La lune


Documentation > Greenhouse effect > Trying to guess the future > How many extra degrees can we get?

How many extra degrees can we get ?

last modified : august 2007

website of the author : - contact the author :


It is a strong conclusion that the average near ground temperature will increase if the concentration of greenhouse gases in the atmosphere increases. But by how much ? The most recent simulations conclude to a bracket of 1,2 to 6,2 °C within a century, depending on the models used and the emission scenarios (graph below). But what exactely explains such a wide bracket ?

Predictions of the average near ground temperature rise between 2000 and 2100 depending on the scenario chosen, for a variety of models. Each emission scenario (named with an abreviation) has a different colour and corresponds to a specific set of hypothesis.

For a given scenario, that is a given colour , the vertical grey bar on the right gives the bracket of possible temperature rises in 2100, depending on the model used. For example, the A1B scenario, in green, would generate a global temperature rise in 2100 varying between 1,8 and 4,2 °C depending on the model used (the "least" sensible model gives "only" 1,8°C, and the most sensible one gives up to 4,2 °C. For each scenario the uncertainty linked to the model used corresponds to the shadowy zone that surrounds the main thick line.

Broadly speaking, the "constant composition commitment" scenario corresponds to a greenhlouse gas concentration that remains constant throughout the 21st century (and therefore CO2 emissions that become nil in 2000), B1 corresponds (roughly) to constant emissions, A1B corresponds to a progressive doubling of the emissions, and A2 to a multiplication fourfold of the emissions to the end of the century.

The average near ground temperature for the 20th century is shown in black, with a grey shading that indicates the 95% confidence interval for the value.

Source : IPCC, 4th assessement report, 2007

But it is necessary to keep in mind two essential things :

A given temperature increase won't happen "no matter what" : a global temperature increase is "just" the part of the response of the climate system to our past and future greenhouse gases emissions. Even if the future evolution of our climate system is driven by many globally known processes, the future is not already totally written. Future emissions are hypothetic by nature, and predictions are not precise forecasts, bound to happen whatever we do from now on,

Though, the more we occidentals intend to generalize our way of living to the whole world, and the higher the emissions will be. The most "pessimistic" IPCC scenario remains very moderate compared to a situation where every human being on earth would tend to live like an average American citizen today. As every human being on Earth experiences living conditions today that are not very far from the average conditions of an Occidental in 1900, such an evolution (tending to the average condditions of an American today) is not impossible as long as no phenomenum causes it to stop (which is what will eventually happen, as the world is finite). If living as an American of 2000 is what the world aims at, emisions in 2050 would be almost double what they are in the "high" IPCC scenario !

when the "large" bracket of 1,2 °C to 6,2 °C (for the mean near ground air temperature increase) is given, half of the uncertainty is due to the various models used, and for the other half it is only the result of the various emission scenarios choosed.

And, at last, it must also be remenbered that no climate simulation for the future has even come to the conclusion that earth might cool, or even stay still, in response to an increased greenhouse gas concentration in the atmosphere.

Furthermore, whatever scenario and model we consider, the temperature rise is extremely brutal compared to the natural evolution of the climate for the past centuries (diagram below).

Evolution of the average near ground temperature, measured or obtained from various analysis from 1000 to 2000, and simulated for the 21st century for various emissions scenarios. The vertical axis does not represent the absolute value of the temperature, but the difference of the annual mean with the mean value for 1990. For example, in 1860 the mean near surface temperature was 0.5 °C lower than it was in 1990.

The red curve corresponds to the most probable value, and the shaded grey zone represents the error range (the actual mean temperature may therefore have been anywhere in the grey zone). Prior to 1860 the curve only represents the northern hemisphere mean temperature.

Beyond 2000, this figure shows the predictions given by various models, depending on the emission scenarios for the 21st century. Each colour line corresponds to a same scenario, with which an ensemble of 15 models is fed, and the median value of the projected temperature rise, all models mingled, is represented.

The brown shaded zone shows the extremes, that is the difference between the least reactive model with the lowest emission scenario for the 21st century (and that is 1.4 °C of global temperature rise) and the mose reactive model with the highest emission scenario for the 21st century (and that is 5.8 °C).

Whatever the case, the rise is much over what we have been accustomed to during the past centuries. And these projections do not take into account the carbon cycle feedbacks, that could add a couple degrees (Celsius)...

From IPCC, 2001

Then what happens beyond 2100 ? Does the climate fit to the roman calendar and celebrates 2100 his way, by returning to normal or stopping to modify ? Unfortunately, no. The greenhouse gases stay for a very long time in the atmosphere, and therefore the climate perurbation that we have put in motion will go on for millenaries no matter what, but, of course, the magnitude of this change still heavily depends on the amount of greenhouse gases that we will emit during the 21st century, as the above figures show well.

And, alas, even if we totally supressed the human emissions tomorrow morning, the greenhouse gas concentrations in the atmosphere would only slowly decrease, and therefore the greenhouse effect induced by these gases will also only slowly decrease. Even with an "extreme" hypothesis such as a brutal stop in emissions we will not see the pre-industrial level of greenhouse gases come back before several thousands years, if not more.

As a consequence, the highest temperature rise is reached much after the maximum greenhouse gas concentration is reached, and the 2100 values for the temperature increase represent only 50% of the absolute maximum to come later (cf. figure below).

For a given atmospheric CO2 concentration (when it ultimately stabilizes) this figure represents two mean temperature increases : the lower one gives the temperature increase in 2100, and the upper one gives the temperature increse at its maximum several centuries later.

For example, if we manage to stabilize the atmospheric CO2 at 750 ppm (which is already tremendously high) within a couple of centuries, the temperature increase will be comprised (with the model used here) between 2 and 3,5 °C in 2100 (careful : we are not at 750 ppm in 2100, but only when we stabilize, later on), and will be comprised between 2,7 and 7°C when the temperature eventually stabilizes, much later.

It is therefore easy to see that the average temperature will keep rising beyond 2100, and that the higher the increase in 2100 (the press only mentions this one) , the higher the residual one after 2100. It is also easy to see that the eventual increase can be 3 times the 2100 increase : troubles won't be over in 2101....

From IPCC, 2001

In other words, if we prolongate the simulations for a couple centuries, the temperature increase might well go over10°C when the maximum is reached if we emit massive quantities of greenhouse gases during the coming century.

Of course the longer the period of time, the more legitimate it is to question the simulation, but still we can reason the following way :

either the climatic system remains about what it is today with such a temperature increase. The prolongation of the simulation is therefore valid, since the model represents the climate as we know it, and such a maximum cannot be excluded. The possible consequences generated by such a temperature increase - probably never experienced on earth since our planet bears life - in such a short time would probably not allow the keeping of our present form of civilization, and it is questionnable whether it would allow a confortable life - or even the survival - of several billion people on earth.

or the climatic system profoundly modifies well before such an increase is reached, as a result of the unforeseen modification of one of its component. As a consequence, the model, which represents the present system, is not valid any more, and we jump into the unknown, which may be more pleasant than what we thought... but also much more unpleasant.

Please allow me to remind a last time that we have not much visibility on what a temperature increase of more than 1 to 2°C may generate, this situation being a "premiere".

Back to climate change index
Back to the top