Recently, climate experts and water planners have spent significant resources to understand if extreme events are increasing (or decreasing) and how to plan for the consequences to urban populations and infrastructure. One piece of this puzzle is computing the likelihood of an event, which, depending on the location and degree of human impact on the watershed, could mean a historic (or future) trend influences how likely a flood, drought, etc, is to occur. Trend analysis is a rigorous process, as persistence in a time series is often difficult to separate from an actual trend (Cohn & Lins, 2005). Nevertheless, suppose evidence suggests an increasing trend in the annual maximum flow series for example, sizing infrastructure and continuing urban development according to the existing 100-year floodplain would not protect against the true 100-year event, as these maps and boundaries need to be updated.
One question that emerges from this discussion is how insurance companies and others that calculate damage costs and bear the financial burden of loss actually compute the probabilities of extreme events? This topic is something we’ve heard about in popular media, but is not explored with near as much attention in the hydrology literature. Literally in the most well-known water resource planning/management/design journals, there is one article, Chen et al., 2013, that outlines how probabilities of events apply to catastrophe (CAT) bonds and impact planning for extreme floods (at least in English, there is one other paper published in Chinese! – Qiu, L. et al., 2008). Though there is a much richer literature on applications for earthquakes and examples from agriculture, these articles are still sparse compared with the plethora of research in the economics and insurance sectors, which detail the complex pricing distributions CAT bonds can follow. I am not claiming that we are ignoring CAT bonds, in fact many researchers in the risk and climate literature have written on their use and application to extreme events like Katrina and now Sandy, as well as popular media like pieces from Michael Lewis (http://www.nytimes.com/2007/08/26/magazine/26neworleans-t.html?pagewanted=all&_r=0). But instead I am concerned that there is an information gap occurring, that we are missing a learning opportunity between understanding how insurance/re-insurance companies compute the likelihood of an extreme event, and how we in hydro-statistics are planning/predicting these. With any hope there is the chance for these fields to learn from one another, and at least make sure we’re on the same page – communicating – when it comes to making claims about trends in extreme events and their potential impacts on society.
Chen, J., Liu, G., Yang, L., Shao, Q., & Wang, H. (2013). Pricing and Simulation for Extreme Flood Catastrophe Bonds. Water resources management, 27(10), 3713-3725.
Cohn, T. A., & Lins, H. F. (2005). Nature’s style: Naturally trendy. Geophysical Research Letters, 32(23).