基于遥感和GIS的日最高最低气温估算

气温是气象要素的重要组成部分,广泛用于全球气候变化、资源环境分析及灾害预警等多个领域.随着卫星遥感技术的发展,气温的估算趋向于遥感或遥感和GIS结合的方法.本文以浙江省为研究区域,利用了36个站点2013年逐日每10min一次的自动气象站气温观测数据和MODIS地表温度及其他参数产品,选用多元线性回归(自变量为地表温度、归一化植被指数、地表反照率、经度、纬度和高程)、温度植被指数以及多元线性回归插值方法进行气温估算,建立了研究区日最高气温最低气温估算模型,并比较了几种气温估算方法在研究区的适用性.结果表明:3种方法最高气温估算的决定系数(R~2)分别为0.96、0.91、0.97,均方根误差(R_(MSE))分别为1.84、2.75、1.49℃;多元线性回归和多元线性回归插值法最低气温估算的R~2分别为0.87、0.91,R_(MSE)分别为3.33、2.93℃,两者均为多元线性回归插值法得到的结果最好.空间分布结果显示,多元线性回归插值法能很好地反映由地形不同所带来的细节差异.

Estimating daily maximum and minimum air temperatures by remote sensing and GIS

Air temperature is an important meteorological factor widely used in the evaluation of global climate change,environmental analysis,and disaster early-warning.With advances in satellite remote sensing technology,air temperature estimation tends to utilize remote sensing data or a combination of remote sensing and GIS.MODIS land surface parameter products and air temperature data in 2013 from 36 automated meteorological stations were used to estimate daily maximum and minimum air temperatures in Zhejiang province,by multiple linear regression (MLR)(variables include land surface temperature,normalized difference vegetation index,surface albedo,longitude,latitudes,elevation),temperature-vegetation index(TVX) and multiple linear regression interpolation (MLRI).R2 of MLR,TVX and MLRI for maximum air temperature were found to be 0.96,0.91,0.97,RMSE were 1.84,2.75,1.49 ℃ respectively.R2 of MLR and MLRI for minimum air temperature were 0.87,0.91,RMSE were 3.33,2.93 ℃ respectively.MLRI performed the best.Spatial patterns indicated that the MLRI method could better reflect temperature differences due to topography in areas with large elevation ranges.