Welcome to the new version of European Tribune. It's just a new layout, so everything should work as before - please report bugs here.

Catastrophe Cacophony

by Nomad Fri Nov 19th, 2010 at 05:24:01 AM EST

In a diary a few weeks ago I indicated that the exaggeration of hurricane science, in the wake of the active 2005-2006 hurricane seasons and hurricane Katrina, didn't do harm to risk assessment and (re)insurance companies. I also indicated that science behind the 5-year projections of risk assessment companies have been shown to be flawed (and that, unsurprisingly, 5-year projections are horribly failing).

While commenting in said diary, I hit upon a feature that drew my interest: how much the world's largest reinsurance company just doesn't care about science, and continues to propagate nonsense, despite knowing better. You know, claims like these:

Munich Re - Two months to Cancún climate summit / Large number of weather extremes as strong indication of climate change

Munich Re's natural catastrophe database, the most comprehensive of its kind in the world, shows a marked increase in the number of weather-related events. For instance, globally there has been a more than threefold increase in loss-related floods since 1980 and more than double the number of windstorm natural catastrophes, with particularly heavy losses as a result of Atlantic hurricanes.

The rise in natural catastrophe losses is primarily due to socio-economic factors. In many countries, populations are rising, and more and more people moving into exposed areas. At the same time, greater prosperity is leading to higher property values. Nevertheless, it would seem that the only plausible explanation for the rise in weather-related catastrophes is climate change.

Except that increasing flood, windstorm and hurricane losses have not been attributed to global warming / climate change so far.


Meet Munich Re: the world's leading reinsurance company, based in Munich, Germany, and serial offender in communicating doom. This was the snippet that grabbed my attention:

ABC The Drum - Bookies buying into climate change race

Mr Rauch's original expertise was in earthquakes, but following devastating losses for the insurance industry from a series of storms in the European winter of 1990, his area of responsibility shifted to the analysis and modelling of meteorological risks.

And the verdict?

"Climate change, we believe, is a fact."

Why?

His pockets are already hurting.

"Based on our own loss experience, climate change we believe is a fact. It triggers natural disasters, atmospheric natural disasters, and the number of these natural disasters worldwide has more than doubled since the 1980s, driven by atmospheric perils, not by earthquakes or volcanic eruptions," Mr Rauch said.

"If we look at the sheer number of losses from natural catastrophes worldwide since the 1980s, more than $US1,600 billion in losses have occured. Most of them were actually weather-related, not earthquakes, not geophysical events."

Or here. Or here.

First of, I do applaud companies speaking out for initiatives to curb anthropogenic greenhouse gas emissions, which Munich Re and other important companies in the reinsurance branch (like the competing Swiss Re) do prominently. Unfortunately, Munich Re seems to frequently suffer from cognitive dissonance between their PR department and their own considerably qualified science group. Let me explain what I mean with that.

Starting with tropical cyclones, that proud icon of climate change, first a backtrack to 2005, when two prominent publications gained wider traction in the public domain. Both articles were released briefly after hurricane Katrina had struck New Orleans, and many people on the world were wondering if climate change had anything to do with that catastrophe.

The Webster et al. (2005) article argued that there were appearing more tropical cyclones in the Atlantic Basin, in short, the frequency was increasing (but the same article also argued against a global increase). On top of that, the Nature paper by Emanuel that had been released shortly before, argued that particularly for the Atlantic Basin, hurricanes were becoming increasingly more violent, that is, an increasing intensity, beyond the scale of natural variation. All of that sounded indeed disconcerting, but the hurricane scientists were heavily divided over these announced results, particularly by how these scientists had gotten their results. Nevertheless, both papers became cornerstones for the conclusions of the IPCC 2007 report on tropical cyclones, summarising them as following:

While other basins do not show overall increases in activity,
observations based on satellite observations of intensity (which start in the 1970s) suggest a shift in the proportion of tropical cyclones that reached the higher intensity (CAT 4 and CAT 5) from close to 20% of the total in the 1970s rising to 35% since the 1990s (Webster et al., 2005). Although challenged by some climatologists based on arguments of observational consistency, as quoted fromTrenberth et al., (2007) Section 3.8.3) "the trends found by Emanuel (2005) and Webster et al. (2005) appear to be robust in strong association with higher SSTs".

