Understanding trends and projections of disaster losses and climate change: is vulnerability the missing link?

The recent IPCC-SREX report demonstrated for the first time comprehensively that anthropogenic climate change is modifying weather and climate extremes. The report also documents, what has been long known, that losses from natural disasters, including those linked to weather, have increased strongly over the last decades. Responding to the debate regarding a contribution of anthropogenic climate change to the increased burden from weather-related disasters, the IPCC-SREX finds that such a link cannot be made today, and identifies the key driver behind increases in losses as exposure changes in terms of rising population and capital at risk. Yet, in the presence of many uncertainties and omissions involved in studying trends in losses, the authors of the IPCC report did not exclude a role for climate change. In particular, one key uncertainty identified has been the incomplete consideration of economic vulnerability to natural hazards, defined as the propensity to incur losses in a hazardous event. Focussing on the role of vulnerability in determining today’s and future disaster loss risk, we critically review the literature on loss trends and projections, and provide context by way of a modeling case study of observed and projected losses from riverine flooding in Bangladesh. We find that research has almost exclusively focused on normalizing losses for changes in exposure, yet not for vulnerability, which appears a major gap given the dynamic nature of vulnerability, and documented evidence regarding decreases in vulnerability in many regions. One such region is South Asia, and of particular interest to us is Bangladesh, a country heavily at-risk, but also with substantial expertise regarding risk management, where we are able to show that economic vulnerability has been substantially reduced over the last decades. In order to understand future flood risk in Bangladesh, we project risk based on past reductions in vulnerability and compare it to a case where vulnerability is not considered explicitly and kept static. In the dynamic scenario, risk would still increase in absolute terms, yet at much smaller increments compared to a static vulnerability case. Thus, a key finding of our analysis is that, absent dynamic quantifications of vulnerability, studies on future losses under climatic change may overestimate future losses. Furthermore, the analysis also suggests that there are substantial benefits to gain by supporting vulnerability-reducing measures in many regions. Finally, we emphasize the need for further taking a risk-based perspective on modelling climate impacts in order to provide robust information on the costs and impacts from extremes in a changing climate. Copyright Springer Science+Business Media Dordrecht 2015