A coastal data model for puget sound

Download 120.22 Kb.
Size120.22 Kb.

Previously presented at Coastal Zone ’07, Portland Oregon, July 26, 2007

Timothy Nyerges*, Scott Dudgeon, Tyanne Faulkes, and Miranda Hett
*Department of Geography

Box 353550

University of Washington

Seattle, Washington 98195

Phone: (206) 543-5296

Email: nyerges@u.washington.edu

Keywords: coastal, data model, geographic information systems, coastal zone management


Coastal resources in the US and around the world are under increasing near, medium and long-term pressure from a variety of stressors. A report from the Renewable Natural Resources Foundation (2005), synthesizing material from the Pew Oceans Commission Report (2003) and the US Commissions on Oceans Report (2004), describes many of the ocean and coastal resources currently at risk. Understanding the many kinds of impacts to coastal resources requires diverse perspectives, not the least of which involves a better understanding of the interactions between land-side and the water-side interactions – or what can be called the nearshore area of the coastal zone.
Coastal zone management (CZM) of ecosystems requires robust geospatial information to be effective everywhere, but particularly in nearshore areas influenced by considerable land development impacting surface water runoff in watersheds that drain into coastal waters (Beatley, Brower, and Schwab 2002). Using geospatial information technologies, particularly geographic information systems technology, can help develop a shared insight about problems, challenges and solutions about how to management coastal resources (Wright and Scholz 2005). CZM applications of geographic information systems (GIS) are not new, but GIS database integration directed at exploring issues associated with nearshore management in hopes of fostering shared understanding among diverse stakeholders is still in its infancy. In this paper we take the reader through the steps we used to develop a coastal data model and report on the considerations and challenges addressed. We present the outcomes of the process and consider next steps for development and use of the data model in the context of a participatory web portal for improving nearshore coastal resources.

The Puget Sound is the 2nd largest estuary in the U.S. In 2005, Washington State’s Governor Christine Gregoire established the Puget Sound Partnership for Nearshore Restoration. The goal of the partnership is to “…ensure that Puget Sound forever will be a thriving natural system, with marine and freshwaters, healthy and abundant native species” (Puget Sound Partnership 2006, p. 10) and the goal of the Nearshore Partnership is to “identify significant ecosystem problems, evaluate potential solutions, and restore and preserve critical nearshore habitat.” (Puget Sound Nearshore Partnership 2006, p. 1). Understanding the complexities of the fish and plant life plus how human activities affect the habitat that support that life within the nearshore can be enhanced through the use of geospatial information technologies.

CZM was used as the motivating theme to teach an intermediate course in GIS within the Department of Geography at the U of Washington (Geography 460 Autumn 2006). The offering combined a set of class lectures from a previous quarter with this new perspective, but the instructor (Nyerges) concluded something more was needed to really enhance the learning experience. Consequently, the instructors and a group of honor students (other co-authors of this paper) agreed that a foundation for learning about the breadth and depth of GIS applied to CZM could come through working with a robust representation of coastal features related to water flows from watersheds and within estuarine ecosystems. Such representations are called data models. A data model consists of the geospatial constructs for structuring data, the operations that can be performed on those structures to derive information from the data, and the rules for maintaining the integrity of data. Li (2000) describes data models for marine and coastal GIS applications, focusing on the structural component which is the more popular aspect of data models.
The core of a GIS approach for managing coastal resources involves the development of a robust coastal data model, and in particular a coastal nearshore data model that should include all three components. As part of an ad hoc honors activity (required of honors students) we focused on the development of the structural and operations components of a coastal data model in general, and a coastal nearshore data model in particular as a basis for enhancing the learning experience. In this paper we report on the schema development process at the conceptual level.

As mentioned earlier, a data model consists of three components: 1) constructs for structuring data, 2) operations for data processing, and 3) rules that maintain the integrity of data (Codd 1981). The popular version of a GIS data model focuses only on the component one, and thus we start with that. We make use of an informal approach to schema integration analysis to synthesize the conceptual contents of the coastal data model as a database design (Nyerges 1989). Conceptual schema integration analysis involves identifying, comparing, and contrasting feature classes and the geospatial data types most appropriate for characterizing those feature classes in order to develop an overall “conceptual data schema” – simply a list of feature classes and potential relationships that form the core of a database design.

