光流法雷达外推产品在突发强降水预报中的应用

Application of optical flow radar extrapolation precipitation products for forecastingsudden heavy rainfall

地形条件复杂的山丘区中小河流洪水突发性强、汇流时间短,高准确率和长时效性降水短时临近预报产品对提高突发洪水预报精度尤为关键。以2021年9月河南省鸭河口水库出现的千年一遇特大洪水为例,利用国家气象信息中心提供的三源融合格点实况降水资料,检验基于改进光流法的雷达外推降水预报产品在本次洪水过程中0~1 h和0~3 h降水预报的TS评分和预报偏差。结果表明:(1)改进光流法在0~1 h的逐小时降水预报上较为精准,累计雨量在50 mm以下时,TS评分在0.45~0.85之间;雨量在50~70 mm之间时,TS评分在0.35~0.70之间;雨量在70 mm以上时,TS评分在0.25~0.35之间。50 mm以上雨量有较高TS评分表现出改进光流法在极端强降水预报中的优势性。(2)改进光流法在0~3 h的降水预报上,累计雨量在50 mm以下时,TS评分在0.55~0.85之间;在50 mm以上时,TS评分在0.35~0.75之间。该降水预报产品不仅对极端性降水预报效果较好,且预报时效长达3 h,可为防洪调度提供更长的决策时间。(3)改进光流法在0~3 h的降水预报产品与融合实况格点降水相比,雨量在20 mm以下的预报结果比较接近,平均绝对误差在10 mm以内;雨量在20 mm以上时,随雨量增大,平均误差、平均绝对误差、均方根误差均逐渐增大。(4)改进光流法在0~3 h的降水预报产品对影响范围小、降水强度大、维持时间长、累计雨量大的极端强降水有较好的预报表现。研究成果可为洪水预报模型提供一种较为可靠的降水输入预报。

The floods occurring in small and medium-sized rivers in complex hills are strong suddenness and short confluence time.The precipitation short-term proximity prediction product with high accuracy and long-term timeliness is particularly crucial for flood forecasting models.Taking the 1000-year return period catastrophic flood of Yahekou Reservoir in Henan Province in September 2021 as an example,the TS score and forecast deviation of radar extrapolated precipitation prediction products based on the improved optical flow method are examined for the 0~1 h and 0~3 h precipitation forecasts during this flood using the three source fusion grid point real precipitation data provided by the National Meteorological Information Center.The results showed that:①The improved optical flow method is more accurate in 0~1 h hourly precipitation forecast.When the accumulated rainfall is below 50 mm,the TS score ranges from 0.45 to 0.85;when the rainfall is between 50 mm and 70 mm,the TS score is between 0.35 and 0.70,when the rainfall is above 70 mm,the TS score is between 0.25 and 0.35.The high TS score of rainfall above 50 mm shows the superiority of the improved optical flow method in extreme heavy rainfall prediction.②In 0~3 h precipitation forecast,when the accumulated rainfall is less than 50 mm,the TS score is 0.55 to 0.85,when it is above 50 mm,the TS score is between 0.35 and 0.75.This precipitation forecast product not only has good prediction effect on extreme precipitation,but also has a prediction time up to 3 hours,providing a longer decision-making time for flood control scheduling.③Compared with the grid point precipitation of the fusion real situation,the precipitation forecast product of the improved optical flow method in 0~3 h is relatively close to the precipitation below 20 mm,and the average absolute error is within 10 mm.The error gradually increases when the rainfall exceeds 20 mm;with the increase of rainfall,the average error,average absolute error and root mean square error increase gradually.④The improved precipitation prediction products of optical flow method in 0~3 h have good prediction performance for extreme heavy precipitation of small influence range,large precipitation intensity,long duration and large accumulated precipitation.The research results can provide a more reliable precipitation input forecast for flood forecasting model.