Pergamon Phys. Chem Xarth (B), Vol. 25,No. 2, pp. 177-181,200O 0 2000 Elsevier Science Ltd All rights reserved 1464-1909/00/$ - see front matter PII: S1464-1909(99)00138-0 Nonseasonal Variability of Monthly Mean Sea Level Pressure and Precipitation Variability .over Europe Budong Qian, Jo&o Corte-Real, Hong Xu ICAT, Faculty of Sciences, University of Lisbon, Campo Grande, 1749-O 16 Lisboa, Portugal zyxwvutsrqponmlkjihgfedcbaZYXW Received 5 May 1999; revised 10 September 1999; accepted 24 September 1999 Abstract. The spatial modes of nonseasonal variability in both monthly mean sea level pressure (MSLP) fields over the northeastern Atlantic and western Europe sector and precipitation over Europe are investigated by using principal component analysis (PCA). The relationships between the two fields are revealed by canonical correlation analysis (CCA). The data sets used refer both to the period from 1911 to 1990. The most important spatial mode in MSLP fields is the NAO pattern and its corresponding principal component (PC) is closely related to the NAO index. However, it is interesting that the NAO pattern seems to be responsible only for the second EOF in precipitation while the most important spatial mode of precipitation corresponds to the third EOF of MSLP (North Sea pattern). Furthermore, the second EOF (Scandinavian pattern) of MSLP is highly associated with the third EOF in precipitation. Significant pairs of canonical correlation patterns between the MSLP fields and precipitation were obtained being coherent to the conclusions above referred. The results can be used to assess potential changes of precipitation over Europe based on variability of MSLP simulated by GCMs. 0 2000 Elsevier Science Ltd. All rights reserved 1 Introduction Because of noisy records of precipitation, signals of precipitation variability are difficult to be clearly detected and to be confidently assigned to corresponding underlying physical processes, especially for nonseasonal variability, although some variability has been related to oscillations of large scale atmospheric circulation such as the North Atlantic Oscillation (NAO) and the El Niiio/Southem Oscillation (ENSO). Precipitation is such an element that is strongly influenced by local factors such as topography. Therefore, temporal variability of precipitation over a large area, e.g., Europe, is more complicated than that if it is discussed over a small region. Nevertheless, precipitation variability must be related to variations in large-scale atmospheric circulation, especially the large-scale variability of precipitation. Correspondence to: ln~o Carte-Real 177 Previous researches revealed that precipitation over Europe is influenced by the large-scale atmospheric variability such as the North Atlantic Oscillation (NAO) and the El NiAo/Southem Oscillation (ENSO) (Fraedrich and Muller, 1992; Hurrel, 1995; Hurrel and van Loon, 1997; Rodo et al, 1997). Undoubtedly, further investigation on variability both of precipitation over Europe and of mean sea level pressure (MSLP) fields can be beneficial not only for further understanding precipitation variability, but also for better evaluating precipitation changes in a warmer climate imposed by greenhouse gases forcing, since climate models seem able to simulate large scale features of MSLP fields while precipitation cannot be properly estimated in the models. In this paper, major spatial modes of MSLP fields over the northeastern Atlantic and western Europe sector, especially those that might be responsible to-variability of precipitation over Europe, are discussed. The spatial modes of MSLP fields and precipitation are obtained by performing principal component analysis (PCA) and canonical correlation analysis (CCA). 2 Data and strategy Monthly mean sea level pressure fields over the northeastern Atlantic and western Europe are extracted for the region (20”-65”N, 60” W-60” E) during the period from 1911 to 1990, from 5’xlO’ monthly MSLP data set (Jones, 1987; Basnett and Parker, 1997). Monthly precipitation in 107 grid boxes well covering Europe during 191 l-1990 are obtained from S’x5’ global gridded monthly precipitation data set over land (G55WLD0095.DAT Version 1 .O), constructed and kindly provided by Dr. Mike Hulme of Climatic Research Unit at the University of East Anglia (Hulme, 1992). Some gaps (a very small number of missing data) in both data sets are filled in with corresponding monthly climatological values. In order to focus on interannual variability, an attempt was done to remove annual cycles in both data sets by normalizing monthly series at each grid, i.e., subtracting monthly mean and dividing by standard deviation in corresponding month. Principal component analysis was performed on both normalized fields, respectively, in order to obtain the most important spatial modes of climatic variability and to condense information in the fields for the subsequent canonical correlation analysis (CCA). CCA was, then,