Application of Crop Modeling for Climate-resilient Agriculture and Remote Sensing for High-throughput Plant Phenotyping

Sustainable cropping systems are important for ensuring long-term agricultural productivity in semi-arid regions like the Texas High Plains, where water scarcity and changing climate pose significant challenges. Assessment of diversified cropping systems through crop modeling can provide valuable insights into improving resource use efficiency, mitigating climate risks, and optimizing crop yield. At the same time, high-throughput cotton maturity assessment using UAS-based remote sensing technologies enables precise quantification of maturity levels and differentiation of cultivars for specific environmental and geographical conditions, contributing to more sustainable cotton production. So, the objectives of this study were to 1) evaluate crop yield, and water productivity in cotton-based cropping systems using DSSAT; 2) compare simulated county-level yield with USDA-NASS data to assess spatial and temporal yield trends; 3) analyze climate change impacts on cropping systems in terms of crop yield and identify climate-resilient cropping systems; 4) predict within-field cotton yield variability using ECa-optimized soil data for precision agriculture and; 5) develop a high-throughput UAS-based approach for cotton maturity assessment and cultivar differentiation. Implementing appropriate cropping systems based on soil types and climatic conditions is important for improving crop yield and conserving water in semi-arid regions. This study used the DSSAT model to simulate yield of cotton, sorghum, and wheat under continuous cotton, cotton-sorghum, and cotton-wheat cropping systems in the Texas High Plains. The model performed well across different soil types and growing seasons, showing good accuracy (R2 ≥ 0.79, NRMSE ≤ 15.74%) in predicting yield in commercially managed farms. The cotton-sorghum system produced the higher cotton yield and water productivity (6.3 kg ha⁻¹ mm⁻¹) than other cropping systems, making it a viable option for managing water resources efficiently. Long-term historical simulations of crop yield and spatiotemporal variability for cotton, sorghum, and wheat showed strong agreement with USDA-NASS reported yield, with R2 values of 0.57 for cotton, 0.72 for sorghum, and 0.71 for wheat. Cotton and sorghum yield remained stable or increased on clay loam, silty clay loam, silt loam, and loam soils, while sandier soils showed slight reduction. Wheat yield remained relatively stable across most soil types. Climate change impacts on cotton, sorghum, and wheat yield were simulated using DSSAT under three SSP scenarios for continuous cotton, cotton-sorghum, and cotton-wheat systems. Cotton yield increased under SSP1-2.6 and SSP2-4.5 (3.4–15.6%) but reduced under SSP5-8.5 (4.1–16.8%). Sorghum saw greater losses under SSP5-8.5, while wheat remained stable in SSP1-2.6 and SSP2-4.5 but declined slightly in late-century SSP5-8.5. Northern counties with clay loam soils showed greater resilience, whereas southern counties with sandier soils, including Cochran, Terry, Garza, Lynn, and Yoakum, faced the most significant yield losses. Cotton-sorghum rotations produced higher cotton yield than other systems. Within-field cotton yield variability was simulated using the DSSAT crop model at a 30 m × 30 m grid resolution by incorporating deep apparent soil electrical conductivity (ECa) data into the soil profile. ECa effectively captured yield variability when soil texture, hydrologic properties, and soil fertility factor were optimized, achieving R2 of 0.43–0.56 and RMSE of 134–343 kg ha⁻¹ over six years. Grid-level predictions better reflected yield variability than soil-type-based aggregation, highlighting the potential of ECa-based simulations for identifying yield-limiting factors and optimizing input applications to improve cotton production. A high-throughput UAS multispectral imagery was used to quantify cotton maturity and differentiation of cotton cultivars. The maturity ratio index (MRI), derived from multispectral imagery, demonstrated a strong correlation with manually counted boll opening rates (R2 = 0.79–0.83). This index effectively distinguished early, mid, and late maturing cotton cultivars in different irrigation treatments. The optimal image acquisition and assessment period was 130–145 days after planting, when boll opening reached 25–60% across different cotton cultivars. So, MRI demonstrated its potential as a scalable tool for precision cotton breeding. Thus, these findings show the critical role of crop modeling in optimizing cropping systems, predicting yield variability, and assessing the impacts of climate change to enhance agricultural resilience in semi-arid regions. Additionally, UAS-based remote sensing proved to be an effective and scalable tool for high-throughput cotton maturity assessment, offering significant potential for precision breeding and cultivar differentiation.