卫星遥感降雨数据在嫩江石灰窑以上流域洪水预报中的可利用性

Availability of satellite remote sensing dataset for Shihuiyao basin of Nenjing River

依据洪水预报对降雨数据时效性要求较高的特点,选取GSMaP_NRT和GSMaP_Gauge_NRT准实时卫星遥感全球降雨产品对嫩江上游石灰窑以上流域进行补充,旨在解决该流域内降雨资料不足的问题。基于流域内及边缘降雨数据评估2种产品在雨量站所在网格上的精度,并输入新安江模型进行洪水预报,验证其可利用性。在汛期的月和日尺度上,GSMaP_NRT和GSMaP_Gauge_NRT产品表现出较高的精度,GSMaP_Gauge_NRT产品的精度总体优于GSMaP_NRT产品,但均存在高估现象。同时设计3种洪水预报输入方案进行应用检验,分别为方案Ⅰ(流域内及边缘雨量站泰森多边形权重降雨)、方案Ⅱ(流域内GSMaP_NRT栅格降雨均值)、方案Ⅲ(流域内GS-MaP_Gauge_NRT栅格降雨均值)。洪水预报结果表明,方案Ⅲ的预报效果最好,产汇流合格率较高,表明GSMaP_Gauge_NRT准实时降雨产品在洪水预报中具有较高的可利用性。

The Shihuiyao basin is located in the upper of the Nenjiang River,with basin area of 17,205 km^(2).Fewer rainfall stations within or near the basin lead to insufficient rainfall data.It is difficult to accurately predict the flooding process at the outlet section of the basin using existing rainfall data,which brings certain difficulties to the downstream river flood control and the Nierji Reservoir′s forecast flood control operations.Based on the characteristic of flood forecasting,the near real-time satellite remote sensing global precipitation products GSMaPNRT and GSMaPGaugeNRT are selected to supplement the Shihuiyao basin,aiming to solve the problem of scarcely gauged precipitation.The two products accuracy on the rainfall stations grid is evaluated based on the precipitation within or near the basin.Three flood forecasting input schemes are designed:scheme I precipitation of rainfall stations within or near the basin by Tyson polygon weighted method,schemeⅡ(average precipitation of GSMaPNRT in the basin),schemeⅢ(average precipitation of GSMaPGaugeNRT in the basin),and combined with the Xin′anjiang model for flood forecasting to verify their availability.On monthly and daily scales for flood season,GSMaPNRT and GSMaPGaugeNRT show high accuracy,but both overestimate the rain gauge observation precipitation,and the accuracy of GSMaPGaugeNRT is generally better than that of GSMaPNRT.The results of 8 flood forecasts during the training and validation stages show that the forecast effect of schemeⅢis the best,and the runoff and confluence results both have higher accuracies.Results indicated that the GSMaPGaugeNRT has high availability in flood forecasting in the Shihuiyao basin,and the GSMaPGaugeNRT can be used as a source of rainfall data in ungauged basins.In future,more near real-time satellite remote sensing global precipitation products can be applied for flood forecasting to further improve flood forecasting accuracy.