It is true that models are not perfect. The problem is that we don't have enough supercomputing capacity to refine them.
Let me compare CC forecasting with weather forecasting (two
entirely different things, BTW, so don't confuse them). With weather forecasting, the grid that is used is far too coarse, typically 100 km x 100 km with 15 slices to the tropopause, say, 800 to 1,000 m each. Local topography is largely ignored with such a large grid. However, the modelling from the data obtained is analysed, changes being weighted to have more effect, and past experience is also used to factor in some of the gaps. The result is that we have a 24 h forecast in most places that is >95% accurate in event/place but is less so in time, with errors of ±10 hours being fairly commonplace. Of course, the longer the forecast, the less accurate it becomes, so that at 5 days it is little more accurate than looking at a bunch of seaweed or asking Aunt Emily how her arthritis is feeling (except under anticyclonic conditions which are somewhat more predictable).
With CC forecasting, we don't have to forecast over 24 hours or 5 days, but over 50 or 100 years. The way this is done is to analyse the global temperature over the last 5000 years, but especially over the last 150 years, when positive data was being collected. We also collected data for all the natural cycles, volcanic eruptions, major forest fires, weather anomalies, El Nino activity etc. Finally, we collected data on the atmospheric gas analyses. Putting all these together, a model was created which gave a remarkably good fit between global temperature and all the known variables, especially over the past 150 years. (See my website at
http://www.cypenv.org/Files/climate_change.htm for graphs.)
Now, what does this mean for the present day? Firstly, there is a
prima facie case that human activity is very contributory to global climate. Many men have been hanged on weaker evidence than this. Secondly, the fact that the fit between the grey and red curves does allow for a certain margin of error, which is known.
And what does it mean for the future? Firstly, we have to forecast scenarios of different rates of different GHG emissions; things like worst case where we burn up all the fossil fuel in n years, best case where we stop emitting GHGs and a number of intermediate scenarios such as with Kyoto in force as it stands, increased measures and so on. We then have to factor in all the natural cycles (obviously, we cannot factor in sporadic events like volcanic eruptions and forest fires, but we can make some educated guesses regarding semi-cyclical ones, like El Nino). From these data, we have to estimate the influence of climate changes on ocean currents, which are the most influential factor governing climate, so that models have to contain variables of positive and negative feedback effects, changes of albedo due to increased evaporation and lowered ocean salinity and so on. Combined with the margins of error, we can calculate the maximum and minimum effects under each scenario.
These models are constantly being updated. What I find interesting is that all the models forecast a global temperature rise due to man-made emissions between now and 2100, even with the minima and a reduction of actual emissions. In some cases, this may be as low as a fraction of a degree, compared with today. The other extreme is over another 10°C rise. Most scenarios oscillate between means of 2°C and 6°C, which are the most likely effects.
That having been said, there is a worrying development; one of the factors is the melting of ice, both polar and oro-glacial. Over the past two years, this has progressed faster than anticipated, in both cases. This has a distinct positive feedback influence and it would seem that the models need a correction to compensate for this. This would give higher global temperatures at a given time but I cannot say by how much at this moment.
The effects of climate change? At this time, we cannot forecast the effects likely at any given place. All we can say is that, as the global temperature rises, we shall definitely see more weather extremes, which includes cold, as well as heat, and more violent episodic events, such as hurricanes, tornadoes, thunderstorms etc., but we don't know where and when. There is anecdotal evidence that this may already have started in recent years.
Finally, I have two quotations for the naysayers:
Although scientific evidence that human activitieswere causing stratospheric ozone depletion was quite robust in the late 1980s, there were a number of sceptics who said, "wait for perfect knowledge; there is uncertainty in the ozone models." Unfortunately, the sceptics were absolutely right. The models were inaccurate. They underestimated the impact of human activities on stratospheric ozone. This means that even with the Montreal Protocol and its adjustments and amendments, society will have to live with stratospheric ozone depletion not only over Antarctica, but over all of the globe, except for tropics and subtropics, for at least another 50 years. Some of the same sceptics are now saying that not enough is known about climate change.
Robert T. Watson, Co-chair Scientific Assessment Panel, Montreal Protocol, former Chairman, Intergovernmental Panel on Climate Change. See
http://www.environmentaldefense.org/doc ... erview.pdf for a short interview of Dr Watson.
Where there are threats of serious or irreversible damage, lack of full scientific certainty shall not be used as a reason for postponing cost-effective measures to prevent environmental degradation.
UNEP: Rio Declaration, Principle 15 (signed by many world leaders, including Pres. George Bush, Sr.)