Stochastic analysis of spatial and temporal variability of rainfall in the wider Mediterranean region

Empirical approach of the role of large scale weather systems in the climate and weather of the Mediterranean basin indicates high impact of cyclones and orography in the precipitation regime of the region. Stochastic analysis of daily precipitation data for the period 1955-2001 of stations along the coastal zone and inland of the Mediterranean is conducted. A network of cross-correlation coefficient weighted edges is constructed focusing on the one day delay pattern. A corresponding probability of rainfall network is developed along with introducing a Markov chain model based on conditional probability theory in order to improve the forecasting ability of observing a wet or dry day. Testing the model with the 2002-2004 daily precipitation values indicates variability in the aforementioned probabilities. The computational model involves spatial and temporal variables, thus the results are compared to the static and dynamic characteristics of cyclone activity in the Mediterranean in order to address the similarities of cyclone intensity and trajectory with the progress and development of rainfall events.