Soil organic carbon (SOC) contents and stocks provide valuable insights into soil health, nutrient cycling, greenhouse gas emissions, and overall ecosystem productivity. Given this, remote sensing data coupled with advanced machine learning (ML) techniques have eased SOC level estimation while revealing its patterns across different ecosystems. However, despite these advances, the intricacies of training reliable and yet certain SOC models for specific end-users remain a great challenge. To address this, we need robust SOC uncertainty quantification techniques. Here, we introduce a methodology that leverages conformal prediction to address the uncertainty in estimating SOC contents while using remote sensing data. Conformal prediction generates statistically reliable uncertainty intervals for predictions made by ML models. Our analysis, performed on the LUCAS dataset in Europe and incorporating a suite of relevant environmental covariates, underscores the efficacy of integrating conformal prediction with another ML model, specifically random forest. In addition, we conducted a comparative assessment of our results against prevalent uncertainty quantification methods for SOC prediction, employing different evaluation metrics to assess both model uncertainty and accuracy. Our methodology showcases the utility of the generated prediction sets as informative indicators of uncertainty. These sets accurately identify samples that pose prediction challenges, providing valuable insights for end-users seeking reliable predictions in the complexities of SOC estimation.

Soil Organic Carbon (SOC) constitutes a fundamental component of terrestrial ecosystem functionality, playing a pivotal role in nutrient cycling, hydrological balance, and erosion mitigation. Precise mapping of SOC distribution is imperative for the quantification of ecosystem services, notably carbon sequestration and soil fertility enhancement. Digital soil mapping (DSM) leverages statistical models and advanced technologies, including machine learning (ML), to accurately map soil properties, such as SOC, utilizing diverse data sources like satellite imagery, topography, remote sensing indices, and climate series. Within the domain of ML, self-supervised learning (SSL), which exploits unlabeled data, has gained prominence in recent years. This study introduces a novel approach that aims to learn the geographical link between multimodal features via self-supervised contrastive learning, employing pretrained Vision Transformers (ViT) for image inputs and Transformers for climate data, before fine-tuning the model with ground reference samples. The proposed approach has undergone rigorous testing on two distinct large-scale datasets, with results indicating its superiority over traditional supervised learning models, which depends solely on labeled data. Furthermore, through the utilization of various evaluation metrics (e.g., RMSE, MAE, CCC, etc.), the proposed model exhibits higher accuracy when compared to other conventional ML algorithms like random forest and gradient boosting. This model is a robust tool for predicting SOC and contributes to the advancement of DSM techniques, thereby facilitating land management and decision-making processes based on accurate information.

An understanding of the key factors and processes influencing the variability of soil organic carbon (SOC) is essential for the development of effective policies aimed at enhancing carbon storage in soils to mitigate climate change. In recent years, complex computational approaches from the field of machine learning (ML) have been developed for modelling and mapping SOC in various ecosystems and over large areas. However, in order to understand the processes that account for SOC variability from ML models and to serve as a basis for new scientific discoveries, the predictions made by these data‐driven models must be accurately explained and interpreted. In this research, we introduce a novel explanation approach applicable to any ML model and investigate the significance of environmental features to explain SOC variability across Germany. The methodology employed in this study involves training multiple ML models using SOC content measurements from the LUCAS dataset and incorporating environmental features derived from Google Earth Engine (GEE) as explanatory variables. Thereafter, an explanation model is applied to elucidate what the ML models have learned about the relationship between environmental features and SOC content in a supervised manner. In our approach, a post hoc model is trained to estimate the contribution of specific inputs to the outputs of the trained ML models. The results of this study indicate that different classes of ML models rely on interpretable but distinct environmental features to explain SOC variability. Decision tree‐based models, such as random forest (RF) and gradient boosting, highlight the importance of topographic features. Conversely, soil chemical information, particularly pH, is crucial for the performance of neural networks and linear regression models. Therefore, interpreting data‐driven studies requires a carefully structured approach, guided by expert knowledge and a deep understanding of the models being analysed.

The growing popularity of vision transformers (ViTs) in remote sensing image classification is due to their ability to effectively capture long-range dependencies. However, their high computational cost and memory footprint limit their applicability, particularly for small-scale datasets and resource-constrained environments. To address these challenges, we propose the Multi-Scale Multi-Head Compact Convolutional Transformer (MSHCCT), a lightweight yet powerful model that integrates convolutional tokenization with small-scale vision transformers to enhance multi-scale feature representation while maintaining computational efficiency. Despite a modest increase in parameters and training time, MSHCCT achieves superior classification accuracy and robustness on high-resolution aerial scenes. Importantly, our approach eliminates the need for model pre-training, additional datasets, or multi-sensor data fusion, ensuring a computationally efficient and practical solution for remote sensing applications. The code will be made publicly available at https://github.com/aj1365/MSHCCT.