Utility of Lidar-Derived Terrain Variables for Wetland Detection.

Wetlands are complex ecosystems that result from a number of environmental factors including hydrologic, geomorphic, and biologic processes. Topography plays a key role in wetland formation. Lidar (Light Detection and Ranging) data can provide detailed and accurate ground elevation information for the identification of micro-topographic features that indicate the presence of wetlands, which would otherwise be undetected. Fine-scale elevation data is particularly helpful in flat, wetland-rich areas where complex interspersion of uplands and lowlands may cause mapping confusion and inaccuracy. The full potential for lidar in wetland related studies is yet to be realized. The objective of this study is to evaluate the predictive power of different terrain indices derived from digital elevation model (DEM) developed from Lidar data for detecting wetlands. The study is carried out in an urbanized agricultural landscape located in Northern Virginia. Different wetland and upland areas were identified through field visits. Different primary and secondary terrain indices mainly, slope, plan curvature, profile curvature, topographic wetness index and topographic position index are calculated from a Lidar-derived DEM. These variables will be classified using pixel-based and object based approaches with Random Forest, a statistical clustering algorithm and assessed for accuracy to determine whether or not the object-based approach improves performance.