基于MODIS数据的云降水估计方法

Estimation method of cloud precipitation based on MODIS data

利用卫星遥感资料定量估计降水及其区域分布对强降水等灾害性天气监测具有十分重要的意义,还可弥补常规气象观测的不足,提供更为丰富的降水信息。(1)利用MODIS卫星的可见光、近红外、短红外、中红外和热红外波段数据,采用云指数法和多光谱综合法对云类型进行了判识,有效地分类出云和冰雪、卷云、水云和高、中、低云。(2)在云分类的基础上定量反演了各类云参数,包括云顶亮温梯度、云顶亮温最小值面积变化、不同量级的云顶亮温面积及变化趋势、对流云的核心位置判别、亮温方差和大气中的水汽含量等因子参数,结合地面实测降水数据和地表高程数据,建立了降雨的强度的定量估计模式。(3)通过对2007年7月16-18日、26-29日的新疆大降水过程进行降水估计,并对估计值与区域站地面实时观测降水值作了比较,结果显示相对误差为31.97%,均方根误差为2.31 mm。降水量级的定量误差在±3 mm以内的占66%,±5 mm以内的占86%。因此,该模式具有较好的降水估计效果。

Application of satellite remote sensing data to precipitation estimation and regional distribution is a meaningful technique for the monitoring of heavy rains as well as other disastrous weathers. It can also compensate the deficiency of the conventional meteorological observation and present more ample rainfall information.( 1) Using the cloud index method and multi-spectrum synthetic method,and based on visible-light,near-infrared,short infrared,mid-infrared and thermal infrared data of MODIS satellite,this paper identifies the cloud type and effectively classifies the cloud,ice and snow,and multilayer cloud( include cirrus cloud and water cloud),as well as high,medium,low cloud.( 2) On the basis of cloud classification,the paper quantitatively inverts the various types of cloud parameters,such as black body temperature gradient of cloud top,the minimum area change of black body temperature of cloud top,area of black body temperature of cloud top,the different magnitude order and their variation trend,the discrimination of core position of convective cloud,the variance of black body temperature andwater vapor content of atmosphere and etc. Combining with the measured ground rainfall data and surface elevation data,the paper quantitatively established an estimation model of rainfall strength.( 3) The estimated value and regional station's real-time rainfall observations were compared. The result is that : relative error is 31. 97%,root mean square error is 2. 31 mm. The regional stations are account for 66% for which quantitative error of precipitation grade is less than 3mm,and 86% for 5mm. So the model has better effect on assessing rainfall.