Satellite Remote Sensing - A Geospatial Technology For Watershed Management
Watershed management, especially across jurisdictional boundaries, requires a great deal of cooperation and understanding among vastly diﬀerent groups and individuals. Geospatial technologies, such as satellite remote sensing, has proven to be an invaluable tool for collecting resource baseline data and establishing common ground to discuss watershed management constructively. This study presents a methodology for extracting watershed baseline information from a time series of satellite imagery for change detection analysis to support watershed restoration, and for detailed watershed modeling to test various management scenarios. The methodology which has been successfully applied to watersheds in Torrance, CA demonstrates a cost effective means of using geospatial technology to map features of the watershed, such as impervious surface, land cover, wetlands, watershed dynamics monitoring, flood mapping, designing optimum water quality sampling locations, water quality mapping, and outfall characteristics.
To capture and describe specific watershed elements, the methodology stresses a mix scale approach involving the integration of primary and secondary data that can easily be obtained through government sources and databases from other organizations, and a measure of descriptive statistics. The main geospatial technologies include WorldView-2 satellite imagery, digital elevation model (DEM), and digital orthophoto quad (DOQ). The methodology is fully automated and can be applied to any watershed. The four general steps can be described as data acquisition, image processing, hydrologic parameter extraction, and change detection and integration with watershed models.
Two scenes from WorldView-2 satellite acquired on different dates were used to explore impervious and land cover changes in the Machado Lake Watershed. The imageries were also used in conjunction with the DEM to identify optimum locations for watershed sampling. For land cover changes in the watershed, supervised and unsupervised classifications were employed to find and compare spectral classes that are attached to the basic land cover types, artificial surfaces, agricultural areas, forests and semi-natural areas and water bodies. Spatial features from the classification were arranged in map layers and stored in a data cube with two dimensions formed by the (x) and (y) spatial axis of the image display, and the third (z) formed by the time of acquisition of the image. The data cube can be complemented by spatial data for watershed modeling and by output layers for visualization of the simulation results. This approach provides both efficient storage of spatio-temporal data and data exchange by multi-band read/write methods.
The methodology proposed here provides a cost effective way to obtain watershed baseline data to update established watershed databases and also data for watershed modeling. Current GIS technology can be used to perform nearly all the tasks focused on watershed modeling, processing of satellite remote-sensing data, and data cube management. The data cube enables efficient storage and management of raster layers containing DEM, satellite images and land-use datasets. However, other datasets such as stream outlets, discharge points, reservoirs, weather data, information about agricultural management, and time series of nitrate concentrations in hundreds of stream sites, require extra storage in various data formats.