摘要
利用国家卫星气象中心处理的NOAA下午轨道卫星的OLR资料,用Xie等在1998年的文章中提出的月降水量计算模式,计算了1991-2008年地理范围在10°-60°N、75°-150°E、分辨率为0.5°×0.5°的中国大陆月降水量,得出:用OLR月距平资料可以计算出月降水量,模式估算出的降水量通过与NCEP提供的18年月降水量陆地观测数据对比,精度为:冬季相对误差49.14%、绝对误差7.97 mm;春季相对误差37.60%、绝对误差14.97mm;夏季相对误差27.37%、绝对误差31.61mm;秋季相对误差37.99%、绝对误差16.95 mm,可见精度效果并不是太好,造成误差的主要原因是降水机制不一,层状云降水特别是逆温层状云和连续阴天不下雨,以及月平均OLR不能完整地反映月内降水云和降水量是造成用OLR月距平估算月降水量的主要误差来源。通过对FY-2C卫星云分类产品的图像分析,得出中国南方冬季主要是层状云降水,OLR月距平值较高,用全球的A、B系数估算出的降水量偏低于实况,因此对中国大陆进行分区、分季节统计A、B系数,是解决OLR月距平估算月降水量精度问题的途径。
By using the model of Xie and the NOAA OLR data in NSMC,the monthly precipitation amount in 1991-2008 was calculated in the area of 10°-60°N,75°-150°E land with data resolution of 0.5°×0.5°latitude-longitude.Therefore we know that monthly precipitation amount could be calculated out by using monthly OLR anomaly.By comparing the model monthly precipitation with gauge precipitation from NCEP in 1991-2008,we obtained the following results：in winter the relative error in China was 49.14%and the absolute error was 7.97 mm;in spring they were 37.60%and 14.97 mm;in summer they were 27.37% and 31.6 mm;and in autumn they were 37.99%and 16.95 mm.The results were not very good.The main errors were caused by stratus cloud precipitation and continuous cloud cover days without precipitation.By analyzing cloud classification products of FY-2C during 2006-2008,we have learned that in winter stratus clouds are common in southern China,and the monthly OLR anomaly there is usually high,and the estimation of precipitation amount obtained from global coefficients A and B is lower than that of ground observation. Therefore,the better way to solve the problem of precipitation underestimation may be in the choice of new coefficients A and B for different areas of China in different seasons.
出处
《气象》
CSCD
北大核心
2011年第1期99-102,共4页
Meteorological Monthly
关键词
OLR
月降水量
绝对误差
相对误差
OLR
monthly precipitation amount
relative error
absolute error
作者简介
吴晓，从事气象卫星遥感反演产品处理．Email：wuxiao@cma．gov．cn
