« Climate Change »

Here We Go Again: Cherry Picking in the IPCC WGII Full Report on Disaster Losses

The IPCC WGII full report is available (hat tip: ClimateScienceWatch). I have had a look at what they say about disaster losses, and unfortunately, the IPCC WG II commits the exact same cherry picking error as did the Stern report.

Here is what IPCC says about catastrophe losses (Chapter 1, pp. 50-51):

Global losses reveal rapidly rising costs due to extreme weather-related events since the 1970s. One study has found that while the dominant signal remains that of the significant increases in the values of exposure at risk, once losses are normalised for exposure, there still remains an underlying rising trend.

The one study? Muir-Wood et al. 2006 that was prepared as the basis for our workshop last year with Munich re on Disaster Losses and Climate Change. Here is what we said when the Stern Report cherry picked this same information:

The source is a paper prepared by Robert Muir-Wood and colleagues as input to our workshop last May on disasters and climate change. Muir-Wood et al. do report the 2% trend since 1970. What Stern Report does not say is that Muir-Wood et al. find no trend 1950-2005 and Muir-Wood et al. acknowledge that their work shows a very strong influence of 2004 and 2005 hurricane seasons in the United States. Muir-Wood et al. are therefore very cautious and responsible about their analysis. Presumably this is one reason why at the workshop Robert Muir-Wood signed on to our consensus statements, which said the following:

Because of issues related to data quality, the stochastic nature of extreme event impacts, length of time series, and various societal factors present in the disaster loss record, it is still not possible to determine the portion of the increase in damages that might be attributed to climate change due to GHG emissions . . . In the near future the quantitative link (attribution) of trends in storm and flood losses to climate changes related to GHG emissions is unlikely to be answered unequivocally.

The Stern Report’s selective fishing out of a convenient statement from one of the background papers prepared for our workshop is a classic example of cherry picking a result from a diversity of perspectives, rather than focusing on the consensus of the entire spectrum of experts that participated in our meeting. The Stern Report even cherry picks from within the Muir-Wood et al. paper.

The full discussion by the IPCC WG II has a bit more nuance, but it is clear that they are reaching for whatever they can to support a conclusion that simply is not backed up in the broader literature. Can anyone point to any other area in the IPCC where one non-peer-reviewed study is used to overturn the robust conclusions of an entire literature? Here is the full discussion:

Economic losses attributed to natural disasters have increased from US$75.5 billion in the 1960s to US$659.9 billion in the 1990s (a compound annual growth rate of 8%) (United Nations Development Programme 2004). Private sector data on insurance costs also shows rising insured losses over a similar period (Munich Re Group 2005; Swiss Reinsurance Company 2005). The dominant signal is of significant increase in the values of exposure (Pielke and Hoppe 2006).

However, as has been widely acknowledged, failing to adjust for time-variant economic factors yields loss amounts that are not directly comparable and a pronounced upward trend through time for purely economic reasons. A previous normalization of losses, undertaken for United States hurricanes by Pielke and Landsea (1998) and US floods (Pielke et al., 2002) included normalizing the economic losses for changes in wealth and population to express losses in constant dollars. These previous national US assessments, as well as those for normalized Cuban hurricane losses (Pielke et al., 2003), did not show any significant upward trend in losses over time, but this was before the remarkable hurricane losses of 2004 and 2005.

A ‘global’ catalogue of catastrophe losses was constructed (Muir Wood et al., 2006) normalized to account for changes that have resulted from variations in wealth and the numbers and values of properties located in the path of the catastrophes, using the method of Pielke and Landsea (1999). The global survey was considered largely comprehensive from 1970–2005 for countries and regions (Australia, Canada, Europe, Japan, South Korea, US, Caribbean, Central America, China, India and the Philippines) that had centralized catastrophe loss information and included a broad range of peril types: tropical cyclone, extratropical cyclone, thunderstorm, hailstorm, wildfire and flood, and that spanned high and low latitude areas.

Once the data were normalized a small statistically significant trend was found for an increase in annual catastrophe loss since 1970 of 2% per year (see Fig. SM1.1). However, for a number of regions, such as Australia and India, normalized losses show a statistically significant reduction since 1970. The significance of the upward trend is influenced by the losses in the US and Caribbean in 2004 and 2005 and arguably biased by the relative wealth of the US, in particular relative to India.

More on the figure that they reference in the next post . . .