Integral Radar Volume Descriptors for Quantitative Areal Precipitation

点击翻译按钮获取中文摘要

The primary goal of the proposal is the specification of Integral Radar Volume Descriptors (IRVD) and their application to quantitative areal precipitation estimates. We hypothesize that the spatio-temporal structure of radar reflectivity contains sufficient information in order to significantly improve the estimation of precipitation at the ground in comparison to currently used local Z-R-relations. The validity of this hypothesis will be proven by an objective scale-dependent analysis of existing radar volume data together with simultaneous surface observations of rain intensity. Secondary goals are objective structure-based clutter elimination, efficient radar volume storage and visualisation. The identification of precipitation process indicators in radar data will be achieved by a wavelet-based analysis of spatial and temporal structures in 4-dimensional radar measurements. Their relation to surface area rainfall on different scales in space and time results in a multi-dimensional, multi-scalar transfer function. For algorithm design and development, we will use the archived C-Band radar data from Karlsruhe (AQUARadar-C). In cooperation with SPP1167/QUEST and AQUARadar-E, a radar simulation tool for weather forecast models will be extended to include polarization. The tool will be applied to the output of a meso-scale weather prediction model (Lokal-Modell of DWD) supplied with new cloud modules (provided by AQUARadar-C) for case studies during the Special Observation Pediod (SOP) to generate pseudo-radar data. The analysis of IRVD-performance on the pseudo-radar data will both help to better understand the performance of the derived algorithms and shed light to deficiencies of the meso-y-scale model in simulating the precipitation process. Finally we will participate in the SOP in order to determine sub-scale variability of radar reflectivity within a radar volume by providing 2 MRR including rain gauges, microwave profiler, disdrometer and ceilometer. Besides the mapping function from 4D radar data to surface rainfall, specific novel objectives, which will be investigated during the project, are pattern based clutter filering, identification of relevant structures and their detection using structure-conserving smoothing operators, wavelets and regularization networks. For the efficient storage and processing of the huge data volumes, compression methods based on hierarchical bases will be used.