dust in the air
by Christina Hulbe
Every winter I teach a sophomore inquiry course in our University Studies program, titled Global Environmental Change. For the first couple of years, my framing idea was revolution and we started with Louis Agassiz's 1837 Discourse at Neuchatel, for the Swiss Society of Natural Sciences (he was meant to talk about fossil fishes but introduced a far-reaching theory of a past ice age instead and something close to a riot ensued). Those poor students also had to listen to me natter on about science in the Age of Enlightenment and read Tennyson's and Percy Shelley's reactions to modern geology. Then the Arctic Climate Impact Assessment was published (free! online!) so it became the central organizing document and we started talking about the cultural impacts of climate change.
Two years ago, I was so impressed by The Worst Hard Time, Timothy Egan's account of the Great Plains droughts of the 1930's, that it became the textbook for the course. Along with climate science, we now talk about such topics as courage, propaganda, and justice. I am, of course, not trained to lecture about Rousseau or von Mises or Walzer but I give it a go, and in an inquiry course, I can count on students who are prepared to discuss philosophy. (Students who can both pronounce and use contextualize are like gifts from the prophets.)
Research papers published over the last few years have provided important insights into the "Dust Bowl" era on the north American Great Plains, making that last change in course seem more provident than it was. I've written about drought here before. It's a normal attribute of climate, though drought severity varies regionally. A variety of paleoclimate proxies (tree rings, pollen, charcoal, lake levels, bison bones...) contain evidence of drought cycles with a range of intensities and time scales throughout the post-glacial record (about the last 10,000 years). Looking to north America, it turns out that the Dust Bowl was not particularly severe compared to other post-glacial events, but it was unusual: human modification of the land surface affected both its pattern and intensity. The 1930's Great Plains droughts would still have happened, but they would have been different without us.
A good place to start this story is an overview of the general circulation of Earth's atmosphere. Intense, year-round solar heating of the land and ocean surface in the tropics warms air near the surface and drives evaporation. The warm, moist air rises because density decreases as temperature and vapor content increase. This results in relatively low atmospheric pressure throughout a globe-circling region called the Intertropical Convergence Zone (ITCZ; more on the convergence thing later).
The moist air cools as it rises, driving condensation of water vapor and eventually precipitation of water droplets. The now-dry air aloft moves north and south toward higher latitudes, cooling and becoming more dense as it goes. Eventually the dry air begins to sink, and as it does, it warms. Atmospheric pressure is relatively high under the descending air mass (around 30 degrees N and S). The pressure gradient drives a circulation back toward the equator, as well as toward higher latitudes. The dry air means low relative humidity and limited cloud formation. Evaporation, not precipitation, dominates in the subtropics.
The air gains water vapor as it moves back toward the equator, eventually becoming moist and warm enough to rise again. The tropical-to-subtropical convective overturning of the atmosphere is called the Hadley cell. A similar cycle transpires at higher latitudes, with moist air rising and driving mid-latitude precipitation (around 60 degrees north and south). This second convective cell is called the Ferrel cell. A third cell, driven by cold, dry air descending over the poles, completes the meridional overturning of Earth's atmosphere.
As you know if you watch or read the weather news, the actual atmospheric circulation is more complicated than this . Those rising and sinking air masses are doing so above a surface that moves beneath them (west to east, due to Earth's rotation). Physicists (and others) call this a rotating frame of reference (as compared to an intertial frame). Rising air parcels experience the effect of this rotating frame via something called the Coriolis force (a frame-dependent "fictitious" force). From the vantage point of the land surface, the rising air parcel appears to veer away. If our air parcel is moving away from the equator and into the northern hemisphere, it appears to veer to the right. In the southern hemisphere, the direction of displacement is to the left.
The Coriolis effect on Earth's atmosphere gives us cyclonic and anticyclonic circulations around low and high pressure centers, jet streams, the easterlies, all the familiar features from the weather news. Let's take a simple example: relatively high pressure in the subtropics and lower pressure in the tropics would tend to drive air motion away from the subtropics and toward the ITCZ. In a non-rotating frame, that's all that would happen but in our rotating frame, the equator-ward motion in the northern hemisphere is deflected the right (from east to west) while equator-ward motion in the southern hemisphere is deflected to the left (from east to west) and voila: the easterly Trade Winds!
Jet streams are relatively narrow bands of strong westerly winds in the upper part of the troposphere. Jets form at the intersections between the meridional overturning cells (top figure here). The subtropical jet stream is at the intersection of the Hadley cell and Ferrel cells, where dry air descends back toward the surface. The paths of the jet streams are made complicated by interactions with high and low pressure systems, topography, and seasonal changes.
These basics, along with varying properties of surface materials and the seasons, generate mean patterns of atmospheric pressure and circulation that we understand pretty well (for more, see a nice article from American Scientist ). Drought is associated with higher than typical pressure over a region, which inhibits cloud formation, lowers relative humidity, and limits precipitation.
Atmospheric circulation patterns vary in recognizable ways on more than the daily and seasonal time scales with which we are all familiar. A source of interannual variability with which most of us are familiar is the El Nino Southern Oscillation (ENSO), a tropical coupled ocean-atmosphere phenomenon. As the tropical easterlies blow across the wide Pacific Ocean, the shallow, wind-driven part of the ocean circulation responds by piling water up under the convergence zone and toward the western side of the basin (the Coriolis effect is important here too). This produces a shallow ocean feature called the western Pacific, or Indo-Pacific, warm pool. Relatively cool water rises up from depth in the east, replacing the warm water moving west. The resulting gradients in ocean surface temperature (warm in the west and cool in the east) are mirrored in the overlying atmosphere, which in turn produces a zonal overturning in the atmosphere called the Walker circulation.
