Return to Website

Number Watch Web Forum

This forum is about wrong numbers in science, politics and the media. It respects good science and good English.

Number Watch Web Forum
Start a New Topic 
Author
Comment
Global Circulation Model Confirmation Criteria

I may be repeating myself, but am I off base for suggesting the following for confirming GCM's.

Let the GCM predict each of the next 5 years. Layout their grid over the world. Gather the data related to the grid from available sources. Compare the results with the predictions.

Am I expecting too little?

Am I expecting too much?

Is anyone actually doing this?

All the results I have seen lately seem point away from their predictions. But even the fact that the predictions are wrong is used to support the "truth" of Climate Change.

Re: Global Circulation Model Confirmation Criteria

Not so fast there Brad ....


I read an interesting piece back an hour or so (just wish I could find it again ...) that pointed out that climate forecasts need to be tested over relatively long timescales (30 years min iirc) but not too long (100 yrs max iirc). But the current models' adjustment parameters blow up using those min and max periods and give runaway results.

Have been reading a lot this evening so the info provided may be off target and finding the reference could be a challenge.

I don't think 5 years would be conclusive for anyone in the wider debate.

That said I am completely in tune with the point you are making.



Grant




Grant

Re: Re: Global Circulation Model Confirmation Criteria

I love the "statistical" aspect of GCMs.

You can't report what is going to happen next year, but you can report what is going to happen on average over the next 30 years.

They don't match what has happened in the past except as an average of the past data.

I just keep thinking of a statement I heard a long time ago. If you want to predict what is going to happen in June in January, your best bet is to check the Farmers almanac.

I don't know how the FA is compiled, but at a guess it would be a compilation of past experience or colloquially the Average.

Do we design heat transfer system using the "average". Oops, we do, but the average we use for heat transfer is more readily quantifiable, and of an entirely different nature. It is not an average over several tens of years, but the average over several tens of centimeters.

The temperature at any given point will likely match the prediction in the model.

The GCM use the same kind of averages within the model. Only their cross sections aren't 10s of centimeters but 10s of thousands of kms cubed (yes I know volume is not a cross section but at the same time it is).

Isn't 30 years about the same period as the Pacific Decadal Oscillation?

I am fairly sure that I would have been fired from the project if I were on it. Rocking the boat does not keep your job.

TNT

Re: Re: Re: Global Circulation Model Confirmation Criteria

Analogy:

NASA is sending more vehicles out to the planets. Planning such missions, which typically use gravitation assist type orbits requires us to model the solar system accurately.

Would we send out a mission if the model needed 30 years to provide us an average to let us know where the destination "might" be?

What would be our chance of success.

The three body problem is complex, but it is infinitely less complex than an eco system.