As more state and other interested parties line-up to contest the EPA’s Endangerment Finding, the EPA is becoming creative in trying to come up with other strategies to justify restricting carbon dioxide (and other greenhouse gas) emissions.
One new strategy is to use the Clean Water Act to justify curbing CO2 emissions because they lead to ocean acidification (an impact which itself seems to be overblown). Another is to explore seeking greenhouse gas emissions controls at a local level, under the guise that concentrated local CO2 emissions (i.e. in cities) change the local environment in such a way as to elevate human mortality there.
Never mind that such an impact will never be detectable.
My colleague Pat Michaels refers to this as the EPA’s “whack-a-mole” strategy—while effort is concentrated on trying to beat down one of its pesky and ill-founded CO2-regulating proposals, the EPA pops up another and another and another. The EPA hopes that after a while, one mole sneaks through unnoticed and manages to grab the prized (CO2-enriched) carrot.
Currently most of the whacking is aimed at trying to halt EPA’s use of the Clean Air Act as a mechanism for sweeping CO2 regulation. The ocean acidification issue is just starting to get some attention. We have briefly touched on ocean acidification here at MasterResource, and found it to be an issue in which it seems that the more that is known the less of a problem it appears to be (i.e. the real world is a pretty adaptable and responsive place). Hopefully, the EPA’s recently announced solicitation of public comment “on what considerations EPA should take into account when deciding how to address listing of waters as threatened or impaired for ocean acidification” will garner some deserved response before the May 21, 2010 deadline.
In this post, I want to take a look at a novel mole that needs attention: the idea that local CO2 produces any sort of impact on local mortality that could be detectably reduced by local CO2 restrictions.
In taking questions from the Senate’s Environment and Public Works Committee last year, EPA Administrator Lisa Jackson insisted that that EPA needed to retain certain performance based standards setting for new and existing power plants in order to address local pollution. Senator Arlen Specter was amazed, saying there is no localized impact from CO2.
But apparently Jackson knew something that Senator Specter didn’t—the EPA was funding a study that was to conclude that, indeed, local CO2 emissions do raise local mortality in the U.S.—by three one-hundredths of one percent—or by about 792 deaths out of 2,700,000 deaths annually (from these numbers it would seem that Senator Specter wasn’t that far off).
The information that Administrator Jackson probably had in her back pocket has now been formalized in the just-published paper by Stanford’s Mark Jacobson in the journal Environmental Science and Technology, titled “Enhancement of Local Air Pollution by Urban CO2 Domes.” In this paper, Jacobson concludes that local CO2 emissions raise the local level of low-level ozone and fine particulate matter, PM2.5, which, in turn, raise the rate of local mortality. This finding prompted Jacobson to remark “This study establishes a basis for controlling CO2 based on local health impacts.” No doubt just what the EPA wanted to hear.
Nowhere did anyone mention that the results of this study were, for all intents and purposes, known a priori, virtually guaranteed by the methodology employed by Jacobson.
After all, climate models (like the one used by Jacobson) produce higher temperatures and higher water vapor content when they are driven with higher CO2 emissions, and air pollution models (like the ones used by Jacobson) produce more low-level ozone and PM2.5 with higher temperature and moisture, and epidemiological models (like the ones used by Jacobson) produce more deaths with higher levels of ozone (O3) and fine particulate matter (PM2.5). So, the result—more CO2=more deaths—is assured.
The only thing that was really missing in this chain of reasoning was whether local CO2 emissions lead to higher local CO2 concentrations (CO2 is often described as being “well-mixed”, that is, evenly distributed in the atmosphere). But, local CO2 “domes” have been documented for more than a decade, so it is a wonder that it has taken so long for the local-CO2-emissions-lead-to-more-local-deaths meme to be hoisted. Seems like a missed opportunity for all these years!
Main Problems of the Jacobson Study
Here is how the Stanford University press release sums up the import of Jacobson’s findings:
In the first study ever done on the local health effects of the domes of carbon dioxide that develop above cities, Stanford researcher Mark Jacobson found that the domes increase the local death rate. The result provides a scientific basis for regulating CO2 emissions at the local level and points out a significant oversight in the carbon dioxide “cap-and-trade” proposal that was passed by the House of Representatives in June 2009 and is awaiting definitive action by the Senate.
But is a relationship between local CO2 emissions and local enhanced mortality in the real world demonstrable, universal, or even detectable?
No, no, and no.
The signal is far too small and the noise is far too great. Thus, the effect of local CO2 on local mortality is simply undetectable and thus, it is impossible to determine whether or not it even exists, much less whether it could be altered by regulating local CO2 emissions.
Take, for instance, the impact of local CO2 emissions on local temperature. CO2 “domes” can and do form over large cities as the CO2 emissions generated from within, under some meteorological conditions, enhance the local atmospheric CO2 concentration faster than it is mixed away. Early work to try to quantify this effect was performed by Robert Balling and colleagues for the city of Phoenix, AZ. They found that under certain meteorological conditions, urban CO2 concentrations could be as high as 550-600 ppm (some 200ppm higher than the surrounding countryside). Running this CO2 enhancement though a radiation model, Balling et al. calculated that local CO2 emissions raised the local temperature by a bit more than one-tenth of one degree Celsius—a very small amount. In fact, Balling et al. commented on just how insignificant this temperature rise is by comparing it to the overall urban heat island in Phoenix which typically adds 5 to 10 degrees C to the urban temperatures, “[w]e conclude that the majority of the surface heating associated with the urban heat island effect is due to forcing by phenomena other than the urban CO2 dome. We suggest that the absorption of solar energy by surface materials and lower soil moisture levels in the urban core would be likely factors in explaining most of the observed surface heating.”
