The image above shows the “absorption spectra” of H₂O (water – in blue) and CO₂ (carbon dioxide – in pink). The absorption is because light (electromagnetic radiation) at each wavelength causes the water or carbon dioxide molecules to change their vibration from one way of vibrating to another. Because water has so many different ways of vibrating, a very large number of wavelengths are absorbed. You can see that the edges are “jagged” – actually, if you zoom in on any one part of the spectrum, you can see that it’s actually made up of lots and lots of lines.

The top image has a wavelength scale in microns (micrometres – 1000 times bigger than the nanometres I’ve used so far). The sun’s spectrum is in the wavelength range from 0.3 microns to around 3 microns. The Earth’s thermal infrared emission spectrum is from 4 microns to 40 microns.

The dominant greenhouse gas is water vapour. Water vapour absorbs more infrared radiation than any other gas because of the many, many different ways the molecule can vibrate. And because that radiation (light) is absorbed, it doesn’t get released into space and the Earth has to heat up to maintain a thermal balance between the incoming solar radiation and the outgoing thermal infrared radiation.

So, what would happen if all 7.5 billion of us boiled a kettle and released water vapour into the atmosphere simultaneously? Well, the simple answer is – it would rain. The atmosphere can only hold so much water (the exact amount depends on temperature and pressure) and when it’s exceeded, the water condenses into clouds and, eventually, rain. The water cycle is a very complex, but also very rapid feedback loop. The exception would be if we boiled those kettles in the upper atmosphere. There it is harder to make clouds, and the extra water vapour creates significant increased warming. This is one of the problems with aeroplanes – they are not only creating carbon dioxide, but also releasing water vapour into the high levels of the atmosphere.

The aeroplane effect is more complex still – aeroplanes burn fuel and the by-products are carbon dioxide, water vapour and nitrous oxides – all greenhouse gases. Emitting water at high altitude creates increased warming – but this, too, eventually falls as rain. The carbon dioxide has a much longer lifetime. If we all stopped flying, the water vapour would disperse quickly – the carbon dioxide would stay around for decades. Note that if we powered our planes with hydrogen, they would still emit water. (See also: this Guardian article from 2010). Of course, planes also make contrails – which means more clouds (see comments below about clouds).

However, hotter air can hold more water than colder air. So if the air temperature increases (for whatever reason), there is more water vapour in the atmosphere, which in turn leads to more “greenhouse effect” heating. This is known as a “positive feedback” – positive in the sense that it makes the effect bigger, rather than that it’s a good thing!

We don’t have a lot of global records of water vapour levels, so it’s hard to put precise numbers on the amount of water vapour in the atmosphere and how that’s changing with time. But there are indications from spot measurements (where it’s been measured in one place for a long time) that water vapour is increasing as the atmospheric temperature increases.

Of course, increased water vapour also leads to more clouds and we’re still not completely sure what more clouds means for the climate. On the one hand, more clouds means more sunlight is reflected – reducing the incoming energy and therefore cooling the Earth. On the other hand, more clouds means that more heat is held in at night (we all know clear nights are the coldest), therefore heating the Earth. For clouds at high altitudes we know that the keeping heat in is the bigger effect (positive feedback), for clouds at low altitudes, reflecting sunlight is a bigger effect (negative feedback). Overall, there’s a lot of uncertainty in what we understand about both the positive and the negative feedback mechanisms. The latest IPCC report concluded that the positive feedback was likely to be more significant than the negative feedback – but it’s still not clear by how much. “Cloud feedback” is the biggest uncertainty in climate models (and better satellite data is needed to improve our understanding of it). The climate modellers have been predicting that doubling the carbon dioxide will change the Earth’s temperature by “something between 1.5 ºC and 4 ºC”. The reason they give a range is almost entirely due to our lack of  – almost all of that range is caused by our lack of understanding of cloud feedback. As we improve our understanding of clouds, we’ll reduce that range (and recent studies suggest the lower end of that range was too optimistic).