Background
With our partner, TerrainWorks, we are developing a GIS-based toolset to systematically compare and test different approaches and data types for remote mapping of wetlands. This toolset will support the determination of the optimal methodology to identify wetlands for a particular region and for particular wetland types. The toolset will quantify the accuracy and precision to which different data sources can resolve wetlands, and create maps delineating probable wetland locations and types that can be calibrated and validated to local conditions.
Our Work
The project will combine results from automated pattern-recognition (object-based) techniques, using high-resolution imagery (e.g., NAIP), with process-based (wetness index) and empirical techniques (e.g., logistic regression), using topographic, geologic, soils, land cover, and climate information. Objectives of the project include:
· Determine the optimal methodology to identify wetlands for a particular region and for particular wetland types
· Determine the accuracy and precision to which different data sources (e.g., LiDAR versus NED DEMs, spectral imagery versus DEM) can resolve wetlands
· Apply a suite of software tools that will use available remote sensing data to construct the wetland indices.
· Develop an open-source tool within ArcGIS to predict wetland locations.
· Create maps identifying probable wetland locations and types that can be calibrated and validated to local conditions.
· Provide a manual describing the wetland identification tool and its use including methods, and data used.
· Provide a report describing challenges encountered and recommendations for future work.
Results
Deliverables
