A regionalization based on Principal Component Analysis (PCA) was made using an up_to_date high-quality climatological database. The set of 175 stations (168 Mexican and 7 in southern USA) contains long-term (1931-2001) monthly precipitation, around half of them computed from daily data. Because more than seventy percent of the annual precipitation totals in México occur between May and October (Mosiño and Garcia, 1974), the study was applied to three different kinds of time series: annual, wet (may-oct) and dry (nov-apr) seasons. PCA orthogonal and oblique rotated solutions with ten significant components that captured about sixty percent of the total variance were used, with the clearest outcomes obtained from oblique methods. Resulting regions strongly agree with the Mexican climatology (García, 1988). Although observed areas in the study reflect particular spatial and seasonal responses, some regions like the Central Highlands and Río Bravo Basin areas are coincident amongst the three temporal resolutions. Overall, good consistency is especially noted between the annual and wet seasons. Trend and extreme analyses are performed on precipitation averages and a representative station for each region used to compare regional against local scales responses, also making use of the full capacity of the long-term daily data.
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner