In the early 1970’s, I noticed the optical depth of the atmosphere at Barrow, Alaska was quite high. Given the isolated location of this station on Alaska’s northern tip (72 degrees N) I had expected that the turbidity would be low, around 0.05. Instead found values twice or three times that magnitude. Moreover the optical depth varied strongly with wavelength, being larger at shorter wavelengths, which suggested that the particles causing the atmospheric contamination were small, probably sub micron in diameter.  From this it could be deduced that the particles were not blown dust from river cuts or suspended ice crystals.  It seemed likely they were related to industrial air pollution, but no sources except a very minor one is within hundreds of miles.  Furthermore, the strange and unknown haze seemed to grow stronger in northerly winds. This was most peculiar indeed because only the Arctic Ocean, with its cover of pack ice, is north of Barrow.

Further investigation, made with a sun photometer carried aloft in a Bush aircraft showed the layers sometimes had complicated vertical structure, sometimes showing distinct bands and were most frequently in the altitude zone from 500m to about 3 km. In addition, the layers were horizontally homogeneous over distances of hundreds of km and showed no signs of being affiliated with leads or cracks in the pack ice. Back trajectory analysis carried out at the 850 mb level suggested that this air had passed over eastern Europe/Western Soviet Union about ten days earlier. At that time, transport of anthropogenic material over such vast distances (3,000 km) was thought to be impossible, except possibly for volcanic plume emissions injected into the upper troposphere/lower stratosphere.

I met Kenneth Rahn of the University of Rhode Island at a meeting in Reno in 1975 and told him the story about the strange layers in the Arctic.  Ken at the time was investigating heavy metal signatures in polluted layers with the neutron activation method at the URI reactor.  He was intrigued and so we wrote a small proposal and submitted it to Office of Naval Research, which had an interest in Arctic matters around the Barrow area.  Under ONR sponsorship we collected more than a year of weekly high volume filters at Barrow and found to our great surprise that indeed the air did seem to be polluted, most especially during the spring months.  In summer the atmospheric aerosol concentrations dropped to very low values.  About this time I also discovered that my colleague Gerd Wendler had made regular semi broad band turbidity measurements with a Linke Feussner Actinometer at or near McCall Glacier in the Brooks Range in northern Alaska. Indeed, these also showed turbidity maxima during the late winter/early spring months. 

In the spring months of 1977 we carried out specific investigations of the haze chemistry using a Cessna 180 aircraft equipped with a high volume sampling system.  We learned from this investigation that the layers were crustal material, resembling desert dust microscopically and minerologically. We speculated, correctly that the layers had formed from dust storms in the Gobi and Takla Makan deserts in Asia. Indeed back trajectory analysis supported this conclusion.  And though this conclusion was published in Nature in what we now call our “Red Herring Paper,” it turned out the sampling had occurred during a rather unusually strong and somewhat rare synoptic system favoring the transport of desert dust to Alaska.  Both the wavelength dependence of the optical depth and the trace elemental signature differed from samples taken during more usual springtime meteorological conditions.

Further investigation of the chemical and physical properties of the aerosols, along with back trajectory analyses, convinced us that the more usual springtime Arctic Haze is primarily emanating from anthropogenic sources, especially from smelting activities in eastern Europe and western and northern USSR.  Of great interest in helping us to firm up this conclusion was the usage of the V/NC Mn ratio, where NC Mn is the estimated non-crustal component of manganese. We had found that this ratio is considerably higher for aerosol sources in North America, which is primarily an oil-based economy, than it is in Eurasia, based more on coal burning technology that time.  

Expansion of the research by investigators in the Canadian Arctic, especially by Len Barrie’s group, and by investigators in Scandinavia, especially by B. Ottar at the Norwegian Institute for Atmospheric Research (NILU) and others eventually showed quite clearly that Arctic Haze is most common and strongest within the Arctic Airmass, roughly bounded by a weak meteorological feature known as the Arctic Front. The Arctic Airmass is quite small in extent and sometimes even disappears in summer, but grows larger and larger throughout winter, reaching a two lobe structure dipping down over Eurasia and North America in late winter/early spring. It is this air mass that becomes rather homogeneously polluted, due to its aerial over-cover of strong pollution sources, such as metal smelters, and to the rather low rate of aerosol removal in this dry, stable air mass system.

