Determining the onset of summer rainfall over Vietnam using self-organizing maps

• Published in: Climate dynamics Vol. 62; no. 9; pp. 9189 - 9206
• Main Authors: Bui-Minh, Tuan, Doan, Quang-Van, Nguyen, Kim-Cuong, Phan-Van, Tan, Trinh-Tuan, Long, Cong, Thanh, Trinh-Minh, Ngoc
• Format: Journal Article
• Language: English
• Published: Berlin/Heidelberg Springer Berlin Heidelberg 01-09-2024; Springer Nature B.V
• Subjects: Algorithms; Climatology; Clustering; Convection; Dynamic height; Earth and Environmental Science; Earth Sciences; Geophysics/Geodesy; Geopotential; Geopotential height; Intraseasonal oscillation; Long wave radiation; Monsoon onset; Monsoons; Oceanic convection; Oceanography; Original Article; Precipitation; Rain; Rainfall; Rainy season; Self organizing maps; Summer; Summer monsoon; Summer rainfall; Weather; Weather patterns; Westerlies; Wet season; Wind; Winds; Vietnam; Intraseasonal oscillation; Tropical disturbance; Rainy season onset; Monsoon index; Foehn effect; Self-organizing maps
• ISSN: 0930-7575; 1432-0894
• DOI: 10.1007/s00382-024-07385-x

The onset of the rainy season in Vietnam displays distinct characteristics compared to other Asian summer monsoon (ASM) regions due to its elongated shape running from south to north, coupled with its complex terrain and the influences of multiple weather systems. Determining the onset is challenging, although numerous empirical indices exist. Here, we proposed an alternative approach to determine the onset by examining the large-scale weather patterns associated with the abrupt increase in rainfall in the country. In this new approach, the evolutions of the ASM system were interpreted as consecutive occurrences of multiple weather patterns, detected by a clustering algorithm Self Organizing Maps (SOMs). Specifically, SOMs were applied on outgoing longwave radiation, 850 hPa wind, and geopotential height over the ASM regions from April to July. The primary features of these atmospheric variables were generalized to 16 patterns. A high-resolution Vietnam Gridded Precipitation (VnGP) dataset was employed to investigate the linkage between these large-scale patterns and local-scale rainfall. The large-scale patterns associated with the onset were distinguished, which are characterized by the development of westerly winds and large-scale convections. Then, the onset of summer rainfall in Vietnam is identified based on the initial appearance of wet-condition weather patterns. This approach offers several advantages: it is objective and does not require additional thresholds of rain or wind to determine the monsoon onset. It also more effectively captures the characteristics of large-scale circulation and deep convection compared to relying on predefined monsoon indices alone. Additionally, this approach helped to elucidate two types of weather patterns associated with the onset: those associated with the northeastward development of deep convection and westerly wind from the Indian Ocean to the Indochina Peninsula and activities of the Boreal Summer Intraseasonal Oscillation.