Twenty years of radar data from a single C-band radar - Potentials and drawbacks in rainfall statistics

In terms of spatial and temporal information, weather radar data offers notable advantages when used for rainfall statistics compared to rain gauge records. In the past, long-term statistics were only possible through the use of rain gauge records. However, the availability of more than one decade of radar data now allows for a more detailed analysis of rainfall variability in both space and time.

For radar-based statistical analyses to be a reliable alternative to rain gauge statistics, it is essential to have high-quality, consistent data that is free of errors. While advancements in technology, such as dual polarimetric estimation, have improved our ability to quantify rainfall intensities with radar, it is still necessary to utilize "ground truth" records from rain gauges to adjust and ensure the accuracy of the radar estimates. In extreme value statistics where long-term continuous records are required, it is necessary to not only utilize the latest technology but also to adjust and verify older data (e.g. pre dual-pol. data or data with lower spatial and temporal resolution)

This abstract present data from a 20-year radar series single radar from Denmark, where several analyses with regard to rainfall statistics have been conducted. We describe and quantify challenges in bias adjustment, advection interpolation to improve temporal resolution, duration-dependent biases, spatial scaling issues comparing points and pixels, subpixel variability, range dependence in rainfall estimates, etc.

We apply the data to develop extreme value statistics based on peak-over-threshold ranking and stochastic storm transposition. Furthermore, we examine spatial variability using radar-based areal reduction factors as well as spatial correlation analyses and severity diagrams. The analyses and application of radar data are conducted with an aim towards application in urban hydrology.