Quantitative Estimation of Precipitation from Radar Measurements: Analysis and Tool Development

Abstract

Current hydrological modeling efforts in Norway usually involve hourly or daily precipitation data obtained from a network of rain gauges. There is often a mismatch between the gauge network density and the spatial scale of precipitation events, causing severe errors in interpolated precipitation. The availability of high-resolution weather radar data from the Norwegian Meteorological Institute (MET) has made radar a potential tool to improve estimation of precipitation between the rain gauges. Despite the recognition of this potential, the use of radar-based precipitation estimates in hydrological research and applications in Norway has been very limited. This can be attributed to both scientific and technical challenges. From a scientific point of view, the necessary steps in estimating quantitative precipitation from radar reflectivity measurements poses several challenges, and involve multiple sources of errors which require correction algorithms. From a technical point of view, the limited availability of flexible radar data processing tools prevents hydrologists from working with radar data. Such tools are necessary for the efficient testing of existing algorithms, the development of new algorithms and for generating user-specific radar precipitation products. As a first step towards addressing the challenges mentioned above, this PhD project aims to identify the main sources of errors in precipitations estimates from a radar in central Norway, to develop correction algorithms for some of these errors and encapsulate these in a software tool where also other algorithms can be conveniently added by the user through a pre-defined interface. First a long-term evaluation of the existing radar precipitation product from MET was carried out using a GIS toolset developed in this study. From this evaluation, non-uniform vertical profile of reflectivity (VPR) was identified as one of the main sources of errors. A correction algorithm for this source of error was developed using the raw radar volume data as input. Finally, a program which includes this correction algorithm and other routines which are necessary for generating precipitation products for hydrological applications was developed.