Simulating pearl millet growth and yield with DSSAT-CERES-millet model: Evaluating climate-smart practices for uptake across diverse environments in Senegal

Abstract

Background: Evaluating crop suitability across diverse agro-climatic zones is crucial for mitigating risks to agricultural yield under climate change, particularly for pearl millet production in Senegal. To facilitate informed tactical or strategic decision making, aimed at enhancing productivity, a modeling approach is essential for comprehending climate impacts and formulating effective adaptation strategies. Methods: The DSSAT model’s capability to simulate four pearl millet cultivars (Souna-3, SL 28, SL 423, Thialack-2) under varying plant populations and micro-dosing fertilization strategies in diverse environments is evaluated in this study. The research also examines the risk associated with management practices on millet productivity in the context of climate variability and changing conditions for the period 1990–2020. To achieve these objectives, the CERES-millet model was calibrated using field data from 2022 to 2023 growing seasons for the four cultivars and subsequently validated with independent datasets from on-farm demonstrations in Senegal under various organo-mineral fertilizer applications for the same seasons. Results: The CERES-Millet model demonstrated high accuracy, simulating key growth stages and yield components across various millet cultivars and environments. For anthesis and physiological maturity, the model achieved low RMSE values (3–5 days and 2–4 days) and acceptable d-index ranges (from 0.6 to 0.9), indicating reliable predictions of these critical developmental milestones. The model’s performance extended to accurately simulating leaf area index (LAI), grain yield, and above-ground biomass, with specific RMSE values reported for different cultivars such as Souna 3 (RMSE of 443 kg ha ⁻¹ for grain yield and 747 kg ha ⁻¹ for AGB), Thialack-2 estimated RMSE of 389 kg ha ⁻¹ for grain yield and 208 kg ha ⁻¹ for AGB. SL 28 recorded RMSE of 257 kg ha ⁻¹ for grain yield and 560 kg ha ⁻¹ for AGB, while SL 423 had RMSE of 698 kg ha ⁻¹ for grain and 640 kg ha ⁻¹ for AGB. A long-term 31-year seasonal simulation revealed that dual-purpose varieties consistently outperformed Souna 3 in terms of yield, despite significant inter-annual variability. Conclusion: Combining these varieties with micro-dosing fertilizer applications led to higher yield thresholds. Notably, micro-dosing fertilizer level 3 (3g/hill of NPK, 2g/hill of Urea, and 50g/hill of cow manure) demonstrated consistent stability, making it a cost-effective approach to mitigate yield variability in semi-arid regions of Senegal. Long-term simulations highlighted the promising potential of combining dual-purpose varieties with optimized micro-dosing strategies to enhance yield stability and resilience. This approach offers a practical method for mitigating the effects of climate variability in Senegal and similar environmental contexts.