Climate change poses an increasing threat to agricultural productivity in the Texas High Plains (THP), a semi-arid region facing both warming trends and declining groundwater resources. This study integrates process-based crop modeling with geospatial analysis to identify spatial zones of climate vulnerability and adaptive potential for four major crops; winter wheat, cotton, maize, and grain sorghum under future climate scenarios. Using the Decision Support System for Agrotechnology Transfer (DSSAT) model, historical (1991–2020) and future yields (2031–2060 and 2070–2099) were simulated across 48 counties under Representative Concentration Pathway 4.5 and 8.5 (RCP 4.5 and RCP 8.5). Spatial clustering techniques, including Global Moran’s I and Getis-Ord Gi* statistics, were applied to classify counties into vulnerable, adaptive, stable, and more stable zones based on projected yield changes. Results revealed that wheat vulnerability was concentrated in southern counties, with projected yield decreases of 10–30% under RCP 8.5 relative to the 1991–2020 baseline. In contrast, northern counties showed yield increases of 30–50% under RCP 4.5 by mid-century and under RCP 8.5 by end-century. Cotton yields are projected to increase by 20–40% across most counties under both RCP 4.5 end-century and RCP 8.5 mid- and end-century relative to the historical baseline, although localized vulnerability may emerge in southwestern THP under RCP 8.5 by end-century. For grain sorghum yields are projected to increase by 10–20% in eastern and northern counties under RCP 4.5, but under RCP 8.5 widespread yield declines exceeding 40% are expected by end-century. These reductions are largely attributed to reduced rainfall and increased temperature stress during the growing season. In this study, sorghum was planted in June, which may have exposed the crop to greater late-season heat stress; earlier planting dates could potentially mitigate some of these adverse effects. Maize, which was planted in April in this study, showed spatially variable yield changes, with several southern counties exhibiting positive trends. These localized yield increases are primarily associated with projected precipitation increases during the maize growing season compared to the historical baseline, rather than an inherent crop-specific resilience to climate change. These spatially explicit findings underscore the need for targeted adaptation strategies, including the deployment of climate-resilient crop varieties, optimized irrigation management, crop diversification, and adaptive land use planning. The study offers actionable insights to support climate-resilient agricultural planning and inform precision adaptation policies for sustaining crop productivity in the THP under future climate scenarios.

Spatially explicit crop-yield information is needed for regional environmental modeling, sustainability assessment, and agricultural decision support, yet official yield statistics are commonly reported only at aggregated administrative scales. This study introduces the nayd R package, a reproducible geospatial workflow for converting county-level harvested area and yield statistics into spatially explicit production units and zonal clusters while preserving consistency with official records. County-level statistics from the USDA National Agri-cultural Statistics Service were integrated with USDA Cropland Data Layer crop masks, multi-sensor NDVI products, and satellite-derived evapotranspiration from OpenET SSEBop. An NDVI-based eligibility filter refined crop masks toward reported harvested area, while normalized NDVI and evapotranspiration layers were combined into spatial weighting surfaces and aggregated into contiguous production units and graph-based clusters. Case studies for cotton and winter wheat in Texas showed that the eligibility filter removed approximately 20–40% of CDL-classified pixels while maintaining consistency with reported harvested area and preserving plausible spatial gradients associated with ir-rigated and dryland systems. Evaluation against independent Texas A&M AgriLife vari-ety-trial data yielded normalized RMSE values of approximately 11–15% for cotton and 9–12% for winter wheat. The workflow provides an open-source framework for generating statistically consistent production zones for regional crop modeling, environmental as-sessment, and sustainable agricultural planning.

Spatially explicit crop yield information is needed for regional environmental modeling, sustainability assessment, and agricultural decision support, yet official yield statistics are commonly reported only at aggregated administrative scales. This study introduces the NAYD R package, a reproducible geospatial workflow for converting county-level harvested area and yield statistics into spatially explicit production units and zonal clusters while preserving consistency with official records. County-level statistics from the USDA National Agricultural Statistics Service were integrated with USDA Cropland Data Layer crop masks, multi-sensor NDVI products, and satellite-derived evapotranspiration from OpenET SSEBop. An NDVI-based eligibility filter refined crop masks toward reported harvested area, while normalized NDVI and evapotranspiration layers were combined into spatial weighting surfaces and aggregated into contiguous production units and graph-based clusters. Case studies for cotton and winter wheat in Texas showed that the eligibility filter removed approximately 20–40% of CDL-classified pixels while maintaining consistency with reported harvested area and preserving plausible spatial gradients associated with irrigated and dryland systems. Evaluation against independent Texas A&M AgriLife variety trial data indicated that the disaggregated clusters reproduced plausible spatial patterns of yield variability while retaining the county-level NASS constraints. The workflow provides an open-source framework for generating statistically consistent production zones for regional crop modeling, environmental assessment, and sustainable agricultural planning.

Diversified crop rotations and cover crops are increasingly promoted as climate-smart management strategies for improving soil organic carbon (SOC) while minimizing nitrous oxide (N₂O) emissions in semiarid agroecosystems. However, regional-scale assessment of SOC–N₂O trade-offs remains computationally challenging because process-based simulations across large spatial and climatic domains are highly demanding. This study combined long-term DSSAT simulations with machine-learning (ML) surrogate models to quantify and spatially predict SOC and N₂O responses to diversified rotations and cover-crop systems across the Texas High Plains (THP). Random Forest models were trained using DSSAT outputs representing multiple crop rotations, climate scenarios, and soil conditions. Two complementary modeling frameworks were developed: (i) BAU-relative responses, which quantified SOC and N₂O changes relative to baseline management, and (ii) added cover-crop effects, which isolated the additional benefits of diversified cover-crop systems relative to a simplified no-cover crop improved rotation system. Model interpretation was conducted using permutation importance and SHAP analysis to identify the dominant environmental and management controls. The surrogate models accurately reproduced DSSAT-derived responses, particularly for BAU-relative SOC and N₂O changes reaching R² ≈ 0.90, while added cover crop SOC was less predictable than added N₂O. Spatial predictions revealed strong geographic variability in climate-smart response zones, with larger SOC benefits generally observed under future climate conditions, particularly during the 2070s. However, some regions also exhibited stronger N₂O trade-offs, highlighting the need for spatially targeted management strategies. The results demonstrate the potential of combining process-based simulations with ML surrogates to generate computationally efficient decision-support tools for climate-smart agriculture in semiarid cropping systems.