Total nitrogen levels as a key constraint on soil organic carbon stocks across Australian agricultural soils

Understanding how pedoclimatic drivers regulate soil organic carbon (SOC) stock is crucial for gaining insights into terrestrial carbon-climate feedback and thus adaptation to climate change. However, current data-driven SOC predictive models often neglect to incorporate total nitrogen (TN) data, thereby constraining our understanding of carbon-nitrogen interactions and their influences on SOC storage mechanisms across large scales. Utilizing an interpretable machine learning technique, we investigated how key drivers (TN, climate, elevation, land use, pH, SiO₂) affect SOC stocks at different soil depths across Australian major agricultural production regions. Incorporating TN into data-based SOC predictive models enhanced the explained variation by approximately 11 %. TN was identified as the predominant factor influencing SOC stocks, accounting for over 47 % of observed variability across all depths and outweighing climate effects in subsurface soils. Furthermore, we identified depth-specific thresholds of TN levels that constrain SOC accumulation: 1.45 mg/g soil for 0–10 cm, 0.80 mg/g soil for 10–20 cm and 0.63 mg/g soil for 20–30 cm. Projections of SOC stocks under different scenarios suggest that achieving these TN thresholds can promote SOC accumulation and help offset SOC losses associated with a 1 °C increase in mean annual temperature. This study underscores TN levels as a key constraint on SOC stocks across Australian agricultural soils, and thus should be explicitly considered when predicting large-scale SOC dynamics and formulating soil carbon sequestration strategies.