Diagnosing warm-season rainfall variability in Mexico: A classification tree approach

Many observational studies provide evidence that connects seasonal rainfall variations over parts of the subtropical and tropical Americas with the behaviour of large-scale climate modes. This paper extends these claims based on analysis of a fairly extensive historical rainfall data set (1943-2006) from several locations in Mexico. The paper applies classification tree (CT) analysis to link warm-season rainfall variability to climate indices, which characterize ENSO state, the NAO, and sea-level pressure in the eastern North Pacific. While the results generally indicate that a significant part of the variability in warm-season rainfall over Mexico is conditioned by large-scale climate modes, there is only limited evidence of simple lag relationships involving Mexican rainfall and antecedent climate behaviour. The main implication here is that advances in warm seasonal rainfall prediction are likely to depend partially on improved capabilities to foreshadow the behaviour of large-scale climate modes a season or more in advance. Copyright (C) 2009 Royal Meteorological Society