Abstract

This paper presents a modelling approach to evaluate carbon retention and climate interaction in dryland farming by integrating process-based biogeochemistry, remote-sensing constraints, and socio-economic decision layers. The framework couples a water-limited crop–soil model tracking soil organic carbon pools, root allocation, and microbial turnover with an energy-balance land–atmosphere scheme that simulates albedo, surface roughness, boundary-layer coupling, and evapotranspiration feedbacks. Parameter estimation uses Bayesian calibration and hierarchical pooling to fuse field trials, eddy-covariance fluxes, and satellite products (soil moisture, land surface temperature, albedo, vegetation indices), while enforcing water and energy balance closure. Management is represented via ensembles spanning tillage intensity, residue retention, cover crops, organic amendments, deficit irrigation, and precision fertilizer timing. Disturbance modules capture drought lengthening, heat extremes, wind erosion, and pest pressure; policy levers include carbon pricing, drought insurance, and stewardship incentives. Model outputs include net ecosystem carbon balance, SOC stock change by pool, methane and nitrous oxide fluxes, water productivity, and radiative forcing equivalents. To quantify climate interaction, biogeochemical and biophysical effects are decomposed with counterfactual simulations carbon-only, biophysical-only, and combined yielding partial contributions to near-surface temperature and vapor-pressure-deficit anomalies. A decision module computes abatement cost curves and reliability-adjusted credits, ranking practices by expected carbon retention, co-benefits for soil health and yield stability, and risk of reversal. Global sensitivity and variance decomposition identify leverage points across uncertain precipitation regimes, soil textures, and management intensities. The approach is tested along precipitation and texture transects using cross-validation against independent SOC resampling, flux towers, and crop-cut data. Case studies show how carbon gains from residue retention and cover crops may be offset or enhanced by albedo shifts and energy-partitioning changes, and how diversified, risk-aware portfolios stabilize carbon while improving drought resilience. The framework supports monitoring, reporting, and verification and aligns with emerging agricultural carbon programs through robust uncertainty quantification and additionality screening. By linking mechanistic processes with decision analytics and observational constraints, this modelling approach provides a basis for planning, incentives, and credible climate claims in water-limited agroecosystems.