Climatic analysis linked to land vegetation cover of Mexico by applying multivariate statistical and clustering analysis

Abstract

The climate regions of Mexico are delimitated using hierarchical clustering analysis (HCA). The data used consists of monthly means of maximum and minimum temperatures and monthly-accumulated precipitation. The dataset was obtained from heterogeneously distributed climatic stations in Mexico for the period from 1961 to 2004. This cluster method assigns precipitation and temperature variables to groups of clusters based on similar statistical characteristics. We carried out a principal components analysis to obtain a standardized reduced matrix to be used in HCA. By applying two clustering criteria (K-means and Ward's method) it was possible to define statistically groups of stations that delimit regions of similar climate. In addition, the applied methodology describes the dominant vegetation distribution for each climate region. This analysis may contribute to the generation of new climate scenarios, where the dynamics of land vegetation cover could be included as a biomarker of climate.