Assessing resilience to drought: defi ning drought and reviewing trends in the Horn of Africa
As a focus area of the Technical Consortium is to assess the resilience of populations to drought, one has to have a clear definition of drought. Climatologists identify three different types of drought: (1) meteorological drought, (2) hydrological drought and (3) agricultural drought (a fourth category of socio-economic drought is sometimes distinguished, but not relevant to this paper). The US National Oceanic and Atmospheric Administration (NOAA) provides succinct descriptions of each of these. Meteorological drought is usually defined as below normal expected rainfall over a defined period based on long-term observed averages. This is the easiest to observe and is, at least in the short-term, wholly exogenous to human systems. According to the NOAA National Weather Service4, hydrological drought usually occurs following periods of extended precipitation shortfalls that impact water supply (i.e. streamflow, reservoir and lake levels, ground water), potentially resulting in significant societal impacts. Because regions are interconnected by hydrologic systems, the impact of meteorological drought may extend well beyond the borders of the precipitation-deficient area. Agricultural drought links various characteristics of meteorological (or hydrological) drought to agricultural impacts, focusing on precipitation shortages, soil water deficits, reduced ground water or reservoir levels needed for irrigation, and so forth4. Most of the datasets and scholarship on Africa focus on rainfall, since irrigation is very limited and most farmers rely on rain-fed agriculture. How one operationalizes the definition of drought can have major effects on the findings for what areas are found to have historically been drought-prone in the past. Many use variations of the Standardized Precipitation Index, although it is unclear how well this captures common conceptions of drought in Africa. The specific operationalization of drought very much depends on the reference period for normal rainfall, the length of the period one is studying to assess deviations from normal rainfall, whether one is using satellite data or rain gauge data, etc.5. For pastoralists and other communities that rely on groundwater wells for livestock, other metrics of water availability might be important. At the same time, when people reference drought in this part of the world, they may also be conflating some measure of negative rainfall anomalies with chronic water scarcity in arid lands, which is a wholly different concept than drought. If, for example, drought is defined as below normal rainfall, it becomes meaningless to discuss drought in areas with very little normal rainfall based on long-term averages. To address these challenges of defining drought, the approach discussed in this paper incorporated two measures to account for both rainfall anomalies and chronic water scarcity.