- AREA 2 - Improve radar products by using NWP products
One of the most significant problems of using radar data in flood forecasting procedure and of integrating radar data with NWP techniques is in ensuring that parameters derived from radar data are free of contamination to which radar data are generally prone. This is also relevant in reducing the impact of unknown biases on the accuracy of parameter extraction from radar data only. The mutual interaction between radar parameters and NWP methods allows a positive feedback to radar products itself. An integrated use of all information available could decrease the interpretational errors due to atmospheric inhomogeneity.
Anomalous propagation (ANAPROP) is a common source of strong non-atmospheric reflectivity in radar measurements. Thus it is very important to eliminate such spurious effects before the radar data can be used for assimilation in NWP or for 'real-time' precipitation estimation. The key tool for achieving this will be a microwave propagation model, which will be used to diagnose and predict the incidence of ANAPROP. The atmospheric conditions required as input for the propagation model will be taken from standard TEMP products and from NWP mesoscale data. The outputs from this model would be provide algorithms and products capable of improving ANAPROP recognition and cancellation in real-time.
In the current practice of operational radar, it is usual to obtain volumetric reflectivity information under the assumption that the thermodynamic phase of hydrometeors is homogeneous (usually liquid water) in the whole measured volume. In the European climate, such a simplification leads to large biases in radar data as the thermodynamic phase in reality varies within the 3D volume. This spatial variation strongly affects the magnitude of attenuation along the beam and it also distorts the relationship between radar reflectivity factor and the precipitation intensity. The novel work here will seek to use 3D atmospheric fields from a NWP model to diagnose the 3D distribution of hydrometeor phase. This, in turn, would enable reliable calculation of the attenuation along the beam and better estimation of precipitation intensity and water content. The direct result of this exercise will be better reflectivity data both for assimilation and nowcasting purposes.
A crucial problem in quantitative application of radar measurements at long ranges is the sampling difference between the actual precipitation reaching the ground and the radar estimate well above the ground level. The radar simply cannot detect the actual reflectivity profile below the lowest elevation beam. This feature of radar observation often introduces strong underestimation of surface precipitation at longer ranges. On the other extreme, intensive frontal precipitation at the radar beam level often evaporates completely before it reaches the ground. To correct for such effects, the measured vertical radar reflectivity profile and the 3D reflectivity distributions parameterised from the 3D NWP model fields will be used. The methods and algorithms will improve considerably the radar estimates of surface precipitation at long ranges. The main advantage of this technique is that it widens the quantitatively reliable coverage area of a radar network.
Radar rainfall retrieving procedures are, mainly, based on averaged (both spatially and temporally) drop size distribution, which lead to a fixed Z-R relationship. A more physical based approach, which takes full advance of polarimetric radar data, is proposed in order to evaluate the errors and uncertainty related with a single Z-R approach.