Several international conferences were held on Arctic Haze and it continues to be investigated and classified by surface based stations, airborne studies, such as the ARCTAS experiment in 2009, and by satellite borne sensors.  

It is believed that generally Arctic Haze has a small positive climate forcing impact (warming), of approximately one watt/ square meter for the earth-atmosphere system. This is mainly due to a combination of slight aerosol absorption from soot aerosol, and multi reflections of the radiation streams between the haze and underlying bright surface.  The climatic effect on the surface itself is one of slight cooling. There has been speculation that Arctic Haze impacts formation and dissipation of thin Arctic stratus cloud systems, due to alterations in the cloud condensation nuclei activation.  This remains to be studied more intensely.  It appears that Arctic Haze has generally decreased in intensity since the breakup of the Soviet Union in 1989, but that there may be increasing quantities of Gobi dust and anthropogenic pollution from the China mainland, no doubt associated with the industrial growth currently taking place in China. 

Further information on Arctic Haze can be found at the wikipedia page on Arctic Haze

References: 

P. K. Quinn, G. Shaw, E. Andrews, E, Dutton, T. Ruoho-Airola and S. Gong., Arctic Haze:  current trends and knowledge gaps, Tellus 59B ,99-114, 2007.

Shaw, G., The Arctic haze phenomenon, Bull. Am Met. Soc, 76, 2403-2413, 1987

Shaw, G., Evidence for a central Eurasian source area of Arctic haze in Alaska, Nature, 299, 815-818, 1983.

Shaw, G., The vertical distribution of atmospheric aerosols at Barrow, Alaaska, Tellus 27, 39-50, 1975.

Shaw, G., Aerosol chemical components in Alaska air masses 1. Aged pollution and 2. sea salt and marine products, J. Geophy Res., 96, 22,357-22,372, 1981

Rahn, K, R. Borys and G. Shaw, The Asian source of Arctic haze bands. Nature 268, 713-715, 1977

Glenn Shaw

October 24, 2007

The CLAW paper ^[1], in which researchers elaborated on the idea that sulfur emissions from phytoplankton might affect the climate, appeared in Nature two decades ago. Its authors are colleagues and long-time friends. It has been rewarding to see all the subsequent attention and development given to this idea after the publication of my paper in 1983 speculating about sulfur modulation of global climate.

I considered then and consider still that the nonlinear nature of the cloud activation is somehow one of the central points necessary to evaluating the idea. The machinery of the cloud activation is such that it often can be a kind of self-limiting process for certain types of aerosol-size distributions, because the cloud albedo is sensitive to parent new aerosols only when the new aerosol is present in low quantities and with a size distribution that is rather “flat.” ^[2].  Conditions like these are most likely to be found in mid-oceanic regions and especially in the clean southern hemisphere, where the organosulphur aerosols activate the marine stratocumulus.

The mid-oceanic regions indeed constitute a large fraction of our planet. But do the biologically produced aerosols really modulate climate?

The jury is still out after two and a half decades. However, the nonlinearities of the cloud activation process are now finally being thought about, measured, and modeled. It is currently fashionable to conduct so-called “closure” experiments, in which various parameters of the parent aerosol are measured (chemical composition, solubility, size, etc) and put into a model to predict microphysical parameters of the cloud (cloud number drop concentration). Those predicted results are then compared with actual measurements of the cloud. People are measuring and thinking about the super-saturation spectrum of cloud condensation nuclei (CCN), which holds the key.

There is no law of physics that insists that the cloud condensation nuclei in a climate feedback loop have to be sulfate, and in this regard Meskhidze and Nenes ^[3] found evidence of cloud albedo modulation connected with phytoplankton blooms in the southern ocean, which probably involved secondary organic aerosol, possibly isoprene. We obviously have to carry out more research on this topic. 

