地理研究 ›› 2017, Vol. 36 ›› Issue (11): 2251-2260.doi: 10.11821/dlyj201711017

所属专题: 气候变化与地表过程

• 研究论文 • 上一篇    下一篇

东北冻土区MODIS地表温度估算

刘世博(), 臧淑英(), 张丽娟, 那晓东   

  1. 黑龙江省普通高等学校地理环境遥感监测重点实验室,哈尔滨师范大学,哈尔滨 150025
  • 收稿日期:2017-05-08 修回日期:2017-09-16 出版日期:2017-11-20 发布日期:2017-11-20
  • 作者简介:

    作者简介:刘世博(1993- ),男,河北石家庄人,硕士,研究方向为冻土与积雪遥感研究。E-mail:327764711@qq.com

  • 基金资助:
    国家自然科学基金项目(41571199)

Estimation of land surface temperature from MODIS in Northeast China

Shibo LIU(), Shuying ZANG(), Lijuan ZHANG, Xiaodong NA   

  1. Key Laboratory of Remote Sensing Monitoring of Geographic Environment, College of Heilongjiang Province, Harbin Normal University, Harbin 150025, China
  • Received:2017-05-08 Revised:2017-09-16 Online:2017-11-20 Published:2017-11-20

摘要:

地表温度作为重要的地表参数是驱动土壤热状态的主要因子,对冻土分布和活动层厚度变化的研究具有重要意义。常规方式获取地表温度数据往往来自气象站点监测,范围小且不连续。NASA官网提供的MOD11A1地表温度产品可以提供大范围地表温度数据,但在冬季由于对云与雪的混淆导致大量的数据缺失,影响该产品在东北冻土区的使用。根据对东北冻土区植被、裸土、水体、积雪等常见下垫面状况的遥感分类结果,利用劈窗算法反演2006年四幅少云或无云的MODIS1B卫星影像,并分别以气象站实测数据和MODIS温度产品进行验证和对比分析。结果表明:该方法得到地表温度结果与气象站点实测数据误差较小,平均绝对误差仅为1.24 ℃。且可根据分类情况较好的得到积雪区域地表温度的空间分布状况,与地表温度产品的一致性较高,弥补地表温度产品因为云和积雪的混淆所导致的数据缺失,得到较为完整的地表温度空间分布数据。

关键词: 东北地区, 冻土, MODIS, 劈窗算法, 地表温度, 弥补数据

Abstract:

Land surface temperature (LST) is an important parameter driving dynamics of biogeophysical processes on Earth surface. It has significant impacts on the distribution of permafrost and the change of the active layer depth. Conventional acquisition of LST data usually comes from weather station monitoring in a small and discrete scope. NASA's MOD11 A1 surface temperature product can provide a wide range of surface temperature data. In winter, however, the confusion of clouds and snow often leads to a large amount of data missing in the MOD11 A1 products in the permafrost region. In this paper, an improved split-window algorithm was selected to re-build the LST products in Northeast China, one of the major permafrost regions in China. Within the common land covers extracted from remote sensing classification results, such as vegetation, bare soil, water and snow. We extracted LST in each cover type from four cloud-free MODIS 1B satellite images in 2006. Both our results and the original MOD11 A1 products were statistically compared with ground measurements at weather stations. The average difference between our results and measurements at meteorological stations was small, reaching a room-mean-square error (RMSE) of 1.24 ℃. In comparison with the original MOD11 A1 products, our results took advantage of land covers and revealed better distributions of land surface temperature in snow area, and had a high consistency with the surface temperature products. This study provides a good approach to filling in the gaps of current land surface temperature products due to confusion caused by the cloud and snow.

Key words: Northeast China, permafrost, MODIS, split window algorithm, land surface temperature, gap filling of data products