Letters from our readers
First of all, let me say that I am very sorry to hear that Chemical Innovation will cease publication at the end of the year. It has always been one of my favorite ACS publications.
As a chemist and resident of an historic mining state, I have always been interested in place names related to chemical elements and really enjoyed Nancy K. McGuires Periodically in need of a vacation in the August issues The Last Word. However, I was a bit disappointed that her tour completely missed the state of Idaho. Please pass on to her the following names of Idaho communities (some of which are quite small): Cobalt, Cuprum, Leadore, Silverton, Silver City, Soda Springs, and, my personal favorite, Stibnitederived from the Latin name for antimony, by way of the mineral stibnite.
Nancy K. McGuire replies:
Thanks for the fan mail! Heres a belated salute to the chemical towns of Idaho. Had I not been facing a tight deadline and space limitations, I might have been more exhaustive in my research. I used Yahoo!, a world atlas I had at home, and my own memory as resources. Even using only those, I had to do some serious cutting to make everything fit on one page.
More on global warming
I believe that some comment is appropriate concerning the rather provocative article by Robert H. Essenhigh (Viewpoint, May 2001). Essenhigh concludes that increased atmospheric levels of CO2 do not drive global warming, but rather the reverse is true. I disagree with this conclusion.
There are several serious flaws in Essenhighs argument. First, it is logically fallacious to arbitrarily claim that, because there is no answer to the question of where CO2 comes from (this in itself is not true; see below), the consideration that CO2 drives the temperature should be ruled out. It is equally fallacious to adopt an obscure Arctic Ocean model to support the notion that temperature drives the CO2 level. In fact, there is a third consideration, which Essenhigh excludes: that CO2 influences the temperature, and temperature can also influence atmospheric CO2 through feedback mechanisms.
Essenhigh invoked the Arctic Ocean model from a book published some 30 years ago, but he did not refer to any more recent publications by experts in the field to explain the role of CO2 in global temperature. First, the timescale was supposed to be millions of years, but then Essenhigh suggested the Arctic Ocean trip which represents a rapid, discontinuous oscillation between two extreme temperatures, with a timescale of 500 years and more recently, 5 years. The mechanism and energetics of the trips, and how the trips reconcile with the million-year timescale are all a mystery. Second, he suggested that when the Arctic Ocean is frozen, heat release from the Earth would be slowed down, leading to the warming mechanism. But a frozen ocean also increases the albedo, reducing the ability of the Earth to absorb solar radiation, its most important heat source. Thus, the Arctic Ocean model does not make much physical sense.
The second flaw regards the Mauna Loa CO2 observation. If Essenhighs hypothesis that the CO2 observed at Mauna Loa were indeed coming primarily from the sea around Hawaii, then the CO2 level should have been found to increase during summer and decrease during winter. This is exactly opposite to what is observed.
The third and fourth serious flaws in Essenhighs article are, respectively, the claim that CO2 emissions associated with fossil fuel combustion can have only a trivial impact on atmospheric levels of CO2, and the claim that IR absorption by CO2 is sufficiently weak that even a doubling of atmospheric CO2 would not have any significance.
Essenhighs claim about emissions is based on a comparison of the ~56 Gt carbon (GtC) emission of CO2 associated with fossil fuel use with the ~150 Gt of carbon that is exchanged annually among the atmosphere, oceans, and biosphere. Discussion of the human impact on the levels of CO2 in the atmosphere is complicated by two factors. First, emissions of CO2 associated with human activities are small when compared with natural fluxes of CO2 associated with photosynthesis, respiration, uptake into ocean water, and release from ocean water. Second, several large reservoirs of CO2 (e.g., atmosphere, upper ocean, deep ocean, biosphere) are continually exchanging CO2. In such a system, one needs to be very careful when using the words source and sink; it is better to refer to net source and net sink. Thus, the oceans are a large (90 GtC/year) source and a large (92 GtC/year) sink of atmospheric CO2 (1). Overall, the oceans provide a net sink for CO2 of 2.0 ± 0.8 GtC/year, and forest regrowth in the Northern Hemisphere 0.5 ± 0.5 GtC/year. Human activities are believed to lead to emissions of 5.5 ± 0.5 GtC/year from fossil fuel combustion and 1.6 ± 1.0 GtC/year from deforestation (2). The atmospheric burden of CO2 is increasing at a rate of 3.3 ± 0.2 GtC/year.
To balance the CO2 budget, other terrestrial sinks invoked have been inferred to account for 1.3 ± 1.5 GtC/year of CO2 (2). Emissions associated with fossil fuel use are large when compared with the net flux of CO2 among the atmosphere, oceans, and biosphere. The current atmospheric CO2 concentration is 370 ppm (30% above the level of 280 ppm observed in preindustrial times) and increasing at a rate (~1.5 ppm/year) consistent with that expected from emissions associated with fossil fuel use.
Essenhighs claim that CO2 is not an important greenhouse gas is based upon overly simplistic arguments. The relative masses of CO2 and H2O are irrelevant to the discussion, and the absorption of CO2 over the entire terrestrial radiation spectrum is important (not just at 5.67.6 µm).
