Big Knob Critique Response

January 23rd, 2006

Posted by: Roger Pielke, Jr.

In 2000, Dan Sarewitz, Bobbie Klein and I published a paper titled “Turning the Big Knob: Energy Policy as a Means to Reduce Weather Impacts” (PDF) in which we calculate the relative sensitivity of future tropical cyclone damages to the independent effects of changes in storm behavior under climate change to changes in societal vulnerability. For the changes in both storm behavior and societal vulnerability we used the assumptions of the IPCC. Brian Schmidt, who works for an environmental organization in San Francisco, and who occasionally has visited our website with always-thoughtful comments, has taken the time to write up a critique of our paper and post it on his blog. We appreciate the engagement. Brian graciously asked us for a response, so here it is.

First a correction, Schmidt states that “Climate will increase hurricane costs by about 40-50%, according to the studies cited in the paper.” This is what those early-1990s studies said, however, they used assumptions about changing hurricane intensity that were later revised downward by the IPCC. Using in the damage projection methods the actual numbers cited by the IPCC SAR for changing “maximum potential intensity” of tropical cyclones the projected increase in damages due only to the effect of changes in storms related to climate change is in fact about 8-10%. For close followers of this subject the numbers used by the SAR were about twice those projected in a subsequent paper by Knutson and Tuleya. This error is repeated throughout Schmidt’s post. It is worth noting that the general conclusions are pretty much insensitive to an error of this magnitude.

Schmidt offers three critiques of our paper. First, he writes, “it ignores the combined effects of climate change and increased economic value.” We have heard this before. The response is that we are performing a sensitivity analysis and not generating a prediction of future damages. Our point in this paper was to identify the independent effects of climate change versus societal change. It Schmidt would like to go ahead and add another factor that recognizes the fact that future damages will indeed be the result of the combined effects, he should go right ahead. This does nothing to change the fact that the independent effect of societal changes is larger than the independent effect of climate change by a factor of between 22 to 1 and 60 to 1. This is the case if combined effects are included or not.

Schmidt offers up this as a second critique, “Another major problem is the implicit assumption that by controlling land-use, one can in effect relocate away from global warming.” We said no such thing. What we did say is that if societal factors are far more responsible than climate change for the expected growing impacts of tropical cyclones, then from a policy perspective it is only logical that decisions related to societal vulnerability are likely to have greater potential to address those impacts. We discuss land use, but also forecasts and warnings (to save lives), reducing environmental degradation, enforcement of building codes and other policies. Relocation is not something that we discuss.

Schmidt’s third critique is even less appropriate, “A third major problem parallels the problem with Bjorn Lomborg’s critique of a lack of economic analysis over global warming, particularly a lack of cost-benefit analysis.” We did not conduct a cost-benefit analysis, nor did we claim to, so it is hard to respond to this. We conducted a sensitivity analysis to ascertain where policy makers might have the greatest ability to influence future hurricane damages. What we found is that using the “big knob” of tuning global atmospheric carbon dioxide concentrations is unlikely to be able to affect anything more than a very small portion of future tropical cyclone losses, and that this finding is very robust under all combinations of scenarios of climate change and societal change present in the IPCC SAR.

Schmidt offers three other minor critiques:

“ Damage estimates are artificially low because they come primarily from a time when hurricanes were at a low point in their long-term cycle. Recalibrating the damages to include more intense hurricane cycles we are experiencing now would give much larger figures.

Let’s accept this as true. It is irrelevant. Our analysis focused on how damages would increase as a multiple of a particular base year. Changing the base has no effect on the multiple, and has no effect on our sensitivity analysis.

“It’s dated. It’s from 2000, and relies extensively on 1996 IPCC assessments, which rely on still earlier studies. This isn’t a flaw of the study itself, but rather than relying on it, the work should be done again with updated information.”

It is a 2000 paper. Nonetheless the analysis holds up exceedingly well as there has nothing that has occurred related to scientific understandings of tropical cyclones or projections of societal vulnerability that would change our basic conclusions. We do intend to update this analysis using the assumptions of the IPCC AR4 (we do have funded research on such sensitivity analyses under our SPARC project.). Better yet would be if the IPCC itself did such a sensitivity analysis, not just for tropical cyclones, but for every climate impact that is the joint function of climate and societal factors.

“It ignores costs resulting from redirecting land use.

We don’t claim to be doing a cost-benefit analysis, nor do we discuss relocation.

The bottom line is that the analysis is robust under a wide range of realistic and unrealistic scenarios for climate change and societal change. Even if we were to simply assume that the IPCC SAR underestimated changes in future tropical cyclone intensity by 100% or 200% the qualitative implications of our paper would remain unchanged. I appreciate Schmidt’s comments that changing land use behavior is difficult. But the reality is that there is no basis for expecting that a global energy policy focused on stabilizing greenhouse gas offers a meaningful tool with which to modulate future tropical cyclone damages.

