Validation of Elevation-derived Channels
We are developing methods for autonomous validation of elevation-derived hydrographic features using remote sensing data to update the NHD.
The goal of this work is to develop methods for the improvement in accuracy and regular update of the National Hydrography Dataset (NHD). The NHD is an important dataset with applications including natural resource management, emergency response, and hydrologic modeling. In the past, the NHD was compiled from local hydrographic surveys and aerial photography with data collections dating from the 1940s to the present. More recent data collection approaches rely on deriving surface water drainage networks and other features from elevation data with interpretation from high-resolution photography or satellite image data. Apart from the variation in hydrographic patterns due to the vast range of geomorphic conditions across the United States, the NHD has heterogeneous coverage and resolution because of the various climate conditions and compilation standards that existed during the years of data collection.
Recent developments in remote sensing tools and methods have furnished opportunities to improve and continually update the NHD. For example, CEGIS researchers have used Digital Elevation Models (DEMs) and GIS in developing automated methods for mapping streams and comparing them to NHD. Results of the comparison indicate that the largest variance between automatically generated hydrography and NHD vector data is caused by missing headwater channels in the NHD. Areas of low topographic relief and heavy vegetation present the greatest challenge to synchronizing NHD to elevation-derived hydrography.
We are testing the use of lidar point clouds and other high-resolution optical data for identification and categorization of headwaters. Optical and lidar data collected with an Unmanned Aerial System (UAS) will be used in studies investigating the potential of high-resolution data fusion for modelling and for remote stream mapping.