The agrarian economy of Bhutan is highly vulnerable to rainfall uncertainties for its typical geographic location and rugged topography. Rainfall time series is also constrained by inadequate rain gauge stations in the country. To complement rainfall data obtained from 12 distinct satellite sources is validated against surface measurements from 12 ground stations. The rainfall obtained from Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks-Cloud Classification System-Climate Data Record (PERSIANN-CCS-CDR) exhibited a strong correlation with the measured rainfall with minimal bias and error. The seasonal mean rainfall is lowest in the winter (3.62%) with high variability (89.18%) and highest in the summer (60.36%) with low variability (33.93%). Using 3- and 12-month accumulation periods, the Standardized Precipitation Index (SPI) was computed to examine the frequency of wetness and drought. In all time frames, the severity of wetness and drought are of the “moderate” type. The majority of the districts witnessed higher levels of wetness in 1987, 1988, 2012, 2015, and 2017, while severe to extreme droughts were found over varying time periods in 1984–86, 1990–93, 1996, 1999, and 2019. In general, the frequency of wetness is 1.17 and 1.33 times higher than drought during June-August and 12-month periods, respectively. A strong positive correlation was observed between the Oceanic Niño Index (ONI) of October-December and the SPI of following June-August as well as the 12-month period. Additionally, the Indian Ocean Dipole Mode Index (DMI) showed a strong positive correlation with SPI, especially during June to August. The low amplitude and rapid fluctuation of DMI implies that the impact of Indian Ocean Dipole may be more seasonal; conversely, ENSO exhibited more noticeable fluctuations on an inter-annual basis. The results also suggests a complex interplay between the East Asian and Indian summer monsoons, the Tibetan plateau, and the local orography of Bhutan, which determines the wetness and drought. A district-level Mann-Kendall test of SPI showed the trends in wetness and dryness in a spatial context. Three districts at the 12-month scales and four districts at the 3-month scale (September-November) exhibited a significant upward trend, indicating increasing wetness. On the other hand, only one district showed a significant downward trend between September and November, suggesting an increasing drought incidence.

中文翻译：

---

厄尔尼诺南方涛动和印度洋偶极子与时间序列 （1983-2022） PERSIANN 降雨量数据对不丹湿旱趋势的遥相关


不丹的农业经济由于其典型的地理位置和崎岖的地形，极易受到降雨不确定性的影响。该国雨量计站不足也限制了降雨时间序列。为了补充从 12 个不同卫星源获得的降雨数据，根据来自 12 个地面站的表面测量值进行了验证。使用人工神经网络-云分类系统-气候数据记录 （PERSIANN-CCS-CDR） 从遥感信息估计降雨量中获得的降雨量与测量的降雨量表现出很强的相关性，偏差和误差最小。季节性平均降雨量在冬季最低 （3.62%），变率高 （89.18%），夏季最高 （60.36%），变率低 （33.93%）。使用 3 个月和 12 个月的积累期，计算标准化降水指数 （SPI） 以检查潮湿和干旱的频率。在所有时间范围内，潮湿和干旱的严重程度都属于“中等”类型。大多数地区在 1987 年、1988 年、2012 年、2015 年和 2017 年经历了较高的湿度水平，而在 1984-86 年、1990-93 年、1996 年、1999 年和 2019 年的不同时期发现了严重到极端的干旱。一般来说，6 月至 8 月和 12 个月期间的潮湿频率分别是干旱的 1.17 倍和 1.33 倍。在 10 月至 12 月的大洋尼诺指数 （ONI） 与随后的 6 月至 8 月以及 12 个月期间的 SPI 之间观察到很强的正相关。此外，印度洋偶极子模式指数 （DMI） 与 SPI 呈很强的正相关，尤其是在 6 月至 8 月期间。 DMI 的低振幅和快速波动意味着印度洋偶极子的影响可能更具季节性;相反，ENSO 在年际基础上表现出更明显的波动。结果还表明，东亚和印度夏季风、青藏高原和不丹当地地形之间存在复杂的相互作用，这决定了湿度和干旱。地区级 SPI 的 Mann-Kendall 检验显示了空间背景下湿润和干燥的趋势。3 个 12 个月尺度的地区和 4 个 3 个月尺度的地区（9 月至 11 月）均呈显著上升趋势，显示湿度增加。另一方面，只有一个地区在 9 月至 11 月期间呈显着下降趋势，表明干旱发生率增加。