According to a team of Indian scientists, while CO2 and other greenhouse gases released into the atmosphere are causing the Earth's surface to warm, they are simultaneously causing the upper atmosphere to cool. I recall reading something about six years ago to this effect, that the absorption of infra-red radiation ("heat") by CO2 present in the lower atmosphere actually restrains it from heating the stratosphere, causing the latter to cool. For ready-reckoning, the surface of the Earth is immediately blanketed by a "boundary layer" which extends to an altitude of about 2 km, above which lies the troposphere extending upwards to anywhere between 7 kilometers (km) at the poles and 17 km at the equator. Then there is the stratosphere, which goes up to around 50 km. Above this is the mesosphere, and finally the thermosphere which begins at around 85 km. The atmosphere is extremely rarefied at such altitudes, and the 1000 K temperatures that may be estimated for individual air molecules, do not correspond strictly to a condition of thermal equilibrium since the average distances ("mean free paths") for collisions between molecules are of the order of kilometers, and so the facile energy-transfer in the "thermal-bath" assumed in physical chemistry is now rather inefficient.
The different regions of the atmosphere are characterised by their differing temperatures. As one moves away from the surface of the Earth, the temperature initially falls, in the manner of moving one's hand increasingly upward from an electric hotplate. This can be seen on the drop-down panels that are common in passenger aircraft now, which read such parameters as outside temperature, altitude and speed. I think these things fascinate me more than non-scientist travellers, but I have struck-up some very pleasant conversations with fellow passengers, based around these readings ... so science can provide good chat-up lines! For example on one flight to the U.S., at 42,000 feet (13 km) I looked out of the window and slightly "upward", and realised that the blue-black canopy I could see some way above us was in fact the "ozone layer". I explained this to my travelling companion, who knew that the loss of this was "the reason why she and her friends couldn't go sunbathing for so long these days in California"! She was a nice lady and as usual it is interesting to compare notes with someone from the other side of "The Pond" as to how our lifestyles differ.
If you go high enough (say in Concorde, before it was withdrawn), at the top of the troposphere (called the "tropopause") the temperature is seen to increase from the (minus) 60 degrees C that it has fallen to in consequence of decreased radiative heating by the surface. This is a result of entering the lower region of the ozone layer, which absorbs ultraviolet (UV) light from the Sun, and converts this to heat by collisions between molecules that have been initially excited by the UV energy. The gas pressure, though low (of the order of a few millimeters; ground-level pressure is 760 mm) is still high enough that energy is effectively redistributed through the gas. The process acts as a "shield" to protect life on Earth from UV which would otherwise harm us, e.g. give us skin cancer, which is on the increase in some parts of the world, such as California where sun-worshipping lifestyles are popular. The region of "heated gas" is still fairly cool, but warmer that it would be without the ozone and warmer than the gas below it, so that it acts as a "lid" on the troposphere and keeps most of the gases contained there from diffusing (convecting) upward, hence maintaining its unique chemistry.
The stratosphere is nonetheless heated to some extent by radiation emitted from the Earth's surface, but as levels of CO2 (and other greenhouse gases, such as methane and water vapour - let's not forget that one, as the Earth warms-up!) increase, more of the heat is absorbed and retained by the troposphere, and so less gets further up, meaning that the stratosphere cools. This encourages the formation of clouds especially in the polar regions, upon the surfaces of which ozone molecules are actually decomposed, thus contributing to a thinning of the ozone shield. Thus it is thought that there may be a link between global warming and ozone loss.
The new study shows that the upper atmosphere is cooling much faster than the surface of the Earth is warming, and in the last three decades it has cooled by somewhere between 5 and 10 degrees C while the surface has increased by just 0.2 - 0.4 degrees C. They are making much of this, but surely just by thinking of the Earth and its atmosphere in terms of an engine which transfers heat between regions of differing density, this is not surprising. If more heat is absorbed in denser regions lower down, it is effectively being taken from those more rarefied regions higher up, and so of course they will get relatively colder, as there is less "up there" to take the heat from, hence the effect will be manifested by a greater fall in temperature.
The study refers specifically to the region between 50 and 100 km above the Earth's surface, which is also known as the ionosphere and is critical to long range satellite communication. The region is physically shrinking (gases contract when they are cooled) because of the reduction in temperature, and it can be said literally that "the sky has fallen", since the upper level of the atmosphere is now 8 to 10 kilometers lower that it was thirty years ago. It is feared that these changes to the ionosphere could lead to a deterioration in the quality of short wave radio reception. It has also been suggested that solar panels used to power on-board satellite systems might become increasingly degraded by high energy particles from space which can increasingly penetrate a thinner ionosphere? I think the latter concern is the less compelling of the two, but only time will tell. However, geostationary satellites, which are used in relaying signals for communications, television broadcasts and GPS (geopositioning satellite) systems are parked very much higher, at about 36,000 km, and should not therefore be affected.