Performing Department
Sustainable Resource Management
Non Technical Summary
In Washington, the existing statewide wetland maps (National Wetlands Inventory [NWI] maps) are out of date and inaccurate in many locations. They are based on imagery and data from the 1980s and do not reflect current wetland location and extent. Additionally, wetlands are missing from the NWI maps. These errors of omission have been recorded to be as high as 50% in some areas, and may be as high as 90% in some forested areas. Inaccuracies and errors of omission are due in part to the difficulty of photo-interpreting certain land cover types (e.g., forested wetlands, wetlands on slopes, and vernal pools), especially when using lower quality spectral, spatial, and temporal resolution imagery from the 1980s. Also, many wetlands on agricultural lands were not mapped. NWI classified wetlands to identify wetland habitat types. It lacks abiotic information such as landscape position, landform, and water flow path, which can be used to predict functions and, in combination with land uses, wetland condition. Remote sensing data sources such as LiDAR, high-resolution aerial imagery, Landsat imagery, digital elevation data, hydrography, and updated soil maps provide an opportunity to address these known shortcomings. Moreover, recent developments in automated remote sensing technologies (e.g., object-based image analysis) allow for more efficient coverage of large areas.An improved, statewide map of wetland location and type is critical to the ability of local governments to protect wetlands. Under Washington State's Growth Management and Shoreline Management Acts, local governments play a critical role in wetland protection and management. They do this through comprehensive planning, zoning, and permit review. Planners and permit reviewers rely on existing NWI maps for these processes. A few local jurisdictions have conducted their own wetland inventories and improved their maps, but these are limited due to lack of resources, and none have predicted functions and conditions of wetlands. This leaves many local governments with inadequate maps and information on local wetlands, and the state with uneven coverage.This project will improve the ability to more efficiently and accurately identify the location, size, and type of Washington's (WA) wetland resource. This will be accomplished using remote sensing data to identify (digitally map) wetland locations and classify wetland types.We will develop a systematic process that will result in an updated, statewide map of wetland locations, with attributes covering multiple classification schemes. For Phase One of this project, we will determine the most efficient and accurate statewide approach for remotely mapping and classifying wetlands. The approach will be applied in areas representative of land use and ecological diversity in Washington. This will allow us the opportunity to estimate the accuracy of mapping in areas where unique challenges have contributed to errors in existing maps.Project results will be available as a publicly accessible, web-based map. Information about the maps, and any analyses using the data, will be disseminated through articles and presentations to state and federal agencies, and local governments and planners' forums.
Animal Health Component
0%
Research Effort Categories
Basic
0%
Applied
100%
Developmental
0%
Goals / Objectives
In Washington, the existing statewide wetland maps (National Wetlands Inventory [NWI] maps) are out of date and inaccurate in many locations. They are based on imagery and data from the 1980s and do not reflect current wetland location and extent. Additionally, wetlands are missing from the NWI maps. These errors of omission have been recorded to be as high as 50% in some areas, and may be as high as 90% in some forested areas. Inaccuracies and errors of omission are due in part to the difficulty of photo-interpreting certain land cover types (e.g., forested wetlands, wetlands on slopes, and vernal pools), especially when using lower quality spectral, spatial, and temporal resolution imagery from the 1980s. Also, many wetlands on agricultural lands were not mapped. NWI classified wetlands to identify wetland habitat types. It lacks abiotic information such as landscape position, landform, and water flow path, which can be used to predict functions and, in combination with land uses, wetland condition. Remote sensing data sources such as LiDAR, high-resolution aerial imagery, Landsat imagery, digital elevation data, hydrography, and updated soil maps provide an opportunity to address these known shortcomings. Moreover, recent developments in automated remote sensing technologies (e.g., object-based image analysis) allow for more efficient coverage of large areas.An improved, statewide map of wetland location and type is critical to the ability of local governments to protect wetlands. Under Washington State's Growth Management and Shoreline Management Acts, local governments play a critical role in wetland protection and management. They do this through comprehensive planning, zoning, and permit review. Planners and permit reviewers rely on existing NWI maps for these processes. A few local jurisdictions have conducted their own wetland inventories and improved their maps, but these are limited due to lack of resources, and none have predicted functions and conditions of wetlands. This leaves many local governments with inadequate maps and information on local wetlands, and the state with uneven coverage.This project will improve the ability to more efficiently and accurately identify the location, size, and type of Washington's (WA) wetland resource. This will be accomplished using remote sensing data to identify (digitally map) wetland locations and classify wetland types.We will develop a systematic process that will result in an updated, statewide map of wetland locations, with attributes covering multiple classification schemes. For Phase One of this project, we will determine the most efficient and accurate statewide approach for remotely mapping and classifying wetlands. The approach will be applied in areas representative of land use and ecological diversity in Washington. This will allow us the opportunity to estimate the accuracy of mapping in areas where unique challenges have contributed to errors in existing maps.Project results will be available as a publicly accessible, web-based map. Information about the maps, and any analyses using the data, will be disseminated through articles and presentations to state and federal agencies, and local governments and planners' forums.
