RESDEM, A VEHICLE FOR DATA MINING AND MODELING OF TEMPORAL REMOTE SENSING DATA (TRMM, SSM/I, AVHRR, AMSR, MODIS, AND QUICKSAT)

Use of extensive public domain global remote sensing data sets for the assessment and modeling of complex environmental problems is often hindered by the extensive data management routines required to make full use of these data sets and by the general absence of user-friendly routines for data extraction. The problem was addressed through the development of the Remote Sensing Data Extraction Model (RESDEM). The main modules of RESDEM are: (1) TRMM, (2) SSM/I, (3) AVHRR, (4) AMSR, (5) MODIS, (6) QuickScat, and (7) Precipitation Verification. RESDEM performs the following functions: (1) sub-setting global dataset by location, (2) selection of temporal range of interest, (3) radiometric calibrations, (4) georeferencing, (5) application of user-defined minimum threshold values (e.g., precipitation), (6) extraction of derived products (e.g., reflectance), (7) verification and quality control, and (8) providing user-defined format for outputs. General advantages for RESDEM applications include: (1) user friendly technology, (2) reduction in processing time (e.g., processing of the entire TRMM for Africa was reduced from 1 year to a week), and (3) reduction in required computer resources. Examples of available RESDEM products include: processed TRMM (1998-2007) and SSM/I (1987 to 1998) data for six continents (www.esrs.wmich.edu). Examples of RESDEM applications in hydrology and LCLUC include development of continuous (1998-2007) rainfall runoff models taking advantage of readily available global remote sensing data sets. Continuous rainfall runoff models were developed (1998-2007) for Eastern Desert and Sinai; 3-hourly TRMM precipitation data was extracted, precipitation verified using automatic detection of cloud coverage from AVHRR, and soil moisture extracted from AMSR. Using temporal (every 12 hours: 1998-2007) wind speed and direction data (from QuickScat) and temporal (1938, 1947, 1960, 1969, 1977, 1986, 1987, 1994, 2001, 2003, 2005) aerial photographs, we demonstrated that dune movement in Cape Cod is largely related to the strong (up to 22 knots) NW to SE winter winds. Applications utilizing temporal global remote sensing data for conducting data extensive research are now within reach.