Estimation of Z-R relationship and comparative analysis of precipitation data from colocated rain-gauge, vertical radar and disdrometer
Abstract
The estimation of precipitation rates and accumulations constitutes an essential input data for distributed hydrological models and for many hydrometeorological applications, such as short term hydro-scheduling, forecasting and monitoring of river floods and inflow forecasting into a catchment, its peak flow and response time. In order to improve quantitative precipitation estimates from C-band weather radar, variable Z-R relationships that represent the local precipitation characteristics and conditions are derived from measured precipitation drop size distribution (DSD) by two ground-based instruments: vertical pointing Doppler radar MRR and the optical disdrometer OTT Parsivel. The variability and robustness of those local Z-R relationships are analyzed, as well as their level of dependency to the storm type (stratiform, convective, air mass convection), to the precipitation phase (rain, snow, mixed precipitation), to each event and to each month. Comparative analysis of the precipitation accumulations measured or estimated by three different local instruments (tipping bucket rain gauge, disdrometer OTT Parsivel, vertical pointing radar MRR) is also performed in order to assess under which conditions those instruments are able to provide reliable precipitation rates and robust local Z-R relationships.MRR provides the most variable and uncertain local Z-R parameters that are highly dependent on the precipitation phase. This dependency leads to high event-to-event variability of DSD-measured Z-R relations and large differences in estimated precipitation rates and accumulations (especially for snow) between the local instruments and compared to precipitation values estimated from the weather radar, using the standard Z-R relationship. The optical disdrometer OTT Parsivel provides the most stable and robust Z-R parameters that are independent of the precipitation phase and the season. In the only case of rain events, both instruments derive similar local Z-R parameters that are comparable to the standard Z-R relationship. The high variability and uncertainty related to Z-R parameters concern mainly mixed precipitations. There is no evident dependency of Z-R parameters on storm type.When comparing the three local precipitation instruments, they all provide similar rain accumulations. In cases of snow and mixed precipitation, accumulations derived from those instruments are quite different. However, the disdrometer OTT Parsivel and the tipping bucket rain gauge agree relatively well in terms of accumulations. In addition that OTT Parsivel may provide robust local Z-R relationships for any kind of precipitation that correspond well to standard Z-R relationship, it may compensate for precipitation losses, catch deficit and low temporal resolution of the conventional rain gauge. However, long periods of instrument instability and breakdowns for the disdrometer reduce significantly the number of valid precipitation data available at any time of the year and any precipitation conditions, hence the data representativity of this instrument. In case of MRR, more investigations and possibly better measurement filtering and correction prior to the precipitation estimation may reduce the uncertainties around the MRR-derived Z-R parameters and precipitation estimates.