The Figure below shows that Paleoclimate evidence gives some guidance of the sensitivity envelope of the earth's climate, and therefore, the range of possible temperature increases we might experience as we increase CO2. However, there is no direct analogue of the current increase in CO2. CO2 changes during the Ice Ages were forced by changes in the latitudinal distribution of solar radiation at the earth's surface from changes in the earth's orbit around the sun.
A doubling of CO2 would give rise to an initial decrease in LW emission to space of 4 W/m2. If the absorbed solar radiation remained constant, and there were no other changes in the atmosphere (water concentrations fixed, no changes in ice cover, clouds), there would have to be a surface temperature increase of about 1.25 C to bring the earth's energy budget back into balance. This 1.25 C increase is called the "No Feedbacks" response.
However, most models and the paleoclimate evidence suggest that the climate system has strong positive feedbacks. The two main positive feedbacks are the water vapor feedback, and the ice albedo feedback. This sign of the cloud feedback is not yet determined with certainty. However, the final steady state warming from these feedbacks likely gives rise to a temperature increase on the order of 2.5 C.
The main climate uncertainty is the role of clouds. How will the cloud distribution change in the future? Will the positive LW (greenhouse) forcing of clouds increase or decrease? Will the negative SW (solar) forcing of clouds increase or decrease? This is partly a cloud height problem. If average cloud height increases in response to a CO2 doubling, one would expect the cloud feedback to be positive. However, for example, if low-level boundary layer clouds or fog increases in response to a CO2 doubling, one would expect clouds to have a negative (stabilizing) effect on climate change. It is hard for climate models to simulate cloud cover accurately, since most clouds are so much smaller than the grid size of climate models (about 100 km).
Note that the sign of the overall current effect of clouds on climate does not determine the sign of the cloud feedback. For example, clouds in the current climate likely cool the earth when one considers both their LW and SW effects. However, this negative forcing could get smaller if you double CO2. In this case, the cloud feedback is positive, so it amplifies the climate response to any forcing.
An obvious similar case is snow and ice. Snow and ice increase the short wave (solar) reflectivity of the planet and therefore cool the earth. However, the ice-albedo feedback is positive.
The size of a feedback depends on the climate system. During warm periods in the past, where there was no ice or snow, the ice-albedo feedback would not have existed.
Examples of forcings: human induced emissions in CO2, changes in CO2 due to plate tectonics/volcanism, changes in the earth's orbit around the sun (tilt, eccentricity of orbit, etc), the solar cycle, volcanic eruptions. These changes drive climate change; they are not a response to climate change.
In the current climate, the CO2 changes are considered a climate forcing: they are not an internal response of the climate to something else. Orbital changes and volcanic eruptions are also examples of forcings. However, sometimes a change can be a forcing in one situation, but a feedback in another. The changes in ice volume during the ice ages were forced by orbital changes. However, CO2 tended to go down during the cold intervals and up during the warm intervals. In this case, the CO2 changes were acting like a feedback, and since they would have been amplifying the orbitally induced temperature changes, they would have been a positive feedback.
We don't have a very good understanding of why CO2 went down when ice volume increased, and vice versa. It probably had to do with an increased effectiveness of the biological pump during cold intervals.
The relationship between CO2 and temperature changes during the ice ages is not a good analogue for how we might expect current changes in CO2 to affect future climate: the ice age CO2 changes were a feedback response, whereas now CO2 is a forcing. However, the long slow tectonic decrease in CO2 over the past 100 million years was a "forcing".
Tropical/Monsoonal (India): wet/dry seasons with little temperature seasonality. Monsoon associated with northward movement of ITCZ in NH summer. Monsoon outbreak cools temperatures a little.
Desert Subtropical (Arizona):
Maritime (Norway) : Appreciable rainfall during all seasons due to proximity to ocean. Ocean also reduces temperature seasonality, despite high latitude and huge seasonal variation in TOA SW energy.
Continental (Moscow, Colorado, Alberta): High seasonality in both rainfall and temperature; most precipitation in the summer; no temperature buffering from the oceans; too cold, and too far from a moisture source to have much precip in the winter. However, any snow that does fall tends to accumulate, so snow depths can be significant.
Mediterranean regime (Naples): higher rainfall in winter when the storm track is closer; dry summers since in the subtropical descending zone of the Hadley circulation during this season when the storm tracks have moved into northern Europe.
