Title: Analysis of daily rainfall series in NW Spain: Searching for evidence of climatic change
Authors: Elena Fernandez Iglesias - University of Oviedo (Spain) [presenting]
Roland Fried - TU Dortmund University (Germany)
Jorge Marquinez - University of Oviedo (Spain)
Abstract: The aim is to analyze time series representing daily rainfall events collected in the Esva river basin in Asturias (northwest of Spain) during the last 50 years, where there is a lack of quantitative data about climatic oscillations. Our interest is a better understanding of the seasonal patterns found in these data and of possible trends which might indicate effects of climate change. An exploratory data analysis leads us to a first order Markov model with alternation between wet and dry periods. The probability of rainfall is modeled by logistic regression, and precipitation values on wet days are described by Gamma distributions with a first order autoregressive dependence structure. Lowess smoothing is used to include seasonality in both submodels because the seasonal pattern could not be covered adequately by standard periodic functions. A tentative result of our analysis is some evidence that dry periods get longer in the course of time, while precipitation values on wet days might be unaffected.