湖泊表面温度的单通道热红外遥感反演研究——以北极滨海平原融冻湖为例  

Water Surface Temperature Inversion of the Thaw Lakes in Arctic Coastal Plain Using Single-channel Thermal Infrared Remote Sensing

作　　者：林征[1,2] 乔纪纲[3] 刘小平[1,2] 

机构地区：[1]中山大学地理科学与规划学院 [2]广东省城市化与地理环境空间模拟重点实验室,广州510275 [3]广东商学院资源与环境学院,广州501320

出　　处：《热带地理》2012年第3期252-259,共8页Tropical Geography

基　　金：国家重点基础研究发展计划(973计划)(11CB707103);广东省自然科学基金项目(S2011040003226)

摘　　要：基于TM影像的单通道热红外遥感数据反演了阿拉北极滨海平原融冻湖的表面温度,利用实地湖泊水体温度传感器所测量的水体温度进行了反演结果的精度验证。通过对比4种不同的单通道热红外遥感数据地表温度反演算法,发现其反演结果均与实地测量湖表温度高度相关,但精度存在差异,其中通用单通道算法(SC^(JM&S))的湖泊表面温度反演结果精度最高,获取均方根误差达0.45K,且该算法输入参数仅需大气水汽含量一项,适用于北极滨海平原地区的湖泊表面温度反演。湖泊表面平均温度的空间插值结果表明研究区的湖泊表面温度受纬度和海洋性气候影响显著,在空间分布上具有沿内陆向滨海地区递减的趋势,且一阶趋势面的等值线与东北方向的波弗特海岸线近乎平行。Water surface temperature of thaw lakes in the Arctic Coastal Plain of northern Alaska is inverted on the basis of single-channel thermal infrared remote sensing data of the TM images, and the accuracy of the inverted results is validated by the measured values using temperature sensors in the field. By comparing 4 different inversion algorithms, it is indicated that all those algorithms are suitable for the studied area, and the results inverted from those algorithms are also highly correlated with the real temperature. The result obtained by SC^JM＆S algorithm has a RSME（Root Mean Square Error） at 0.45K, which was applied to analyze the spatial pattern of lake surface water temperature for the entire study region. Spatial interpolation of mean temperature of lake surface shows that： water surface temperature of lakes have decreased along the inland to coastal area which are impacted significantly by latitude and marine climate, and isoline of first order trend are almost parallel to the coastline of northeast Beaufort Sea.

关 键 词：湖表温度 热红外遥感 融冻湖 SC^JM&S算法 阿拉斯加 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置]