Abstract
This study examined the spatial and temporal variability of drought occurrences over Pakistan during 1960-2013.For this purpose, standardized precipitation index on 3 monthly time scale (SPI3) was calculated form the observed precipitation data. The prevailing wetness and dryness conditions on provincial scale were identified by respective SPI3 time series. The spatio-temporal variability of SPI3 values over the entire country was analyzed by principal component analysis (PCA).We choose first three dominant modes those explain 45.8 % variability in the data. The first dominant spatial pattern (EOF1) scores are positively correlated with the regions getting higher precipitation amount in Pakistan. The second dominant pattern represented by component EOF2 divi des the whole country into northern and southern regions. The distribution of third spatial pattern (EOF3) divides the whole country into north, south and central parts. All the PCA time series did not display any significant linear trend but however there are some cyclic patterns present in the data as revealed by spectral analysis. There is a major cycle of 10.7 years in the occurrence of drought intensive period possibly related to solar cycle of 11 years in sunspot activity in which a weak sunspot activity leads to a strong rainfall variability. The second dominant cycle of 6.1 years is present in SPI3 time series that could be explained with the help of El Nino southern oscillation whose periodicity lies in the range of 3-6 years. Another cycle of 2.7 years in rainfall variability could be explained with the help of Quasi-biennial oscillation (QBO) that is a cyclic phenomenon of reversal of winds between easterlies to westerlies over the equatorial stratosphere. These highlighted spatio-temporal patterns of drought incidences and their periodic behavior would help policy makers to implement well-coordinated water resources planning and drought preparedness strategies over the country to mitigate its possible adverse impacts