The lure of a “big knob” that can be tuned to an ideal state is indeed appealing, but in the case of tropical cyclones the sooner we recognize that effective policy will take place on the ground in thousands of vulnerable locations around the world that experience damage, then the more effective policy responses will be.

4 Responses to “Big Knob Critique Response”

  1. Brian S. Says:

    Thanks Roger – I’ll respond more fully later, but several questions if you have a chance to respond. The “22 to 1 and 60 to 1″ ratio – is that from Figure 2 of your paper? It doesn’t match the $4-5 billion annual figure for climate change and the $22-48 billion for growth. Is that a cumulative, 50-year figure? And what’s the justification for stopping in 2050? My limited understanding is that the thermal effect on oceans is delayed, so fun times are ahead for us the further out we project. Will you move the time frame forward in your revised paper?

    More later, but I’ll just note squeakily here that the alleged error of a 40-50% increase in intensity from warming “repeated throughout Schmidt’s post,” is taken from your original article.

  2. 2
  3. Roger Pielke, Jr. Says:


    Thanks for the follow up.

    You correctly saw the 40-50% figure referenced in our paper to the various integrated assessment studies, but we did not use these figures in our calculations because the IPCC provided updated and more authoritative estimates for projected hurricane intensity.

    Specifically, on p. 260 of the article we describe the methodology used by each of the integrated assessment models for calculating the sensitivity of losses to climate change influences on hurricanes, and summarized in Table 1. Each of these methods assumed a 40-50% increase in the intensity of hurricanes to 2050. (We used 2050 because that is what the IPCC SAR used. Obviously one could use whatever end date one wanted to in such an analysis, providing that the proper information is available, e.g., from the IPCC.) The IPCC SAR WG I concluded that hurricane intensity would increase by 10% by 2050, so we used this figure in the damage formulas presented by Tol, Fankhauser, and Cline under the assumption that the IPCC represented the most updated and accurate state of the science. On page 264 we show that the damages under such a scenario would increase by 8-10% under the three integrated assessment methods.

    Thus, the 22 to 1 figure comes from (32 billion (A2 scenario) – $10 billion (base))/1 billion (Cline estimated increase). Similarly the 60 to 1 figure comes from (58 billion (A1 Scenario) – $10 billion (base))/0.8 billion (Tol estimated increase).

    As far as updating the analysis, currently the state of the science would lead to numbers more like 50 to 1 and 130 to 1! Consider that Knutson and Tuleya (2004) provide a scenario that suggests a 5% increase in hurricane intensity by 2080, which is 30 years later than IPCC SAR estimated and half the intensity increase. Of course, scientists may come up with other estimates that are larger in the near term, but these results are not really sensitive to whether it is 5% or 50%, given that population and wealth are doubling in coastal locations every 7-15 years. Redoing this analysis when the next IPCC makes sense, but because the IPCC has not updated the SRES scenarios, the wealth and population data will not change from the 2000 study, and because the IPCC is only supposed to use peer-reviewed work published as of Dec, 2005, I think that the Knutson and Tuleya article will be among the most recent and authoritative. Thus, our next analysis may make the 2000 study look conservative by comparison.


  4. 3
  5. Steve Bloom Says:

    Roger, unless I’m not recalling correctly the terms of the deadline for the AR4 are that a paper must be accepted for publication by December 2005, which means it will be some months before we see the last of the potentially significant papers.

    I need to do a longer comment on all of this, but if your new effort both uses a 100-year horizon and considers the enhancement of storm surge by sea level rise (did the first one?), can you be so confident about the conclusions?

  6. 4
  7. Roger Pielke, Jr. Says:


    Thanks. Yes, you are correct about the acceptance. I am assuming that if there are radically novel papers in the pipeline I would have heard about it among the hurricane community, but you are right, there always may be unanticipated surprises.

    The longer the time horizon, the greater that the analysis falls in favor of the societal factors. This is because vulnerability is doubling every 7-15 years, while the projected increase in hurricane intensity is measured in tens of percent over the same time frame. So as we go from 2050 to 2100 wealth and population will double at least 2 times (a 200% increase), while hurricane intensity is currenty expect to increase by less than 5% over that same time period. But assume that these estimates of intensity increases are in error by a factor of 3 (i.e., 15%), or 5 (25%) even (unrealistic given current understandings, but assume it anyway). It does not change the qualitative interpretation of the results. This is a pretty robust conclusion.

    Yes, I think that sea level rise was considered in the integrated assessment model studies that we surveyed, but it can’t be too significant of a factor in damages. Consider that the IPCC projects sea level increases of 0.25 to 0.45 meters by 2100, while a storm like Katrina had a storm surge of 10 meters in some places. I doubt that damage would have been measurably different if this had instead been 9.5 or 10.5 meters. If you can identify a signal of sea level rise in the damage record over the past 100 years, then you might have a case for its importance.

    In any case I look forward to your longer post.