Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This study introduces SpTGNN, a multi-modal spatio-temporal graph neural network designed for predicting top-soil organic carbon (SOC) by integrating heterogeneous graph representations of soil measurements with advanced feature extraction from satellite data. By utilizing relational graph attention and a Mixture-of-Experts module, the model effectively captures complex spatial and temporal patterns while addressing the limitations of traditional methods that overlook the irregular spatial structure of field data. The results demonstrate a significant improvement over existing approaches, achieving an R² of 0.762 on the Africa test split and showcasing the model's ability to provide robust uncertainty quantification through a combination of heteroscedastic regression and deep ensembles.
SpTGNN achieves a remarkable R² of 0.762 for soil organic carbon prediction, outperforming traditional methods by effectively leveraging multi-modal data and advanced graph neural network techniques.
Top-soil organic carbon (SOC) prediction is fundamental to agricultural sustainability, land use policy and fertilization planning. Existing approaches face two limitations: they pair hand-crafted covariates with classical ML or single-modal deep models that miss rich spectral and temporal information, and grid-based architectures ignore the irregular spatial structure of field measurements. We introduce SpTGNN, a multi-modal spatio-temporal graph neural network addressing both. SpTGNN represents soil measurements as nodes in a heterogeneous graph with three edge types (spatial proximity, spectral similarity, elevation), and applies relational graph attention to learn separate patterns per relation. A fine-tuned TerraMind encoder extracts node features from Sentinel-2, Sentinel-1 and DEM signals, combined with per-sample environmental covariates and learned positional and temporal embeddings. A sparse Mixture-of-Experts module fuses the four streams via top-$k$ routing. Uncertainty is captured by pairing heteroscedastic regression (aleatoric) with deep ensembles (epistemic), and a Moran's $I$ penalty regularizes spatial autocorrelation. We evaluate on a global SOC corpus split into three regional instances ($\sim$49k samples globally, Africa $\sim$26k, Europe $\sim$14k). Our 5-member deep ensemble reports $R^2=0.762$, RMSE $=3.51\pm0.48$ g/kg and MAPE $=22.9\%$ on the Africa test split, improving over a tabular XGBoost baseline; the best single checkpoint reaches validation $R^2=0.864$. Ablations confirm the heterogeneous graph, MoE fusion and fine-tuned backbone each contribute substantively, and the ensemble UQ stack achieves post-calibration ECE of $0.031$ (hybrid) and $0.026$ ($β$-NLL). To our knowledge, this is the first framework to unify foundation-model feature extraction, heterogeneous graph attention and decomposed uncertainty quantification for SOC estimation.