More on Carbon Sinks

Since the authors neglect atmospheric reactions of CO₂, the article ascribes all sinks of atmospheric CO₂ to either terrestrial or oceanic sources. We suggest that the models should also include atmospheric reactions.

Atmospheric CO₂ and included carbonate aerosols form carbonic acid in cloud droplets in the atmosphere. The carbonic acid undergoes dissociation to bicarbonate and carbonate ions sufficiently rapidly to influence rainwater pH. That natural mechanism removes atmospheric CO₂ from the atmosphere. The magnitude of that gaseous scavenging process is not characterized well enough for a quantitative estimate of the magnitude of the sink, but large volumes of air flow through clouds and are processed by precipitating clouds during their life cycle. Land and water sinks for atmospheric CO₂ provide surfaces that interact only with the air in the immediate proximity.

Although atmospheric CO₂ has been regarded by many as chemically nonreactive, formate ions and formaldehyde are produced by its reduction in single ice crystals in mixed clouds--those having regions that contain both liquid and solid phases--when the cloud droplets contain low concentrations of sodium chloride.² It is assumed, but not determined, that methyl alcohol is also a product of the reduction process. A number of laboratory cloud chamber and field experiments have demonstrated that growing ice crystals containing low concentrations of calcium carbonate or magnesium carbonate derived from terrestrial dust also absorb and reduce CO₂. This class of chemical reactions occurs at growing ice interfaces in the presence of appropriate solutes in the system. The presence of formate ions in natural precipitation, which is normally attributed to the oxidation of methane by hydroxyl free radicals, may actually be the result of reduction of CO₂.

In their Figure 3, Sarmiento and Gruber show a strong correlation of incidences of El Niño years with higher CO₂ accumulation rates in the atmosphere. The authors suggest that the correlation may be the result of terrestrial vegetation's response to climatic variability. The presence of the growing ice phase in the atmosphere is highly variable and may contribute to the observed correlation shown. During El Niño years, there is a significant reduction in oceanic precipitation over a significant portion of the globe's tropical oceans, thus reducing both the incidence of maritime clouds and the opportunity for CO₂-reducing reactions in the growing ice phase of those clouds.

Sarmiento and Gruber mention the uncertainty associated with CO₂ uptake in the Southern Ocean. At 60° S, oceans exist at all longitudes; the climatic high frequency of storms at that latitude is reflected in the rising air in the Ferrell Cell (the region near 60 S) of the general circulation. Consequently, this region contains both maritime clouds and storms, in which cloud glaciation occurs and frequently leads to precipitation. Thus, the uncertainty of results that Sarmiento and Gruber find in the region near 60° S is probably partially due to weather patterns leading to chemical reduction and removal of carbon dioxide in precipitation.

Before reading the article, we believed that the atmospheric reactions to remove CO₂ could not compare in magnitude to the estimated oceanic and terrestrial sinks. Inclusion of atmospheric reactions that chemically reduce CO₂ in mixed clouds, followed by removal of the resulting products in precipitation, might advance the modeling technique presented in Sarmiento and Gruber's article.

1. C. E. Junge, Air Chemistry and Radioactivity, Academic Press, New York (1963).

2. W. G. Finnegan, R. L. Pitter, B. A. Hinsvark, J. Colloid Interface Sci. 242, 373 (2001).

In their article, Jorge Sarmiento and Nicolas Gruber emphasize a premise that carbon dioxide is the most prominent greenhouse gas causing global warming. In contrast, NASA, in a newspaper article two months earlier (New York Times, 31 May 2002, p. A16), stated that water vapor is the "dominant natural heat-trapping gas." Telemetry that NASA installed recently on the satellite Aqua is intended for making a worldwide study of water vapor.¹ We already know from previous satellite measurements summarized by B. J. Mason² that, on average, more than 50% of Earth's surface is covered by clouds.

