The global effort to combat climate change highlights the critical role of storing organic carbon in soil to reduce greenhouse gas emissions. Traditional methods of mapping soil organic carbon (SOC) have been labour-intensive and costly, relying on extensive laboratory analyses. Recent advancements in unmanned aerial vehicles (UAVs) offer a promising alternative for efficiently and affordably mapping SOC at the field level. This study focused on developing a method to accurately predict topsoil SOC at high resolution using spectral data from low-altitude UAV multispectral imagery, complemented by laboratory data from the Nogalte farm in Murcia, Spain, as part of the LIFE AMDRYC4 project. To attain this objective, Python version 3.10 was used to implement several machine learning techniques, including partial least squares (PLS) regression, random forest (RF), and support vector machine (SVM). Among these, the random forest algorithm demonstrated superior performance, achieving an R2 value of 0.92, RMSE of 0.22, MAE of 0.19, MSE of 0.05, and EVE of 0.71 in estimating SOC. The results of the RF model were then visualised spatially using GIS and compared with simple spatial interpolations of soil analyses. The findings suggest that a multispectral sensor UAV-based modelling and mapping of SOC can provide valuable insights for farmers, offering a practical means to monitor SOC levels and enhance precision agriculture systems. This innovative approach reduces the time and cost associated with traditional SOC mapping methods and supports sustainable agricultural practices by enabling more precise management of soil resources.
As the largest carbon reservoir in terrestrial ecosystems, soil organic carbon (SOC) plays a critical role in the global carbon cycle and climate change mitigation. A promising approach to swiftly procuring geographically dispersed SOC data is the amalgamation of UAV-based multispectral imagery at the local scale and Sentinel-2 satellite imagery at the regional scale. This integrated approach is particularly well-suited for precision agriculture and real-time monitoring. In this study, we evaluated the performance of UAVs and Sentinel-2 imagery in predicting SOC using four machine-learning models: Multiple Linear Regression (MLR), Support Vector Regression (SVR), Random Forest (RF), and Artificial Neural Networks (ANNs). UAV imagery outperformed Sentinel-2, achieving more accurate detection of local SOC variability thanks to its finer spatial resolution (5–10 cm versus 10–20 m). Among the models tested, the Random Forest algorithm achieved the highest accuracy, with an R2 of up to 0.85 using UAV data and 0.65 using Sentinel-2 data, along with low RMSE values. All models confirmed the superiority of UAV imagery based on key error metrics (SSE, MSE, RMSE, and NSE). Although Sentinel-2 remains valuable for regional assessments, UAV imagery combined with Random Forest provides the most reliable SOC estimates at local scales. The spatial SOC maps generated from both UAV and Sentinel-2 imagery showed more nuanced spatial variability than standard interpolation techniques. While prediction accuracy using UAV-based models was slightly lower in some cases, UAV imagery provided greater spatial detail in SOC distribution. However, this is associated with higher acquisition and processing costs compared to freely available Sentinel-2 imagery. Given their respective advantages, we recommend using UAV imagery for detailed, site-specific SOC estimations and Sentinel-2 data for broader regional-to-global SOC mapping efforts.