地表温度和近地表大气温度是地球系统、大气系统以及地—气相互作用物理过程的重要参量。在陆地—大气的相互作用过程中，水汽含量、NDVI指数、下垫面变化等因素会对地—气热量传输造成一定的影响。本文首先利用地表温度产品（MYD11A1）以及气温站点数据（GSOD）获得全国尺度下地表温度年最大值、近地面气温年最大值。在此基础上，使用趋势分析法分析2003年—2018年地、气温度年最大值时空分布特征及变化趋势，以及地—气温差气候倾向率变化趋势。最后，结合大气总水汽含量产品（MYD05）、NDVI指数（MYD13A3）、二氧化碳平均浓度增长率分析导致地表温度年最大值与近地面气温年最大值趋势发生变化的原因。研究结果表明：（1）在全国尺度下，2003年—2018年地表温度年最大值呈现北高南低的空间分布特征。近地面气温年最大值的空间分布与地表温度年最大值相反。大气总水汽含量年最大值在热带、亚热带季风气候区内总体较高。水汽含量既影响近地面气温的大小，同时也受到近地面气温的影响，因此，水汽含量年最大值与近地面气温年最大值表现出一定的空间分布一致性特征。（2）在2003年—2018年期间，地表温度年最大值的气候倾向率在空间上表现出北高南低的分布特征。近地面气温年最大值的气候倾向率在空间上也表现为北高南低，与地表温度年最大值的气候倾向率变化基本一致。但地表温度年最大值的变化幅度要大于近地面气温年最大值，并且在个别区域表现不一致。主要分布在天山地区、三江平原以及秦岭南侧地区，地—气年最大值变化趋势相反即地—气差减小。（3）大气总水汽含量年最大值的增加可造成近地面气温年最大值的增加，而植被覆盖度的上升可造成地表温度年最大值下降。但在天山地区大气总水汽含量与地—气差的响应不明显，但天山地区的近地面气温年最大值与CO₂平均浓度增长率的关系较为明显。（4）遥感数据反演的地表温度年最大值和站点观测的近地面气温年最大值空间分布表现出差异，但时间变化趋势基本一致。

Land Surface Temperature (LST) and Air Temperature (AT) are important parameters of land-atmosphere interaction system. In the process of land-atmosphere interaction, Water Vapor Content (WVC), NDVI index, carbon dioxide concentration and other factors will have great impact on land-atmosphere heat transfer. Therefore, we investigate the spatiotemporal variation trend of annual maximum of LST(AMLST) and annual maximum of AT(AMAT) from different data source including remote sensing and meteorological station in China in recent 20 years.The trend analysis method is used to analyze the spatiotemporal variation trend of AMLST and AMAT from 2003 to 2018. Combined with the annual maximum of WVC(AMWVC) and NDVI index, the driving factors leading to the variation of AMLST and AMAT were analyzed.The results showed that: (1) The spatial distribution characteristics of AMLST in China from 2003 to 2018 is higher in the north and lower in the south, and the highest AMLST appears in Turpan Basin, Xinjiang. The spatial distribution of AMAT in China is higher in the south and lower in the north. In virtue of interaction between AMWVC and AMAT, the two impact factors appear a certain coincident characteristics. (2) During the period of 2003-2018, the spatial distribution of climate inclination rate of AMLST is higher in the north and lower in the south, the AMAT climate inclination rate have the same characteristics with AMLST. From the perspective of variation range, the AMLST is larger than that of AMAT. But it is inconsonant in Tianshan Mountains, Sanjiang Plain and the south side of Qinling Mountains. (3) The increase in AMWVC makes the AMAT rise, while the increase of vegetation coverage can decrease the AMLST. In the Tianshan area, the response of AMLST-AMAT interaction system to the AMWVC is not obvious, but the relationship between AMAT and the growth rate of carbon dioxide concentration is obvious. (4) The spatial distribution of the AMLST retrieved from remote sensing data is different from that observed by stations, but the temporal variation trend is basically the same.