versão impressa ISSN 0187-6236
Autoregressive Integrated Moving Average (ARIMA) models have been devised for long-term monthly series of maximum and minimum temperatures from south-western Spanish observatories. The original series were transformed into stationary ones, and the orders (p, d, q) for each monthly series were obtained. These were validated using series of residuals, and the parameters of the functions were estimated from the monthly stochastic models. The forecasts of change formulated from the ARIMA models suggest that the minimum temperatures in the study area are in a mean warming phase of the order of 0.2 °C over the next decade. The change in maximum temperatures is not significantly defined. These estimations correspond with those obtained by the authors in earlier studies using other methods, but the level of precision is much higher in this case.
Palavras-chave : Stochastic processes; time series; climate change; Iberian Peninsula.