(source)

While being perfect fodder for Munich Re's press releases, the papers exacerbated divisions in the hurricane science community, with a classic divide running between modellers and data scrutinizers. Yet, unlike the Climate Wars over temperature proxies, experts kept talking and discussing with each other. And the result? An authoritative, key paper in Nature Geosciences published early this year, written by hurricane experts encompassing several camps, including Emanuel. So what does it read?

Tropical cyclones and climate change : Abstract : Nature Geoscience

Large amplitude fluctuations in the frequency and intensity of tropical cyclones greatly complicate both the detection of long-term trends and their attribution to rising levels of atmospheric greenhouse gases. Trend detection is further impeded by substantial limitations in the availability and quality of global historical records of tropical cyclones. Therefore, it remains uncertain whether past changes in tropical cyclone activity have exceeded the variability expected from natural causes.

Emphasis mine. That single sentence negates Webster et al. for the time being. How does it work out for Emanuel's paper?

The most significant cyclone intensity increases are found for the Atlantic Ocean basin, but the relative contributions to this increase from multidecadal variability (whether internal or aerosol forced) versus greenhouse-forced warming cannot yet be confidently determined.

And with one stroke, the IPCC chapter on tropical cyclones needs to be rewritten.

But wait! If we actually can't observe a global warming trend in tropical cyclones, how do we attribute a global warming signal to damage losses from Atlantic hurricanes - as Munich Re blithely announces? Well. We can't. And even before the recent Nature paper, damage losses from hurricanes in the Atlantic Basin had been analysed at least twice, with identical results:

Pielke et al. (2008):


[T]here is no remaining trend of increasing absolute damage in the data set, which follows the lack of trends in landfall frequency or intensity observed over the twentieth century.

(source)

Schmidt et al. (2009):


There is no evidence yet of any trend in tropical cyclone losses that can be attributed directly to anthropogenic climate change.

(source)

That last paper was actually written by scientists working for Munich Re... That's what I'd call cognitive dissonance - although the authors do stress that there is a higher chance climate change is playing a part in tropical cyclone losses. Except: it can't be shown.

Right, increasing hurricane damage for the USA can't be contributed just yet to climate change. Can happen. But USA is only a country. Doesn't say anything about the rest of the world! And I didn't even mention floods so far.

The analysis of twenty-two disaster loss studies shows that economic losses from various weather related natural hazards, such as storms, tropical cyclones, floods, and small-scale weather events such as wildfires and hailstorms, have increased around the globe. The 10 studies show no trends in losses, corrected for changes (increases) in population and capital at risk, that could be attributed to anthropogenic climate change. Therefore it can be concluded that anthropogenic climate change so far has not had a significant impact on losses from natural disasters.

(source)

You know, I could next point everyone to this study on flood losses in Europe showing no correlation between increased flood damage and climate change, based on data amongst others from Munich Re but I think the point I wanted to make is clear by now.

All listed studies are prominent, solid science publications, all available in the public domain. I will post a diary the moment I find studies that conclusively show a linkage between climate change and catastrophe losses. Promise. Because, in all fairness, that there will be increasing (normalised) catastrophe losses with increasing global temperatures does look guaranteed - except that this is a long term prediction. But it's simply not the reality of today: in terms of major catastrophes losses, a climate change signal has not be attributed so far by scientific research. The subjects are complex, many factors contribute, the variability in the data has been too large so far to determine a trend.

Yet, incredibly convenient, unsubstantiated doom is sold by the world's leading company selling reinsurance. Whom would you believe?

