An AI-Driven Cradle-to-Farm-Gate Life Cycle Assessment Framework for Energy and Carbon Sustainability in Crop–Soil–Climate Systems

İngilizce (OpenAlex)

This research introduces a framework for assessing sustainability in agricultural production, enhanced by AI, that merges environmental, agronomic, and management aspects into a cohesive decision-support system. Drawing on data from Karnataka, India, the study examined six critical parameters including crop type, soil pH, rainfall, temperature, soil classification, and nitrogen-phosphorus-potassium (NPK) fertilizer inputs. A Composite Sustainability Indicators (CSI) was created to assess the multifaceted performance of various crops through five normalized sub-index computation: fertilizer efficiency, pH suitability, rainfall adequacy, temperature compatibility, and soil quality. Specific optimal thresholds were established for key crops like wheat, barley, and maize to align with their unique ecological needs. An efficient Life Cycle Assessment (LCA) component calculated energy equivalents (N = 60.6 MJ kg⁻¹, P₂O₅ = 11.1 MJ kg⁻¹, K₂O = 6.7 MJ kg⁻¹) and greenhouse gas emissions (N = 6.6, P₂O₅ = 1.0, K₂O = 0.6 kg CO₂e kg⁻¹) associated with fertilizer usage. An AI-driven regression model (R² = 0.93; MAE = 0.024) was employed to predict CSI under diverse environmental conditions, to facilitate scenario analyses for variations in soil pH (+0.5), rainfall (±100 mm), temperature (±2 °C), and fertilizer rates (±20%). The findings indicated that moderate reductions in fertilizer use improved sustainability scores, whereas climatic variability elicited varied responses based on the crop. Overall, this AI-LCA-enhanced framework offered a data-driven and flexible solution for precision agriculture, melding resource efficiency with environmental responsibility. It fosters sustainable production planning, adaptive management amidst climate variability, and informed policymaking aimed at building resilient agricultural systems.