Use of the standardized precipitation index (SPI) and a modified SPI for shaping the drought probabilities over Turkey

This paper suggests a new methodology for the Standardized Precipitation Index (SPI), for studying different aspects of drought events in Turkey, including intensity, frequency and identifying the spatial and temporal patterns. The classical SPI is locally defined; for instance, for monthly scale or running means, a drought event would begin when the precipitation amount is below its long-term average. However, the proposed method considers the local-time means of the precipitation series by fitting an upper and a lower envelope to precipitation data to obtain the SPI values, thus it explicitly assesses a dry (or wet) spell of weather. When we consider the climatological/meteorological features of precipitation and the physical geographical controls of the climate over Turkey, it is found that the proposed method is more capable of estimating the probability of dry or wet conditions for a station's monthly precipitation totals (or running means) than if the SPI values are obtained using the classical method. Comparisons of the probability values calculated for the long-term series and various time scales have revealed that the proposed method successfully estimates and/or well represents the various precipitation anomalies, particularly the wet and dry extremes. For instance, the probabilities of monthly values obtained by the proposed SPI method being above normal, below normal, near normal (i.e. normal) and extremely dry are more realistic and reasonable than those found by the classical SPI method. In conclusion, one can apply the proposed SPI methodology for determining and monitoring droughts of other semi-arid, dry sub-humid and semi-humid climate regions, where the precipitation series show high seasonality and year-to-year variability, such as in the Mediterranean macro-climate region. Copyright (C) 2009 Royal Meteorological Society