Triple-Frequency Doppler Retrieval of Characteristic Raindrop Size

A retrieval for characteristic raindrop size and width of the drop size distribution (DSD) based on triple-frequency vertical Doppler radar measurements is developed. The algorithm exploits a statistical relation that maps measurements of the differential Doppler velocities at X and Ka and at Ka and W bands into the two aforementioned DSD moments. The statistical mapping has been founded on 7,900 hr of disdrometer-observed DSDs and their simulated Doppler velocities. Additionally, a retrieval of D-m based only on DDVX-W measurements is also presented, and its performance is compared to the analogous algorithm exploiting DDVKa-W data. The retrievals are tested using triple-frequency radar data collected during a recent field campaign held at the Juelich Observatory for Cloud Evolution (JOYCE, Germany) where in situ measurements of the DSD were carried out only few meters away from the vertically pointing radars. The triple-frequency retrieval is able to obtain D-m with an uncertainty below 25% for D-m ranging from 0.7 to 2.4 mm. Compared to previously published dual-frequency retrievals, the third frequency does not improve the retrieval for small D-m (< 1.4 mm). However, it significantly surpasses the DDVKa-W algorithm for larger D-m (20% versus 50% bias at 2.25 mm). Also compared to DDVX-W method, the triple-frequency retrieval is found to provide an improvement of 15% in terms of bias for D-m = 2.25 mm. The triple-frequency retrieval of sigma(m) performs with an uncertainty of 20-50% for 0.2 < sigma(m) < 1.3 mm, with the best performance for 0.25 < sigma(m) < 0.8 mm.