Display:
Of course all of this is just personal procrastination to prevent me from writing about the pending failure that Cancun most likely will be.
by Nomad on Thu Nov 18th, 2010 at 09:03:44 AM EST
European Tribune - Catastrophe Cacophony
That's what I'd call cognitive dissonance

Except, of course, there's another harmonic involved, which puts everything into profit-making tune.

by afew (afew(a in a circle)eurotrib_dot_com) on Thu Nov 18th, 2010 at 11:33:27 AM EST
I was reminded in the last diary, several times, to not be as silly to suggest such a thing.

So I won't, and just stick to pointing out BS.

But I don't think you are silly.

by Nomad on Thu Nov 18th, 2010 at 12:56:14 PM EST
[ Parent ]
Are we to assume, for the purposes of science, that Munich Re is not in the profit optimisation business, that their PR function is not primarily tasked with supporting that objective, and that the science they fund is not primarily geared to supporting that objective even if it is objective within it's own terms of reference?  (it is, for instance, important that Munich Re can accurately assess relative risks when writing insurance and managing risk exposure).

Is it not the job of PR functions to selective promote favourable data?  Is it not reasonable to assume, a priori, that that is what Munich Re does - even if we do not have evidence of specific such behaviour by specific management officials at specific times?

The argument that business actionable data has a lower standard of proof that scientific publications  can also be turned on its head: The analysis of how businesses operate does not require scientific standards of proof for every specific action they engage in.  Normal economic models/theories for business behaviour can be applied a priori in the absence of specific evidence to the contrary.

Index of Frank's Diaries

by Frank Schnittger (mail Frankschnittger at hot male dotty communists) on Fri Nov 19th, 2010 at 03:56:44 AM EST
[ Parent ]
decision making is, I think we can agree, quite different from that required for scientific decision-making.

Nothing in this diary indicates that you understand this point, Nomad.

To be clear : scientific studies with sparse data and short time series are going to have difficulty drawing a trend line with a high confidence interval, and are therefore deontologically obliged to declare "no significant correlation".

I read the short paper on European disasters you linked above (Normalised flood losses in Europe: 1970-2006). The graphs on the fifth page offer prima facie evidence for an increasing number of major flood events, and increasing damage, over the period. But the authors say :

Figure 3 shows the 5-year moving average of normalised losses. This figure does not reveal a
clear trend over time. Rather, a sequence of alternating periods with losses below and over the annual average can be observed. This could be attributed to natural variability of extreme floods.

Well, it could. Perhaps we should wait another thirty years so that the trend line will be statistically significant with a 95% confidence interval. (The inclusion of the past five years might be edifying. Or not. But I suspect it would.)

The same paper quotes a number of studies which demonstrate, without question, increased winter rainfall overall, and higher rainfall per rainy day, over the examined period (as well as a warming effect which is bigger than the global average). It is not unimagineable that these increases have not provoked any increase in major flooding; but it is powerfully counter-intuitive.

Two take-away points :

a) absence of proof is not proof of absence. Your rhetoric would lead one to believe that science has demonstrated an absence of an increase of insureable risks. I'm sure none of the authors linked would agree with that.

b) reinsurers do not require evidence that would stand up in a court of law, or in a scientific publication, before making projections or fixing premiums. I suspect they would do very little business if they did.

None of this argues against the possibility (which appears to be your subtext) that reinsurers are scamming us with inflated premiums. Perhaps they are. But the evidence you present is unpersuasive in that respect.

It is rightly acknowledged that people of faith have no monopoly of virtue - Queen Elizabeth II

by eurogreen on Thu Nov 18th, 2010 at 03:33:10 PM EST
Sorry to do this in JakeS style but I see little other option here.

scientific studies with sparse data and short time series are going to have difficulty drawing a trend line with a high confidence interval, and are therefore deontologically obliged to declare "no significant correlation".

Not-withstanding deontologically, agreed entirely. I don't think ethics are of a concern here.

The same paper quotes a number of studies which demonstrate, without question, increased winter rainfall overall, and higher rainfall per rainy day, over the examined period (as well as a warming effect which is bigger than the global average). It is not unimagineable that these increases have not provoked any increase in major flooding; but it is powerfully counter-intuitive.