Project Methods
TASK 1) Technical assistance to identify and evaluate the tools and approaches available to efficiently and accurately map wetlands in Washington.Task 1.1. Assist the State of Washington, Department of Ecology (ECOLOGY) with the identification of locations of study areas based on availability of remote sensing data (e.g., LiDAR), level of threat (e.g., development, forest practices) to wetland condition and functions, and other data resources (e.g., county-level delineation maps). Demonstration areas will be identified in areas representative of land use and ecological diversity of Washington. The areas will include locations east of the Cascades and west of the Cascades. Study areas will be at the county or watershed scale, depending on resources and data availability in the two regions. Ecology and the University of Washington (UW) will use a stratified approach as follows:1. Identify areas with sufficient LiDAR coverage to ensure improved wetland mapping.2. Identify areas that span a gradient of urban to wildlands where development pressure and managed forest lands represent different sources of stress.3. Identify areas where local information may be available to augment the field verification effort.Task 1.2.Assist Washington State Department of Natural Resources (DNR) with the development of a crosswalk between wetland classification systems to guide classification algorithms. UW will provide feedback and consultation to DNR on the crosswalk as needed. The following classification systems will be included in the crosswalk:Federal Geographic Data Committee (FGDC) Wetlands Classification Standard (adapted from and referred to herein as Cowardin).NWIPlus LLWW.WA HGM Classes.WA Natural Heritage Types.Forest Practices Water Types.TASK 2) Develop a Draft Quality Assurance Project Plan for remote sensing and provide technical assistance to identify map dissemination processes Task 2.1.Develop a Draft Quality Assurance Project Plan (QAPP) that describes the methods and data sources that will be used in the remote sensing process for identifying wetlands. Data standards will be designed to meet those of state and federal partners, including the Federal Geographic Data Committee (FGDC), with the intent that wetland maps from this project will be of appropriate accuracy for inclusion in the national data layer of the National Wetlands Inventory. UW will determine what remote sensing data are needed and where there are data gaps to improve the ability to identify wetland locations and apply classifications. UW will identify the remote sensing data that are available for use in this project's semi-automated approach to mapping wetlands, and where data gaps might limit application of the method. The QAPP will detail the data standard goals for the project.TASK 3) Create maps with wetland extent and classification in the Phase One study areas.Task 3.1. Develop a preliminary map of wetlands in demonstration areas identified in Task 1.1 using remote sensing data and an automated approach. UW's automated approach will use a combination of data, which may include LiDAR, soil maps (SSURGO), elevation, high-resolution imagery, hydrography, to improve wetland identification through Object-Based Image Analysis (OBIA), and/or Random Forest Modeling.Task 3.2. Information Technology Specialists at ECOLOGY will develop a web-based map platform, initially for use as a way to share data exclusively among project partners during map development. UW will provide technical assistance to ensure that the platform developed by ECOLOGY will be accessible to UW and other project partners. ECOLOGY will ensure project partners will be able to access and provide feedback to the unrefined maps. ECOLOGY will invite project partners to review the maps and provide feedback.Task 3.3. UW will apply feedback from project partners for a first level verification and refinement of the preliminary wetland maps. UW will also use locally-verified wetlands data and State-generated classifications to refine the accuracy of the wetland maps. For example, Pierce County maintains a County Wetland Inventory (CWI) that includes a verified wetland layer. Verified wetlands data will be used as an iterative step to refine the mapping algorithms by comparing how well the automated wetlands map corresponds to verified wetland and non-wetland data.Task 3.4. UW and ECOLOGY will meet with EPA prior to map generation to collectively determine a statistically appropriate sampling plan.TASK 4) Refine mapping and classification from field data.Task 4.1. UW will use the field verification data generated by ECOLOGY to adjust mapping algorithms and apply a mechanistic model. Data collected in the field will help to identify areas where the automated mapping process identifies wetlands where none occur (false positive) or misses wetlands where they do occur (false negative). UW will use the data from field verifications to test the accuracy of the automated map and adjust the algorithm wherever possible to make improvements to the automated process. UW will estimate the probability of wetland occurrence will be estimated during this task to help inform the photo-interpretation process in later tasks.Task 4.2. Submit mapping results from automated remote-sensing process for further refinement via photo-interpretation (PI) and addition of NWIPlus classification. UW will coordinate with the contractor who will do the photo-interpretation and application of NWIPlus to ensure that the NWIPlus contractor has the materials (refined map data) needed to complete the PI and classification.Task 4.3. UW will complete the final map product by applying the three different Washington state wetland classifications to PI refined maps. UW will use the crosswalk developed in Task 1 to guide the algorithm that will apply the wetland classifications. UW will use the field data collected by ECOLOGY to verify the classifications and adjust the algorithm as needed.TASK 5) A report summarizing results of Phase One, with recommendations for improving the approach in subsequent phases. Task 5.1. Work with ECOLOGY and NWIPlus project partners to load maps to publicly-accessible web-based mapper. The maps will have interactive features that allow users to explore the locations and classifications of wetlands.Task 5.2. UW will submit the draft final report for review by project partners. The draft final report will include estimates of the accuracy of wetland locations and classifications in the final map using the data collected by ECOLOGY in the field. These estimates will describe overall and per class error rates relative to errors of omission (false negatives) and commission (false positives). Depending on the structure of validation site selection, results for individual classes may require weighted adjustments based on frequency of occurrence. Confidence intervals for accuracy estimations will be provided. The Draft Final Report will include the following:The methods (process and data) used to complete Phase One.The results of mapping and classification.Findings of the accuracy and limitations of the map products.Recommendations for applying the mapping and classification methods to the entire state or in high priority areas in phases.Feasibility and costs associated with applying mapping approach on a statewide scale.Task 5.3. Submit Final Report including the following:The methods (process and data) used to complete Phase One.The results of mapping and classification.Findings of the accuracy and limitations of the map products.Recommendations for applying the mapping and classification methods to the entire state or in high priority areas in phases.Feasibility and costs associated with applying mapping approach on a statewide scale.Final maps and metadata