"Mediterranean" climates would be expected to occur in roughly the 30 N - 40 N latitude band, since 30 N is perhaps the furthest south you might expect winter rainfall to be enhanced by the winter storm track, and 40 N might be the furthest north you might be expect to be influenced by the northward movement of the Hadley circulation in NH summer (and so have reduced rainfall due to having subtropical descent). However, there are many places in this latitude band that do not have a Mediterranean climate. For example, the southeast US and Florida have significant rainfall in the summer, (and some of the strongest convective storms on earth) despite being at that time of the year "subtropical". Nice, (in southern France) is at 45 N, and perhaps even Victoria BC (50 N), could be considered to have Mediterranean climates (hot drier summers, cool moist winters). The dry summers are not so much due to being in the descending zone of the Hadley circulation, but downstream of a high at this time of year (Pacific and Bermuda high respectively).
The climate at any location is determined by a multiple of factors in addition to latitude: proximity to the ocean, temperature of the ocean, altitude, the existence of nearby mountain ranges, and its location relative to nearby climatological highs and lows. For example, the lack of summer rainfall in Nice would be partially due to it being influenced by subsidence (sinking) from the Bermuda High. This sinking aloft would heat the atmosphere, and tend to suppress the development of convective storms. Similarly, being downstream of the Pacific High in summer would reduce the summer rainfall of Victoria, and make it effectively "subtropical" at this time of year (of course the Pacific Ocean prevents Victoria from having the classic hot Mediterranean summer). Rainfall in the southeast US is strongly affected by the heat low that develops over the interior of North America during summer. This heat low, in combination with the Bermuda high, helps transport warm moist air from the Gulf of Mexico northward and moistens the eastern half of North America. Italy is a long narrow peninsula with quite high mountains, surrounded by warm Mediterranean waters: you might think that these factors would favor the development of a convergent sea breeze circulation and orographic thunderstorms over Italy in summer, and increase summer rainfall in Italy. Apparently, this tendency is sufficiently offset by subtropical and Bermuda high subsidence, which would again tend to suppress the development of thunderstorms, so that it doesn't increase summer rainfall by very much. However, Florida has a very strong convergent sea breeze circulation in summer which gives rise to afternoon thunderstorms almost every day.
This figure shows that the earth has warmed up by about 0.6 C over the last 30 years, or about 0.2 C/decade. This rate of warming is roughly in line with the rate of warming climate models project will continue for the next 100 years (at present CO2 emission rates). The rate of global warming will likely accelerate if CO2 emissions go up, sea ice and/or Greenland ice goes down faster than expected (appears to be the case), and cooling from sulphate aerosols goes down (i.e. less burning of sulphur rich coal). Also note that, in this record, the warmest year to date has been 1998 (other records suggest that 1998 and 2006 were basically tied). 1998 was about 0.2 C above the smoothed curve suggesting that the 1997/98 El Nino warmed the earth by about 0.2 C. This is about equal to one decades worth of current CO2 induced global warming (we believe), so one would expect the 1998 record to be convincingly beaten in the next few years, especially if there is another El Nino event, as the trend in surface temperature from CO2 overtakes that expected from a big El-Nino. Some projections suggest that 2007 will be higher than 1998. We will see.
The reason an El Nino warms the earth is that the thick layer of warm water that builds up in the western Pacific during La Nina years spreads (sloshes) back the central and eastern Pacific. This increases global average SST's, which then increases temperatures in the global atmosphere (with some time lag of course). Anyway, it is important to see the future global warming as a slow trend (on our timescales, fast compared with geological timescales), which is superimposed on faster natural variability. There is always a multiplicity of both natural AND human processes affecting climate, and it is extremely difficult to disentangle them. CO2 induced global warming can contribute to a trend in global temperatures but does not by itself help you understand questions like why 2007 might be colder or warmer than 2006. The year-to-year noisy interannual variability in climate will always be dominated by the ENSO phase, volcanoes, solar induced variability, other facts we do not understand, as well as just "chaos". Similarly, higher levels of CO2 in the atmosphere will tell us nothing about the year to year variability in hurricanes - changes in CO2 will only have the ability to affect the long term trend in hurricane frequency (a very controversial topic ...).