During an El Nino event (the "warm" phase of ENSO), the easterlies weaken, eastern upwelling declines, and the warm pool extends eastward. During La Nina (the "cool" phase), the easterlies are stronger than average and the warm pool shifts westward. As you might expect, this variability has a profound effect on precipitation in the tropics. The effects of these changes extend far beyond the tropics as well.
Here in the US, El Nino years are associated with wetter than average winter in the Plains and across the south, and milder than normal winters farther north. La Nina years are associated with drier than average winter and spring in the plains and southern states, a condition that may persist into summer, and relatively cool winter farther north.
It's worth noting that other normal modes of variability are recognized to drive drought cycles elsewhere.
So why does a tropical oscillation affect weather in the extratropics? First, heat loss from the sea surface is an important source of the energy driving the Hadley cell. When surface temperature cools in the tropical Pacific (as happens during the La Nina phase of ENSO), the temperature and pressure gradients between low- and mid-latitudes are reduced. The subtropical jet stream weakens, becomes more variable than normal, and experiences an overall poleward shift. Portland gets more rain than usual and the Llano Estacado gets less.
The pattern in tropical sea surface temperature also affects the location of the Intertropical Convergence Zone. The ITCZ follows the warmest surface water (the lowest sea surface pressure is here). When the equatorial Pacific is relatively cool (La Nina), the ITCZ moves north to warmer temperatures and the global atmospheric circulation must adjust. In the story of Great Plains drought, an important change is to the location of the Bermuda, or Azores, high.
The Bermuda high is a large, semi-permanent atmospheric high pressure center over the subtropical north Atlantic. Seasonal shifts in the center of the feature drive large seasonal weather patterns in Iberia, France, northern Germany, and the southern UK. Anticyclonic (clockwise) circulation around the high can also influence weather in the US, bringing moisture onshore from the Gulf of Mexico. When the ITCZ shifts north, the Bermuda high adjusts, disrupting the flow of moisture from the Gulf (Central America gets it instead). (Try here for more detail and some graphics.)
Here's a nifty animated time series of anomalies (difference from the mean) in sea surface temperature and terrestrial vegetation.
So if we understand modes of atmospheric variability, we can explain drought in north America. Mostly. A complication with the Dust Bowl is that compared to other 20th Century droughts, it lasted a long time and its areal extent was unusual. It doesn't help that meteorological and oceanographic data grow scarce as we look backward in time. What to do? Use a climate model of course. This is just what a research group at NASA's Goddard Space Flight Center did, publishing the results in 2004 (Schubert et al., 2004, Science v 303 n 5665; abstract).
Starting from a model with demonstrable skill in reproducing 20th Century climate and a 1930's sea surface temperature anomaly data set (using temperature observed by ships at sea), Schubert and colleagues designed a series of experiments that could ID the cause of the Dust Bowl droughts. First, they divided the global ocean into regions based on their known importance in climate variability: tropical oceans, Indian Ocean, Pacific Ocean, Atlantic Ocean. Next, they conducted a series of model runs in which either the 20th Century mean climatology or the 1930's anomaly (difference from the mean) was applied in the various sectors. The results were compared to a 1902 to 1999 "control" run in order to determine the time scales of variability and region of the ocean most closely linked to the Dust Bowl droughts. In a sense, they were trying to fingerprint the coupled Dust Bowl sea surface temperature and precipitation anomalies.
Schubert and colleagues concluded that cooling in the tropical Pacific and tropical Atlantic drove the Dust Bowl droughts and that year-to-year variability in the climate system was secondary in shaping the drought event. An overview of the work is here, though I would note that the headline (NASA Explains...) is a bit of an overstatement. In fact, while this work was a huge step forward, in was not enough. The droughts produced in these models and others (abstract) were not severe enough and were limited to the southern high plains, while the reach of the Dust Bowl droughts included most of the Great Plains region. Something was missing.
That something turned out to be dust. Mineral dust from terrestrial surfaces affects climate in many ways. I'd need another entire post (or more) to cover it all. For our story, what matters is the effect of these fine particles on radiative transfer in the atmosphere. In short, dust in the atmosphere scatters and absorbs incoming shortwave radiation from the sun and radiates the energy back out at longer wavelengths. This reduces the total amount of energy arriving to the surface beneath the dusty atmosphere, which in turn reduces evaporation and subsequent precipitation (a map of global dust forcing on atmospheric radiative balance can be found here). It is possible, then, for the particularly vigorous soil erosion that characterized the Dust Bowl era to have played a role in shaping the drought event.
In a paper published earlier this year, climatologists at Lamont-Doherty Earth Observatory and NASA's Goddard Institute for Space Studies (Cook et al., 2008, Geophysical Research Letters v 35; abstract and science brief) tested this idea. Using a model that included the 1930's sea surface temperature anomalies, Cook and colleagues compared drought events simulated with and without dust emissions from the Great Plains. They found that including this extra climate forcing moved the center of the drought northward (due to southwesterly winds carrying the dust) and intensified the moisture deficit, compared to the no-dust scenario. The match still isn't perfect, but it's better. Were the details of dust production better known (records are mostly anecdotal), the simulations would likely be improved. (Check co-author Richard Seager's website for more.)
the Dust Bowl
The Dust Bowl was exceptional. While the paleoclimate record indicates that the drought was not especially intense compared to other post-glacial droughts, it was a singularly destructive event in our particular cultural context. We have long understood that poor farming practices coupled with overcapitalization in the 1920's and the economic decline of the 1930's were critical factors in the collapse of the plains agricultural economy. What has become clear thanks to recent climate research is that our own modification of the landscape is also part of the story. While the 1930's droughts would have happened without us, poor land management intensified the climate event and expanded it northward, engulfing most of the Great Plains. As we all know, the economic and social dislocation extended much farther.