Apply Cluster Analysis to Study Daily Maximum Temperature Variation Over Vietnam

Studying the trends of maximum temperature variation over Vietnam is essential for assessing the impacts and developing climate change adaptation strategies. However, previous studies have relied on linear trend analysis at individual stations, which does not capture short-term fluctuations or abrupt changes in the data series. In this study, the Self-Organizing Map (SOM) method is applied to monthly averaged daily maximum temperature anomalies to provide a more detailed analysis of the spatiotemporal variability of maximum temperature across Vietnam. The results indicate that the variability of maximum temperature can be classified into six primary patterns, with significant differences among climatic regions. Patterns characterized by below-average maximum temperatures show a strong decreasing trend throughout the study period, while those with above-average maximum temperatures exhibit a marked increase in recent decades. Seasonal analysis further reveals that during summer, patterns with a temperature increase of 1-2 °C occur more frequently, whereas in winter, patterns with increases exceeding 2 °C are more prevalent. Notably, the frequency of months within each cluster is closely linked to ENSO, with above-average temperature months tending to occur more frequently during El Niño phases, and the opposite during La Niña phases.