How far out do these computer models project and have they been spot on in the short term.?

Question

I guess my point is this. In order for a computer model to be trustworthy wouldn't it have to predict climate very precisely in the short-term? At what point would a model become untrustworthy?

Answer 1

There are several predictions going out a hundred years or so.  They vary, depending on what we do about greenhouse gases.

Predicting short term weather is completely different.  In the short run, temperatures jump up and down a lot unpredictably.  The data is "noisy".

But, over the long run, the average temperature is very predictable these days.  This data shows it nicely:

http://data.giss.nasa.gov/gistemp/2007/

discussed in detail, with confirmation, at:

http://tamino.wordpress.com/2008/01/24/g...

This situation occurs often in science.  A classic case is radioactivity.  Put a Geiger counter over a radioactive substance and the noise from the counter seems completely random, as individual atoms decay.  But we know exactly when one-half of them will decay, the "half-life".

Answer 2

Since they can't reliably predict weather even 3 days ahead in even a limited way, such as a particular region or city, how in the world could they predict global temperatures 100 years out? It's simple, they can't. They use computer models that are unproven and when the data doesn't fit they hammer at it and 'correct' for new variables until it does fit their predictions.

The main argument is that CO2 is inextricably linked to global temperatures when even a cursory look at the long-term record show it's not true. See http://www.geocraft.com/WVFossils/Carbon... The only relation is that they sometimes intersect, CO2 levels are sometimes far higher than today, but almost never lower. If they did drop very much, all plant life on Earth would die because we are at the extreme low end of historic levels of CO2. Before the Industrial Age it was barely high enough to provide for plant growth. If you don't believe me, click the link above. Temp doesn't seem be steady for millions of years at a time while CO2 levels fluctuate from 10 to 18 times the current level and  temp doesn't go up just because CO2 does, despite what you constantly hear on tv.

They've already downgraded the Little Ice Age and Medieval Warming Period to 'local events', even though most of the world enjoyed the warm period (2 degrees warmer than today) and suffered during the Little Ice Age (2 degrees cooler). Maybe they'll eventually get around to fossil records and those records will be 'adjusted' to show CO2 always drives temperature, high CO2 always means high temp (despite high levels during past ice ages), and that we're destroying the world.

Why does James Hansen still work for NASA? How much manipulation of the numbers can he get away with? Even the raw temp data is 'massaged' until it fits predictions or they relocate monitoring stations from parks to parking lots, which tend to be much warmer and then fail to adjust for that difference. 1934 was the warmest year in the US but the numbers have been changed over and over until now it's a 3-way tie between 1934, 1998 and 2007. You'll never hear a AGW advocate mention 1934 of course, or that six of the warmest years on record were in the 30s, mainly because GISS and James Hansen have now 'corrected' the numbers so many times that they only reflect what he wants them to show.

Answer 3

Scientists generally extend climate models to the year 2100, as you can see in some examples here:

http://en.wikipedia.org/wiki/Image:Globa...

Models so far have been very accurate.  James Hansen (head of NASA Goddard Institute of Space Studies) made a climate model in 1988 which predicted the subsequent warming to high accuracy, as discussed here:

http://www.realclimate.org/index.php/arc...

Answer 4

climate models are quite accurate over the short term but the problem is there is nearly no change in temperatures over a few years and unpredictable events like volcanoes can affect global temp for a few years.

Answer 5

You have to understand the way these models are done. They run them, then tweak them to match previous temp. But I have yet to see any that are "spot" on, much less close at actually predicting the future. They are great at predicting the past (one it is the past). Of all the models that are out there, some are bound to be right for a short period of time, while most are wrong. I bet none of them predicted the extreme global temperature drop of recent.

With this single drop, we are now back to temps of 100 years ago. So much for global warming.

Answer 6

The first link shows trends to 2100.

The second is a 2005 study showing models spot on.

Third link is a GISS model study in 2007.

Answer 7

To predict the average temperature on Pacific regions known as Nino 3, Nino 4, Nino 3.4, ......one year ahead the accuracy is about 0.8 degrees C

Answer 8

Models are based upon assumptions.   Dr. Hansen's model had three assumptions and three different results - if CO2 levels continue to increase, remain flat, are reduce, etc...

The model output can be explained like this:

If A then X, if B then Y, if C then Z.

Assumption A occurred, but result C occurred.

So while result C was "spot on" with one of the scenarios, it wasn't "spot on" with the scenario that actually happened.

This enables people like Dana to say "Dr. Hansen's prediction was spot on."

But it's a bit like my saying if Chara sustains a season-ending injury the Bruins will lose 15 out of 20 and finish out of the playoffs again but if he stays healthy the Bs will make the playoffs - if he says healthy but the Bs still lose 15 out of 20, I wasn't right.

Answer 9

NO. They can not even show past known temps and climates accurately so they are a waste of time because they can not compute for water vapor or solar patterns from the SUN.