This study proposes an improved temperature and emissivity separation (TES) algorithm for simultaneously retrieving the land surface temperature and emissivity (LST&E) from the Advanced Himawari Imager (AHI) data, including a modified water vapor scaling (WVS) method and a calibrated empirical relationship over vegetated surfaces. The modified WVS algorithm is comparable to the original WVS algorithm in deriving the LST&E but expands the application scope of the original WVS algorithm. The calibrated empirical relationship improved the LST&E and retrieval accuracy over vegetated surfaces by up to 0.165 K and 0.004, respectively. Comprehensive validation and evaluation are conducted in this study. In situ measurements from three networks are collected for the temperature-based validation. The bias and RMSE are 0.19 and 2.93 K in the daytime, and -0.43 and 1.95 K in the nighttime, respectively. Radiance-based LST validation shows that the bias and RMSE are 0.25 and 1.88 K, respectively. In addition, the AHI LST is evaluated using the MYD11 LST over large inland lakes, and the bias and RMSE are 0.25 and 1.12 K, respectively. The AHI LST is also compared to the MYD21 LST. The spatial distributions of the two LSTs are similar, and the LST differences are mostly within 4 K. The bias of the AHI LST ranges from -0.57 to 0.36 K, and the RMSE ranges from 1.7 to 2.64 K. The retrieved AHI LSE is compared with the latest MYD21 LSE. The biases and RMSEs are smaller than 0.005 and 0.014, respectively, for the three AHI bands. The improved TES algorithm is proven to be capable of obtaining accurate LST and LSE from AHI data.

Land surface temperature (LST) is crucial for understanding surface energy budgets, hydrological cycling, and land–atmosphere interactions. However, cloud cover leads to numerous data gaps in existing remote sensing thermal infrared (TIR) LST products, seriously restricting their applications. This article provides a comprehensive review concerning both LST recovery methodologies and 26 emerging all-weather products derived from polar-orbiting and geostationary (GEO) satellites. Clarifying product distinctions will enable end users to select suitable options for diverse research. Methodologies are categorized into spatiotemporal interpolation, surface energy balance (SEB)-based physical estimation, passive microwave (PMW)-based methods, and simulated temperature-based approaches. Historical research trajectories, strengths, limitations, and potential research directions of the methodologies and products are discussed. The review reports that existing all-weather LST products generally exhibit root-mean-square errors (RMSEs) of <4 (2.5) K at instantaneous (daily mean) scales based on extensive ground measurements, comparable to clear sky retrievals. Deep learning (DL) models prominently feature in state-of-the-art interpolation and fusion approaches [e.g., long short-term memory (LSTM) and extreme gradient boosting (XGBoost)]. Product intercomparisons in various application scenarios reveal that interpolation-based products offer better texture details; however, noticeable biases exist compared to fusion-based products, especially in arid and semiarid regions, despite the high availability of clear sky samples. The bias shifts to negative at higher latitudes, due to ignored cloud radiative effects. The review emphasizes the underexplored recovery of diurnal temperature cycles (DTCs) from GEO satellites. This focus will benefit heat exposure monitoring for public health, understanding circadian rhythm responses of ecosystems to environmental changes, and harmonizing existing and forthcoming high-resolution TIR missions.

This paper extends a new temperature and emissivity separation (TES) algorithm for retrieving land surface temperature and emissivity (LST and LSE) to the Advanced Geosynchronous Radiation Imager (AGRI) onboard Fengyun-4A, China’s newest geostationary meteorological satellite. The extended TES algorithm was named the AGRI TES algorithm. The AGRI TES algorithm employs a modified water vapor scaling (WVS) method and a recalibrated empirical function over vegetated surfaces. In situ validation and cross-validation are utilized to investigate the accuracy of the retrieved LST and LSE. LST validation using the collected field measurements showed that the mean bias and RMSE of AGRI TES LST are 0.58 and 2.93 K in the daytime and −0.30 K and 2.18 K at nighttime, respectively; the AGRI official LST is systematically underestimated. Compared with the MODIS LST and LSE products (MYD21), the average bias and RMSE of AGRI TES LST are −0.26 K and 1.65 K, respectively. The AGRI TES LSE outperforms the AGRI official LSE in terms of accuracy and spatial integrity. This study demonstrates the good performance of the AGRI TES algorithm for the retrieval of high-quality LST and LSE, and the potential of the AGRI TES algorithm in producing operational LST and LSE products.

The high-frequency all-weather land surface temperature (LST) product generated from the thermal-infrared (TIR) observations of the geostationary meteorological satellite is of great significance to study the diurnal variations in the LST and the land surface energy balance. However, the TIR sensor cannot penetrate the clouds and obtain the desired LST under cloudy conditions. In this study, we developed a physical-based framework for generating high-frequency (hourly) all-weather LST data by synchronizing geostationary satellite TIR observations and simulations of the land surface model (LSM). There are three parts in the developed framework. First, the clear-sky LST was retrieved from the Advanced Himawari Imager (AHI) onboard the geostationary satellite Himawari-8 using our newly developed temperature and emissivity separation algorithm. Second, the Advanced Microwave Scanning Radiometer 2 (AMSR2) observations were assimilated into the Noah land surface model with multiple parameterization options (Noah-MP) model to generate the all-weather LST. Finally, the retrieved clear-sky AHI LST and Noah-MP assimilated LST were fused using the Ensemble Kalman filter (EnKF) algorithm. In situ measurements from three networks were collected to evaluate the Noah-MP assimilated LST and EnKF fused LST. The bias/RMSE of the Noah-MP assimilated LST and EnKF fused LST were –0.16/3.01 K and 0.15/2.68 K, respectively, under all-weather conditions. Compared to the Noah-MP free-run LST, the absolute values of the bias were reduced by 0.64 K and 0.68 K for the Noah-MP assimilated LST and EnKF fused LST, while the RMSEs were reduced by 0.33 K and 0.65 K, respectively. In addition, the spatial distribution of EnKF fused LST was in good agreement with the retrieved clear-sky AHI LST. The proposed framework in this study was demonstrated to be capable of obtaining accurate high-frequency (hourly) all-weather LST data.