Global climate models include detailed treatment of the IR absorption of H2O and CO2 and show that CO2 is an important greenhouse gas (23). Satellite observations show that CO2 is a major greenhouse gas (4). The importance of CO2 as a greenhouse gas is not a new or controversial scientific discovery. As far back as 1896, Svante Arrhenius (5) estimated that the increased greenhouse effect resulting from a doubling of atmospheric CO2 would warm the Earth by 56 °C; current climate models predict a 1.54.5 °C rise from doubling of CO2 (2).
Ole John Nielsen
Robert H. Essenhigh replies:
Because informed and pertinent argument is always valuable, I have no problem with Nielsens disagreement with my conclusion, that CO2 levels do not and cannot drive global warming, if it leads to constructive discussion. But it is disappointing that his objections do not seem to be well informed. In some cases, they substantially misrepresent what I wrote, and in others, they come through as assertions without substance. In particular, there seems to be a lack of fundamental knowledge of the physics of radiative absorption and emission. Nielsens statement in his penultimate paragraph, The relative mass of CO2 and H2O is not relevant . . . , is completely wrong. The standard charts on this, numerically showing the importance of concentrationbased on the data of Hottel and Egbert (1, 2) and later supported by detailed numerical computation (e.g., 3)have an experimental and theoretical history going back >70 years.
The radiation data are most commonly set out graphically in terms of the product pL (concentration p [in atm] times path length L). The relative concentrations of the two gases are then a central factor in determining their importance in the atmosphere, and this further influences the accuracy of the models (Nielsens last paragraph) if the concentration effect is improperly included in the model. I have >40 years experience of radiative modeling and testing the predicted behavior experimentally (e.g., 4). That experience supports the standard view that model predictions should always be treated with reserve until they have been validated against critically structured, independent experiments. Clearly, such validation testing is a particularly difficult task in climate model studies, a point which further reinforces the need for care in advancing model predictions. (This phenomenon exemplifies the definition of a tragedy as a theory killed by a fact or a model destroyed by an experiment.)
An additional problem would seem to be Nielsens confusion between the temperature oscillation period (~100,000 years) and the term trip time (5500 years) that destroys his argument on the Arctic Ocean model. His arguments on the relative insignificance of carbon from combustion and the Mauna Loa behavior require that he accurately reread the original article, and there is an update on Mauna Loa in my response to an earlier letter (In Box, November 2001).
Nielsens most interesting point is the dismissal of the relative masses (concentrations) of the two governing gases, H2O and CO2. If his view is supported by the majority of those knowledgeable on global warming, this may be the source of many or most of the disagreements that I have seen in the past 30 years on the presumed importance of atmospheric CO2. Certainly, many detailed studies completely omit water, notably in Rotmans recent book (5). Water may have been omitted from the earlier physical models because it is not included in standard (engineering) analyses of the atmosphere; it was omitted because the water is so variable. This shortcoming has now been realized, but many later studies evidently still failed to appreciate that point.
There is then the question of how to evaluate the water, and this requires agreement, still generally lacking, on what base condition to use. The value I use is 60% relative humidity (RH) at 60 °F (Case 1 in the figure, which is the standard ASHRAE psychrometric chart ).
For comparison, I also indicate a desert condition (Case 2, 5% RH at 125 °F) and a Gulf Coast condition (Case 3, 80% RH at 95 °F). Also included for further comparison is CO2 at 400 ppm (corrected to a water equivalent for this comparison); this line is just above the zero of the humidity ratio (y-axis) line. Its small magnitude in comparison with the other values makes its own point.
In evaluating the atmospheric radiative behavior, the governing equations, as I noted originally, are the Schuster Schwarzschild radiative equations of transfer (79). This is an extension of the BeerLambert radiative-absorption equation, which contains the governing group G = kpL; the pL pair is defined above and k is the radiative-absorption coefficient for either a specific gas or for a specific band of a specified gas. Values of k for the different bands for H2O and CO2 and also for methane are tabulated in the November response. The important result here is that the k values for comparable bands of H2O and CO2 are very close (as commonly assumed in engineering calculations ), so that the governing factor, for a given common path-length L is, as emphasized earlier, the concentration p.
As shown in the figure, the atmospheric mass-ratio values of p for H2O and for CO2 are very different; the molar ratio for the base case is about 25:1 but can exceed 100:1. At higher altitudes, the H2O concentration can drop because of rainouts, and a 1:1 ratio is possible at the tropopause. The average concentration for H2O from ground level upward, nevertheless, is somewhere between 0.5 and 1.5 orders of magnitude higher than that of CO2.
The proportional influence of concentration is then shown more directly (see November response), using 99% saturation distances (Lsat values) for comparison. Full interpretation of those results is not possible here; it requires evaluation by solving the equation of transfer (10). However, the numbers show very comparable values of Lsat for both gases at the same concentration (400 ppm), but drastically different values for H2O at the base case (1%). These numbers show that the statement the relative mass . . . is not relevant is clearly wrong. If this factor is not properly included in the models, they will be worthless. This needs further careful consideration.
In summary, Neilsens commentary focuses attention on what has been substantially obscured in many of the past discussions, which is the importance of the water component in the models, and the clear need for clarification of the relative magnitudes of these factors in the models and their effects on predictions. This does highlight a problem of potentially major impact that has not, evidently, been given at all the attention that it needs. Once again, it depends on the numbers!