By itself, a doubling of CO₂ in the atmosphere has only a modest effect on global temperatures – an increase of around 1ºC. Even a quadrupling of CO₂ in the atmosphere would only cause a direct temperature increase of about 2ºC (1). However, this modest increase can be either amplified or reduced depending on the feedback effects.

Some feedback effects are clearly positive. For example, warmer climates cause ice to melt, and the darker surface exposed by melting ice is much less reflective than ice and would therefore tend to absorb more heat, thus amplifying the initial warming.

Another positive feedback effect is related to the solubility of CO₂ in water, which is negatively related to temperature. Thus, when oceans warm, they release CO₂ to the atmosphere, amplifying the initial effect. This is the effect responsible for the close relationship between temperature and CO₂ over the last several hundred thousand years, with changes in CO₂ concentrations lagging behind changes in temperature by a few hundred years (see Figure 1).

Other feedback effects are negative, however. For example, higher temperatures cause more evaporation and thus more clouds. Clouds have a net cooling effect, which means that they would tend to reduce the initial heating effect.

The whole global warming debate is really a debate about these feedback effects. If, all together, the feedback effects are negative, then the planet will warm less than 1ºC over this century, which is hardly anything to despair about. However, if the net feedback effects are positive, then the planet may get progressively warmer, which would clearly be a cause for concern.

Our theoretical understanding of the feedback effects is very incomplete, which means that we rely on empirical data to calibrate the climate models. So far, most climate models have been calibrated with positive net feedback effects. The Stern Report even includes studies that ten-double the initial effect, thus predicting warming of more than 10ºC during this century.

The ice core data from Figure 1 is useful to evaluate these feedback effects. If they were dramatically positive, as suggested by the Stern Report, then at times of high CO₂, we would expect ever increasing temperatures and ever increasing CO₂ concentrations due to the positive feedbacks from melting ice and CO₂ released from the warming oceans. In fact, every time temperatures reached a bit above the current level, both temperatures and CO₂ levels plunged dramatically, instead of continuing upwards.

Similarly, during cold periods with low CO₂ concentrations, we would expect ever decreasing temperatures due to the positive feedbacks of increasing ice coverage causing higher reflectivity and colder oceans absorbing ever more CO₂, thus further reducing the temperature. But in fact, although the downward trend in both temperatures and CO₂ concentrations has clearly been dominant over the last 400,000 years, the planet has recovered from every single ice age ever experienced, instead of getting progressively colder.

If the net feedback effects were positive, the climate system would be unstable, and we would be heading towards an ever warmer planet or an ever colder planet. But for hundreds of millions of years, the global temperature has always oscillated within a 10ºC range, even during periods when atmospheric CO₂ concentrations were more than 10 times higher than today (2). This is incredibly strong evidence of a negative feedback effect.