A few years ago, Tony Clarke and I noted that the aerosol in the clean oceanic boundary layer was bimodal. The bimodality probably arises through the process outlined by Bill Hoppel and colleagues ^[4], wherein sulfates are produced inside the liquid reactors in cloud droplets, then released into the atmosphere when the cloud evaporates. Marine stratocumuli undergo about a dozen activation-evaporation cycles before precipitating. This process probably produces most of the mass of sulfate in the atmosphere and creates more effective CCN, and it has to be considered in future planetary-climate feedback cycles. It is interesting because of the circular nature of that argument that aerosols produce, or at least seed clouds while at the same time clouds build aerosols. In this regard, while in Hawaii a few years ago conducting experiments, Rich Benner, Will Cantrell, Dave Veazey, J. Ji, and I managed to compare the aerosol size distributions within the boundary layer with those above the boundary layer at the Mauna Loa Observatory, finding a single-mode aerosol distribution at the higher elevation. This led our small team to conclude that mid-troposphere aerosol probably serves as the parent aerosol which, mixed downward into the boundary layer, seeds the marine cloud systems. This makes me think that the sulfate aerosols may well be produced in the mid-troposphere in bursts in the region of strong convection by homogeneous, possibly heteromolecular nucleation, and then go on to grow by coagulation. They then can get mixed downwards into the boundary layer where they serve as CCN for the lower cloud decks, introducing a climatic cooling. We need to investigate this whole convection process, including atmospheric chemistry in the mid-troposphere, nucleation of new particles followed by re-injection of particles into the marine boundary layer and cloud activation. 

The sulfur-modulation idea for my 1983 paper had its genesis in the early 1970’s when the opportunity arose to visit Barrow, Alaska, the northernmost city in the United States at latitude 72 degrees. Always having been interested in measuring the spectral transparency of the atmosphere, I hauled along one of my home-built precision sun photometers and found the arctic sky to be turbid. Furthermore, when seen from an airplane, the aerosol was found to exist in distinct dark layers. This made no sense. One would expect the atmosphere at this remote, snow-covered location to be clean.

Kenneth Rahn and I investigated the chemical composition of the arctic haze and were surprised to learn that the aerosols had an annual cycle, peaking in late winter and spring and disappearing in summer and autumn. On the basis of the chemical signatures, we deduced that the haze was industrial, having most likely originated from Eurasian sources. The latter allegation was based on the rather low vanadium to manganese ratios ^[5].  The strange arctic haze evidently had traveled to the Arctic in substantial quantities over pathways several thousand kilometers long.

This early work on arctic haze, along with the almost simultaneous work by Prospero in Miami on transport of Sahara Dust across the Atlantic, ^[6] led to the idea that aerosol residence times are longer than had been thought and that that transport of aerosols can reach global scales, indicating that aerosols are more global in importance than we had thought.

We followed up this work at the other end of the planet, asking whether pollution haze might be present over the Antarctic Ice Sheet. It would be surprising to find an Antarctic haze since the southern hemisphere is lightly populated, there is little continental area and, moreover, the ice sheet is surrounded by intense washing machine-like storm systems that would remove pollutants diffusing in from lower latitudes. Indeed, these expectations were met: the atmosphere was very clean, almost void of aerosols. The optical depth arising from aerosols was so small it was difficult to measure; we estimated it to be on the order of 0.01 for green light. There was practically no aureole around the sun. It was close to being a pristine atmosphere.

On close inspection there were some slight deviations in the spectral extinction of sunlight passing through the polar ice sheet’s atmosphere. The optical depth decreased slowly from the edge to the center of the ice sheet, and fell off with altitude. This was evidence for an imported aerosol or gaseous aerosol precursor from unknown sources flowing through the troposphere. We used a diffusion model to estimate that the residence time of the aerosol must be about 30 days. Chemical fingerprinting confirmed that there were only slight traces of human industrial contamination. The majority of the Antarctic aerosols were sulfates, often occurring as sulfuric acid droplets, but occasionally as crystals of ammonium sulfate. In summary, though there was no evidence for an Antarctic pollution haze. There was, however, a ubiquitous aerosol presence notable for its high sulfate content.

