青藏高原地区被动微波遥感反演地表温度算法验证

Validation of Surface Temperature Inversion Algorithm Using Passive Microwave over Qinghai-Tibet Plateau

近年来,青藏高原地区的水热平衡成为关注的焦点,而地表温度是陆表过程模型的重要输入参数之一。被动微波遥感在地表温度反演上已经取得了一些进展。本文重点用被动微波数据反演地表温度算法对青藏高原地区的数据做不同下垫面的地面验证和分析,包括Mao(2005)、Richard(2003)、Zhao(2011)3种算法。研究表明:Richard(2003)的单通道算法能够适应低矮植被地区,反演精度高;Zhao(2011)算法在裸土地区的反演精度更高;而Mao(2005)算法出现了低估的情况。研究发现3种算法的绝对误差随不同时间降雨的变化呈现相同的波动趋势,即反演精度受到降雨的影响,降雨量增大,温度反演误差变大;降雨之后,随着地表逐渐干燥,土壤水分逐渐减小,误差随之减小。

Surface soil temperature is an important prognostic variable in land surface processing models and an input parameter in radiative transfer models.Therefore,there is a distinct need for accurate estimates of surface soil temperature in environmental modeling.This paper firstly reviews three algorithms of surface temperature inversion which are Mao（2005）,Richard（2003）and Zhao（2011）,and then makes a comparative analysis on overall accuracy.The result indicates that the Richard（2003） algorithm appears to be performing well at low vegetated land area,and the precision of Zhao＇s algorithm is better at barren and sparsely vegetated area.It also indicates that the absolute errors of three algorithms show the same sequence trend of fluctuations over time,i.e.,the precision decreases as precipitation is increased;after the rain,as the surface is drying,soil moisture gradually decreased and the error then decreases.