Fragmented spatial and temporal data leads to unreliable predictions of soil degradation and biomass yield. Improving predictive accuracy prevents resource misallocation and agricultural land loss.