A comparative study of land surface temperature retrieval methods from remote sensing data

The main purpose of this paper is to describe, compare, and analyze the various extraction methods for land surface temperature (LST) in terms of their computational algorithms, their different input parameters, and their relative accuracy to make them more readily usable by a broader cross-section of nontechnical practitioners. Due to the heterogeneity of most natural land surfaces, the atmospheric influence, and a wide variety of satellite sensors, the estimation and validation of LST can be difficult. Furthermore, the large number of algorithms developed to deal with this heterogeneity has led to widespread confusion on how and when to use one algorithm versus another. This paper provides a concise, but thorough, overview of the different algorithms used for the estimation of land surface temperature as well as a comparative list of methods and associated parameters that facilitate, to the general user, the selection and application of the most appropriate method for LST extraction given the situation at hand. We restricted our analysis for the single-channel algorithms to two models. We included a two-channel algorithm (or split-window when it is applied in the region 10–12.5 µm) according to the literature. The Temperature Emissivity Separation algorithm was also taken into account. The determination of the key parameters needed to execute these algorithms is presented.