- author = "Julio J. Valdes and Antonio Pou",
- title = "Climatic variation of the structure of maximum daily temperatures in Spain: A combined statistical and computational intelligence approach",
- booktitle = "International Joint Conference on Neural Networks, IJCNN 2009",
- year = "2009",
- month = "14-19 " # jun,
- address = "Atlanta, Georgia, US",
- pages = "3172--3179",
- keywords = "genetic algorithms, genetic programming, Iberian Peninsula, Mediterranean climates, Spanish meteorological stations, classical optimization, climatic variation, computational intelligence approach, geographical distribution, maximum daily temperatures, multimodal empirical distribution function, statistical approach, artificial intelligence, climatology, geophysics computing, meteorology, statistical distributions",
-
DOI = "
doi:10.1109/IJCNN.2009.5178649",
- ISSN = "1098-7576",
-
abstract = "Two blocks (1904-1921 and 1990-2007) of daily maximum
temperature data from seventeen Spanish meteorological
stations exhibit a multimodal empirical distribution
function (EDF). Most of the stations show important
differences in their EDF for each one of the considered
periods of time, a fact that reveals the complexity of
climatic changes within the accepted general warming
trend of the Iberian Peninsula.
As a tentative approach to understand the underlying structure of data, each EDF has been decomposed on two normal distributed functions. The parameters describing these functions for each station and for each time period have been space-optimized and visualized using classical optimization and genetic programming. The changes in the geographical distribution of the classes derived from the analysis point towards a recent greater role of Mediterranean climates, spreading its influence to the interior of the Peninsula. The general picture, however, is much more complex than a linear warming and a number of stations even show negative trends.
This study is considered to be a preliminary methodological exploration of future procedures destined to close the gap between data driven analysis and what models based upon first principles may tell.",
- notes = "Also known as \cite{5178649}",
