The impact of surface circulation on significant daily rainfall patterns over Mallorca

Abstract

This study is a preliminary attempt to associate the distribution of significant rainfalls over the island of Mallorca with recognized dominant surface circulation types. Based on the extraction of five prominent PCs from a principal components analysis using a T-mode (time) correlation matrix, six dominant daily rainfall spatial patterns have been isolated using cluster analysis for significant rainfall days (area mean 2 5.0 mm) for the island of Mallorca in the western Mediterranean. These six 'pattern groups', and a number of important subgroups, are linked closely to the prevailing near-surface circulation. Seven dominant circulation types have been identified, which are closely, although not perfectly, associated with rainfall distribution, so that a conceptual spatial rainfall model for the island seems possible. Such a model would be most useful if used in conjunction with NWP forecasts of probable rainfall amount issued for the area, and would provide valuable local information on likely heavy rainfall occurrence. AIMS OF THE STUDYThe main aim of this paper is to test the strength of association between the regional surface circulation and synopsis, and resultant rainfall patterns on days with significant falls for a part of the western Mediterranean: the island of Mallorca in the Spanish Balearic Islands. A Cyear run of daily rainfall data was available (19861989), so that derived associations must be regarded as 'pointers' to indicate the directions that future in-depth studies may take. Results from the study thus aim to provide a first approximation to rainfall patterns over the island of Mallorca linked to the more common synoptic situations and resultant near-surface airflows. Given that NWP forecasts yield mean rainfall predictions for areas larger than is generally considered the locul scale, knowledge of the relutiue distribution of rainfall within these NWP forecast areas is all that is required to provide the potential for more detailed local area forecasting, superimposing a predicted rainfall pattern on the NWP forecast area mean.An additional important aspect of this study is that surface circulation should be based on pre-existing classification schemes for the purposes of comarability and clarity. Further, the full range of synoptic patterns within a given classification may yield comparatively few distinctive rainfall distributions. Closely similar rainfall patterns may result from a number of different surface circulations, and, indeed, many surface CCC 0899-841 81951060673-24 0 1995 by the Royal Meteorological Society

Developments and prospects in synoptic climatology

“…van Regenmortel (1995) used PCA to create a 9-year realization of Botswana rainfall patterns. Sumner et al (1995) extracted principal components from surface circulation patterns to establish a linkage to daily rainfall for the Mallorca region of the Mediterranean. Serra et al (1998) adopted the S-mode PCA in a synoptic climatology of daily precipitation for Catalonia, and Romero et al (1999) classified synoptic patterns based on daily rainfall for the entire Spanish Mediterranean area.…”

Variation in leaf longevity of Pistacia lentiscus and its relationship to sex and drought stress inferred from leaf δ13C

“…(a) Characterization and location of the sampling sites. (b) Precipitation is given as approximate annual ranges for the precipitation zone of each sampling site from rainfall distribution and affinity maps (Sumner et al 1993;Sumner, Guijarro & Ramis 1995). Additional data on precipitation in 1993 and 1994, approximately corresponding to the time period of water supply to the C+1 leaves analysed for δ 13 C, are given for six sites at which rainfall was measured at a station near the location where samples were collected …”

Improved Cluster Partition in Principal Component Analysis Guided Clustering

“…In applications such as environmental sciences and climatology, Pearson correlation is typically used in PCA for calculating its eigenvectors and eigenvalues [8], [9], [10], [11].…”