Geo-Spatial Analysis of NEXRAD Rainfall Data for Central and South Florida

The South Florida Water Management District (District) is responsible for managing water resources in 16 counties over a 46,439-square kilometer (17,930 square-mile) area. The area extends from Orlando to Key West and from the Gulf Coast to the Atlantic Ocean and contains the second largest freshwater lake that is wholly within the United States — Lake Okeechobee and the world famous Everglades wetlands. The District operates approximately 3,000 kilometers (~1,800 miles) of canals and more than 420 water control structures. Near-real-time rainfall data are used in operation of these pumps and water control structures. The District uses a network of approximately 280 active rain gauge stations that cover the more populated and environmentally sensitive areas under its management and provide data for this purpose. Since 2002, the District has being acquiring regional NEXRAD (Next Generation Weather Radar) data from a vendor that collects the radar data from four NEXRAD sites operated by the National Weather Service. Corporate access of 15-minute, rain gauge-adjusted NEXRAD data, for each of the 2 x 2 km cells in the grid covering the District, was a major objective of the acquisition. The District is receiving the NEXRAD rainfall data as two types of datasets — near-real time (NRT) and end-of-month (EOM). The District developed ArcIMS-based application for NEXRAD data retrieval which provides aggregated datasets that are varied spatially and temporally and available in tabular and image formats. The District's geo-spatial analysis of the NEXRAD rainfall data was performed using the data that was collected between 1995 and 2005 at 2 x 2 km resolution. A regular array of analysis blocks was used to account for the variability of rainfall processes and local rainfall patterns; the analysis block size used was 20 x 20 km. The spatial autocorrelations of the radar rainfall were identified using a semivariogram approach at hourly and daily time scales. The exponential model was fitted to the semivariograms of the detrended radar data from 1998 to 2005. The exponential model verification was performed for 1995 and 1996 data. The range parameters of the semivariograms (correlation distances) for hourly and daily time scales of wet season were found to be stable across the District and median value varied between 6 and 7 km.