The land surface temperature (LST) downscaling is a valuable technique for obtaining high-spatiotemporal resolution LST data. By incorporating object-level or pixel-level residuals, the spectral information of the downscaled LST can be better restored, leading to improve the prediction accuracy. However, pixel-level residuals may introduce uncertainty in the residual results, whereas object-level residuals may overlook detailed land cover change information within objects, potentially resulting in discontinuities and abrupt changes in the predicted results. To address these challenges, this study proposed an integrated object- and pixel-based residual compensation framework (IOPRCF) to effectively restore the land cover change information within fine-resolution LST, thereby generating more accurate high-temporal and spatial resolution LST data. The IOPRCF was rigorously tested in 12 geographical diverse regions in China using three methods: thermal sharpening algorithm (TsHARP), random forest (RF), and simple and effective downscaling (SED). These tests encompass multiple time intervals and scaling scales to comprehensively validate the effectiveness and applicability of IOPRCF in the field of LST downscaling. The results demonstrate that, when compared with the original benchmark algorithms, TsHARP, RF, and SED with IOPRCF achieved average increases in$R^{2}$of 0.035, 0.035, and 0.028, respectively, average decreases in root mean square error (RMSE) of 0.110, 0.120, and 0.133 K, respectively, and average increases in SSIM of 0.017, 0.014, and 0.016, respectively. Moreover, the downscaled LST data produced by these algorithms with IOPRCF exhibited more similar spatial structures and captured more detailed information. Furthermore, the proposed IOPRCF demonstrated strong robustness and transferability across different LST downscaling algorithms and segmentation models. Finally, IOPRCF is expected to enhance the capabilities of LST downscaling algorithms for recovering land cover information, effectively supporting the generation of high-quality, high-spatiotemporal resolution LST globally.

Accurate spatial quantification of cropland soil organic carbon (SOC) in plateau lake basins is crucial for assessing the carbon sequestration potential in ecologically fragile regions. This study developed a machine learning (ML) framework that integrates multi-source satellite-derived environmental covariates (topography, climate, vegetation, soil properties, and parent materials) to estimate SOC distribution in the Erhai Lake basin. Using 432 topsoil samples (0–20 cm), we systematically compared 15 models, including conventional ML approaches (e.g., random forest, support vector machine, and light gradient boosting machine) and deep learning (DL) models (e.g., long short-term memory, recurrent neural network, and multilayer perceptron). The results showed that DL models achieved higher predictive accuracy than conventional ML models, reducing RMSE by 0.1680 g kg⁻¹ and increasing R², RPIQ, and CCC by averages of 0.0225, 0.1143, and 0.0253, respectively, although conventional ML models exhibited greater robustness. Elevation and temperature were identified as dominant factors controlling SOC spatial patterns, with higher concentrations clustered in the western and northern subbasins. Greater prediction uncertainty in the northwestern and eastern margins was associated with complex terrain heterogeneity. Spatially explicit SOC mapping derived from the integration of multi-source satellite data and ML models offers innovative approaches for carbon management in ecologically fragile lacustrine agroecosystems.

Land surface temperature (LST) is a key factor in monitoring and improving thermal environments. However, conventional LST retrieval algorithms have not sufficiently accounted for the cavity and adjacency effects caused by the three-dimensional (3D) structures in urban settings. In this study, we propose an urban multiple scattering radiative transfer model (UMS-RTM). Based on this model, we develop an urban nonlinear split-window (UNSW) algorithm to retrieve urban land surface temperature (ULST) from Landsat-9 satellite data. The UMS-RTM optimizes the thermal radiation transfer process by correcting the cavity and adjacency effects. Analysis shows that land surface emissivity (LSE) and sky view factor (SVF) are the primary factors influencing these effects. The cavity effect increases the effective LSE by 0.01 to 0.08, while the adjacency effect raises the ground-leaving brightness temperature (BT) by 0.82 K to 3.51 K. The UNSW algorithm’s coefficients were calibrated across various LST, water vapor content (WVC), and SVF groupings to eliminate atmospheric effects and correct for cavity and adjacency effects. Sensitivity analyses of instrument noise, WVC, effective LSE, and SVF uncertainties demonstrated the reliability of the UNSW algorithm. Validation using simulated data showed that the ULST retrieved by the UNSW algorithm had a root-mean-square error (RMSE) of 0.35 K. When applied to Landsat-9 satellite data, the UNSW algorithm revealed that conventional algorithms and LST products overestimate ULST by 0 K to 2 K, with overestimations exceeding 1 K in areas with low SVF. The UNSW algorithm provides more accurate ULST retrieval and finer spatial distribution details.

This study presents a framework for predicting the total carbon storage of wetlands based on machine learning models that integrate Sentinel-1 (S1), Sentinel-2 (S2), the Digital Elevation Model (DEM), and climatic data. The results indicated that the Random Forest (RF) model outperformed the Support Vector Machine (SVM) and Extreme Gradient Boosting (XGB) models in prediction accuracy. With an R-squared value of 0.67 for aboveground biomass carbon density, 0.75 for belowground biomass carbon density, and 0.65 for soil organic carbon density. The predictive performance utilizing multi-source data is significantly superior to that of single indicators, and the inclusion of climate data enhances the model’s predictive capabilities. The total carbon storage of wetlands in the Yunnan-Kweichow Plateau is estimated at 55.5 MtC, comprising 13.8 MtC in aboveground biomass carbon, 7.6 MtC in belowground biomass carbon, and 34.1 MtC in soil organic carbon.

Mountainous land surface temperature (MLST) is a key parameter for the research of mountainous climate change. In this study, an improved method was developed to retrieve the MLST by considering the multiple scattering between pixels over rugged mountainous surfaces based on sky view factors (SVF). The method was applied to the Landsat-9 thermal infrared (TIR) remote sensing data by using the single channel (SC) algorithm. Due to the lack of measured data, the discrete anisotropic radiative transfer (DART) model is used to validate the accuracy of MLST retrieved by the proposed method. The results show that the MLST retrieved has high precision. The findings demonstrated the need for multiple scattering between pixels was taken into account in the retrieval of high-precision MLST under certain conditions.