In other words, other processes associated with urbanization are 50 to 100 times more influential than local CO2 emissions. And not to mention that the atmospheric conditions aren’t always conducive to a big local CO2 build-up.
The situation is little different in Los Angeles, the city that Jacobson most intently examined with his climate model and determined that the peak air temperature increase from local CO2 emissions was about 0.1°C—about the same as in Phoenix. But the climatalogical conditions in L.A. are much different than those in Phoenix and Jacobson’s climate model indicated an increase in atmospheric moisture to accompany the temperature rise, something the Balling et al. did not document in Phoenix. However, as with the temperature increase, the moisture increase was small as well. Typical values of atmospheric water content are measured in terms of kilograms/m2 and can vary by several tens of kilograms on a daily basis—Jacobson reports changes resulting from local CO2 emissions to be in tens of grams/m2—a thousand times less than daily variability. This renders such changes observationally undetectable.
Virtually undetectable, too, are all of the knock on air-pollution impacts in the L.A. basin predicted by Jacobson’s models: higher low-level ozone (projected to increase by 0.1 to 0.2 ppbv), more fine air particulates, PM2.5, (projected to increase by less than 1 microgram/m^3), and enhanced peroxyacetyl nitrate, PAN, (projected to increase by less than 0.02ppbv).
But, the beauty of the mortality models is that no matter how small the changes in air pollutants like O3, PM2.5 and PAN, they will produce fractionally more mortality. Sum these fractional changes over a large enough population and you start to produce whole numbers.
For instance, for L.A., summing over the population of 17.3 million people, Jacobson gets about 14 extra deaths/yr from the O3 enhancement and 81 extra deaths/yr from the PM2.5 enhancement. So, according to the models, that is 95 extra deaths as a result of local CO2 emissions out of the approximately 144,109 deaths that occur in L.A. every year. Try finding that in the observations.
Which is the overall problem in a nutshell. All of Jacobson’s reported changes, with the exception of the transient existence of the CO2 dome itself, are only detectable in computer-model world, where all other factors are controlled for and each model in each step is presumed correct. In the real world, it is impossible to find these extra deaths, even if they did occur exactly as Jacobson describes—there are just too many other confounding factors.
Further, all throughout the Jacobson analysis, the changes are so small that tiny tweaks to any of the employed models—climate, air pollution, or epidemiology—will have the potential to significantly alter the results as currently reported. And such tweaks are always occurring as none of these model-types are set in stone and each is an active and evolving field of scientific study.
Other Problems of the Jacobson Study
And the problems don’t stop here. In fact, there are myriad other areas of concern. Here are brief summaries of a few:
• Jacobson uses his models to scale the results to the state and national level (he reports about 792 extra deaths nationally from local CO2 emissions), but in scaling up, the spatial resolution of his climate model drops, and thus local (and complex) meteorological processes are not well-handled. As Jacobson described these local processes as being important in L.A., and as Balling et al. have shown the effect to be different in Phoenix, not having a good handle on local meteorology—which itself is influenced by the very existance of the city—makes state and national impact estimates unreliable.
• Jacobson does not discuss the notion of “mortality displacement”—that is, some deaths from a particular event may simply be moved up a few days, and so the impact on total mortality averaged over a longer period of time may be much lower. Mortality displacement has been shown to account for a significant portion (~10-30%) of excess mortality associated with heat waves. So it is a potential issue in Jacobson’s analysis as well. In other words, his total elevated mortality estimates which are already too low to be detected, probably should be even lower.
• The epidemiological models are far from being certain, and even further from being universally applicable. There are many studies which suggest that different populations show different responses to environmental stressors. This limits the ability of up-scaling to provide reliable results
• Despite rising urban temperatures and urban CO2 domes, air quality has been getting demonstrably better in American cities. So any local CO2 negative effect is being overwhelmed by other phenomena.
• Despite rising urban temperatures and urban CO2 domes, heat-related mortality (to which most of the projected elevated CO2 mortality is associated) has been demonstrably declining in American cities. So any local CO2 negative effect is being overwhelmed by other adaptive measures.
Hopefully, the Jacobson study was just supposed to get folks thinking about this issue. However, I hope that my article gets the same people to understand that the issue is far more complex than the Jacobson study seems to indicate.
Let this serve as a shot across the bow to the EPA that pursuing the idea of regulating local CO2 emissions in the name of decreasing local mortality is not going to go uncontested—when/if this mole sticks its head up, it is going to get whacked.
Balling, Jr., R.C., Cerveny, R.S., and C. D. Idso, 2001. Does the urban CO2 dome of Phoenix, Arizona contribute to its heat island? Geophysical Research Letters, 28, 4599-4601.
Jacobson, M. Z., 2010. Enhancement of local air pollution by urban CO2 domes. Environmental Science and Technology, 44, 2497-2502.