Each of the three steps in our method used a different source of “community of practice” knowledge to perform conceptual schema integration. In a first step, we explored how to integrate watershed data and marine data into GIS using the ArcHydro and ArcMarine Data Models. In a second step we identify coastal feature classes described within a popular reader about coastal zone management (Beatley, Brower, and Schwab 2002), and add them to the feature class list for the coastal data model. In a third step we make use of the recommendations put forward by the Puget Sound Nearshore Partnership to further contextualize the coastal data model for anticipated use. We used these different “communities of practice” knowledge because they are all considered “vetted knowledge” and because they are convenient to us.
Integrating ArcHydro and ArcMarine Data Models – First Step

The ArcHydro Data Model describes geospatial and temporal data on surface water resource features of the landscape (Whiteaker, Schneider, Maidment 2001). The data model addresses three issues. First, it addresses the principal water resource features on a landscape. Second, it offers a description about how water moves from feature to feature, that is through connectivity. Third, the data model provides for a description of time patterns of water flow and water quality associated with these features.

The ArcMarine Data Model represents a new approach to spatial modeling using improved integration of many important features of the ocean realm, both natural and manmade (Wright 2006). The model considers how marine and coastal data can be most effectively integrated in 3D and 4D space and time and includes an approach towards a volumetric model to represent the multidimensional and dynamic nature of ocean data and processes.
Drawing the two data models together to examine what is in common resulted in recognizing that even the idea of “shoreline” is not common between the two data models. Although a “shore zone” can be constructed, it was not explicitly evident in either one. This result heightened our interest about what should be included in a coastal data model. We concluded that feature classes from both are useful (See Table 1), but many more would be needed if we were to really focus on “coastal features”. Understanding the similarities and differences in the ArcHydro and ArcMarine data models leads to a better understanding of how to develop a coastal data model. However, to gain a better sense of what feature categories should be considered, we developed a follow-on activity using a second source of information.
Table 1. ArcHydro and ArcMarine Data Models in Terms of Specific Geospatial Data Types

Geospatial Data Types

ArcHydro Data Model

ArcMarine Data Model

Fixed point

Drainage area centroids

Marker, buoy, transponder

Instantaneous point

Discharge measurement, dissolved oxygen value

Raw bathymetry



Sediment transport line



Habitat, marine boundaries

Time duration points


Current meter

Time duration vectors

Temperature at one point to temperature at another point

Algae bloom trawl

Time duration areas

Water surface elevation

Oil spill

Feature classes

Drainage, network, channel, hydrology

Watershed, waterbody, monitoring points, streams

From the comparison and contrast of the two data models it was clear that the coastal zone was not characterized with much nuance. Therefore, a second step was taken to provide more thorough insight.

Feature Classes from a Coastal Zone Management Book – Second Step

A collection of feature classes and several attributes were compiled from a text reader about coastal zone management, assuming this textbook reader was evidence of another form of expert knowledge. The left-most column of Table 2 contains a list of feature categories, attributes and processes. Students used one of the two course textbooks (Coastal Zone Management - Beatley, Brower, and Schwab 2002), to compile a more comprehensive list of feature classes. We argue that authors of a textbook are themselves experts in a topic, and that topic is peer reviewed by other experts familiar with the topic. Because this particular text is published by Island Press, a well known environmental publisher, we expect the slant on the information is more environmentally-oriented. However, because the book was used as a reader in the GIS course, we know from experience that it has reasonably well-balanced perspective as it speaks to sustainability issues about economic, social, and ecological aspects of communities.

Table 2. List of Feature Categories and Some Important Attributes

(compiled from Beatley, Brower, and Schwab 2002)

  • Barrier Islands

  • Estuaries

  • Coastal Marshes

  • Coral Reefs

  • Rocky Shores

  • Bluffs

  • Tides (dynamic, temporal, presents a difficult to map in a static environment)

  • Currents (same issue as above)

  • Wind (Currents/Patterns) (same issue as above)

  • Erosion and Accretion (same issue as above)

  • Pollution and Toxic Contaminants (also a challenge, show w/ zones?)

