Assessing Agricultural Production within Planetary Boundaries using an Integrated Monitoring and Hybrid Modelling Approach

Planetary Boundaries (PBs) define biophysical limits that safeguard Earth system stability. Exceeding these limits undermines ecosystem services, food security, economic stability, and climate resilience. Humanity is currently transgressing several of the PBs, demanding integrative and transformative research approaches that connect biophysical monitoring, sustainability targets, and societal decision-making. Despite its conceptual strength, the PB framework remains difficult to operationalize for regional agricultural systems. Global-scale assessments obscure the pronounced spatial heterogeneity of farming landscapes, where localized exceedances in nitrogen cycling, freshwater use, climate sensitivity, and biosphere integrity accumulate to drive broader Earth system risks. Consequently, there are limited guidance on where, how, and under which biophysical constraints agriculture can remain productive without breaching local environmental limits. This study proposes an integrated monitoring and modelling paradigm to assess regional agricultural production within planetary boundaries.

Our method moves beyond static, indicator-based assessments toward a dynamic, process-aware evaluation of local biophysical variables. We integrate high-resolution climate, soil, and land-use data with a spatially explicit crop model (SIMPLACE) to define regional control variables, including yield thresholds, nitrate leaching, and water-stress limits. To address structural uncertainties and capture non-linear climate-crop-soil interaction, we develop a hybrid modelling approach that couples SIMPLACE with machine learning algorithm (XGBoost).

Using SSP5-8.5 projections, we quantify specific yield and environmental constraints for Winter Wheat and Silage Maize in the Berlin–Brandenburg region in Germany. Hybrid simulations significantly outperform standalone process-based models, reducing mean absolute percentage error by ~9% for Winter Wheat and yielding consistently higher skill for Silage Maize. Our results reveal that emerging local boundaries are increasingly governed by compound climate extremes, particularly heat stress and precipitation deficits during flowering and early grain filling.

By framing PBs at the regional scale, hybrid modelling approaches enable the identification of conditions under which agricultural productivity, climate adaptation, and environmental integrity remain compatible—and where biophysical limits impose fundamental constraints. This approach offers a transferable pathway for embedding planetary stewardship into regional agricultural planning, climate adaptation strategies, and land-system governance.