Environmental characterization of maize yield variability in a major Corn Belt state

Abstract Environmental characterization offers a robust framework for explicitly studying spatio-temporal variability in crop yields and for designing region-specific management strategies. Indiana, a major maize-producing state in the U.S. Corn Belt, exhibits substantial spatial and interannual yield variability of rainfed maize driven by climatic gradients. This study developed a yield-based environmental regionalization framework, in which long-term county-level maize yield data (2000–2024) were first used to delineate environmental regions, while climatic records were subsequently used to characterize those regions and identify the climatic variables most strongly associated with interannual yield variability for each region. A multi-criteria evaluation of fuzzy c-means solutions based on cluster validity indices, spatial coherence, and agronomic interpretability identified five yield-based environmental regions (NW: north-west, NE: northeast, WC: west-central, SW: south-west, and SE: south-east), revealing a marked productivity gradient between the high-yielding WC region (ca. 11.2 Mg ha − 1 ) and the lower-yielding NW region (ca. 8.5 Mg ha − 1 ). Principal component analysis of long-term climatic records explained nearly 84% of the total variance across the first two components, with vapor pressure deficit (VPD), reference evapotranspiration (ET₀), and a heat-stress index (maximum air temperatures > 35 °C) as the main variables differentiating climatic conditions across regions. In all regions, there was a positive trend in grain yield (Mg ha − 1 ) over time, but relative gains (% yr − 1 ) were larger in lower-yielding environments (e.g., NW region, 2% yr − 1 ) than in higher-yielding environments (e.g., WC region, 1.6% yr − 1 ), suggesting a tendency toward relative convergence in maize performance, though absolute yield gaps may persist. Correlation analyses between (i) yield residuals from linear regressions fitted to yield and year for each region and (ii) climatic anomalies (i.e., deviations of each county–year observation from the corresponding county-specific long-term mean [2000–2024] for each climatic variable) demonstrated that VPD around flowering showed the strongest and most consistent association with interannual yield variability in the NE, SE, and SW regions. In contrast, in the NW and WC regions, interannual yield variability was more strongly associated with the heat-stress index during the same maize period, although the strength of these relationships was lower compared to those observed for VPD in other regions. Overall, results highlight VPD and heat stress around flowering as key climatic factors for maize yield variability across Indiana and provide a potentially transferable framework for linking environmental gradients to yield stability in rainfed cropping systems and region-specific adaptation strategies.