Relevance of merging radar and rainfall gauge data for rainfall nowcasting in urban hydrology

摘要

Accurate Quantitative Precipitation Forecasts (QPF) at high spatial and temporal resolution are a perquisite for urban flood prediction. Commonly Lagrangian extrapolation of the rainfall patterns recognized by radar data, forms the basis of such forecasts for the near future (up to 2 h lead time) - referred here as nowcasting. Nevertheless, due to the intermittent nature of the rainfall at such fine scales, the predictability of storms is limited to about 20 min. The predictability loss is caused mainly by the inability of the radar to capture the true rainfall field and because the Lagrangian Persistence is unable to model the temporal evolution of the rainfall field. In this study we focus on the first problem, on how to extend the predictability limit of rainfall at such scales by improving the rainfall field fed into the nowcast model. To overcome the errors associated with the radar intensities, merging techniques between radar and gauge measurements are advised. Among different employed techniques (mean field bias, kriging with external drift and quantile mapping based correction) the conditional merging between the radar and rainfall-gauge measurements seems to capture at best the spatial and temporal patterns of the rainfall at the desired fine scales (1 km(2) and 5 min). Moreover, when fed to two nowcast models, the conditional merging doesn't only increase the predictability of storms from 20 min to longer than 1 h, but as well it improves the agreement of radar based QPF with the gauge measurements. The results are drawn from 110 events observed in the period 2000-2018 by the Hannover Radar (Germany) in an area with a radius of 115 km, where 100 recording gauges were available. As the urban hydrological models are commonly validated on gauge measurements, nowcasting with conditionally merged data, provides a useful tool for urban flood prediction for lead times up to 1 h.