Cotton yield increases while risk decreases with bahiagrass integration in a cropping system

Context Integrating perennial phases into annual crop rotations may improve productivity and reduce downside yield risk, but long-term evidence from the southeastern U.S. Coastal Plain remains limited. Methods A 21-year field experiment compared three cotton systems: (1) cotton following bahiagrass (C_BG) within a bahiagrass-bahiagrass-peanut-cotton rotation, (2) cotton following peanut (C_P), and (3) cotton following cotton (C_C) within a conventional peanut-cotton-cotton rotation. Mixed-effects models, climate and period-stratified analyses, Finlay-Wilkinson regression, resilience metrics, and extreme-outcome probabilities were used to evaluate system performance. Results Median lint yield was highest under C_BG (1427 kg/ha), exceeding C_P and C_C by 8.1% and 14.9%, respectively. Model-estimated means showed the same ordering (C_BG = 1414, C_P = 1289, and C_C = 1227 kg/ha; p < 0.001), and differences persisted after detrending (p < 0.01). Yield advantages for C_BG were maintained across rainfall, temperature, and temporal categories. Finlay-Wilkinson analysis indicated intermediate responsiveness for C_BG, whereas C_P was more responsive than C_C. During drought years, C_BG exhibited the highest median resilience (1.02) and probability of resilience ≥ 1 (0.59). The probability of high-yield outcomes was less variable among rotations (25%, 22%, and 21% for C_BG, C_P, and C_C, respectively), whereas the probability of low-yield outcomes was lowest under C_BG (34%) compared with C_C (39%) and C_P (44%). Conclusion Integrating bahiagrass into crop rotations increased cotton lint yield across a wide range of environmental conditions. These benefits were associated primarily with a lower probability of low-yield outcomes, indicating reduced downside yield risk alongside improved average productivity.