Climate change is projected to have a global impact that affect food production and security. The objectives of this study were to determine the potential impact of climate change on sorghum yield for rainfed production systems and to evaluate the potential of irrigation and shifting planting dates as adaptation options for two major sorghum production regions in Ethiopia. The Decision Support System for Agrotechnology Transfer (DSSAT) Cropping System Model (CSM)-CERES-Sorghum model was used to simulate the impact of climate change on sorghum yield for two Representative Concentration Pathways (RCPs; RCP 4.5 and RCP 8.5) and for three future periods including the 2025s (2010–2039), 2055s (2040–2069), and 2085s (2070–2099). The Agricultural Model Improvement and Inter-comparison Project (AgMIP) framework was used to select five representative GCMs for hot/dry, cool/dry, middle, hot/wet, and cool/wet climate scenarios. Two climate change adaptation practices including supplemental irrigation at two levels (deficit and full) to the current rainfed production system and shifting planting dates were evaluated. The CSM-CERES-Sorghum model was calibrated and evaluated using eight years of experimental data from Meisso, eastern Ethiopia. The model was then run for Kobo and Meisso under different climate change and crop management scenarios. Based on model evaluation results, the model performed well for simulating sorghum yield (R² = 0.99), anthesis (R² = 0.86, RMSE = 1.3), and maturity (R² = 0.79, RMSE = 4.4). The results showed that the average temperature for Kobo and Meisso is expected to increase by up to 6 °C under RCP8.5 in 2085. For the rainfed production systems without adaptation practices, drought stress is projected to intensify during anthesis, which was reflected by projected yield reductions by up 2 t ha⁻¹ for the two sites. Full irrigation was effective in reducing moisture stress and, thereby, increasing sorghum yield by up to 3 t ha⁻¹ for Kobo and 2 t ha⁻¹ for Meisso. On average, full irrigation resulted in a 1 t ha⁻¹ yield increase compared with deficit irrigation. Early planting dates also resulted in an increase in yield compared to the baseline planting dates, especially when combined with supplemental irrigation, although late planting was consistently disadvantageous even with supplemental irrigation. This study highlighted that the CSM-CERES-Sorghum model can be effectively used to simulate climate change effects on sorghum yield and evaluate different climate change adaptation practices. The outcomes of this study can also help to implement management decisions towards climate change adaptation for the current subsistence and fragile rainfed crop production system in Ethiopia and similar ecoregions across the globe.

预计气候变化将对全球粮食生产和安全产生影响。本研究的目的是确定气候变化对雨养生产系统高粱产量的潜在影响，并评估灌溉和改变种植日期作为埃塞俄比亚两个主要高粱生产区的适应选择的潜力。农业技术转让决策支持系统 (DSSAT) 种植系统模型 (CSM)-CERES-高粱模型用于模拟气候变化对两个代表性浓度途径（RCP；RCP 4.5 和 RCP 8.5）和三个的高粱产量的影响未来时期，包括 2025 年代（2010-2039）、2055 年代（2040-2069）和 2085 年代（2070-2099）。农业模型改进和比对项目 (AgMIP) 框架用于为热/干、冷/干、中、热/湿和冷/湿气候情景选择五个具有代表性的 GCM。评估了两种气候变化适应做法，包括对当前雨养生产系统进行两个水平（亏缺和全额）的补充灌溉和改变种植日期。CSM-CERES-Sorghum 模型使用来自埃塞俄比亚东部迈索的八年实验数据进行校准和评估。然后在不同的气候变化和作物管理情景下为 Kobo 和 Meisso 运行该模型。根据模型评价结果，该模型在模拟高粱产量方面表现良好（评估了两种气候变化适应做法，包括对当前雨养生产系统进行两个水平（亏缺和全额）的补充灌溉和改变种植日期。CSM-CERES-Sorghum 模型使用来自埃塞俄比亚东部迈索的八年实验数据进行校准和评估。然后在不同的气候变化和作物管理情景下为 Kobo 和 Meisso 运行该模型。根据模型评价结果，该模型在模拟高粱产量方面表现良好（评估了两种气候变化适应做法，包括对当前雨养生产系统进行两个水平（亏缺和全额）的补充灌溉和改变种植日期。CSM-CERES-Sorghum 模型使用来自埃塞俄比亚东部迈索的八年实验数据进行校准和评估。然后在不同的气候变化和作物管理情景下为 Kobo 和 Meisso 运行该模型。根据模型评价结果，该模型在模拟高粱产量方面表现良好（然后在不同的气候变化和作物管理情景下为 Kobo 和 Meisso 运行该模型。根据模型评价结果，该模型在模拟高粱产量方面表现良好（然后在不同的气候变化和作物管理情景下为 Kobo 和 Meisso 运行该模型。根据模型评价结果，该模型在模拟高粱产量方面表现良好（R ²  =  0.99）、开花期（R ²  =  0.86，RMSE  =  1.3）和成熟期（R ²  =  0.79，RMSE  =  4.4）。结果表明，2085 年，在 RCP8.5 条件下，Kobo 和 Meisso 的平均温度预计将升高 6 °C。对于没有适应措施的雨育生产系统，预计在开花期间干旱胁迫会加剧，这反映在预计这两个地点的产量将减少 2 t  ha ^-1。充分灌溉可有效减少水分胁迫，从而使Kobo 的高粱产量增加多达 3 t  ha ^-1和 2 t  ha ^-1对于迈索。平均而言，与亏缺灌溉相比，充分灌溉导致 1 t  ha ^-1产量增加。与基线种植日期相比，早期种植日期也导致产量增加，尤其是与补充灌溉相结合时，尽管即使有补充灌溉，延迟种植也始终处于不利地位。本研究强调，CSM-CERES-Sorghum 模型可以有效地用于模拟气候变化对高粱产量的影响并评估不同的气候变化适应做法。这项研究的结果还有助于为埃塞俄比亚和全球类似生态区当前的自给和脆弱的雨育作物生产系统实施气候变化适应管理决策。