在雷达测雨和洪水预报中自适应卡尔曼滤波法的应用

Application of Self-Adaptive Kalman Filter Method to Radar Rainfall Estimation and Flood Forecasting

采用自适应的卡尔曼滤波方法,运用少量雨量计对雷达估测降水进行了校正,结合新安江水文模型和GAME/HUBEX项目1998年在史灌河流域观测的水文资料,对6月28日08:00～7月5日20:00史灌河流域的一次洪水过程进行预报,并与雨量计网测量的面雨量预报结果进行对比.比较结果表明,经过校正后的雷达估测雨量和雨量计测量的面雨量都能较好地预报出流域出口蒋家集的洪水变化趋势、洪峰时刻以及洪水总量,雷达预报的结果同实际观测更加接近,而雨量计预报结果明显偏高.

Rainfall estimation from radar data has been increasingly used for flood forecasting applications, but the errors of rainfall estimation from radar data often led to the bad performance of hydrological model. In this paper, we use self-adaptive Kalman filter method to calibrate rainfall estimation from radar data which then are applied to Xin＇anjiang model to forecast flood of SHIGUANHE catchment during at 08：00 on 28 June to 20：00 on 5 July 1998 in GAME/HUBEX experiment period. By comparison of the predicted flood from radar data and that from gauge data, the results show that both the value of the flood forecasted from radar data and that from rain gauge data agreed well with the observed value. Among them, the result from radar data is more closer to the observed value than that from rain gauge data.