A comprehensive study is carried out to investigate the sensitivity of various combinations between cumulus (CU) and microphysics (MP) schemes using the Weather Research and Forecasting (WRF) model for simulating Extreme Rainfall Events (EREs) over the hilly terrain of Nagaland (25.1 N – 27.2 N; 93.2 E– 95.3 °E). The study explores fifty-four distinct simulation combinations resulting from six CU and nine MP schemes. The simulated rainfall is compared with ground-based rain gauge network and Global Precipitation Measurement (GPM) satellite observations. The investigation focuses primarily on finding the optimal CU-MP scheme combination based on spatiotemporal precision achieved by the identified schemes. The C (New Simplified Arakawa Schubert) cumulus scheme is found to be the optimum individual CU scheme. Similarly, the M (WDM5) microphysics scheme performs as the optimum individual MP scheme. The study highlights the significance of evaluating the performance of combined schemes over individual ones. Among the fifty-four combinations, sensitivity analysis identified six top-performing combinations, out of which the C-M (Multiscale Kain Fritsch - New Thompson) combination stands out to be optimum with a robust correlation coefficient (CC) of 0.75, lower standard deviation (STD) of 0.76 mm, and RMSD of 0.51 mm. A variation in lead/lag time is found to vary from (−) 3.2 to (+) 5.0 h, with a median value of (+) 1.3 h, suggesting that, on average, the model tends to predict the occurrence of peak rainfall events slightly earlier than observed one. The absolute difference in latitudes (longitudes) of peak rainfall variation ranges from 0.42 - 0.8 (0 - 0.51) with a median value of 0.66 (0.37), indicating that the model could better represent the longitudinal peak rainfall positions with notable accuracy. The spatial correlation coefficient (SCC) ranges between 0.31 and 0.68 with a median value of 0.50. From the overall analysis, it is also found that double moment MP schemes (New Thompson, Morrison (2 M), WDM5, WDM6) consistently perform better than the single moment MP schemes (Kessler, Lin et al., WSM3, WSM5, WSM6), whether in combination with CU schemes or individually. Spatiotemporal variations observed across different cases underscore the pivotal role of accurate topography representation in complex terrains, emphasizing the significance of terrain data in shaping rainfall patterns.

中文翻译：

---

WRF 模型中积云和微物理方案模拟那加兰邦丘陵地形极端降雨事件的敏感性分析

进行了一项全面的研究，以调查积云（CU）和微物理（MP）方案之间各种组合的敏感性，使用天气研究和预报（WRF）模型来模拟那加兰邦丘陵地形上的极端降雨事件（ERE）（25.1）北纬 – 27.2 北纬；93.2 东经 – 95.3°E）。该研究探索了由 6 个 CU 和 9 个 MP 方案产生的 54 种不同的模拟组合。将模拟降雨量与地面雨量计网络和全球降水测量 (GPM) 卫星观测结果进行比较。研究的重点是根据已确定的方案实现的时空精度寻找最佳的 CU-MP 方案组合。 C（新简化荒川舒伯特）积云方案被发现是最佳的个体 CU 方案。类似地，M (WDM5) 微物理方案作为最佳单独 MP 方案执行。该研究强调了评估组合方案相对于单个方案的绩效的重要性。在 54 个组合中，敏感性分析确定了 6 个表现最好的组合，其中 CM（多尺度 Kain Fritsch - New Thompson）组合脱颖而出，具有 0.75 的鲁棒相关系数 (CC)、较低的标准偏差 ( STD）为 0.76 毫米，RMSD 为 0.51 毫米。超前/滞后时间的变化范围为 (−) 3.2 到 (+) 5.0 小时，中值为 (+) 1.3 小时，这表明平均而言，该模型倾向于预测峰值降雨量的发生事件比观察到的事件稍早。降雨峰值变化的纬度（经度）绝对差范围为0.42~0.8（0~0.51），中值为0.66（0.37），表明该模型能够较好地反映纵向降雨峰值位置，且具有较高的精度。空间相关系数（SCC）范围在 0.31 至 0.68 之间，中值为 0.50。从总体分析来看，还发现双矩 MP 方案（New Thompson、Morrison (2 M)、WDM5、WDM6）始终优于单矩 MP 方案（Kessler、Lin 等人、WSM3、WSM5、WSM6） ），无论是与 CU 方案结合还是单独使用。在不同情况下观察到的时空变化强调了复杂地形中准确地形表示的关键作用，强调了地形数据在塑造降雨模式中的重要性。