TRMM卫星降水数据在淮河息县流域径流模拟的适用性

Hydrological Simulation Capability of TRMM Satellite Precipitation Data in Xixian Catchment,Huai River Basin

卫星遥感降水由于高时空分辨率优势为现代水文模拟预报提供了有力的数据支持。结合地面观测数据,在中纬度淮河息县流域定量评估TRMM卫星降水（3B42RT和3B42V7）的精度,并采用栅格新安江模型进行卫星数据的径流模拟适用性分析。结果表明：3B42RT与3B42V7存在系统偏差,分别高估流域降水量25.25%与14.88%,ABIAS分别达到70.57%和67.61%,但二者与地面站点数据相关性较好,CC值达到了0.75以上;月尺度上精度有明显提高,ABIAS大幅度下降,CC值有较大提高,3B42V7的CC值为0.97。径流模拟方面,对卫星数据设计两种不同模拟情景,情景Ⅰ利用雨量站点观测数据率定模型参数,情景Ⅱ利用卫星降水数据重新率定模型参数,3B42RT在情景Ⅱ下模拟结果较情景Ⅰ下精度有所提高,但仍不及3B42V7模拟结果;3B42V7在情景Ⅰ、Ⅱ下都有较好的径流模拟表现,NSCE达到0.62以上,BIAS在±10.48%之间,CC达到0.79以上,表明TRMM卫星降水具有较好的径流模拟适用性。本研究可为TRMM和GPM卫星降水数据的水文应用提供参考借鉴。

Satellite precipitation products with high temporal-spatial resolution provide strong data support for modern hydrological simulation and forecast. Combining with the ground observation data, this study aims to evaluate the accuracy and hydrological simulation capability of the latest Tropical Rainfall Measuring Mission （TRMM） satellite precipitation product （TRMM 3B42RT and 3B42V7） in mid-latitude Xixian catchment, Huai River basin. The Gridded Xinanjiang model was used for streamflow simulation. The results show that, TRMM satellite precipitation has different extent overestimation of the benchmark precipitation, with BIAS of 25.25% for 3B42RT and 14.88% for 3B42V7; in the absolute error sense, TRMM satellite precipitation has large values; for the correlation sense, TRMM satellite precipitation has good values more than 0.75. At the monthly time scale, the precision of the TRMM satellite precipitation has greatly been improved in terms of declined ABIAS values and increased CC values. With the model parameters first benchmarked by the rain gauge data, the behavior of the streamflow simulation from TRMM 3B42RT is poor, but it still could characterize the diurnal variation of the streamflow. When the model parameters were recalibrated by each individual satellite data, the performance of TRMM 3B42RT was significantly improved. Compared with the TRMM 3B42RT, the TRMM 3B42V7 has good streamflow simulations in both parameter calibration scenes, with NSCE values more than 0.62, BIAS values between ±10.48% and CC values more than 0.79, which demonstrated good hydrological simulation utility of the TRMM data. The present work will hopefully be a reference for future hydrological utilizations of TRMM and GPM （Global Precipitation Measurement） satellite precipitation products.