IPCC Report.pdf - Adam Curry

Managing the Risks from Climate Extremes at the Local LevelChapter 5triangulate, obtaining better primary data to reduce uncertainty, anddeveloping tools for estimating the impact of Hurricane Andrew onFlorida using reconstruction scenarios. The lack of research on disasterloss estimates in developing countries creates problems of underreportedeconomic losses or overestimation of disaster losses depending onpolitical or other interests. This is a big research gap.5.5.2.2. Adaptation and Risk Management – Present and FutureStudies on the costs of local disaster risk management are scarce,fragmented, and conducted mostly in rural areas. One study estimated thebenefit/cost ratio of disaster management and preparedness programs inthe villages of Bihar and Andra Pradesh, India to be 3.76 and 13.38,respectively (Venton and Venton, 2004), suggesting higher benefits thancosts. Research undertaken by the Institute for Social and EnvironmentalTransition on a number of cases in India, Nepal, and Pakistandemonstrated that benefits exceed the costs for local interventions(Dixit et al., 2008; Moench and Risk to Resilience Study Team, 2008).For example, they note that return rates are particularly robust forlower-cost interventions (e.g., raising house plinths and fodder storageunits, community-based early warning, establishing community grain orseed banks, and local maintenance of key drainage points), whencompared to embankment infrastructure strategies that require capitalinvestment (Moench and Risk to Resilience Study Team, 2008). Thestudies demonstrated a sharp difference in the effectiveness of the twoapproaches, concluding that the embankments historically have not hadan economically satisfactory performance in that study area. In contrast,the benefit/cost ratio for the local-level strategies indicated economicefficiency over time and for all climate change scenarios (Dixit et al.,2008). In developed countries, there are cost differences in adaptationstrategies between urban and rural areas. For example, in Japan disasterdamage is several hundred times more costly in urban than in ruralareas, necessitating different disaster risk management strategiesdepending on the benefit to cost analysis (Kazama et al., 2009).Though disaster risk management and adaptation policies are closelylinked, few integrated cost analyses of risk management and adaptationare available at the local level. One example draws from recent studiesof the cost of city-scale adaptation. Rosenzweig et al. (2007, 2011)developed a sophisticated analytical response to a projected fall in wateravailability in New York. This frames adaptation assessment within astep-wise decision analysis by identifying and quantifying impact risksbefore identifying adaptation options that are then screened, evaluated,and finally implemented. Another series of studies used simplifiedcatastrophe risk assessment to calculate the direct costs of storm surgesunder scenarios of sea level rise coupled with an economic input-outputmodel for Copenhagen and Mumbai (Hallegatte et al., 2008a,b, 2011;Ranger et al., 2011). The output is an assessment of the direct and indirecteconomic impacts of storm surge under climate change includingproduction, job losses, reconstruction time, and the benefits of investmentin upgraded coastal defenses. Results show that the consideration ofadaptation is an important element in the economic assessment ofFAQ 5.3 | Is it possible to estimate the costof risk management and adaptationat the local scale?Studies on the costs of local disaster risk management arescarce, fragmented, and conducted mostly in rural areas. Mosteconomic data (e.g., input-output table, income data) areavailable at the national scale. Moreover, there is a clear lack ofresearch on disaster estimates in developing countries, whichpresents a big gap in need of further research. In developedcountries, there are cost differences in adaptation strategiesbetween urban and rural areas. The reliability of disastereconomic loss estimates is especially problematic at the locallevel due to factors associated with the global nature ofspatial coverage and resolution. In addition there is someambiguity on impact and adaptation costs that affect localleveleconomic analyses, such as the lack of consensus onphysical impacts of climate change and adaptive capacity andon the evaluation of non-market costs (e.g., biodiversity orcultural heritage), which creates some uncertainty about localimpact and adaptation costs.extreme disaster risks related to climate change (Hallegatte et al.,2011). Ranger et al. (2011) show that by improving the drainage systemin Mumbai, losses associated with a 1-in-100 year flood event could bereduced by as much as 70%. This means that the annual losses could bereduced in absolute terms compared with the current level, even withclimate change. Full insurance coverage of flooding could also cut theindirect cost by half. These analyses highlight the fact that adaptation toextreme events and climate change can focus on reducing the directlosses (e.g., through the upgrade of coastal defenses) or indirect lossesby making the economy more robust, utilizing insurance schemes, orenacting public policies to support small businesses after the disaster.5.5.2.3. Consistency and Reliability of Cost andLoss Estimations at the Local LevelThere are inconsistencies in disaster-related economic loss data at alllevels – local, national, global – that ultimately influence the accuracyof such estimates (Guha-Sapir and Below, 2002; Downton and Pielke Jr.,2005; Pielke Jr. et al., 2008). The reliability of disaster economic lossestimates is especially problematic at the local level. First, the spatialcoverage and resolution of databases are global in coverage, but onlyhave data that represent the entire country, not sub-units within it suchas provinces, states, or counties. Second, thresholds for inclusion, whereonly large economically significant disasters are included, bias the datatoward singular events with large losses, rather than multiple, smallerevents with fewer losses. Third, what gets counted varies betweendatabases (e.g., insured versus uninsured losses; direct versus indirect; Gallet al., 2009). Moreover, disaster loss estimates have various purposes318