Given the critical issues that are arising as a direct result of climate change, the need to monitor the status of our environment is self-evident. One key parameter in monitoring climate change is the land surface temperature (LST). It is a measurement of the temperature of the land itself, as opposed to the more common temperature measurements that we receive as weather reports, which are readings on the temperature of the air above the ground.
The difficulty with LST measurements is that over 60% of the earth's terrestrial area is mountainous and the complex topography of these regions makes getting an accurate LST very difficult. This is because the current models that calculate LST are based on remote sensing data gathered from flat surfaces. In montane areas, angled surfaces, shaded surfaces and differing vegetation can drastically alter the LST in ways that are not accounted for in these models. Critically, the data gathered by remote sensors is dependent on the angle at which the measurements are taken. These are called directional anisotropies.
Chinese researchers have been working on building a model that incorporates the effect of angles in the measurement of LST to more accurately estimate LST, especially in topographically complex areas. "The challenge arises from the conflict between the topographic effect and the assumption of flatten surface in many existing studies. In the absence of a simple and practical model for the topographic effect on the directional anisotropies of LST over mountainous areas, our equivalent slope method is introduced to bridge the gap between studies conducted on flat surfaces and complex terrain," said paper author Tengyuan Fan, a researcher from the State Key Laboratory of Remote Sensing Science, Aerospace Information Research Institute, Chinese Academy of Sciences and the University of Chinese Academy of Sciences.
Their research was published in the Journal of Remote Sensing on Sept. 4.
The importance of directional anisotropy, or the angle effect of surface temperature is startling. Measurements used to calculate LST are often gathered using thermal infrared measurements via remote sensing. Two different satellites with different angles of measurement can show relative biases differing by as much as 8 Kelvin.
To address this, the researchers built a new model that took the angle effect into account. They believed that this would result in a more accurate estimation of the LST from the thermal infrared measurements.
Tests were carried out in Chongqing, in the southwestern region of China. An area with a complex topography as much of the area is either mountainous or at least hilly. They validated their model (TESKD) using data gathered by remote sensing using unmanned aerial vehicles and a computer simulation using a 3-D model. They then compared their models results with the results from a more commonly used flat model (LSF-Li).
They found that "TESKD consistently outperforms the original LSF-Li method in terms of accuracy in all scenarios," said Zunjian Bian, Associate Professor at the Aerospace Information Research Institute, Chinese Academy of Sciences.
"In this study, the proposed TESKD model effectively addresses the impact of internal terrain variations within pixel, which have a resolution of 50 m for UAV data and 1 km for simulated data on directional infrared observations. Both UAV observational data and simulated data demonstrate its superior performance compared to the existing kernel-driven models developed based on flat terrain," Bian said.
Looking to the future, "the next step is to apply the model to the angle effect correction of thermal infrared surface temperature remote sensing products in mountainous areas. The ultimate goal is to develop a set of angle normalization products for further quantitative analysis of changes in surface temperature in mountainous areas and their impact on other land surface processes," Fan said.
Other contributors include Jianguang Wen, Biao Cao, Hua Li, Yongming Du from the State Key Laboratory of Remote Sensing Science, Aerospace Information Research Institute, Chinese Academy of Sciences; and Qing Xiao and Qinhuo Liu from the State Key Laboratory of Remote Sensing Science, Aerospace Information Research Institute, Chinese Academy of Sciences and the University of Chinese Academy of Sciences; and Zhonghu Jiao at the State Key Laboratory of Earthquake Dynamics, Institute of Geology, Earthquake Administration; and Shouyi Zhong at the Faculty of Geographical Science, Beijing Normal University; and Wenzhe Zhu from the University of Chinese Academy of Sciences and the Key Laboratory of Water Cycle and Related Land Surface Processes, Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences.
This work was supported in part by the Chinese Natural Science Foundation Project.
