In 1967 Journal of Atmospheric Sciences published the paper: Thermal Equilibrium of the Atmosphere with a Given Distribution of Relative Humidity by Manabe and Wetherald.
Here is one interesting model projection:
The corresponding note says:
Can this be true? How can “greenhouse” gases reduce temperature? Is this another “global warming causes more snow storms” type story?
First, a little about the stratosphere.
The stratosphere is the region of the atmosphere from around 10km to 50km. In pressure terms it’s the pressure between about 200mbar and 1mbar.
Ultraviolet radiation is almost completely absorbed in the stratosphere. The high energy photons of wavelength less than 0.24μm can break up molecular oxygen, O2, into atomic oxygen, O+O.
O2 and O combine to create O3, or ozone, which is again broken up with absorption of more ultraviolet.
Ozone production is greatest at a height around 25km. At higher levels, there are too few oxygen molecules to intercept all of the photons. At lower levels, there are few high energy photons left.
Here’s an interesting way of seeing how the absorption of solar energy at different wavelengths changes as thicker sections of the atmosphere, especially the stratosphere, are traversed:
The reason why the troposphere (lower atmosphere) warms from the bottom is that once the UV is absorbed the atmosphere is mostly transparent to the rest of the solar radiation. Therefore, the radiation passes straight through and is absorbed by the earth’s surface, which warms up and consequently warms the atmosphere from beneath.
Air that warms expands, and so rises, causing convection to dominate the temperature profile of the lower atmosphere.
By contrast, the stratosphere is warmer at the top because of the effect of solar absorption by O2 and O3. If there was no absorption by O2 or O3 the stratosphere would be cooler at the top (as it would only be heated from underneath by the troposphere).
Just about everyone has heard about ozone depletion in the stratosphere due to CFCs (and other chemicals). Less ozone must also cause cooling in the stratosphere. This is easier to understand than the model results at the beginning (from increased “greenhouse” gases). Less ozone means less ability to absorb solar radiation. If less energy is absorbed, then the equilibrium stratospheric temperature must be lower.
Stratospheric Temperature Trends
Temperature measurements of the stratosphere are limited. We have satellite data since 1979 which doesn’t provide as much vertical resolution as we need. We have radiosonde data since the 1940s which is limited geographically and also is primary below 30hPa (around 25km).
Lots of painful work has gone into recreating temperature trends by height/pressure and by latitude. For example, in the 2001 review paper by Ramaswamy and many co-workers (reference below), the analysis/re-analysis of the data took 23 of the 52 pages.
Here is one temperature profile reconstruction from Thompson and Solomon:
From Thompson & Solomon (2005):
From 1979 to 1994, global-mean stratospheric temperatures dropped by 0.75 K / decade in the stratosphere below 35 km and 2.5 K / decade near 50 km
Before explaining why more CO2 and other trace gases could cause “stratospheric cooling”, it’s worth looking at the model results to understand the expected temperature effects of less ozone – and more CO2.
Observations and Recent Model Results
Notice that in the 1967 paper the predicted temperature drop was larger the higher up in the stratosphere. The effects of ozone are more complex and also there is more uncertainty in the ozone trends because ozone depletion has been more localized.
Here are model results for ozone – the best estimate of the observed temperature changes are in brown but aren’t expected to match the models because ozone is only one of the factors affecting stratospheric temperature:
Note that the effect of ozone depletion has a projected peak cooling around 1hPa (50km) and a second peak cooling around 80hPa.
Now the same paper reviews the latest model results for stratospheric temperature from changes in “greenhouse” gases:
The same paper reviews the model results for changes in stratospheric water vapor. This is a subject which deserves a separate post (watch this space):
Finally, the model results when all of the effects are combined together:
The model results are a reasonable match with the observed trends – but a long way off perfect. By “reasonable match” I mean that they reproduce the general trends of decadal cooling vs height.
There are many uncertainties in the observations, and there are many uncertainties in the changes in concentration of stratospheric ozone and stratospheric water vapor (but not so much uncertainty about changes in the well-mixed “greenhouse” gases).
A couple of comments from A comparison of model-simulated trends in stratospheric temperatures, by Shine et al, first on the upper stratosphere, reviewing possible explanations of the discrepancies:
None of these potential explanations is compelling and so the possibility remains that the discrepancy is real, which would indicate that there is a temperature trend mechanism missing from the models.
and then on the 20-70hPa region:
Nonetheless, assuming that at least some part of this discrepancy is real, one possible explanation is stratospheric water vapour changes. Figure 3 indicates that an extra cooling of a few tenths of a K/decade would result if the Boulder sonde-based water vapour trends were used rather than the HALOE water vapour trends. If this were one explanation for the model–observation difference, water vapour could dominate over ozone as the main cause of temperature trends in this altitude region.
Why Is the Stratosphere Expected to Cool from Increases in “Greenhouse” Gases?
