GPM卫星在台风降水过程中的应用评估

Evaluation on application of GPM satellites during typhoon-precipitation process

【目的】为了解新一代气象卫星对气象灾害预警的监测水平,【方法】以长三角地区约6 000个地面雨量站数据为基准,利用相关系数(CC)、相对偏差(RB)、均方根误差(RMSE)以及分级评分指标等方法定量评估GPM多源卫星融合产品(V6B)IMERG(E_Cal、L_Cal、F_Cal、F_Uncal)在2015—2021年期间影响长三角地区8个台风(“灿鸿”“温比亚”“云雀”“利奇马”“米娜”“黑格比”“烟花”“灿都”)的降水估测能力。【结果】结果表明:(1)IMERG的4个产品基本可以捕获台风期间的降水分布,它们在长三角地区的平均CC分别为0.70、0.72、0.78、0.73,但容易低估该区域的累积雨量。(2)IMERG在降水中心的雨量估测表现更差,对应的CC分别为0.42、0.44、0.53、0.48。(3)IMERG可监测出长三角和降水中心小时降水的变化趋势,平均CC均达到0.8以上,但它们对短时降水峰值有不同程度的低估和滞后。【结论】IMERG对小于5 mm/h的降水估测能力较好,但随着降水强度(5~20 mm/h)的增加,估测能力逐渐降低,几乎无法估测出大于20 mm/h的极端降水。总体上,F_Cal、F_Uncal、L_Cal、E_Cal在台风期间的降水估测表现依次变差,它们对强降水的估测精度仍需进一步提升。

[Objective] In order to understand the monitoring levels of the new generations of meteorological satellites for meteorological disaster early warning,[Methods]the precipitation estimation capabilities of GPM multi-source satellite fusion products,i.e.(V6B)IMERG(E_Cal,L_Cal,F_Cal,F_Uncal,during 8 typhoons:Chan-hom,Rumbia,Jongdari,Lekima,Mitag,Hagupit,In-fa and Chanthu from 2015—2021 are quantitatively assessed with the methods of correlation coefficient(CC),relative bias(RB),root-mean-squared error(RMSE),grading and scoring indexes,etc.through taking the data from about 6000 rain-gauge stations in the region of the Yangtze River Delta as the basis.[Results] The results show that(1) four products of IMERG can basically capture the precipitation distributions during the periods of the typhoons,of which the mean CC are 0.70,0.72,0.78 and 0.73 respectively within the region of the Yangtze River Delta,but the accumulated rainfalls in the region are prone to be underestimated;(2) IMERG products are poorly performed for the rainfall estimation within the precipitation center with the corresponding CC of 0.42,0.44,0.53,0.48 respectively;(3) the four products of IMERG can accurately monitor the changing trends of the hourly precipitations in the whole region of the Yangtze River Delta and the precipitation centers with all the mean CC of over 0.8,but they have different extents of underestimations and laggings for the peaks of short-term precipitation.[Conclusion] IMERG products have better estimation capabilities for the precipitation less than 5 mm/h,but the estimation capabilities are to be gradually decreased along with the increase of the precipitation intensity(5~20 mm/h) and then almost impossible to estimate the extreme precipitation over 20 mm/h.The precipitation estimation performances of F_Cal,F_Uncal,L_Cal,E_Cal generally become worse in turn during the typhoons,thus the accuracies of their estimation capabilities for intensive rainfall are necessary to be further improved.