Something being "powerfully counter-intuitive" has led to some poor scientific thinking before.

Your rhetoric would lead one to believe that science has demonstrated an absence of an increase of insureable risks. I'm sure none of the authors linked would agree with that.

Except that's not what the science authors write, nor what I write. It seems that's what you presume. I'll repeat my key-point: no increase for normalized hurricane, storm or flood damages could yet be attributed to climate change.

In case that was too subtle, let me spell that out: The inability to attribute a trend does not equate to argumentation of an absence of trend. And actually, I explicitly argued from the absence of proof is not proof of absence POV.

European Tribune - Catastrophe Cacophony

Except that increasing flood, windstorm and hurricane losses have not been attributed to global warming / climate change so far.

increasing hurricane damage for the USA can't be contributed just yet
 
has not be attributed so far

Emphasis mine.

reinsurers do not require evidence that would stand up in a court of law, or in a scientific publication, before making projections or fixing premiums. I suspect they would do very little business if they did.

Nice rhetorical trick. Of course I am not talking about projections or fixing premiums, you are. I am talking about claims released in the public domain and press by a company proclaiming the presence of a trend, without providing any evidence to show for.

That is a practice just as acceptable as, let's say, a pharmaceutical company selling anti-cancer drugs, announcing by press release a global increase in liver cancer rates for 60+ years old males, and then not providing the evidence.

None of this argues against the possibility (which appears to be your subtext) that reinsurers are scamming us with inflated premiums. Perhaps they are. But the evidence you present is unpersuasive in that respect.

Nice rhetorical trick again - point me the specific sentence where I argue that "reinsurers are scamming us with inflated premiums". It is no wonder I am unpersuasive in that respect - I am not even arguing for it.

by Nomad on Thu Nov 18th, 2010 at 08:22:14 PM EST
[ Parent ]
The same paper quotes a number of studies which demonstrate, without question, increased winter rainfall overall, and higher rainfall per rainy day, over the examined period (as well as a warming effect which is bigger than the global average). It is not unimagineable that these increases have not provoked any increase in major flooding; but it is powerfully counter-intuitive.

Something being "powerfully counter-intuitive" has led to some poor scientific thinking before.

While studies of damage up to now may be ambiguous about a link, all the studies with future projections I found show a link.

Increased crop damage in the US from excess precipitation under climate change

We compute that US corn production losses due to this factor, already significant under current climate, may double during the next thirty years, causing additional damages totaling an estimated $3 billion per year.

ingentaconnect The Impacts of Socioeconomic Development and Climate Change on Se...

...The climate change scenarios respectively project 13.5% and 21.5% increases in annual precipitation. The first simulation increases only the mean value of annual precipitation; the second simulation assumes that the mean and standard deviation of annual precipitation change in the same proportion. Results show that the growth in reported losses from weather-related natural disasters is due mainly to three socioeconomic factors: inflation, population growth and growth in per capita real wealth. However, weather variables such as precipitation and the number of hurricanes per period also clearly affect losses. The three stage least squares (3SLS) simultaneous equation model shows that a 1% increase in annual precipitation would enlarge catastrophe loss by as much as 2.8%...

The IPCC mentions the last one, and also: AR4 WGII Chapter 3: Fresh Water Resources and their Management - 3.5.2 How will climate change affect flood damages?

Another study examined the potential flood damage impacts of changes in extreme precipitation events using the Canadian Climate Centre model and the IS92a emissions scenario for the metropolitan Boston area in the north-eastern USA (Kirshen et al., 2005b). They found that, without adaptation investments, both the number of properties damaged by floods and the overall cost of flood damage would double by 2100 relative to what might be expected with no climate change, and that flood-related transportation delays would become an increasingly significant nuisance over the course of the century. The study concluded that the likely economic magnitude of these damages is sufficiently high to justify large expenditures on adaptation strategies such as universal flood-proofing for all flood plains.

