Chris Peterson wrote:sid113 wrote:So, how do we test the models? Well, we could run them backwards and see if they predict the past (they all fail).
Where do you get this. In fact, the models all do a reasonable job of predicting the past over the ranges they are designed around (that is, there are long range models- thousands of years or more, and short range models, centuries).
Well, okay, I have not looked at the exact, fully functional model outputs, and run the models through my own systems to completely validate all of the data. However, for the dozen or so models I have looked at, they show climate variations that more or less meet the variations that we see in the data. However, the error bars in the models are as big as or bigger than the error bars in the data. And the error bars in the data, when we get more than 200 years in the past, are approaching or exceeding the level of temperature change we are looking in the future.
For example, the medieval warming period in the 1300-1500 time frame was warmer than the temperatures now. Unfortunately, we do not know exactly how much the world-wide temperature average was above today, with estimates the I have seen in the 0.2-0.9 degree range. So, models going back to that time show temperatures back then with error bars in the 1-2 degree range. So, while the model can say they "agree" with observation, their errors dwarf the observations so how can the model be said to work back then.
We could test to see if the predictions on data from 1900-1999 entered into the model will predict accurately the results of 2000-2009 (especially 2008-2009 timeframe). As far as I know, they all fail with this test as well.
Again, I'd ask where you get this. On the whole, it makes little sense as the time frames are associated with weather, not climate. No model can be used to test if a particular year or two fits. Even a single decade is barely within measurable trends.
[/quote]
I agree with you - a few years, or even a few decades does not a climate model make. However, politicians and many scientists would argue that the 2-decade warming trend of the 1980's and 1990's do so a trend - a trend so strong they are predicting a 2-78 degree temperature increase in climate over the next 80 years. A pretty bold prediction when the 4 previous decades had shown a bit of a cooling trend culminating in the 1970's cooling such that many scientists of the time predicted a new ice age.
Yes, the last two years were quite a bit warmer than the midline of the models suggests. But those sorts of fluctuations are completely normal, and meaningless within the context of climatic change.
I am not sure where you get this data. It turns out that the original data posted by NASA for 2006-2008 was stilted due to local changes for many of the measuring sites. When sites where moved, and the data were corrected, 2007 and 2008 showed a temperature drop greater than anything expected in any model. However, the drop was consistent with the fact that the solar output (as indicated by sunspot activity) was the lowest in 90 years.
All of the major models include the Earth's dynamics- orbital variations, precession, etc. All of the major models include solar cycles.
While they may take these various processes into account, the question is do they weigh them enough or do they apply a greater weight to CO2 then it deserves. Considering that in the past CO2 levels were up to 20 times greater than today, yet the temperatures are on the order of the predicted temperatures from many models (5-7 degree increases) and the models are looking at a doubling of CO2 levels, it makes one wonder about the models.
They do not include volcanic activity because it is unpredictable, and for short term climate- a few centuries- almost certainly unimportant.
Interesting. So, a process where a single event may affect weather over many years to the tune of a few degrees is not worth considering, yet models that are predicting changes over the next decade must be heeded, even though the error bars grow for each decade the model extends. Consider when Krakatau exploded, it sent so much aerosol into the atmosphere it reduced global temperatures by an estimated 1.2 degrees for at least 5 years, and maybe longer. Granted, a Krakatau or Pinatubo are relatively rare, but there are a number of smaller volcanoes that put both significant CO2 and aerosols into the atmosphere - all of which greatly increase the error bars of the models.
Naturally the models do not include effects that are not yet understood. A big part of climate research involves trying to better understand geophysical processes so they can be included or improved in models. That's the "good science" bit again.
There is nothing wrong with continuing the study to make the models better and better. However, considering not only are the geophysical processes not fully understood, but the biological processes are also not well understood. For example, I was reading an article (I wish I could find it again) that indicated as CO2 levels increased, plant growth and thus plant absorption of CO2 also increased, resulting in a much less increase in atmospheric CO2 concentrations. Also, considering that CO2 levels in the atmosphere have not grown as fast as the models (and actually decreased over the last few years, even though humans are still using fossil fuels), maybe the models need to adjust their biological assumptions as well.
I would add that we're mostly concerned with short term climate change.
And now we come to the true problem with all of the models. The long-range actions seem to dwarf the CO2-based elements in the real climate by a significant amount. So, trying to change affect the climate with CO2 legislation when CO2 concentrations may have a relatively negligible affect is a waste of time and money. Yet it is this very outcome that the UN wants - and it wants the effort to kick in to the tune of several trillion dollars over the next 10-20 years.
This is an example of science that is continuing to evolve being used to effect policy when there is little or no basis in the science to support the requested policy changes.
I wonder what Galileo would say to public policy being made based on incomplete scientific study and analysis.