news & views
ATMOSPHERIC SCIENCE
Rise in upper-atmospheric carbon
Carbon dioxide cools the upper atmosphere. Satellite measurements suggest that concentrations of this
greenhouse gas have risen in the thermosphere over the past decade, with implications for the energy balance of
the upper atmosphere.
Stefan Noël
C
848
120
Thermosphere
100
Mesopause
80
Altitude (km)
arbon dioxide contributes significantly
to the heating of the lower layer of the
terrestrial atmosphere, the troposphere.
Here, carbon dioxide molecules are present
at high concentrations, and thus absorb
outgoing infrared radiation and re-emit it
towards the surface, producing the wellknown greenhouse effect. In the middle
and upper atmosphere, and in particular
in the thermosphere above 90 km altitude,
the situation is different: at these altitudes,
the concentrations of carbon dioxide are
much lower, such that carbon dioxide is
essentially transparent to infrared radiation.
Here, the dominant effect of carbon dioxide
on temperature comes from collisions with
atomic oxygen, which excite carbon dioxide
molecules. This excitation leads to the
emission of radiation to space, cooling the
thermosphere (Fig. 1). Writing in Nature
Geoscience, Emmert and co-authors1 present
satellite-based evidence for an increase in
thermospheric carbon dioxide concentrations
between 2004 and 2012.
Tropospheric carbon dioxide
concentrations have increased as a result
of human activities. Some of this carbon
dioxide has propagated into higher
atmospheric layers, resulting in a cooling
— and concomitant contraction — of the
thermosphere2. Thermospheric density has
declined as a result, weakening atmospheric
drag on satellites3 and space debris4. Thus
changes in thermospheric carbon dioxide
concentrations have implications not only for
the energy balance of the thermosphere, but
also for the movements of man-made objects
orbiting the Earth.
Unfortunately, measurements in the
thermosphere are very sparse, because
balloons and aircraft do not reach these
altitudes, and ground-based measurements
and rockets provide only limited temporal
and spatial coverage. Therefore, satellitebased remote sensing is probably best suited
for global observations of the thermosphere.
Specifically, satellite measurements
of solar occultation, that is, sunlight
attenuation through the atmosphere, can
yield information on the concentration of
greenhouse gases such as carbon dioxide.
Mesosphere
60
Stratopause
40
Stratosphere
20
Tropopause
Troposphere
0
180
200
220
240
260
280
300
320
340
360
Temperature (K)
Figure 1 | Atmospheric layers as defined by temperature. The atmosphere is divided into distinct layers,
characterized by sharp gradients in temperature. The greenhouse gas carbon dioxide, which warms
the surface of the Earth, cools the upper atmosphere, and contributes significantly to the decline in
temperatures with altitude in the thermosphere, the uppermost layer of the atmosphere pictured. Emmert
and colleagues1 show that carbon concentrations in the thermosphere have risen more than expected
over the past eight years, potentially due to an increase in vertical mixing and transport. The actual
vertical extent of the atmospheric layers changes with time and location (US standard atmosphere of
1976; the thermosphere is cut off at the top, it reaches up to 500 to 1,000 km).
However, only a few satellite instruments are
capable of performing these measurements.
Emmert and co-authors1 make use of solar
occultation data obtained by one of these
instruments — the Atmospheric Chemistry
Experiment Fourier Transform Spectrometer
(ACE-FTS) on-board the Canadian
SCISAT-1 satellite5,6 — to assess recent
changes in the amount of carbon dioxide and
carbon monoxide in the mesosphere and
thermosphere. After the removal of seasonal
and latitudinal variations, the amount of
carbon dioxide still varied over the Sun’s
11-year cycle of waxing and waning activity,
showing a reduced increase after the solar
minimum. This solar dependence stems from
the fact that the partitioning of thermospheric
carbon into carbon monoxide and carbon
dioxide depends on solar radiation; solar
radiation triggers the breakdown of carbon
dioxide into carbon monoxide, leading to
an increase in the amount of carbon dioxide
when solar radiation declines. This makes it
difficult to determine trends in carbon dioxide
levels, particularly given that the dataset spans
just eight years.
Emmert and co-authors overcome this
problem by calculating the trend in total
carbon, derived from the combination of
the amount of CO and CO2, which they
term COx. The amount of COx increased
steadily between 2004 and 2012, with a
linear trend of 23.5 ± 6.3 ppm per decade.
Assuming a constant relative proportion
between carbon monoxide and carbon
dioxide on timescales longer than the solar
NATURE GEOSCIENCE | VOL 5 | DECEMBER 2012 | www.nature.com/naturegeoscience
© 2012 Macmillan Publishers Limited. All rights reserved
news & views
cycle — thereby avoiding the effects of solar
radiation variability — they consider this
trend proportional to the long-term trend
in thermospheric carbon dioxide levels. This
inferred thermospheric trend in COx exceeds
that predicted by an upper atmospheric
model by around 10 ppm per decade.
