Simulation and verification of a flood inundation event in Linhai induced by super severe typhoon based on CMPAS data

Taking the flood inundation event in Linhai, Taizhou, Zhejiang Province caused by the super severe typhoon (1909) Lekima for example, we have verified and evaluated the 5km real-time data (CMPAS-5km) from the CMA Multi-source Precipitation Analysis System by using the precipitation data obtained by hydrological and meteorological stations over Zhejiang Province. In addition, we haveconductedflood simulations under the rainstorm scenarios based on the spatial distribution with weighted grid of area precipitation and accumulated precipitation calculated by CMPAS-5km to drive the FloodArea two-dimensional hydrodynamic inundation model, and performed verification of the simulated flooding by using Sentinel-1 SAR images and representative flooded points. The results show that the precipitation from CMPAS-5km and that from the stations have good temporal and spatial consistency, and the correlation coefficient (R) ofprecipitation from CMPAS-5km with those from hydrological stations and"hydrological + meteorological"stations are both 0.89. Rand root mean square error (RSME) of precipitation from CMPAS-5km with the hourly averaged precipitation from hydrological stations are 0.79 and 1.7 mm·h^-1, respectively. The hourly area rainfall among CMPAS-5km, hydrological stations and meteorological stations have more consistent variation trend in Taizhou flooded simulation area, and the accumulated area rainfall in the whole simulation process is 305.5 mm, 304.1 mm and 283.7 mm, respectively. CMPAS-5km is superior to the other two sets of data from ground stations in stability, timeliness, precision and data interface. The newly-increased water area in SAR images matches the FloodArea simulation results well, and the relative bias between the maximum simulated water depth and the actual water depth at the four verification points is within ±22%.