This is a difficult one to answer with a 30-second soundbite. You can find a few “explanations” on the web which don’t really explain it, and others which appear to get the explanation wrong.
The simplest approach to explaining it is to say that the physics of absorption and emission in the atmosphere – when calculated over a vertical section through the atmosphere and across all wavelengths – produces this result. That is – the maths produces this result..
You can see an introduction to absorption and re-emission in CO2 – An Insignificant Trace Gas? Part Three.
[Note added to this article much later, the series Visualizing Atmospheric Radiation has an article Part Eleven – Stratospheric Cooling – from January 2013 on why the stratosphere is expected to cool as CO2 increases. It is quite involved but shows the detailed mechanism behind stratospheric cooling].
After all, this approach is what led Manabe and Wetherald to their results in 1967. But of course, we all want to understand conceptually how an increase in CO2 – which causes surface and troposphere warming – can lead to stratospheric cooling.
The great Ramanathan in his 1998 review paper Trace-Gas Greenhouse Effect and Global Warming (thanks to Gary Thompson of American Thinker for recommending this paper) says this:
As we mentioned earlier, in our explanation of the greenhouse effect, OLR reduces (with an increase in CO2) because of the decrease in temperature with altitude.
In the stratosphere, however, temperature increases with altitude and as a result the cooling to space is larger than the absorption from layers below. This is the fundamental reason for the CO2 induced cooling.
In Ramaswamy (2001):
For carbon dioxide the main 15-um band is saturated over quite short distances. Hence the upwelling radiation reaching the lower stratosphere originates from the cold upper troposphere. When the CO2 concentration is increased, the increase in absorbed radiation is quite small and the effect of the increased emission dominates, leading to a cooling at all heights in the stratosphere.
Are they saying the same thing? Yes (probably).
If these explanations help – wonderful. If they don’t, refer to the maths. That is, the mathematical result provides this solution and overall “hand waving” explanations are only ever a second-best “guide”. Also check out The Earth’s Energy Budget – Part Three for explanations about emissions from various levels in the atmosphere.
Understanding stratospheric temperature trends is a difficult challenge. Understanding the mechanisms behind this changes is much more of a conceptual challenge.
But over 40 years ago, it was predicted that the upper stratosphere would cool significantly from increases in CO2.
The depletion of ozone is also predicted to have an effect on stratospheric temperatures – in the upper stratosphere (where CO2 increases will also have the most effect) and again in the lower stratosphere where ozone is the dominant factor.
Stratospheric water vapor also has an effect in the lower stratosphere (where more water vapor leads to more warming and vice-versa), but more on this in a later post.
For some, who feel/believe that CO2 can’t really significantly affect anything in climate – this post isn’t for you – check out the CO2 – An Insignificant Trace Gas? series.
There will be others who will say “Ozone is the reason the upper stratosphere has cooled“. True, but increases in CO2 are also an important factor. The same calculations (maths and physics) that lead to the conclusion that less ozone will cool also lead to the conclusion that more CO2 will cool the upper stratosphere.
This subject also has two other possible consequences. One is about attribution. Global temperatures have increased over the last 40 years and many people want to understand the cause.
If solar heating was the direct cause (see Here Comes the Sun) the stratosphere would not be cooling. However, other effects could possibly also cause stratospheric cooling at the same time as tropospheric and surface heating. It’s a complex subject. But something to question for those other potential causes – would they also cause stratospheric cooling?
The other consequence is about GCMs. Some say that stratospheric cooling is a “vindication” of GCMs. In so far as we have covered the subject in this post we couldn’t reach that conclusion. The modeling of tropospheric and stratospheric temperature profiles can be done (and was by Manabe and Wetherald) with 1D radiative-convective models. Certainly 3d GCMs have also been used to calculate the effect by latitude but these results have more issues – well, the whole subject is much more complex because the change of ozone with height and latitude are not well understood.
But it is important to understand the difference between a GCM solving the general climate problem and a more constrained mathematical model solving the temperature profile against height through the atmosphere.
However, stratospheric cooling while the surface and troposphere are warming does indicate that CO2 and other “greenhouse” gases are likely influencers.
Thermal Equilibrium of the Atmosphere with a Given Distribution of Relative Humidity, Manabe and Wetherald, Journal of Atmospheric Sciences (1967)
Trace-Gas Greenhouse Effect and Global Warming, Ramanathan, Royal Swedish Academy of Sciences (1998)
Stratospheric Temperature Trends: Observations and Model Simulations, Ramaswamy et al, Review of Geophysics (2001)
A comparison of model-simulated trends in stratospheric temperatures, Shine et al, Q. J. R. Meteorol. Soc. (2003)
Recent Stratospheric Climate Trends as Evidenced in Radiosonde Data: Global Structure and Tropospheric Linkages, Thompson & Solomon, Journal of Climate (2005)
An update of observed stratospheric temperature trends, Randel, Journal of Geophysical Research (2008)