Research Description:

Stochastic Modelling of Rainfall using Historical Point-Based Data with NEXRAD Spatial Data

The goal of this study is to develop a stochastic model for the spatial and temporal distribution of rainfall using historical rain gauge data in conjunction with data from the Next Generation Radar (NEXRAD). Rainfall totals and intensities have been monitored by rain gauges at point locations over the last century. With the introduction of Next Generation Radar (or NEXRAD) in 1988, high-resolution (1 km by 1 km pixel) images of precipitation can be produced. Recent developments use a combination of these methods: NEXRAD produces a spatially filled picture of rain; this data is "corrected" or enhanced, using the point measurements from the rain gauges. Since NEXRAD data is only available starting in the late 1990's, it may not be considered adequate for the purpose of probabilistic hydrologic forecasting. However, since the NEXRAD data is spatially complete, this improvement over rain gauge data should be exploited. Individually, both radar and gauge data have weaknesses, which could propagate down to affect the adjusted data. To temporarily circumvent this problem, this research will consider the set of data as ideal. By using spatial analysis of NEXRAD data together with temporal analysis of rain gauge data, an improved understanding of the stochastic description of the precipitation is sought. Such a development could ultimately lead to more accurate assessment and quantification of flood risks.