Data Fusion, Data Mining, Pattern Recognition and Regional Classification of the Water Quality Data in the Mid-Western United States
College: Illinois State Water Survey
Award year: 2005-2006
The objective of the proposed research is to investigate the nutrient and sediment data of the Mid-Western streams and their spatial and temporal attributes using new technologies, such as data fusion, data mining and pattern recognition. The goal is to develop methods for regional classification of watersheds, regional regression equations for predicting water quality parameters as a function of various habitat factors. The research will also address optimum sampling frequency and bias reduction for load calculation of selected nutrients and sediment. The proposed multi-disciplinary research will be based on the water resource expertise at Illinois State Water Survey (ISWS) and computational expertise at National Center for Supercomputing Applications (NCSA).
The classification and regression methods will relate geo-spatial GIS-based watershed attributes with water quality parameters to find homogeneous geographical partitions using artificial neural networks, hierarchical clustering and segmentation, advanced K-Means and decision tree algorithms. The NCSA computational resources for memory demanding data management and computationally intensive processing will facilitate data analyses and access to the heterogeneous database of input and output data, including federal agencies, such as, NASA, NRCS, NWS, USACE, USDA, USEPA, USGS, and state agencies, for example, IEPA and ISWS.
The results of this largely data-driven research will enhance our ability to evaluate the variations in water quality as a function of the spatial attributes, land-use and climatic variability. The proposed research will investigate patterns related to the fundamental water quality and quantity processes, such as the processes governing sediment and nutrients yields from agricultural watersheds. The results will also enhance our ability to evaluate the effects of a range of future land-use/climate scenarios on nutrients and sediment. Overall, the research will facilitate effective, efficient and sustainable water management strategies that promote protection of surface water from nutrient and sediment contamination.