These and other published measurements resulted in a review paper on the Antarctic aerosol ^[7] trying to piece together the observations into a kind of model. This was difficult with the very rough measurements and reported data then available, but worthy of attempting since the area is so “simple,” clean, devoid of aerosol sources, and presenting such a beautiful example of circular symmetry. From this work it was deduced that there must be significant sources of sub-micron sulfate aerosol or sulfur-bearing precursor gases being produced from biogenic sources of sulfur in the surrounding oceans. 

The aerosols found in Antarctica consisted largely of removal-resistant particles in an accumulation mode. Larger and smaller particles had been evidently removed during the long transport by inertial impaction and diffusive removal.  It was interesting to note that the diameter of this removal-resistant accumulation mode (about 0.2 micron) was almost synonymous with that of particles that interact most efficiently by electromagnetic scattering with sunlight. This accidental (insofar as we know) near coincidence between particles that live for the longest time in turbulent atmospheres and particles that scatter light most efficiently was the basis for my sulfur-modulation proposal in 1983. It was but a short step, having known of Jim Lovelock’s work on the Gaia hypothesis, to put all these facts together into a hypothesis for biogenic involvement. At that time it was imagined that a kind of Gaia Machine might have caused a slow cooling that would mitigate what otherwise would be a constant global warming, as the sun burns on the Main Sequence and converts its hydrogen to helium. Perhaps this was accomplished by life forms slowly pulling out the strong greenhouse gas carbon dioxide in the atmosphere. My proposal had the duel advantage of including high thermodynamic efficiency and being in no danger of “running out,” which would soon to happen if CO2 continues to be removed by the evolutionary mechanism. The idea was that biogenic sulfate aerosols with diameters of a few tenths of a micron would not only remain in the atmosphere for extended times, but prove to be especially efficient at modulating climate. It  was further elaborated on by Lovelock ^[8].

The sulfur modulation manuscript ^[9] was submitted to Stephen Schneider’s journal Climatic Change in the hope that he might be maverick enough to publish something related to the Gaia hypothesis. Publishing work on Gaia was at the time risky; it was only some years later that a Chapman conference on the subject brought the topic into more general scientific legitimacy. 

^[1] R. J. Charlson, J. E. Lovelock, M. O. Andreae, S. Warren, Oceanic phytoplankton, atmospheric sulphur, cloud albedo and climate, Nature 1987, 326, 655.

^[2] S. Twomey, The nuclei of natural cloud formation, Part II: The supersaturation in natural clouds and the variation of cloud droplet concentration, Geofis, Pura Appl. 1959. 43,243.

^[3] N. Meskhidze, A. Nenes, Phytoplankton and cloudiness in the southern ocean, Science 2006, 314, 1419.

^[4] W. A. Hoppel, G.M. Frick, R.E. Larson, Effects of non-precipitating clouds on the aerosol size distribution in the marine boundary layer, Geophys. Res. Lett 1986, 13, 125.

^[5] K. Rahn, The Mn/v ratio as a tracer of large-scale sources of pollution aerosol for the Arctic, Atmos. Environ. 1981, 15, 1457

^[6] J. M. Prospero, T. N. Carson, Vertical and areal distribution of Saharan dust over the western equatorial North Atlantic Ocean, J. Geophys. Res. 1972, 77, 5255.

^[7] G. E. Shaw, Consideration on the origin and optical properties of the Antarctic aerosol, Rev Geophys. Spa. Sci. 1979,171983.

^[8] J. Lovelock, The Ages of Gaia, A biography of our Living Earth, W. W. Norton & Co., Inc, New York, N. Y. 1988 , 251 pp.

^[9] G. E. Shaw, Bio-controlled thermostasis involving the sulfur cycle, Climatic Change 1983, 5, 297.