« Climate Change »

Climate Change and Reinsurance, Part 2.5

My recent posts on climate change and reinsurance (Part 1 and Part 2) led to a wide-ranging and fruitful discussion with a number of colleagues in the U.S. and Europe. (Thanks all!)

To bring you up to date, Part 1 made the case for a clear conflict of interest when reinsurers attribute or project increasing disasters because of climate change. Part 2 sough to evaluate the merits of such claims, which to be fair are made by many people well beyond just some in the reinsurance industry. The central question is, when looking to the past, to what degree is climate change responsible for the growth in disasters and disaster costs up to the present? I asserted in Part 2 that the answer is "not at all."

In this post, Part 2.5, I'd like to expand on the information presented in Part 2 drawing on additional information and analyses drawing from my recent discussions with colleagues. Specifically, there seems to be a strong consensus within the climate impacts community that the trends of increasing damage related to storms (whether tropical, extra-tropical, thunder, hail or other extreme weather) is completely the result of trends in societal impacts. Questions were raised about trends in impacts related to floods, heat waves and drought. Let's consider each in turn.

Floods

I noted in Part II that while the IPCC identified some regional trends in the incidence of what it calls "extreme" precipitation, it did not find similar trends in extreme streamflow (i.e., floods). After talking with colleagues and reading up on some of the more recent literature, it seems clear that there is no evidence for global trends in floods, although there may be some regional trends. There is considerable mixed evidence depending upon definitions of "flood" and the time period selected. It will be interesting to see how the next IPCC comes out on this. It is abundantly clear that flood damage is increasing around the world.

While it is conceivable that trends in precipitation and streamflow are responsible for some part of the growing impacts of floods, I await data and analyses making such a case quantitatively and globally. It is both logical and shown in our own research that flood damage tends to increase with aggregate precipitation (i.e., imagine a graph of such a relationship; it necessarily starts at the graph's origin - no precip, no flood damage), but exactly how much and how significant are open questions. My hypotheses, resulting from my work with Mary Downton on this paper .

Pielke, Jr., R.A., and M.W. Downton, 2000: Precipitation and Damaging Floods: Trends in the United States, 1932-97. Journal of Climate, 13(20), 3625-3637,

. are that (a) trends in damage will be most closely connected to trends in precipitation and streamflow at the regional and local level, and (b) attempts to aggregate regional data to national or global levels will necessarily result in a much smaller connection because of the contextual nature of flood impacts. Even if a signal could be found in the noise, it would be tiny when compared with the effects of societal vulnerability on flood damage.

Heat Waves

I am not an expert on heat waves or their impacts, so I called up someone who is, Larry Kalkstein from the University of Delaware. Here is what I asked him:

1. Is there evidence for an increase (globally) in heat waves? 2. Is there evidence for a trend of increasing societal impacts related to heat waves? 3. Is there evidence for growing societal vulnerability to heat waves?

His answers:

1. No. There is no evidence of an increase in the number of heat waves, in the U.S. or globally, but there are regional variations (e.g., Great Britain may have seen an increase). (Note that what he reported to me is completely consistent with the IPCC TAR.)

2. Yes and no. In the United States heat mortality has decreased since the 1960s by about 20-25%, which is less than what others have reported in the literature. Europe, by contrast has seen an increase in heat related mortality over the past decade. But there are large data gaps around the world.

3. Yes. Vulnerability has increased largely because of two factors -- demographics and the costs of energy. There are growing numbers of elderly people, particularly in Europe, and the elderly are more vulnerable to heat. Air conditioning is ever present in the United States, so its availability is less of a problem than is the cost of running the units. Many poor people choose not to run their air conditioners because of the cost, leading to an increase in their vulnerability. The urban structure of Europe makes its populations much more vulnerable to heat related impacts.

My conclusions: There is a lot that we don't know about global trends in the impacts of heat waves. Any trends that exist in the incidence of heat waves would seem to be regional in nature. If so, then we can narrow our focus on the role of climate changes as a factor in trends in impacts to those places that have seen (a) an increase in heat waves, and (b) an increase in impacts. This area would seem to be far smaller than global. Even so, better data will be useful from around the world. But clearly there is no evidence that would allow for a connection to be made between trends in heat waves and their societal impacts. I'll stick with my hypothesis that any trends in impacts are the result of increasing societal vulnerability and ask for falsification.

Drought

What we know about the societal impacts of heat waves seems to similar to what we know about the societal impacts drought. Drought has tremendous societal impacts, yet there seems to be no systematic collection of data on those impacts and how they have changed over time and space. For example, Donald Wilhite's two volumes on drought provide no trend data (Drought: A Global Assessment, 2000. Natural Hazards and Disasters Series, Routledge Publishers, London).

A recent NCAR study argues that one measure of drought shows a significant increase in areal extent over the past 20 years. But a connection to societal impacts remains to be made.

My conclusions on drought echo those on heat waves: There is a lot that we don't know about global trends in the impacts of droughts. There may be robust trends in the incidence of drought that are global in nature. Better data on impacts will be useful from around the world. But clearly there is no evidence that would allow for a connection to be made between trends in drought and their societal impacts. I'll stick with my hypothesis that any trends in impacts are the result of increasing societal vulnerability and ask for falsification.

So the bottom lines:

1. Anyone making assertions that changes in climate (whether human caused or not) are responsible for any part of the global trend of increasing disaster losses had better provide some new scientific evidence to back up such claims. Future research may tell a different story, but my reading of the current state of science is that, today, such claims are groundless.

2. This series should be viewed as an intellectual challenge to the IPCC WG2 and the climate impacts community. I propose that we in this community first begin with a hypothesis, namely, "All trends observed in recent decades indicating growing damage related to weather and climate can be explained through the growth of societal vulnerability to those trends." Then, the second step is to conduct research that seeks to falsify this hypothesis.

It is important to reiterate that the discussion thus far has been retrospective, focused only on the attribution of factors responsible for the global trend in disasters. I will next turn to Part 3 which will explore the question, in the future to what extent should we expect climate changes to be responsible for increasing disasters and disaster costs?