  • Wetlands (Protected/Unprotected)

  • Habitats – endangered species

  • Land use and zoning of areas

  • Building code

  • Soil Composition/make-up

  • Catch Basins/ catchments

  • Watershed areas

  • Streams/Rivers/Water Flow

  • Ports – Freight and Passenger

  • Ferry Systems/Water Taxi

  • Continental Shelf/Slope

  • Water Depth/Slope

  • Land Cover – (e.g. Beach/Dunes)

  • Present Buildings/Structures

  • Infrastructure (on land, underneath)

Puget Sound Nearshore Partnership – Third Step

In 2005, Washington State’s Governor Christine Gregoire formed and established the Puget Sound Partnership for Nearshore Restoration. Her goal of the partnership is to “ensure that Puget Sound forever will be a thriving natural system, with marine and freshwaters, healthy and abundant native species,” (See Draft Recommendations). On October 13th, 2006, the partnership executive committee released recommendations for focusing efforts in the Puget Sound area. These recommendations are useful to a) identify fundamental theme for improving the health of Puget Sound, b) identifying features that can corroborate the list identified from reviewing Beatley, Brower, and Schwab (2002) as well as those in the integration of the ArcHydro and ArcMarine Data Models, and c) identify primary and secondary processes that encourage a type of GIS data analysis to derive information as a basis for decision support to restore the Sound.

When building a data model, GIS analysts must also consider the GIS data processing operations to be supported as part of an application. In this case, the processes and applications involving coastal nearshore features. The following is a summary of feature themes and processes collated from the Nearshore Partnership recommendations (Puget Sound Nearshore Partnership 2001, 2006, 2007). Those themes and processes motivate a general set of GIS data processing opportunities.

Table 3. GIS Data Themes and Data Processing Opportunities

Protect Existing Habitat and Prevent Further Losses

Salmon Recovery plan along nearshore

- GIS opportunity: Map (model) flow and movement of salmon

Critical Areas Ordinances support and guidance

- GIS opportunity: Identify and Map these spaces

Evaluate effectiveness of funding within protection plans

- GIS Opportunity: Multicriteria decision evaluation
Restore the Amount and Quality of Habitat and Reduce Fragmentation

Restore freshwater and marine habitats and 100 miles of marine shoreline

Other restoration actions: remove derelict vessels and creosote logs

- GIS opportunity: Identify, locate and map “pollutants”

Improve habitat mitigation programs and methods

- GIS opportunity: Identify, map, and provide multicriteria decision support capabilities

Reduce toxics entering Puget Sound

Conduct a characterization of toxic sources and loadings

- GIS Opportunity - Map presence/absence in areas, and potential areas of proximity that may be at risk

Clean-up all sites on the state’s priority list by 2020

- GIS Opportunity – Monitor and map presence/absence in areas, and potential areas of proximity that may be at risk
Reduce Pollution from Human and Animal Wastes into Puget Sound

Characterization of nutrient and pathogen sources and loadings

- GIS Opportunity – Identify areas of concern; Identify piping (utility) network availability; Model interaction among urban and agricultural land uses

Promote and support new and existing treatment facilities (financial and technological support)

- GIS Opportunity – Where are the treatment plants? Use criteria established for most effective citation of new facilities to benefit
Improve Water Quality and Habitat by Managing Stormwater Runoff

Identify, prioritize, and implement retrofits where stormwater runoff is causing environmental harm; mitigation strategies

- GIS Opportunity – Model interaction between CAO (wetlands for instance) and runoff occurrences; watershed characterization - fill, flow direction, flow accumulation, basin,
Provide Water for People, Fish and Wildlife, and the Environment

Improve stream flows

- GIS Opportunity – Identify these Spaces of flow for freshwater; Identify water treatment plants/reservoirs; Characterize utility (sewage/water supply) networks

Flow restoration actions in priority basins

Protect Ecosystem Biodiversity and Recover Imperiled Species

Implement existing recovery plans and create recovery programs for species at risk of extinction lacking current recovery plans

- GIS Opportunity - Identification of species and animals and any possible migration patterns, plus the land features important to habitat for these species.
The above themes and processes encourage inclusion of the following additional feature classes within the Coastal Data Model.

  1. Utility networks

  2. Freshwater sources/treatment plants for freshwater

  3. Waste treatment locations

  4. Critical Area Ordinance (CAO) spaces


Feature classes identified in steps 1, 2, and 3 are collected together in Table 4. The feature classes are grouped into feature datasets (bold face text in left column). The step numbers 1, 2, and/or 3 are identified next to each category. Furthermore, along with the list of features and grouping, we identify the most likely geospatial data type to act as a database representation. The resulting table provides a target for the conceptual schema of a coastal data model. Clearly, not all features would be used in all applications, so it is important to identify which feature classes and processes are implemented by what data operations.
Table 4. Coastal Data Model Features/Processes and Potential Geospatial Data Types


