数字农科院2.0

A practical two-step framework for all-sky land surface temperature estimation

文献类型： 外文期刊

作者： Zhang H.;Tang B.-H.;Li Z.-L.

作者机构：

关键词： All-sky land surface temperature (LST);Cloud radiative forcing effect;Cloudy-sky conditions;Hypothetical clear-sky LST;Surface energy balance

期刊名称： Remote Sensing of Environment

ISSN： 0034-4257

年卷期： 2024 年 303 卷

页码：

收录情况： SCIE（2024版） ; ; EI（2024版）

摘要： Land surface temperature (LST) is a key parameter in global ecological and climate system, while its accurate estimation under cloudy-sky conditions remains a challenge. In this paper, a practical two-step framework has been proposed for all-sky LST estimation, which effectively combined physical mechanisms and machine learning algorithms. The first step was the estimation of hypothetical clear-sky LST. The Bayesian optimization-based Extreme Gradient Boosting (BO-XGB) model was developed to establish the implicit and complex relationship between hypothetical clear-sky LST and corresponding independent variables, including hypothetical clear-sky radiation energy input, surface biophysical parameters, atmospheric conditions, as well as spatial and temporal features. The second step was the cloud radiative forcing effect (CRFE) correction for hypothetical clear-sky LST under cloudy-sky conditions. Based on the surface energy balance (SEB) equation and conventional force-restore method, the analytical expressions of CRFE correction terms were derived, which speeded up the calculation and facilitated error analysis. Based on several remote sensing products, including Advanced Baseline Imager (ABI) LST and auxiliary data from Moderate Resolution Imaging Spectroradiometer (MODIS) and Global LAnd Surface Satellite (GLASS), as well as ERA5 reanalysis data, the proposed framework was applied to recovered spatially continuous hourly all-sky LST information over the contiguous US (CONUS) in 2018. The accuracy of the proposed method has been validated by ground measurements from 60 representative sites operated by the Surface Radiation Budget (SURFRAD) and AmeriFlux. For clear-sky conditions, the overall root mean square error (RMSE) was 2.67 K, with a bias of −0.47 K and R2 of 0.96. For cloudy-sky conditions, the corresponding accuracy metrics were 2.60 K, −0.49 K, and 0.96, respectively, and RMSEs for all sites were below 4 K, which indicated the stability of the proposed method. Besides, daily average LST has also been calculated based on the estimated hourly LSTs with RMSE of 1.68 K and bias of −0.48 K. © 2024 Elsevier Inc.

分类号：

  • 相关文献

[1]Spatial-scale effect on the SEBAL model for evapotranspiration estimation using remote sensing data. Tang, Ronglin,Tang, Ronglin,Chen, Kun-Shan,Li, Zhao-Liang,Li, Zhao-Liang,Jia, Yuanyuan,Li, Chuanrong,Sun, Xiaomin.

[2]A novel surface energy balance-based approach to land surface temperature downscaling. Mohammad Karimi Firozjaei,Naeim Mijani,Majid Kiavarz,Si Bo Duan,Peter M. Atkinson,Seyed Kazem Alavipanah. 2024

[3]A framework of crop water productivity estimation from UAV observations: A case study of summer maize. Minghan Cheng,Ni Song,Josep Penuelas,Matthew F. McCabe,Xiyun Jiao,Yuping Lv,Chengming Sun,Xiuliang Jin. 2025

作者其他论文 更多>>