INTERNATIONAL JOURNAL OF CLIMATOLOGY Int. J. Climatol. 25: 549–568 (2005) Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/joc.1147 GLOBAL ANALYSIS OF RUNS OF ANNUAL PRECIPITATION AND RUNOFF EQUAL TO OR BELOW THE MEDIAN: RUN MAGNITUDE AND SEVERITY MURRAY C. PEEL, a, * THOMAS A. MCMAHON a and GEOFFREY G. S. PEGRAM b a Centre for Environmental Applied Hydrology and the Cooperative Research Centre for Catchment Hydrology, Department of Civil and Environmental Engineering, The University of Melbourne, Victoria 3010, Australia b Civil Engineering, University of KwaZulu-Natal, Durban 4041, South Africa Received 23 March 2004 Revised 15 November 2004 Accepted 15 November 2004 ABSTRACT Fluctuations of wet and dry years have long been investigated in the climatology and hydrology literature. In this, the second of two papers investigating runs of consecutive dry years, the magnitude, also known as the intensity, and severity (length × magnitude) of dry runs are investigated. In the first paper the length of dry runs was investigated. Periods of consecutive dry years are associated with drought and the attendant physical and economic stresses that are placed on society. Run magnitudes of consecutive years equal to or below the median were analysed for 3863 precipitation and 1236 runoff stations from around the world. For both annual precipitation and runoff, run magnitude was found to be predominately related to interannual variability and to a lesser extent skewness. Run magnitude of annual runoff was observed to be greater than that for annual precipitation, due to annual runoff having a higher coefficient of variation than annual precipitation. Continental differences in run magnitude of annual runoff were observed and were consistent with continental differences in interannual variability reported previously. Annual run severity was also investigated and found to be independent of run length and strongly related to run magnitude. These findings differ from previously published work; this difference is primarily due to the methodology of comparing run metrics between stations (used in this paper) rather than at a station (previous research). The relationships between run magnitude, severity and interannual variability highlight the importance of adequately reproducing interannual variability within global climate models for future modelling of drought scenarios, as well as having economic implications for drought relief and management policy-making. Copyright  2005 Royal Meteorological Society. KEY WORDS: comparative hydrology; run magnitude; run severity; runs analysis; annual precipitation; annual runoff 1. INTRODUCTION Investigations into the nature of extended periods of consecutive dry years, which are associated with drought, have formed the basis of considerable research in the climatology and hydrology literature. Drought is a significant physical and economic phenomenon that can severely impact ecosystems and societies (Wilhite, 2000). Owing to its multifaceted nature, a generally accepted definition of drought remains elusive (Dracup et al., 1980a; Wilhite and Glantz, 1985; Frick et al., 1990; Bonacci, 1993; Wilhite, 2000; Heim, 2002; Keyantash and Dracup, 2002; Panu and Sharma, 2002; McMahon and Finlayson, 2003). This paper does not argue for any particular definition of drought, but simply uses runs analysis as a tool for investigating drought behaviour. Applying runs analysis, Yevjevich (1967) and Dracup et al. (1980a) defined drought severity as the sum of the negative deviations from a threshold (mean, median, etc.) for a given length of negative deviations. * Correspondence to: Murray C. Peel, Centre for Environmental Applied Hydrology, Department of Civil and Environmental Engineering, The University of Melbourne, Victoria 3010, Australia; e-mail: mpeel@unimelb.edu.au Copyright  2005 Royal Meteorological Society