Geospatial Data Types






Physical/Natural Shoreline

2Barrier Islands




2Coastal Marshes


2Coral Reefs


2Rocky Shores










2Soil Composition


2Land Cover


Human Infrastructure/Impact

12Pollution and toxic containments




2Land Use and Zoning


2Building Code





2Ferry Systems/Water Taxi


2Present Buildings/Structures



2Roads Network



3Sewage Utility Piping Network



2Sea Walls



Dynamic Natural Phenomena






2Winds Patterns/Flow


2Erosion and Accretion


2Migratory Animals (e.g. birds)


Water and Water Bodies

12Catch Basins/catchments



12Watershed areas


12Streams/Rivers/Water Flow


Underwater Topography

12Continental Shelf/Slope


12Water Depth/Slope


12Canal Shape/Depth/Slope


3Critical Area Ordinances (CAO) Spaces


3Freshwater sources/treatment plants for freshwater



3Waste Treatment locations




1 - Step 1 features compiled from ArcHydro and Arc Data Model information

2 - Step 2 features compiled from textbook information

3 - Step 3 features compiled from Nearshore Task Force information

Below we identify the most common GIS data processing operations important for deriving information of various processes. These operations would be the basis of deriving data to gather information for higher knowledge on problem solving in the coastal zone.
Migratory animal movement (for example: Birds, Whales, Salmon, or Turtles)

Operations: Network Analyst tools

The construction and mapping of networks can establish areas which these migratory animals pass through. It can characterize the distances in which they travel and the times in which they arrive in those areas and the total time in takes them to move from area to area.

Transportation interaction with the coast (for example: Ferry systems/road systems for automobiles)

Operations: Network Analyst tools, Flow direction/Accumulation (Hydrology)

As with migratory animals, these networks will allow us to point to areas of concern that these transportation system may pass through affecting the coast. These are strongly linked to discovering the pollution that comes from these transportation systems, as we can use traffic counts and VMT on particular road segments (or travel segments with say the Ferry system) to show how much pollution is coming from road segments and also where mitigation retrofits need to be added.

Tidal Currents and pollution interaction with tides (for example: dump points of sewage and how it moves with these tides in the water)

Operations: Flow direction/accumulation (hydrology), movement on top of water

This complicated look at how water flows over water (or how pollution interacts in currents and tides when entering bodies of water) can be done with certain water flow/accumulation operations, as well as digitizing and creating new shapefiles for directionally of tides/currents. Understanding how water interacts with itself is important to understand how different substances of pollution would move within it and affect specific zones.

Pollution runoff/stormwater runoff (for example: Finding where it occurs (i.e. identification of critical areas of concern)

Operations: Hydrology operations for flow, identification of spaces

The hydrology tools (fill, flow direction, flow accumulation, basin tool, and watershed tool) allow for showing how and where water and other pollutants would flow from one area to another. This is useful again in establishing where and how runoff occurs, and finding areas where new infrastructure for this type of runoff needs to be placed and where mitigation retrofits need to be applied to already present infrastructure. This is a serious goal for the Puget Sound Partnership as detailed in their draft recommendations.

Interaction of particular spaces (for example: Dairy Farms and Urban Sites, or Tsunami impacts upon various types of soil composition (erosion hazards and land cover)

Operations: Buffering, flow accumulation, overlays

These are more basic interaction operations which can help to show what kinds of areas may be affecting one another. The projects which both of our groups worked on for our finals projects dealt with isolating areas that may have adverse externalities or affects on one another.

From the above analysis we developed a common set of process and operations for Nearshore coastal applications (See Table 5).
Table 5. Common GIS Data Processing Operations

  1. Mapping the movement of animals/species and also simple identification of presence

  2. Find/Identify at risk to toxic materials and pollutants spaces

    1. Proximity of spaces (that is buffering)

    2. Interconnectedness between spaces (what else connects them?)

  3. Network Analyst tools

    1. For Utility networks (piping, sewage, freshwater)

  4. Hydrology tools

    1. Fill tool

    2. Flow Direction tool

    3. Flow Accumulation tool

    4. Basin tool

    5. Watershed tool

The above schema and operations can be further detailed when embarking upon a particular application, which was the next step in the project.


