Cosmos Q&A: Cooling the upper atmosphere

The same greenhouse gases that are warming Earth’s surface are cooling the upper atmosphere (the mesosphere) 90 kilometres above Antarctica.

Research released this week precisely measured this cooling rate (putting it at 10 times faster than the average warming at the planet’s surface) and revealed an important discovery: a new four-year temperature cycle in the polar atmosphere.

Cosmos put some questions to John French and Andrew Klekociuk from the Australian Antarctic Division, who carried out the research with Frank Mulligan, from the National University of Ireland Maynooth

In simple terms, the same process that is warming the Earth is cooling the atmosphere less than 100 kilometres above us. How does that work? 

This works because of the ability of carbon dioxide (CO₂) to absorb infrared radiation, exchange energy with the gases in the atmosphere, and re-emit radiation.

The Sun heats the Earth over a range of wavelengths of the solar spectrum which penetrate through the atmosphere: some wavelengths are absorbed by ozone and water vapour. The warm Earth then re-radiates a lot of this energy as long-wave (infrared or heat) radiation, otherwise we would continue to heat up. While most of the atmosphere (78% nitrogen, 21% oxygen) is transparent to the re-radiated infrared radiation, carbon dioxide is not, and is able to absorb it.

The CO₂ molecule can hold on to this absorbed energy only for a short time – its “radiative lifetime” – before it re-emits the radiation in all directions. However, in the dense lower atmosphere it can impact with other molecules in the air (the nitrogen and oxygen), exchanging energy through collisions, thereby increasing the kinetic energy of the air (ie the temperature). This is the “greenhouse effect” of global warming. CO₂ essentially transfers more energy to other molecules in the lower atmosphere. 

On the other hand, in the upper atmosphere, where the density is very low, the chance of collision is much lower, so the molecule more effectively radiates the energy it absorbs to space, resulting in an overall cooling of this region.

With increasing CO₂, both processes occur more effectively. More warming in the lower atmosphere, more cooling in the upper atmosphere. The concentration of CO₂ now exceeds 410 parts per million compared to 280 ppm in pre-industrial times. On current trends, this generation will see a doubling of the CO₂ concentration compared to the maximum it has ever been over the last 800,000 years – as old ice core data has shown.

Australian scientists have been monitoring the Antarctic for 20+ years. Was this work a Eureka moment or a little more routine? 

No Eureka moments were encountered in this discovery! We continually analyse the spectra we measure, meticulously calibrate to derive temperatures, compute nightly and winter averages, and compare with previous years and other measurements to verify the quality of the data. 

The variations in temperature we measure are merged responses from seasonal variations, the 11-year solar cycle and the response to changing atmospheric composition (greenhouse gases). We need to untangle these responses to separate the contributions from each. We extract the mean climatology and use a solar activity indicator (the 10.7cm solar radio flux) to extract the solar component. A linear fit to the remainder is the long-term trend. 

We have previously reported on the solar cycle and long-term trends in the data. The first of our new papers is an update of these trends with better precision.

On top of that, the new four-year cycle we call the Quasi-Quadrennial Oscillation (QQO) became apparent. As an observer, your eye seeks patterns in the data in order to help forecast what might be coming. After a few cycles of the QQO we could generally predict warmer and colder years. But we needed to verify the QQO with independent data, to have confidence it was not an instrumental anomaly or local effect.

There are not many long-term measurements of temperatures in the mesosphere, particularly in Antarctica. In recent years two satellite instruments have provided global temperature measurements through the atmosphere. These are NASA’s Microwave Limb Sounder (MLS) on the Aura satellite and an instrument called SABER (Sounding of the Atmosphere using Broadband Emission Radiometry) on NASA’s TIMED satellite (Thermosphere Ionosphere Mesosphere Energetics Dynamics).

Although these have fewer years of measurements, both satellites confirmed the QQO pattern over the last three cycles. The second paper is our investigation of the properties, extent and source of the QQO.   

Were the results a surprise? 

A four-year cycle is kind of a surprise. The atmosphere undergoes many different cyclic oscillations. Some are obvious like the summer-winter seasonal cycle, some reasonably well know like the El Niño southern oscillation (ENSO) and the solar cycle, and some less well known like the Quasi-Biennial Oscillation (QBO; a two-year variation in equatorial pacific winds), the Pacific Decadal Oscillation (PDO) or the Southern Annular Mode (SAM).

A four-year cycle is unusual in terms of weather and climate modes, and we don’t yet fully understand the source or the mechanism that causes it. 

Should we be concerned? 

We have measured a three degree Celsius cooling of the upper atmosphere just during the 25-year course of this Australian Antarctic Division project. A consistent 1.2 degrees per decade is a big change in climate terms and is a symptom of the same carbon dioxide problem causing the warming in the lower atmosphere. If the trend continues at the current rate, what will the atmosphere look like in another 25 years?