This finding is supported by a scenario study of the damage due to river and coastal flooding in England and Wales in the 2080s (Hall et al., 2005), which combined four emissions scenarios with four scenarios of socio-economic change in an SRES-like framework. In all scenarios, flood damages are predicted to increase unless current flood management policies, practices and infrastructure are changed. For a 2°C temperature increase in a B1-type world, by the 2080s annual damage is estimated to be £5 billion as compared to £1 billion today, while with approximately the same climate change, damage is only £1.5 billion in a B2-type world. In an A1-type world, with a temperature increase of 2°C, the annual damage would amount to £15 billion by the 2050s and £21 billion by the 2080s (Hall et al., 2005; Evans et al., 2004).

The impact of climate change on flood damages can be estimated from modelled changes in the recurrence interval of present-day 20- or 100-year floods, and estimates of the damages of present-day floods as determined from stage-discharge relations (between gauge height (stage) and volume of water per unit of time (discharge)), and detailed property data. With such a methodology, the average annual direct flood damage for three Australian drainage basins was projected to increase by a factor of four to ten under conditions of doubled atmospheric CO2 concentrations (Schreider et al., 2000).

I also found a more recent study from Spain, Flood risk and climate change. An estimation for the catchment area of the River Urola from 2009. It has a crappily worded summary so I won't quote it, but in the scenarios they investigated, not only did climate change increase damage significantly, but extra flood damage due to climate change dwarfed extra flood damage due to socio-economical change.

*Lunatic*, n.
One whose delusions are out of fashion.

by DoDo on Fri Nov 19th, 2010 at 03:25:59 AM EST
[ Parent ]
Quibble: Since when are projections capable of "proving" a link? A underlying trend in those projections is assumed - not that I have an immediate problem with these what-if assumptions. Except that projections are not evidence for something - although they are useful data points to quantify scientific understanding in hindsight.

Of course I am not dealing with projections at the moment, and I thought I had already made clear in the diary that projected increased losses due to climate change are indeed likely, and that at some point in time there will be consequences. Quantifying these consequences, however, is a topic rife with speculation and assumption.

by Nomad on Fri Nov 19th, 2010 at 05:35:57 AM EST
[ Parent ]
If only all ET argumentation could be so good! (A comment from someone with no expertise in the science under review)

Index of Frank's Diaries
by Frank Schnittger (mail Frankschnittger at hot male dotty communists) on Fri Nov 19th, 2010 at 03:35:06 AM EST
[ Parent ]
Your chosen quote from the American Met Society paper

Therefore it can be concluded that anthropogenic climate change so far has not had a significant impact on losses from natural disasters.

... is a little bit truncated. It continues :

Considerable uncertainties remain in some of these studies, as exposure and vulnerability that influence risk can only be roughly accounted for over 15 time. In particular the potential effects of past risk reduction efforts on the loss increase are often ignored, because data that can be used to correct for these effects is not available.

To translate : This study (and the others quoted) show quite clearly that the major driver in the well-documented increase in weather-related insurance payouts is socio-economic change :  for example, new construction in flood plains and on steep hillsides is vulnerable to flooding and landslides. This phenomenon is well-explored in this study, and the others quoted in the diary. HOWEVER, the opposite phenomenon -- risk reduction -- is entirely unexamined in this study.

For example, I would be very surprised if protection against extreme climate events had not been greatly improved in western Europe over the past 20 years. This would, all else being equal, result in a lessening of insured damage... and the data shows no such lessening. Perhaps this trend counteracts completely, or partially, a climate-change signal.

But we don't know, based on the scientific studies presented here.

However, I would be very surprised indeed if Munich Re were not interested in this phenomenon, unexamined by the scientists. They have a huge database, much of it probably proprietary and therefore unavailable for scientific research, and a lower standard of proof than the scientists.