None of these lines of evidence by themselves would constitute a "proof". There is no way to absolutely disprove the possibility that the current climate change is "just" natural variability. Further, models are easy to criticize, especially with respect to how they represent sub-gridscale phenomenon like clouds. On the other hand, we do believe weather forecasts to some extent, climate (average weather) should be easier to predict than weather, and models have demonstrated skill in predicting the climate response to external forcings such as volcanoes. They must represent reality to some degree. Together, all these indirect lines of evidence make an extremely strong, but still circumstantial, case that increased CO2 levels are causing climate change.
Keep in mind, however, that as with all geophysical changes, the earth is not a system we can isolate and do an experiment as in a laboratory. We are running the experiment in real life. Critics of global warming often focus on weaknesses in one particular line of evidence, as if the theory would then collapse like a stack of dominos. But a better analogy is a jigsaw puzzle to which we are slowly adding more pieces, so the final picture gradually comes into focus. Or perhaps a spider web of interlocking self consistent arguments.
I believe this is a better image of how most scientific theories come to be accepted. There are some theories that are sufficiently theoretically compelling, or have sufficient predictive power, that all reasonable people are forced to immediately accept them. But this is the exception. In many cases, the experimental evidence is for a long time sufficiently ambiguous that it takes a generation (sometimes longer) for a theory to be accepted. (I am thinking of theories such as the theory of evolution, black holes, general relativity, quantum mechanics, and even issues one might think of as being fairly straightforward such as the charge of the electron, about which experimentalists took a long time to come into agreement.)
In the case of global warming from the impact of the current CO2 increases, I think we are now at the latter stage of this process. However, the details of global warming will continue to unfold over the next few decades (and longer) as additional observations come in, and the idea that the current climate change is due to natural variability will probably become increasingly implausible. Models will continue to slowly improve, but will always be imperfect. There is unlikely to be any Eureka moment when all disagreements suddenly finally end.
To quote a colleague: "There have been 33 deglaciations in the last few million years, and mammoths and mastodons survived all but one - the one in which we are around. Coincidence? And the last five glacial/interglacial cycles have been rather similar to one another, so it cannot be claimed that the most recent one was the most intense. Of course it may be that various species were fewer in number due to a harsher climate and that made it easier for humans to inadvertently wipe them out. But we are still the immediate cause in that case. I saw a talk on the diets of mammoths and mastodons just prior to their extinction. Mammoths ate C3 plants and Mastodons ate C4. C3 and C4 plants are very different, and their tissues have quite different C13/C12 isotope ratios. This is then reflected in the bones of the animals that eat them. When food is scarce, however, herbivores will eat plants they won't normally touch. So high browsers (like Elephants and, I think, Mammoths) will bend down to eat low plants which may be C4, instead of browsing on C3 leaves from trees. Similarly the C3 eaters will branch out to C4. So if hunger played a role in the extinctions, we would expect the carbon isotope ratios in the bones of mammoths and mastodons to converge as each began to eat basically the same diet. But they do not. That is pretty convincing evidence that climate change was not the primary reason for their demise. The California rock seal is not extinct but confined to a few islands off the coast. Evidence shows that before the advent of man, this had a huge on-shore population. Humans ate the lot of them, as we can tell from buried camp fires just packed with seal bones. And the fact that they were wiped out on land within a century or so of our arrival."
Similary in Nova Scotia, we once had a large onshore population of walrus (where the name Cow Bay comes from). They would have been ideally suited for digging up scallops. In this case, it appears that European settlers wiped them out. It is unlikely that humans were the direct cause of the extinction of large predators such as the giant bear and saber toothed tiger. These probably couldn't compete with a rival top predator which was killing their prey. It is curious that African lions survived the competition with humans while those in North America did not. It should be noted, however, that in many societies, the killing of a large predator (polar bear among the Inuit of Greenland, lions among the Masai) is associated with the transition to manhood, so direct human hunting pressure can't be ruled out. It is interesting that the the running ability of the pronghorn antelope in the American midwest evolved in response to a sabre tooth tiger that no longer exists. It can now run much faster than any current predator, so is wasting its speed.
For more, there is an interesting review article by Barnosky et al., Science, 306, October 1, 2004. You will find that many museums, including the Nova Scotia Museum in Halifax, continue to give what I think are outdated attributions of the mastodon extinction to climate change. Generally, predators which enter a land where large mammals have no pre-existing experience or fear of those predators, can be very successful, especially a predator such as man that had evolved sophisticated hunting weapons and social structures. This hunting success probably helps explain the rapid spread of humans throughout North and South America soon after their arrival. Of course, they then would have had to adapt to the elimination of the animals on which their earlier success depended.
