Abstract
Abiotic stress and low resource-use efficiency are among the main challenges in agricultural production systems. Stress management is key to sustainable production. It is still challenging to identify and manage prevailing stresses under field conditions due to limited knowledge of the mechanisms of multiple abiotic stressors in crops. Crop models are becoming popular in agriculture because of their diversified nature in identifying multiple abiotic stresses and resource-use management in the complex nature of agricultural production systems. This study combined field measurements and crop modeling to improve the understanding of below- and above-ground resources use (water, nutrients, and light) and their impact on crop productivity and stress management under various planting conditions. A two-year field trial was conducted in the Thai uplands, comparing six treatments: (T1) maize sole crop with tillage and fertilization; (T2) maize-chili intercropping with tillage and fertilization; (T3) same as T2 but with minimum tillage and Canavalia ensiformis relay cropping; (T4) same as T3 plus Leucaena hedgerows; (T5) same as T3 without fertilization; and (T6) same as T4 without fertilization. The Water Nutrient and Light Capture in Agroforestry Systems (WaNuLCAS) model was calibrated using data from T1, T4, and T6 and evaluated against independent observations from T2, T3, and T5. Row-wise aboveground biomass, grain nitrogen (N) and phosphorous (P) concentrations, δ¹³C values, soil volumetric water content, and root length density were measured over two growing seasons. Grain δ¹³C values were significantly less negative in rows near the hedge (-10.33‰) than in distant rows (-10.64‰). More negative grain δ¹³C values (-9.32‰, p ≤0.001) were observed in T6. Both field observations and model simulations showed reduced maize biomass and lower grain N and P concentrations in rows closest to the hedgerows, driven by root competition for nutrients. Soil moisture was consistently higher in intercropped systems, and hedgerow height control prevented shading, indicating no water or light limitations. From the results it is concluded that WaNuLCAS model accurately reproduced spatial biomass patterns (EF = 0.95, RMSE = 0.98, and R2 = 0.96) and correctly identified nitrogen and phosphorus stress in maize rows planted closely with leucaena hedgerows. Scenario simulations demonstrated that balanced increases in both N and P inputs most effectively alleviated nutrient competition and improved the long-term system productivity. This integrated field-model approach provides a robust framework for diagnosing resource competition and optimizing nutrient management in hedgerow-based agroforestry systems under upland conditions.