hazard index

Erosion Hazard Score
Erosion Hazard Rank
Length of Coast (%)
>= 1 to < 9 Very low 161 km (10 %)
>9 to < 11 low 270 km (16 %)
>= 11 to < 13 moderate 344 km (21 %)
>= 13 to < 16 high 393 km (24 %)
>= 16 to 24 Very high 470 km (29 %)

Table 5: The distribution of erosion hazard scores and erosion hazard class rank show that about half of the coastline is characterized by high to very high erosion hazard.

References

Acknowledgements
Tracy Horsman, Tanya Kohler and Bernie McGuire have conducted most of the interpretation and data entry for the CIS data sets over the decade of CIS operations under contract with the Geological Survey of Canada. Their dedication is largely responsible for the high quality of data stored in the CIS.