HR: 08:45h
AN: SF11A-04    [Abstracts]
TI: Jointly Retrieving Surface Soil Moisture from Active and Passive Microwave Observations Using Cubist Data-Mining
AU: * Zhan, X
EM: xzhan@hsb.gsfc.nasa.gov
AF: UMBC-GEST/NASA-GSFC, Code 974.1 NASA-GSFC Hydrological Sciences Branch, Greenbelt, MD 20771 United States
AU: Houser, P R
EM: Paul.R.Houser@nasa.gov
AF: NASA-GSFC Hydrological Sciences Branch, Code 974 NASA-GSFC Hydrological Sciences Branch, Greenbelt, MD 20771 United States
AB: With the successful launches of NASA's Earth Observing Satellites (e.g. Terra & Aqua) and several environmental satellites (e.g. NPOESS, NPP, SMOS and HYDROS) being planned to launch in the near future, huge amounts of satellite remote sensing data are being collected every day. Maximizing the use of this wealth of data sets is a pressing issue for the Earth system science community. Data mining is extracts patterns from large system data sets. These patterns provide insight into system characteristics that enable outcome prediction for future situations that aids decision-making. The Cubist data-mining algorithm is a powerful tool for generating rule-based models that balance the need for accurate prediction against the requirements of intelligibility. Cubist models generally give better results than those produced by simple techniques such as multivariate linear regression, and are generally easier to understand than neural networks. The NASA's Hydrosphere States (HYDROS) mission, an Earth System Science Pathfinder, will use both L-band microwave coarse resolution radiometer and fine-resolution radar to make the first space borne observations of global soil water availability. These new observations will enable new scientific investigations of atmospheric predictability and global change processes. To assess the potential accuracies of retrieving land surface soil moisture from the radiometer and radar observations, the HYDROS science team has created an Observing System Simulation Experiment (OSSE) that includes a complete land surface geophysical properties data set (soil moisture, surface temperature, vegetation temperature, etc), the associated atmospheric variables, and the simulated HYDROS radar and radiometer observations for the Red-Arkansas river basin. We have applied the Cubist data-mining algorithm to this OSSE data set to evaluate its soil moisture retrieval skill using the active and passive microwave observations simultaneously. The resulting simple rules and models provide insights into how soil moisture soil is related to land-surface geophysical and meteorological variables. The potential to use this data mining tool for analyzing other NASA satellite observations will also be discussed.
DE: 9820 Techniques applicable in three or more fields
DE: 1899 General or miscellaneous
DE: 1699 General or miscellaneous
SC: Special Focus: Advances in Data Acquisition, Management, Analysis and Display [SF]
MN: 2004 AGU Fall Meeting