The Northeast monsoon (NEM) exerts a significant influence on the climate system of Peninsular Malaysia (PM) and has profound implications for water resource management, flood control, and disaster preparedness in the region. This study employed CHIRPS rainfall data from 1981 to 2022 at a spatial resolution of 0.25° to generate climate zones. The performance of global climate models (GCMs) in replicating monthly rainfall was evaluated using multiple statistical metrics, with the top-ranked ones being used for the generation of future projection downscaled ensemble models. The NEM months (November, December, January, February, and March) were used as input parameters to form clustered homogeneous groups using the Ward hierarchical method. The optimal number of climate zones was determined using the elbow method, silhouette width values, and spatial evaluation. Ultimately, a 7-cluster solution was chosen due to its consistency and meaningful cluster formation. The zones were categorized into two regions: the West Coast and the East Coast, based on their geographical location and distinctive monthly NEM rainfall patterns. The analysis of the probability distribution function (PDF) showed that rainfall in the West Coast region had a more spread-out pattern with varying amounts, while rainfall in the East Coast region was more concentrated and consistent. The Northeast climate zone received the highest rainfall, attributed to its proximity to the South China Sea, mountainous areas, and the NEM convergence zone. In contrast, the Northwest and Lower Central-West climate zones received the lowest rainfall due to the rain shadow effect, limited moisture availability, and distance from major moisture sources. The CMCC-ESM2, CMCC-CM2-SR5, EC-Earth3, and CanESM5 consistently ranked among the highest performing models. After linear scaling bias correction, validated GCMs showed significant improvements, with improvements in MAE, NRMSE, PBIAS, RSR, NSE, MD, CP, R2, and KGE by 11–45%, 15–47%, 96–100%, 15–47%, 10–79%, 5–37%, 68–113%, 11–73%, and 9–51%, respectively. Analysis of NEM season rainfall patterns across climate zones indicated consistent mid-future increases for the Northwest and Lower Central-West zones (12.4–19.8%) and the Upper Central-West and Southwest zones (0.2–5.4%) across all scenarios. In the far-future, the Southeast zone saw an overall increase in rainfall (4.9–11.4%), while the Central-West zone exhibited mixed results (−4.0 to 5.6%). The results also highlighted spatial variability and the need to consider specific climate zones when assessing the impacts of climate change on the NEM, emphasizing the importance of incorporating these projections into long-term planning and decision-making processes.

中文翻译：

---

基于CMIP6下马来西亚半岛聚集性气候带的未来东北季风降雨特征

东北季风（NEM）对马来西亚半岛（PM）的气候系统产生重大影响，对该地区的水资源管理、防洪和防灾具有深远的影响。本研究采用 1981 年至 2022 年空间分辨率为 0.25° 的 CHIRPS 降雨数据来生成气候区。使用多种统计指标评估全球气候模型（GCM）在复制每月降雨量方面的性能，其中排名最高的指标用于生成未来预测缩小比例的集合模型。 NEM 月份（11 月、12 月、1 月、2 月和 3 月）用作输入参数，使用 Ward 分层方法形成聚类同质组。使用肘法、轮廓宽度值和空间评估来确定气候区的最佳数量。最终，由于其一致性和有意义的集群形成，选择了 7 集群解决方案。根据地理位置和独特的 NEM 每月降雨模式，这些区域被分为两个区域：西海岸和东海岸。概率分布函数（PDF）分析表明，西海岸地区的降雨量较为分散，数量不一，而东海岸地区的降雨则较为集中且一致。东北气候带由于靠近南海、山区和NEM辐合带，降雨量最高。相比之下，由于雨影效应、有限的水分供应以及距主要水分源的距离，西北和中西部下部气候区的降雨量最低。 CMCC-ESM2、CMCC-CM2-SR5、EC-Earth3 和 CanESM5 始终名列性能最高的模型之列。经过线性标度偏差校正后，经过验证的 GCM 显示出显着的改进，MAE、NRMSE、PBIAS、RSR、NSE、MD、CP、R2 和 KGE 提高了 11–45%、15–47%、96–100%、15分别为 –47%、10–79%、5–37%、68–113%、11–73% 和 9–51%。对跨气候区 NEM 季节降雨模式的分析表明，在所有情景中，西北和中西部下部地区 (12.4-19.8%) 以及中西部上部和西南地区 (0.2-5.4%) 中未来降雨量将持续增加。在遥远的未来，东南地区的降雨量总体增加（4.9%至11.4%），而中西部地区的结果好坏参半（-4.0%至5.6%）。结果还强调了空间变异性以及在评估气候变化对 NEM 的影响时考虑特定气候带的必要性，并强调将这些预测纳入长期规划和决策过程的重要性。