【摘要】：氣溫是氣象要素的重要組成部分,廣泛用于全球氣候變化、資源環(huán)境分析及災(zāi)害預(yù)警等多個(gè)領(lǐng)域.隨著衛(wèi)星遙感技術(shù)的發(fā)展,氣溫的估算趨向于遙感或遙感和GIS結(jié)合的方法.本文以浙江省為研究區(qū)域,利用了36個(gè)站點(diǎn)2013年逐日每10min一次的自動(dòng)氣象站氣溫觀測數(shù)據(jù)和MODIS地表溫度及其他參數(shù)產(chǎn)品,選用多元線性回歸(自變量為地表溫度、歸一化植被指數(shù)、地表反照率、經(jīng)度、緯度和高程)、溫度植被指數(shù)以及多元線性回歸插值方法進(jìn)行氣溫估算,建立了研究區(qū)日最高氣溫最低氣溫估算模型,并比較了幾種氣溫估算方法在研究區(qū)的適用性.結(jié)果表明:3種方法最高氣溫估算的決定系數(shù)(R~2)分別為0.96、0.91、0.97,均方根誤差(R_(MSE))分別為1.84、2.75、1.49℃;多元線性回歸和多元線性回歸插值法最低氣溫估算的R~2分別為0.87、0.91,R_(MSE)分別為3.33、2.93℃,兩者均為多元線性回歸插值法得到的結(jié)果最好.空間分布結(jié)果顯示,多元線性回歸插值法能很好地反映由地形不同所帶來的細(xì)節(jié)差異.
[Abstract]:Temperature is an important component of meteorological elements, which is widely used in many fields such as global climate change, resource and environment analysis, disaster warning and so on. With the development of satellite remote sensing technology, the estimation of temperature tends to the method of remote sensing or the combination of remote sensing and GIS. Taking Zhejiang Province as the research area, the temperature observation data of automatic weather station and MODIS surface temperature and other parameter products of 36 stations in 2013 were used to select multiple linear regression (independent variable is surface temperature). Normalized vegetation index, surface albedo, longitude, latitude and elevation), temperature vegetation index and multivariate linear regression interpolation method were used to estimate air temperature. The applicability of several temperature estimation methods in the study area was compared. The results show that the determination coefficient (R2) of the maximum temperature estimation of the three methods is 0.96 ~ 0.91 ~ 0.97, and the root mean square error (R- _ (MSE) is 1.84 ~ 2.75 ~ 1.49 鈩，

本文編號(hào)：2273122