The humid and semi-arid area of the Argentinian Pampas has experienced marked inter-decadal climate variability and significant increases in annual precipitation until recently. Recent increases in precipitation have expanded the boundary of rainfed agriculture towards drier regions and contributed to major changes in land use. However, it is unclear whether the climatic variations form part of a longer term gradual trend or arise from "regime shifts" which would make these evolutions in land use unsustainable if the climate returns to a drier epoch.
While there are few statistically significant trends in seasonal and annual precipitation statistics, hidden states reflecting dry and wet years are apparent in the >60 year time series. Representation of this data is significantly improved when mixtures of two or more Gaussian or Poisson distributions are fitted, supporting the hypothesis of "regime-like" behaviour. This research is a preliminary step towards modelling the temporal persistence of the hidden states in the mixture models using hidden Markov models (HMM), together with dependence on atmospheric covariates. This talk will present the evidence for multiple regimes in the data and the novelty and utility of HMMs to simulate such regimes. The ultimate focus of this research is on the high impact weather phenomena which can have catastrophic consequences for agriculture. Therefore, ideas to extend the mixture models and HMMs to temperature and precipitation maxima using Extreme Value Theory will also be explored.