So when you say

Unfortunately, Munich Re seems to frequently suffer from cognitive dissonance between their PR department and their own considerably qualified science group

... I  find it quite likely that there is, in fact, no contradiction. They may well know things, as insurers, that are not publishable science, but are actionable knowledge.

It is rightly acknowledged that people of faith have no monopoly of virtue - Queen Elizabeth II

by eurogreen on Thu Nov 18th, 2010 at 04:44:47 PM EST
eurogreen:
For example, I would be very surprised if protection against extreme climate events had not been greatly improved in western Europe over the past 20 years.

Perhaps you should tell this to the Brits who got caught in the 2007 UK floods, because about half of all the British houses built after World War II were built, knowingly, in river floodplains, and housing construction continued unabated even after the 1947 flood.

Or consider, although not related to climate, the extensive urbanisation around the Vesuvius.

I've a sketch of a muse on disasters in modern Europe that simply have not been anticipated yet, either because of the habit of humans to forget about disasters after a generation, or simply because they haven't occurred through recent history.

eurogreen:

They may well know things, as insurers, that are not publishable science, but are actionable knowledge.

See above my argumentation why that is unacceptable to release such knowledge without evidence.

by Nomad on Thu Nov 18th, 2010 at 08:39:51 PM EST
[ Parent ]
I've a sketch of a muse on disasters in modern Europe that simply have not been anticipated yet

I'd like to see that!

*Lunatic*, n.
One whose delusions are out of fashion.

by DoDo on Fri Nov 19th, 2010 at 02:47:34 AM EST
[ Parent ]
I always have my work cut out for me... :)

Will see if I can make it fly.

by Nomad on Fri Nov 19th, 2010 at 05:36:55 AM EST
[ Parent ]
The analysis of twenty-two disaster loss studies shows that economic losses from various weather related natural hazards, such as storms, tropical cyclones, floods, and small-scale weather events such as wildfires and hailstorms, have increased around the globe. The 10 studies show no trends in losses, corrected for changes (increases) in population and capital at risk, that could be attributed to anthropogenic climate change. Therefore it can be concluded that anthropogenic climate change so far has not had a significant impact on losses from natural disasters.

I note that the Conclusions section of that article continues, with my emphasis:

Considerable uncertainties remain in some of these studies, as exposure and vulnerability that influence risk can only be roughly accounted for over 15 time. In particular the potential effects of past risk reduction efforts on the loss increase are often ignored, because data that can be used to correct for these effects is not available. More insight in the relative contribution from climate change on disaster losses could potentially be gained from studies that attempt to project future losses. These studies can assess the impact of future climate change, which is projected to be much 20 larger than the change so far observed. The discussion above shows the need to include exposure and vulnerability changes in future risk projections, which clearly contribute substantially to changing risks.



*Lunatic*, n.
One whose delusions are out of fashion.
by DoDo on Fri Nov 19th, 2010 at 03:32:12 AM EST
I want to note that losses measured in monetary terms has the standard problems of valuation. Naturally monetary losses are interesting for insurance companies, and the scientists are aware of the risks of comparing monetary losses over time and space. But still, it has some serious problems, and when it comes to evaluate the consequences of climate change, non-monetary losses like number of people hurt or number of houses destroyed should be better. Which should data that can be collected from archives.

Even better still would be data directly on wind or water, avoiding as far as possible social changes over time. But that needs measured data, and is very hard to reconstruct later on.

This does not really add to the topic at hand, but I thought it should be noted.

Sweden's finest (and perhaps only) collaborative, leftist e-newspaper Synapze.se

by A swedish kind of death on Fri Nov 19th, 2010 at 03:59:35 PM EST
Fascinating.

It seems almost as if the problem is analogous to peak oil (or more accurately, Peak Oil): we will not know if we are there until long after that point has come and gone.

The fact is that what we're experiencing right now is a top-down disaster. -Paul Krugman

by dvx (dvx.clt št gmail dotcom) on Sun Nov 21st, 2010 at 02:59:12 PM EST


Display:
Go to: [ European Tribune Homepage : Top of page : Top of comments ]