Rather than simply “think up ideas” about coastal zone data, and not having ready access to a community of experts in person, we decided to use three rather different sources of “community (expert) knowledge” about coastal zone data. Each of the three sources of information has been thoroughly reviewed, and is thus considered “vetted knowledge”. Our three sources provide us a “triangulation of knowledge” of sorts, relying upon three different “communities of practice”. We recognize that these “communities” are not the popular use of community (as in placed-based community), but we intend to pursue that type of knowledge integration in the future.

Learning about data models as the underlying frameworks for GIS data analysis – was a way for students to learn details about data management while grounding this learning process in a major challenge facing the Puget Sound region at this time. Group learning was a fundamental part of the schema integration process as students negotiated the interpretations of various feature categories and then developed a set of GIS data analysis operations motivated by the recommendations appearing in the Governor’s Task Force Report. Since operations are part of a full blown data model, elucidating the operations in a conceptual manner can help prioritize what is to be known next – as building the data model is only the start. Applying a data model in data analysis leads to new information about the structures and processes of nearshore resources.
We conclude that development of a coastal nearshore data model definitely enhances a student’s appreciation for the underpinnings of GIS information technology. Constructing suitable data models for GIS applications are the foundation of those applications. Once understanding the character of data models, GIS analysts can more readily develop GIS applications. Data models are what enable and limit GIS applications. Although we undertook this exercise in a classroom setting, we hypothesize that participatory data model development might also enhance stakeholder understanding about what is known about nearshore resources.
Data models constructed using participatory processes naturally lead to analyses performed from participatory perspectives. Such data models can form the foundation of analytic-deliberative decision processes that draw together diverse stakeholders into a discussion about how to improve precious resources. Research about the development and use of participatory GIS to support broad-based analytic-deliberative decision processes, like those for prioritizing beach rearmament, is currently underway at the U of Washington. Such research is being considered for how it might improve the learning experience of students if we cast learning to use GIS within the context of multi-stakeholder role-play scenarios.


Beatley, T., D. J. Brower, and A. K. Schwab 2002. Coastal Zone Management. Island Press. Washington D. C.
Codd, E. F. 1981. “Data models in Database Management” SIGMOD Record 11(2):112-114.
Li, R. 2000. Data Models for Marine and Coastal Geographic Information Systems, in D. J. Wright and D. J. Bartlett (eds.) Marine and Coastal Geographic Information Systems, London, Taylor & Francis, pp. 25-36.
Nyerges, T. L. 1989. Schema Integration Analysis for GIS Database Development, International Journal of Geographic Information Systems, now Intl Journal of Geographic Information Science, 3(2):153-183.
Pew Oceans Commission Report, 2003. America's Living Oceans: Charting a Course for Sea Change. Last accessed July 10, 2006 at http://www.pewtrusts.org/pdf/env_pew_oceans_final_report.pdf.
Puget Sound Nearshore Partnership 2006. Home Page, http://www.pugetsoundnearshore.org/index.htm, accessed October 27, 2006.
Puget Sound Nearshore Partnership 2007. Nearshore Ecosystems Conceptual Model http://www.pugetsoundnearshore.org/techical_papers/conceptual_model.pdf, last accessed January 26, 2007.
Puget Sound Nearshore Partnership 2001. Puget Sound Nearshore Ecosystem Restoration Study
Sept 25, 2001, http://www.pugetsoundnearshore.org/program_documents/project_plan.htm, last accessed January 26, 2007.
Puget Sound Partnership 2006. Sound Health, Sound Future, http://www.pugetsoundpartnership.org/reports/final/final/Final_wAPPx_lr.pdf, last accessed January 26, 2007.
Renewable Natural Resources Foundation (2005), Building Capacity for Coastal Solutions, Renewable Resources Journal, 23(1):5-30. Last accessed July 12, 2006 at: http://www.rnrf.org/.
U.S. Commission on Oceans Report (2004) An Ocean Blueprint for the 21st Century, Last accessed July 10, 2006 at http://www.oceancommission.gov/documents/full_color_rpt/welcome.html.
Whiteaker, T., Schneider, K. and Maidment, D. 2001. Applying the ArcGIS Hydro Data Model, http://www.crwr.utexas.edu/gis/gishydro01/support/schematutorial.pdf, last accessed November 15, 2006.
Wright, D. J. 2006. ArcMarine Data Model. http://dusk2.geo.orst.edu/djl/arcgis/, last accessed October 2006.
Wright, D. J. and A J. Scholz 2005. Place Matters. Ecotrust. Portland OR.

Download 120.22 Kb.

Share with your friends:

The database is protected by copyright ©ininet.org 2024
send message

    Main page