You talk about the cooling happening at “the edge of space”. Is that significant? Is the location as important as the process itself? 

The process is important to demonstrate that it is CO₂ that is causing global warming and not some other process. Heating the Earth due to changes in the Sun for example would not lead to cooling in the upper atmosphere.  

The upper atmosphere provides a sort of “integrator” of effects occurring closer to the surface. There is a lot of stored energy in the oceans, and this drives variability in weather and climate near the surface that can mask slow changes. The effects of this variability subside at higher altitudes, allowing the longer-term changes to be more clearly seen.

What are the mechanics of the work? What and how do you monitor and how are you able to make such detailed findings? 

These observations are based on the remote sensing of hydroxyl airglow. Airglow is similar to aurora, but rather than being produced by charged particle impact on oxygen atoms and nitrogen molecules in the upper atmosphere, airglow is produced by a photochemical reaction. Hydroxyl airglow is actually incredibly bright – about five times brighter than an aurora – but we just can’t see it because it is in the infrared part of the spectrum.

Our spectrometers at Davis Research Station in the Antarctic observe several of these emission lines in the infrared. Simply put, the relative heights of the emission lines in the spectrum are dependent on the temperature of the emission region – kind of like a DNA fingerprint of the airglow, but in temperature.

The technique has been used for decades after the hydroxyl airglow was discovered in the 1950s. We scan a spectrum and obtain a temperature measurement every seven minutes during night hours. Typically, 25,000 measurements per year, for 25 years now.  

A large part of the mechanics is continued, careful calibration of the instruments. Running the same instrument for 25 years has its challenges but we do not want to integrate instrumental changes into the temperature results when evaluating long-term changes.

The project also involves monitoring noctilucent or “night shining” clouds. What are they and where do they fit into the picture? 

Noctilucent clouds are extremely tenuous, filamental, ice crystal clouds that form in extremely cold temperatures (about minus 130 degrees Celsius) at around 83 kilometres altitude. They are the highest clouds in the atmosphere and only occur at high latitudes in summer when the coldest temperatures occur near the mesopause – the boundary between the mesosphere and the thermosphere and the coldest region anywhere on Earth. Seasonal temperatures are inverted in the mesosphere due to the large scale transport of air, rising and cooling in the summer pole, descending and warming at the winter pole.

Noctilucent clouds sightings are very rare in the Southern Hemisphere due to the fact that there are not many observers in the Southern Ocean or Antarctica where they mostly form. We have about 10 observations of them in our collection between the first one that John saw at Davis in 1998 and this year’s observation by Ashleigh Wilson at Macquarie Island.  

They are a phenomenon of the modern world. The first reported sighting of them anywhere was in 1885 after Krakatoa erupted and injected a lot of water vapour into the upper atmosphere.

As a consequence of the CO₂ cooling in the region, it is expected that the occurrence of noctilucent clouds will increase, both in brightness and extent. This is the “miner’s canary” reference in two papers by Gary Thomas: the first “Is the polar mesosphere the miner’s canary of global change?” and the second “Are noctilucent clouds harbingers of global change in the middle atmosphere?”

The new findings have global implications. What has been the response from the international science community? 

The cooling response of this region to carbon dioxide increases was predicted as far back as 1989 by Roble and Dickinson. Since then, several laboratories around the world have been making measurements to test that forecast.  It has taken the best part of three decades to obtain data with the precision needed to quantify the prediction with certainty.  

The rates of cooling are an important quantity to measure and monitor and this is of great interest to atmospheric scientists and modellers to confirm and improve our understanding of the physical world. The information is also vital to governments and policy makers as a means of setting and adjusting emission targets based on evidence.

The QQO will be of interest to climate scientists in a wide range of disciplines, from the atmosphere to oceans, and also to the Antarctic ice. As our papers have only recently been published, the international community has yet to apply our results, but with time we expect to see follow-up work investigating this particular aspect further.

Is there any upside to what we now know? 

Improved understanding of atmospheric processes and responses is always an upside. 

Confirmation of the predicted cooling at 90 kilomere altitude gives us confidence that our global models include the more significant physical mechanisms that control the temperature in our atmosphere.  Quantification of the cooling rate enables these models to be refined, which helps to reduce the uncertainty of future model predictions. 

Identification of the QQO highlights that processes in the climate system are not fully understood or anticipated. As the processes responsible for the QQO appear to lie near the surface, climate models should be capable of capturing this variability.

We have shown that one leading climate model does not appear to show the signal of the QQO (at least of the magnitude we have observed it). This suggests that the model does not properly capture the physical processes that drive the QQO.

We rely heavily on modelling to predict future climate and to quantify the effect of any mitigation measures that may be adopted. The importance of getting them to capture the atmospheric processes and variability at all levels as accurately as possible cannot be overstated.