Both CO₂ and water vapor are considered from a historical perspective by Spencer R. Weart in Physics Today, January 1997, page 34. Experimental spectroscopic studies of infrared absorption in laboratory air cells are cited; the most recent of those studies go back to 1911. John Tyndall made spectroscopic studies in 1861 on air mixtures and concluded that water vapor was a factor of 10 stronger than CO₂ in its IR absorption.

Our purpose here is to encourage revisiting and modernizing those early experiments using modern spectrographic methods to investigate a range of gas mixtures, radiation wavelengths, pressures, and temperatures. This activity would provide IR absorption coefficients having a higher confidence level. Such scientific data will help investigators and funding agencies evaluate where money is best spent to understand global warming.

1. For more information on the Aqua mission, see http://aqua.nasa.gov.

2. B. J. Mason, Contemp. Phys. 43, 1 (2002).

Sarmiento and Gruber reply: We appreciate the opportunity McGuire and Argyle have given us to clarify two distinctions underlying our assertion¹ that CO₂ accounts for more than half the increase in direct radiative trapping. The first is between "natural" greenhouse gases and the "increase" in trapping that has occurred since preindustrial times. The preindustrial (natural) forcing is about 60 times greater than the anthropogenic perturbation, and water vapor accounts for well over half of that.² Our assertion applies to the perturbation in the radiative balance.

Second, climate scientists draw a distinction between radiative forcing and feedback effects. The increase in atmospheric water vapor that occurs as Earth warms is a feedback response to anthropogenic radiative forcing. Model simulations suggest that such feedback will increase global warming by two to three times more than would occur if the water vapor were held constant at its preindustrial concentration.³ That feedback effect is a dominant one in determining how much warming will occur. However, the fundamental driver of the warming is the increase in greenhouse gases due to human activities. There are more recent measurements of the greenhouse gases' absorption properties than those that McGuire and Argyle refer to, but further measurements are still needed, particularly regarding the optical properties of clouds and water vapor (see the article by Thomas Ackerman and Gerald Stokes, Physics Today, January 2003, page 38). We emphasize, however, that these are unlikely to lead to a significant change in climate sensitivity estimates.¹

The letter from Pitter, Finnegan, and Hinsvark raises some points regarding the removal of CO₂ from the atmosphere by rainfall and by chemical reactions that convert CO₂ to formate ions and formaldehyde on ice crystals in clouds. Dissolution of CO₂ in rain removes an estimated 0.08 petagrams of carbon per year from the atmosphere.⁴ That amount is negligible compared to the atmosphere-ocean and land-air anthropogenic CO₂ fluxes of about 2 Pg C/yr, and also compared to interannual atmospheric variability, which is several Pg C/yr. Besides, most of this flux (~0.06 Pg C/yr) is "natural" and thus has no impact on the anthropogenic transient. Rainfall also contains a large amount of dissolved organic carbon (DOC),⁴ about 0.4 Pg C/yr. The role of this in anthropogenic carbon removal is difficult to assess because identifying the sources of the organic carbon is difficult. Formaldehyde accounts for only 3% of the DOC, and formic acid is only one of a long list of organic acids that together account for 40% of the DOC.⁴ We thus believe that direct removal of CO₂ in the atmosphere by the reactions that Pitter and colleagues propose is unlikely to contribute significantly to either the anthropogenic perturbation or interannual variability.

1. Our assertion is supported by V. Ramaswamy et al., in Climate Change 2001: The Scientific Basis, J. T. Houghton et al., eds., Cambridge U. Press, New York (2001), p. 349.

2. R. E. Dickinson, R. J. Cicerone, Nature 319, 109 (1986).

3. I. M. Held, B. J. Soden, Annu. Rev. Energy Environ. 25, 441 (2000).

4. J. D. Willey, R. J. Kieber, M. S. Eyman, G. B. Avery Jr, Global Biogeochem. Cycles 14, 139 (2000).