Detecting spatio-temporal precipitation variability in Portugal using multichannel singular spectral analysis

Abstract

A spatio-temporal variability study of annual precipitation in Portugal has been performed using the multichannel singular spectral analysis (MSSA). The spatio-temporal variability patterns, which can represent progressive signals, reveal significant periodic characteristics that have not been unambiguously detected in previous studies using other methods. In this study, MSSA detects periodic behaviour in annual precipitation of about 3 and 9 years. Similar periodic signals of about 4 and 8 years are detected when singular spectral analysis (SSA) is applied to a series of a North Atlantic oscillation (NAO) index. The strong correlation between the NAO and precipitation variability in Portugal reinforces the importance of detection of these signals with similar periodic characteristics. Copyright © 2006 Royal Meteorological Society