Water Resources Research Act Program

Principal Investigators: Yuming Wen, Leroy Heitz, Shahram Khosrowpanah

Abstract: Land cover change (LCC) is caused by human disturbances and/or natural events such as climate variation and flooding. The quantitative analysis of LCC has been a main concern to scientists and researchers around the world over the past few decades. Identifying land cover change in a watershed can help determine ecological and hydrological changes in the watershed over time. Satellite remote sensing, geographic information systems (GIS), geo-statistics, and global positioning system (GPS) can all be used to identify LCC of watersheds. These technologies provide the basis for developing landscape composition and pattern indicators as sensitive measures of environmental change and thus, may provide an effective and economical method for evaluating watershed conditions related to disturbances from human and/or natural stresses. Landsat observations have evolved from an experimental system in 1972 to a sophisticated means of monitoring changes in the Earth's surface using a multi-date satellite imagery database, which includes Landsat Multi-Spectral Scanner (MSS) imagery, from July 23, 1972 to June 15, 2001, Landsat Thematic Mapper (TM) imagery, from June 16, 1982 to current, and Landsat Enhanced Thematic Mapper Plus (ETM +), from April 15, 1999 to current, at a high spatial resolution. Recent surveys indicate that land cover/use changes have a direct and enormous effect on water quality and environmental change. Watershed water quality and ecosystem are threatened constantly by human impacts such as forest fires and development, as well as natural phenomena like storms and droughts. In addition, the combined uses of GIS, remote sensing and GPS tools have been highlighted with respect to their advantages in watershed applications and management. Spatial and temporal modeling of changes in wetlands and badlands in Southern Guam watersheds was one of the highest priority research needs recently identified for Guam in the Research Advisory Council Meeting of September 26, 2006. Previously, five temporal Landsat images, including Landsat MSS image of November 14, 1973, Landsat TM images of June 18, 1989, September 22, 1989, May 11, 1993, June 12, 1993, and March 15, 2001, were used to extract land cover information for watersheds in Southern Guam. The 1978 topographic map and 1973 digital raster graphics (DRG) for Guam were also utilized with the Landsat MSS image of 1973 to obtain land cover information. The classification process now complete, the next step involves an accuracy assessment to compare the classification to ground truth or other data. Recent IKONOS and QuickBird imagery will be used as reference data to assess Landsat imagery extracted land cover accuracy. Aerial orthophotos of 1975 are also available and can be used as a reference data. GPS and a GPS compatible digital camera will be used to collect field data for ground truthing and land cover classification accuracy assessment. By this means, a land cover accuracy assessment of each watershed in Southern Guam can be conducted. The objectives of this project are to: 1. Select reference pixels at random for classification accuracy assessment; 2. Use a GPS unit to collect field data as reference for accuracy assessment; 3. Use a GPS compatible digital camera to take photos as reference for the GPS collected data for ground truthing; 4. Conduct accuracy assessment, and achieve results.