**Youn-Young Choi and Myoung-Seok Suh \***

Department of Atmospheric Science, Kongju National University, 56, Gongjudaehak-ro, Gongju-si, Chungcheongnam-do 32588, Korea; choiyoyo@smail.kongju.ac.kr

**\*** Correspondence: sms416@kongju.ac.kr; Tel.: +82-41-850-8527

Received: 13 August 2020; Accepted: 16 September 2020; Published: 18 September 2020

**Abstract:** Land surface temperature (LST) is an important geophysical element for understanding Earth systems and land–atmosphere interactions. In this study, we developed a nonlinear split-window LST retrieval algorithm for the observation area of GEO-KOMPSAT-2A (GK2A), the next-generation geostationary satellite in Korea. To develop the GK2A LST retrieval algorithm, radiative transfer model simulation data, considering various impacting factors, were constructed. The LST retrieval algorithm was developed with a total of six equations as per day/night and atmospheric conditions (dry/normal/wet), considering the effects of diurnal variation of LST and atmospheric conditions on LST retrieval. The emissivity of each channel required for LST retrieval was calculated in real-time with the vegetation cover method using the consecutive 8-day cycle vegetation index provided by GK2A. The indirect validation of the results of GK2A LST with Moderate Resolution Imaging Spectroradiometer (MODIS) LST Collection 6 showed a high correlation coefficient (0.969), slightly warm bias (+1.227 K), and root mean square error (RMSE) (2.281 K). Compared to the MODIS LST, the GK2A LST showed a warm bias greater than +1.8 K during the day, but a relatively small bias (<+0.7 K) at night. Based on the results of the validation with in situ measurements from the Tateno station of the Baseline Surface Radiation Network, the correlation coefficient was 0.95, bias was +0.523 K, and RMSE was 2.021 K.

**Keywords:** land surface temperature; GK2A; split-window method; MODIS; BSRN
