Accurate digital soil organic carbon mapping is of great significance for regulating the global carbon cycle and addressing climate change. With the advent of the remote sensing big data era, multi-source and multi-temporal remote sensing techniques have been extensively applied in Earth observation. However, how to fully mine multi-source remote sensing time-series data for high-accuracy digital SOC mapping remains a key challenge. To address this challenge, this study introduced a new idea for mining multi-source remote sensing time-series data. We used 413 topsoil organic carbon samples from southern Xinjiang, China, as an example. By mining multi-source (Sentinel-1/2) remote sensing time-series data from 2017 to 2023, we revealed the temporal variation pattern of the correlation between Sentinel-1/2 time-series data and SOC, thereby identifying the optimal time window for monitoring SOC using Sentinel-1/2 data. By integrating environmental covariates and a super ensemble model, we achieved high-accuracy mapping of SOC in Southern Xinjiang, China. The results showed the following aspects: (1) The optimal time windows for monitoring SOC using Sentinel-1/2 data were July–September and July–August, respectively; (2) the modeling accuracy using multi-source sensor data integrated with environmental covariates was superior to using single-source sensor data integrated with environmental covariates alone. In the optimal model based on multi-source data, the cumulative contribution rate of Sentinel-2 data is 51.71% higher than that of Sentinel-1 data; (3) the stacking super ensemble model’s predictive performance outperformed the weight average and simple average ensemble models. Therefore, mining the optimal time windows of multi-source remote sensing data and environmental covariates, driven a super ensemble model, represents a high-accuracy strategy for digital SOC mapping.
Soil salinity severely threatens global ecosystems and agriculture, making accurate monitoring an ongoing priority. Currently, efficiently utilizing multi-source datasets to enhance monitoring accuracy while minimizing computational resources remains a critical challenge. This study evaluated several modeling strategies, including full-dataset modeling, variance inflation factor (VIF), Boruta, particle swarm optimization, ant colony optimization and recursive feature elimination (RFE), and validated results across diverse regions (Almaty, Kazakhstan; Shandong, China). We further validated the results using multiple algorithms, including linear regression, partial least squares regression, extreme gradient boosting, k-nearest neighbor and random forest (RF), with topsoil (0–20 cm) electrical conductivity inverted via the optimal method. Results indicate that input feature numbers substantially impact model performance: regional-scale feature selection is indispensable, with RFE outperforming full-dataset modeling (R2 improves by up to 0.28, while RMSE decreases by 2.21 dS m−1) and VIF performing the worst. Transferability is also demonstrated in Almaty and Shandong. Additionally, the RF algorithm shows superior performance in soil salinity mapping (overall accuracy = 0.73; kappa coefficient = 0.65). And, the RFE and SHAP results highlight CRSI, BI, and MSAVI2 as particularly important predictors for estimating soil salinity in our study area. Collectively, this study highlights the critical importance of feature optimization and interpretability in soil attribute mapping through the integration of multi-source remote sensing data.
Laboratory visible–near-infrared (Vis-NIR) spectroscopy is an efficient and practical approach for measuring soil properties. However, its predictive accuracy is often limited by spectral variability resulting from soil heterogeneity. Such variability poses challenges for traditional global spectral modeling methods in capturing the highly complex mapping relationships between spectral features and soil attributes. To address this issue, this study introduces a novel spectral similarity-based partitioned modeling strategy and compares it with a soil type-based grouping approach. The proposed method clusters soil samples according to their spectral characteristics to minimize within-group variance. Subsequently, individual inversion models are developed for each cluster, enabling more precise learning of stable relationships between specific spectral patterns and soil properties, thereby improving both model accuracy and interpretability. To validate this methodology, the study systematically evaluates the performance of five traditional clustering algorithms (TCAs) and five deep clustering algorithms (DCAs) using the Chinese Soil Spectral Library (CSSL) and the national soil spectral library of China, supplemented with local laboratory Vis-NIR data from Xinjiang, targeting soil organic matter (SOM). Results indicate that both spectral similarity-based and soil type-based grouping strategies significantly surpass conventional global modeling methods. Among them, the spectral similarity-based partitioning approach yielded the most notable improvements. Compared with traditional global modeling, it increased R2 from 0.65–0.71 to 0.81–0.85 (relative increase: 14.08–30.77%) and reduced RMSE from 2.49–3.24 g kg−1 to 1.07–1.39 g kg−1 (relative reduction: 44.18–66.98%). Compared with the soil type-based grouping strategy, it further increased R2 from 0.70–0.74 to 0.81–0.85 (relative increase: 9.46–21.43%) and reduced RMSE from 1.91–2.56 g kg−1 to 1.07–1.39 g kg−1 (relative reduction: 27.23–58.20%). These outcomes suggest that conventional soil taxonomic information alone is insufficient to capture critical spectral variations. In the spectral grouping stage, DCAs were found to identify and aggregate soil samples with high spectral similarity more effectively than TCA. In particular, the inversion model constructed using Automatic Encoding Clustering (AEC) delivered the best performance. Additionally, the study revealed considerable spectral diversity, even among soils of the same type. In the evaluated dataset, spatial proximity was not consistently associated with spectral similarity. These findings confirm that spectral similarity-based grouping and modeling substantially improve the predictive accuracy of soil property estimation, offering a more refined and effective modeling pathway for hyperspectral soil detection. All spectral data analyzed here are laboratory-based Vis-NIR spectra from the national soil spectral library of China. These findings are based on the Chinese Soil Spectral Library and the Xinjiang dataset and should be validated in other regions and for other soil properties before broad application.