2025 年 21C 巻 Special_Edition 号 p. 10-20
Existing climate zones in the Philippines are based solely on rainfall. However, recent studies suggest that other variables, such as temperature, should be considered for a better representation of the regional climate. Thus, this study aims to provide a novel representation of the regional climate in the Philippines by combining gridded rainfall and temperature data from the APHRODITE dataset from 1979 to 2015 using a two-phase Self-Organizing Map (SOM). This method is less computationally expensive and less sensitive to random noise from the input data than direct clustering approaches. When applying SOM to pentad-sum rainfall only, three distinct clusters are found, which is consistent with present studies that applied different clustering approaches on gridded and station-based rainfall data sets. When pentad-mean temperature is clustered only, two clusters are found, demonstrating the strong influence of elevation. When rainfall and temperature are combined, five clusters are found. The characteristics of the five clusters are discussed and validated against weather station data within each cluster. The results of this study offer a new perspective on climate zones in the Philippines when both rainfall and temperature are taken into account.
