This paper analyses the spatial and temporal variability of winter droughts in a semi-arid geographic gradient in Northeast Spain, from the Pyrenees in the north to the Mediterranean coastland in the south. Droughts that occurred between 1952 and 1999 were analysed by means of the Standardised Precipitation Index (SPI). The influence of the weather-type frequency and of the general North Atlantic atmospheric circulation patterns was analysed. The results indicate that winter droughts show an important spatial variability in the study area, differentiating three well-defined patterns. These correspond to the Pyrenees, the centre of the Ebro Valley, and the Mediterranean coastland. General negative trends in winter SPI have been found, which are indicative of the increase in winter drought conditions in the study area. Nevertheless, important spatial differences have also been recorded. Dominant north–south gradients in the influence of weather types are shown. Moreover, the negative trends in winter-SPI values agree with the negative trend in the frequency of the weather types prone to cause precipitation, such as the C, SW and W weather types and the increase in the frequency of A weather types. Nevertheless, in the Mediterranean coastland, the positive trend in SPI values agrees with the increase in the frequency of weather types of the east (E, SE), which are prone to cause precipitation in this area. Interannual variations in the frequency of the different weather types have been highly determined by different general atmospheric circulation patterns, mainly the North Atlantic Oscillation (NAO). Nevertheless, the correlation between the time series of weather-type frequency and the winter SPI is higher than that found between the SPI and the NAO. Thus, although the interannual NAO variability explains a high percentage of the interannual differences in the frequency of different weather types, it is not sufficient to explain the spatial and temporal variability of droughts, which respond better to atmospheric variability at more detailed (synoptic) spatial scales. Copyright © 2006 Royal Meteorological Society.