ISPRS Journal of Photogrammetry and Remote Sensing 169 (2020) 44–56 Available online 11 September 2020 0924-2716/© 2020 International Society for Photogrammetry and Remote Sensing, Inc. (ISPRS). Published by Elsevier B.V. All rights reserved. Global comparison of diverse scaling factors and regression models for downscaling Landsat-8 thermal data Pan Dong a , Lun Gao b , Wenfeng Zhan a, c, * , Zihan Liu a , Jiufeng Li a , Jiameng Lai a , Hua Li d , Fan Huang a , Sagar K. Tamang b , Limin Zhao d a Jiangsu Provincial Key Laboratory of Geographic Information Science and Technology, International Institute for Earth System Science, Nanjing University, Nanjing, Jiangsu 210023, China b Saint Anthony Falls Laboratory, Department of Civil Environmental and Geo-Engineering, University of Minnesota, Minneapolis, MN 55414, USA c Jiangsu Center for Collaborative Innovation in Geographical Information Resource Development and Application, Nanjing 210023, China d State Key Laboratory of Remote Sensing Science, Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100101, China A R T I C L E INFO Keywords: Thermal remote sensing Land surface temperature Downscaling Spatial resolution Landsat-8 ABSTRACT Statistical downscaling of land surface temperature (SDLST) algorithms with diverse scaling factors and regression models have been used to produce high spatial resolution LSTs based on Landsat-8 LST. However, the optimal choice of scaling factors and regression models and their associated combinations over various land surfaces, especially from a global perspective, remain unclear and even controversial. To cope with this issue, we compare 35 SDLST algorithms derived from a combination of seven scaling factors and fve frequently used regression models over 32 geographical regions worldwide. The seven scaling factors, at varying degrees, make use of the LST-related information embedded within the visible and near-infrared and short-wave infrared bands of Landsat-8 data. The fve regression models involved are multiple linear regression, partial least squares regression, artifcial neural networks, support vector regression, and random forest (RF). Our main fndings are: (1) The performance of the scaling factors and regression models are highly dependent on each other. Never- theless, for most scaling factors, especially for high-dimension scaling factors with numerous LST-related vari- ables, the downscaling algorithms that use RF as the regression model achieve the highest accuracy. (2) RFT21, a newly proposed SDLST algorithm based on the comparison results and further optimization, has high global operability and suffciently high accuracy. RFT21 requires only Landsat-8 data as the inputs, and achieves the highest accuracy in comparison with the thermal sharpening (TsHARP) and high-resolution urban thermal sharpener (HUTS) algorithms, with the mean root-mean-square error (RMSE) reduced by more than 15%. These fndings will facilitate the generation of high spatial resolution LSTs worldwide and associated applications. 1. Introduction Land surface temperature (LST) has been widely used in surface energy balance modeling (Bastiaanssen et al., 1998; Gan et al., 2019), landscape ecological processes analysis (Quattrochi and Luvall, 1999; Kalma et al., 2008; Olivera-Guerra et al., 2017), and urban thermal environment monitoring (Voogt and Oke, 2003; Weng, 2009; Zhou et al., 2019). Satellite thermal remote sensing provides excellent op- portunities for mapping LSTs simultaneously over large surfaces (Li et al., 2013). However, there exists a tradeoff between the spatial and temporal resolution of satellite-derived LSTs. This tradeoff gives rise to the downscaling of LST (DLST) (Agam et al., 2007; Zhan et al., 2013), which has become an essential strategy for obtaining LSTs with high spatial or spatiotemporal resolution (Weng et al., 2014; Fu and Weng, 2016). Various types of DLST algorithms have been proposed for improving the spatial or spatiotemporal resolution of satellite-derived LSTs, such as statistical DLST (Agam et al., 2007; Jeganathan et al., 2011; Gao et al., 2012; Hutengs and Vohland, 2016; Peng et al., 2019; Wang et al., 2020b), data fusion-based DLST (Huang et al., 2013; Wu et al., 2013, * Corresponding author at: Nanjing University at Xianlin Campus, No.163 Xianlin Avenue, Qixia District, Nanjing, Jiangsu 210023, China. E-mail addresses: dongpan0920@foxmail.com (P. Dong), gaoxx996@umn.edu (L. Gao), zhanwenfeng@nju.edu.cn (W. Zhan), liuzihan_trs@foxmail.com (Z. Liu), Jiufengli@smail.nju.edu.cn (J. Li), NJULJM@126.com (J. Lai), lihua@radi.ac.cn (H. Li), nju_huangfan@163.com (F. Huang), taman011@umn.edu (S.K. Tamang), zhaolm@radi.ac.cn (L. Zhao). Contents lists available at ScienceDirect ISPRS Journal of Photogrammetry and Remote Sensing journal homepage: www.elsevier.com/locate/isprsjprs https://doi.org/10.1016/j.isprsjprs.2020.08.018 Received 8 March 2020; Received in revised form 28 July 2020; Accepted 24 August 2020