Análise multivariada aplicada ao vento na área costeira do Nordeste do Brasil.

Abstract

In this work multivariate techniques (factorial analysis by principal components and cluster analysis) are used for identification of homogeneous regions and seasonal patterns of the surface wind on the coastal area of Northeast Brazil. Hourly surface wind data available on the electronic page of the Meteorological Network of the Brazilian Air Force Command (REDEMET) for the eight coastal northeastern metropolis, covering the period 2003-2009, are used. The monthly mean surface wind fields show the São Luís wind direction (east-northeast) to be in contrast with the direction on the other seven metropolis (east-southeast). The wind speed varies throughout the year on the entire coastal area; the values are lower (higher) in the rainiest (less rainy) months, varying from 4 to 6 m/s (1 to 3 m/s). The seasonal variability found on the mean surface wind patterns is mainly due to the South Atlantic subtropical high, and other large scale atmospheric systems that occur during the higher rainfall months. Three homogeneous regions are identified on the coastal area by performing a temporal seasonal analysis. The first region is formed by São Luís only, due to the east-northeasterly winds there. The second one, formed by João Pessoa and Maceió, is characterized by low wind speeds. The third one, comprised by Fortaleza, Natal, Recife, Aracaju and Salvador, is characterized by higher wind speeds. Similar results are obtained by performing a temporal analysis of four three-month periods. Two rotated principal components accounting for 95% of the explained variance are obtained by means of a seasonal and interannual spatial analysis. The PC1 (PC2) is well explained by Fortaleza and the six metropolis located on the east coast (São Luís). The factor loadings identify two wind regimes on the coastal area of Northeast Brazil as a consequence of the South Atlantic subtropical high circulation. Homogeneous groups representing distinct seasonal patterns are identified by applying cluster analysis techniques. Maxima Speeds Groups, Mixed Groups, Minima Speeds Groups and Equal Directions Groups are obtained among which months of extreme climatic conditions are highlighted as, for example, January 2004 in the Groups of Minima Speeds.