Defining Philippine Climate Zones Using Cluster Analysis

Philippine climate zones traditionally have been based on the Modified Coronas Classification (MCC). The MCC is simple in concept as it uses only average monthly rainfall totals to define four climate zones, referred to as Types I-IV. Types I and III have wet and dry seasons, and types II and IV have wet season and no dry season.

            The present study provides redefined Philippine climate zones using a more sophisticated approach, with a cluster analysis being applied to the average monthly rainfall amounts obtained from 53 stations for the 55-year period 1961-2015. To determine the optimal number of climate zone clusters, both single-linkage hierarchical and K-means cluster analysis algorithms were used, together with known characteristics of Philippine rainfall distributions and attributes.

            The most appropriate number of clusters is determined using dendrogram, cophenetic correlation coefficients, and silhouette coefficients as objective measures for selecting the optimum number of clusters. Employing single linkage hierarchical and K-means methods in tandem identified six different Philippine climate zones, which is two climate zones more than the currently adopted MCC climate classification.  The current study has identified meteorological factors that contributed to the six distinct climate zones. The study also is intended to help agricultural stakeholders with their planning and farming decisions.