Statistics of hot and cold events in Catalonia (NE Spain) for the recording period 1950–2004

Abstract

Spatial patterns and statistical models for hot and cold events affecting Catalonia (NE Spain) are obtained from series of daily maximum and minimum temperatures recorded at 65 meteorological stations throughout the period 1950–2004. The study is based on the crossing theory, taking care that daily temperatures are normally distributed and previous data treatments (removal of trends and periodicities) assure their stationary character. With the aim of facilitating comparisons among different thermometric records, hot and cold events are defined as large departures, given in standard deviations, of daily temperatures from daily averages. From the statistical point of view, the assumption of a normal distribution of the initial date of the events is questionable and a simple assignment of hot events to summer and cold events to winter must be discarded according to the averaged initial dates and their standard deviations. The event magnitudes, defined as the absolute value of their maximum departures, follow an exponential distribution and event lengths can be modelled by an autoregressive Markov process with a Gaussian noise component. The number of events per year fits a Poisson distribution well only for high departures and the whole number of hot and cold events decays exponentially with the increasing departure for every temperature series. This property permits an estimation of the expected maximum departure for every thermometric station during the recording period. Even though spatial features observed for the number of events, their mean initial date, average event magnitude and average event length depict quite complex patterns due to the orography of the country and the vicinity to the Mediterranean Sea, the results obtained improve the knowledge on the hot and cold events in Catalonia.