Emmert et al. suggest that this model–
measurement discrepancy could result from
an increase in vertical mixing and advective
transport between the thermosphere and
underlying atmosphere — which contains
more carbon dioxide — due to a change in
atmospheric dynamics. Indeed, they show
that temporal trends in COx and CO2 can
be reproduced in a global mean climate
model when eddy diffusion — a surrogate
for vertical mixing and transport — grows
by 15% per decade. What’s more, the low
thermospheric mass densities previously
inferred from satellite drag data3 can also
be accounted for by an increase in eddy
diffusion, which would lead to further
cooling and contraction of the thermosphere.
Emmert and co-authors1 present a
dataset from a little-measured layer of the
Earth’s atmosphere that provides a glimpse
into the impact of anthropogenic carbon
dioxide emissions on the upper reaches of
the atmosphere. The reported discrepancy
between observations and model simulations
indicates that our knowledge of this part
of the atmosphere is far from complete.
Although the study covers just a short period
in time, the ACE-FTS instrument is still
operational, so there is a good chance that
this unique carbon monoxide and carbon
dioxide dataset can be extended in the
future. With continuing measurements, our
understanding of thermospheric trends in the
context of climate change is set to improve. ❐
Stefan Noël is at the Institute of Environmental
Physics, University of Bremen, Otto-Hahn-Allee 1,
D-28359 Bremen, Germany.
email: [email protected]
References
1. Emmert, J. T., Stevens, M. H., Bernath, P. F., Drob, D. P. &
Boone, C. D. Nature Geosci. 5, 868–871 (2012).
2. Qian, L., Laštovička, J., Roble, R. G. & Solomon, S. C. J. Geophys.
Res. 116, A00H03 (2011).
3. Emmert, J. T., Lean, J. L. & Picone, J. M. Geophys. Res. Lett.
37, L12102 (2010).
4. Lewis, H. G., Saunders, A., Swinerd, G. & Newland, R. J.
J. Geophys. Res. 116, A00H08 (2011).
5. Bernath, P. F. et al, Geophys. Res. Lett. 32, L15S01 (2005).
6. Beagley, S. R. et al. Atmos. Chem. Phys. 10, 1133–1153 (2010).
Published online: 11 November 2012
HYDROLOGY
Southwest Africa’s Okavango River does
not flow into the ocean. Instead, it ends in
a swamp in the Kalahari Desert. The vast
majority of the water evaporates in the flat
delta region at the mouth of the river that
supports one of the richest concentrations
of wildlife in Africa. Botswana’s Moremi
Game Reserve located at the eastern flank
of the delta is a prime destination for ecotourism in the region.
The Okavango River drains an area of
over 140,000 km2, including the uplands of
central Angola. Here, ample precipitation
falls in the Southern Hemisphere summer
months, and then flows towards the delta
and fan. The Okavango River is thus subject
to large seasonal variations in flow. Water
level and extent in the delta region ebbs
and floods accordingly. In addition to the
seasonal cycle, river flow variations on
longer timescales of 60 to 80 years have
also been documented.
Using statistical analyses and
hydrological modelling, Piotr Wolski and
colleagues find that these multidecadal
swings between wet and dry phases
mainly stem from variations in rainfall,
with little influence from temperaturedriven evaporation (J. Hydrol. http://doi.
org/jrx; 2012). They attribute these rainfall
variations to internal feedback mechanisms
between the ocean, atmosphere and land,
as opposed to external influences, for
example from humans or solar variability.
According to climate model projections,
these multidecadal oscillations are likely to
continue at similar amplitudes throughout
© NASA
Complex water future
the twenty-first century. At the same
time, conditions are expected to become
progressively drier in the long term, as a result
of higher temperatures and thus evaporation.
The multidecadal swings are therefore likely
to alternatingly intensify and offset the longterm anthropogenic drying trend.
Water management strategies often
assume a stationary basic state, with the
implication that departures from this mean
state — often on a five-year planning
horizon — need to be countered. Wolski
and colleagues suggest that a river basin
naturally exposed to significant multidecadal
oscillations is not well managed by these
NATURE GEOSCIENCE | VOL 5 | DECEMBER 2012 | www.nature.com/naturegeoscience
© 2012 Macmillan Publishers Limited. All rights reserved
traditional assumptions. As an example,
the water supply infrastructure of the
Botswana town of Maun was redesigned
in response to dry conditions in the 1990s
when precipitation was low — only to be
overwhelmed by floods in 2008 to 2010
when the basin returned to a wetter phase.
To keep residents, wildlife and tourists in
the Okavango Delta healthy and watered,
managers will need to design and build an
infrastructure for the supply of drinking water
that can cope with frequent swings between
a wealth and a dearth of rain and river flow.
HEIKE LANGENBERG
849