 |
Global Assessment Report on Disaster Risk Reduction 2011
Revealing Risk, Redefining Development |
  |
3.5 From drought hazard to drought risk
Given that drought impacts are not systematically recorded and the data constraints for modelling drought hazard, it is still not possible to develop global drought risk models. Building such models at all scales is important to increasing the visibility of the risk and for building political and economic imperatives for drought risk management.
In the same way that meteorological drought is not synonymous with drought hazard, agricultural and hydrological drought hazard are not synonymous with risk. As with other hazards, the translation of drought into risk depends on factors related to vulnerability and exposure. Developing models for drought similar to those already used to analyse risk trends for tropical cyclones and floods (see Chapter 2) is still not possible due to lack of sufficient and suitable data, and previous attempts to model global drought risk (see Box 3.4) produced unsatisfactory results. Box 3.4 Modelling global drought risk
The mortality drought risk index proposed by UNDP (UNDP, 2004 UNDP (United Nations Development Programme).. 2004. Reducing disaster risk: A challenge for development: Geneva, Switzerland: United Nations Development Programme, Bureau for Crisis Prevention and Recovery. ) was unsuccessful because most droughts do not produce fatalities, and most internationally recorded drought mortality is concentrated in countries experiencing conflict or political crisis. Only weak correlations were found between the population exposed to meteorological drought and the mortality attributed to drought (UNDP, 2004. UNDP (United Nations Development Programme).. 2004. Reducing disaster risk: A challenge for development: Geneva, Switzerland: United Nations Development Programme, Bureau for Crisis Prevention and Recovery. ). Drought impacts on human development could provide more suitable criteria than mortality for calculating human risk. However, while such impacts are sometimes recorded in certain locations
(de la Fuente and Dercon, 2008. de la Fuente, A. and Dercon, S. 2008. Disasters, growth and poverty in Africa: Revisiting the microeconomic evidence. Background paper prepared for the 2009 Global Assessment Report on Disaster Risk Reduction. Geneva, Switzerland: UNISDR. ), systematic national data is not available to calibrate a global risk model.. A World Bank study (Dilley et al., 2005 Dilley, M., Chen, R., Deichmann, W., Lerner-Lam, A.L. and Arnold, M. 2005. Natural disaster hotspots. Washington, DC, USA: The World Bank.. ) was more successful in that it produced global risk maps for both mortality and economic loss risk. Risk was calculated as a function of the exposure to meteorological drought of population density and national agricultural GDP, with a proxy indicator of vulnerability calibrated using recorded mortality and economic losses for each geographic and income region. The accuracy of the results is questionable, however, given that meteorological drought is not a good representation of hazard and, as described above, mortality is not an adequate metric to model impacts on humans.
Initiatives such as the National Drought Monitor in the United States of America, FEWS Net, AGRHYMET, and the Sahara and Sahel Observatory (OSS) in Africa, the International Water Management Institute’s (IWMI) PODIUM and FAO’s AquaCrop models, and studies by the World Bank in India (Box 3.5), show how drought risk can be modelled in specific contexts when data is available. Systematically accounting for drought losses and impacts and building credible drought risk models at all scales, from local to global, is important to increasing the visibility of drought risk and building political and economic imperatives for its reduction. Box 3.5 Modelling agricultural drought risk
A study by the World Bank (Lvovsky et al., 2006 Lvovsky, K., Mahul, O., Makino, Y., Noble, I., Krovvidi, A., Francis, S. and Priya, S. 2006. Overcoming drought: Adaptation strategies for Andhra Pradesh, India. Washington DC, USA: The World Bank. ) quantified long-term agricultural and macro-economic impacts of droughts in Andhra Pradesh, India, using catastrophe modelling techniques with a range of drought risk management strategies. By analysing meteorological and agricultural data over 30 years, the effect of mild, moderate and severe droughts was measured on five different crops (rice, groundnut, sunflower, maize and sorghum) in the eight most drought-prone districts of Andhra Pradesh, including average annual and probable maximum losses.. First, the frequency and severity of meteorological drought at different locations was modelled using historic data and a stochastic weather generator (WXGEN) simulating 500 years of weather. Modelled droughts were classified using a seasonal (June–December) SPI computation and validated against historical data. Vulnerability and exposure were analysed using crop-yield and planting-area models to quantify damages to each crop based on the intensity and duration of droughts. Drought impacts on livestock production were also tested but results were inconclusive. The crop-yield model incorporated 47 parameters calibrated to the crops and environmental conditions in each district. The planting-area model was used to capture rainfall variability, including both irrigated and rain-fed cultivation. Average yield and average annual losses for each crop for the 500-year time series were then computed, and the effect of drought intensity and duration on each crop converted to monetary losses based on market prices. Compared to simulated ‘normal’ years, analysis revealed that production losses exceeded 5 percent every 3 years, 10 percent every 5 years, 15 percent every 10 years and 25 percent every 25 years. Individual farmers and especially small farmers may experience much greater losses depending on their crop mix and the severity of drought in their particular location. (Source: Lvovsky et al., 2006 Lvovsky, K., Mahul, O., Makino, Y., Noble, I., Krovvidi, A., Francis, S. and Priya, S. 2006. Overcoming drought: Adaptation strategies for Andhra Pradesh, India. Washington DC, USA: The World Bank. ). As this chapter has shown, drought risk is at least in part socially constructed, and characterized by numerous feedback loops between the different drivers. For example, the lack of systematic recording of drought losses and impacts, particularly those affecting poor and vulnerable rural households, contributes to its reduced political and economic visibility, reflected in only weak imperatives to address underlying risk drivers and strengthen risk governance. Policies to promote economic and urban development in water-scarce areas may transfer drought risk to smallholder farmers. Drought-relief programmes that compensate for short-term impacts may increase dependence on relief and increase vulnerability in areas that may become more drought-prone with climate change. International efforts to develop and apply standards for drought identification and monitoring are an important starting point to address drought risk. They need, however, to go alongside the development of mechanisms to systematically account for drought losses and impacts, and that comprehensively assess and estimate drought risks as a crucial next step to raising the profile of drought risk. Forecasting, early warning and compensatory measures such as insurance are critical elements of drought risk management. However, to address the underlying drivers of drought risk, countries will have to strengthen and reorient other risk governance capacities, particularly those related to development planning and land and water management. There are often powerful political disincentives against addressing issues such as water rights and land use, but with ever-increasing drought impacts and losses, the imperative to seriously manage drought risk may soon outweigh these disincentives. |
| |
GAR 2011 Contributing Papers