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
Author
Comment
YouTube Video (Referenced in January 2011 edition of NW)

Near the end, they show a comparison of "model" prediction to "real" data.

I have to call BS on the model depiction.

There is no way their models are able to create definition that tight. I know that we can get Petabytes of data stored now, but to get the granularity they represent on that picture is still not feasible. Geometries of grid size make the data storage issue really challenging. The iteration length (how much time happens between each iteration in the model) is based on the size of the grid and how long it will take a piece of air to move from one side to another. A 100km X 100km grid means an interval of 1 hour (potentially). A 10km X 10km grid means an interval of < 6 minutes. Approximate calculations for the map shown in the story. US is approximately 30 centimeters. 5000km / 30 = 166 km/cm. A millimeter would be about 16km. (I don't think they are down to 10km cells though).

Hmmm. Maybe I am wrong on this. The image on the screen can be deceptive. I use to fret that may pie crusts looked nothing like the ones they made on television. I could never get my dough so homogenous. If I did, the crust would be really tough. I finally realized that the image onscreen was both color shifted and color blended by the process of transmission. Pixels get averaged.

I still think there is a bunch of misdirection going on.

Re: YouTube Video (Referenced in January 2011 edition of NW)

You're referring to the bit around 35 minutes in where the simulation and real data for movement of clouds was being compared on a giant split-screen. A lot of the BS was coming from Paul Nurse who was applauding and expressing delight at the comparison, but I didn't think the comparison being shown on the screen actually looked that good. Nurse is a geneticist and cell biologist, and probably has no experience of running or validating computer models. Also I suspect that a weather model might be being used rather than a climate model. Weather forecasting with computer models supposedly works reasonably well for a period of up to about 4 days from a given starting condition.

The only AGW sceptic blogger who seemed to notice that the comparison was a bit deficient was Lubos Motl in this blog post: