Potentiality Of Satellite Rainfall Products In Flood Forecasting

Flood forecasting can help human controling flood resources scientifically and efficiently.Satellite precipitation products which are new sources for precipitation data overcome spatial and temporal limitations,and these products are important for no or little data area.This study evaluates and compares the comprehensive performance of three satellite precipitation products(TRMM 3B42RT,TRMM 3B42 and PERSIANN)on multiple time scales and spatial scale to characterize the precipitation patterns over the Luanhe River basin above the Panjiakou Reservoir(Semi-arid and semi-humid climatic region).Meanwhile,this study investigates the capability of using satellite precipitation products as inputs for HEC-HMS model to forecast flood.The main results are as follows:(1)The performances of satellite precipitation products(TRMM 3B42RT,TRMM3B42,and PERSIANN)are evaluated by using some statistical indexes,including correlation coefficient(CC),mean error(ME),root mean squared error(RMSE),mean absolute error(MAE),relative bias(BIAs),probability of detection(POD),false alarm ratio(FAR)and critical success index(CSI).The results show that all these three satellite precipitation products have strong correlations with observed precipitation data,but TRMM 3B42 performs better than others.TRMM 3B42RT and TRMM 3B42overestimate precipitation,with BIAs of 62.80%and 20.17%,respectively,while PERSIANN underestimates precipitation with BIAs of-6.38%.Meanwhile,The error of TRMM 3B42 is the smallest among these three satellite precipitation products(RMSE and MAE on both daily and monthly scales are the smallest).Moreover,TRMM3B42 can 3B42 can detect more precipitation events,and it can accurately reflect the occurrence frequencies of different rainfall events and their relative contributions to the total rainfall.Overall,TRMM 3B42 performs better than other two satellite precipitation products over the Luanhe River basin above the Panjiakou Reservoir.(2)This study sets up the HEC-HMS hydrological model by using DEM,land use,soil and observed flood data.The outlets of hydrological model are Kuancheng and Sandaohezi hydrological satations,respectively.And then several indexes such as coefficient of determination(R~2),deterministic coefficient(DC),flood peak flow error(RE_p),flood runoff error(RE_v)and peak time error(△T)are used to evaluate the performance of model.The results show that the performance of HEC-HMS model is good enough.The qualified rate of indexes(except△T of model of Sandaohezi hydrological satation)are 100%.After comprehensively evaluating,the level of flood forecasting can reach B.As a result,this model can be further applied for flood forecasting over the study area.(3)This study further validates the performance of TRMM 3B42 on a 3-hour scale.The results show that the TRMM 3B42 satallite precipitation data has a strong correlation with observed precipitation data,with mean CC of 0.66.However,there are some errors and deviations,but they are acceptable.And TRMM 3B42 can be able to detect and distinguish much more rainfall events correctly.Furthermore,TRMM 3B42is used as the inputs for HEC-HMS model which is calibrated by gauge precipitation data to forecast floods.The simulated floods has a strong correlation with observed floods because the qualified rate of R~2 is 77.8%.However,there are large errors and deviations.The qualified rate of single index(except R~2)is no more than 50%,and of all indexes is only 22.22%.And then TRMM 3B42 is used as the inputs for HEC-HMS model which is calibrated by TRMM 3B42 precipitation data to forecast floods.The performance of HEC-HMS is greatly improved.The qualified rate of single index(except△T of 72.2%)is more than 85%,and of all indexes increases from 22.2%to61.1%.Therefore,although there are large errors and deviations,TRMM 3B42 has a great potential for flood forecasting.