Decades of research investment have produced significant advances in the predictive skill of natural hazard forecasts such as tropical cyclone track and intensity and flash flooding. Such advances have arisen from enhanced understanding of the Earth system, advances in Earth observing systems, and the growth of computational power.
Yet catastrophic infrastructure failures are becoming more frequent. The world is entering a new era of natural disasters that are causing more damage than in the past. Economic losses attributed to natural disasters have ballooned from $75.5 billion in the 1960s to $659.9 billion in the 1990s, a compound annual growth rate of 8 percent. Total worldwide insured losses are now dominated by natural catastrophes.
The principal driver of these increasing losses is increasing exposure, meaning that natural disasters are anything but “natural.” The population of Florida, for example, rose 690 percent between 1950 and 2010, putting many more people at risk given the state’s extensive coastal exposure and subtropical location. Moreover, the potential for the hazards themselves to become more damaging in the future with climate change will compound the effects of increasing exposure.
The speakers in this session laid out frontiers in forecasting natural disasters and looked into a future of useful forecasts, effective messaging, and avoidance of catastrophic failure. The first speaker, Ning Lin (Princeton University) outlined the future of probabilistic, quantitative natural hazard risk assessment in support of risk management. Next, Jeffrey Czajowski (University of Pennsylvania) outlined the critical importance of accounting for human behavioral biases to move from