When does a meteorologist become skeptical about the information that is provided by a model?

Question

I have to know this for a science project. It is about meteorology.

Answer 1

Good question. Whenever he or she knows the physics limitations of the particular model in question. Each model has such limitations such as its space and time scales, or the type of cloud physics that are being represented (called parameterizations), or perhaps its terrain model. The manner in which turbulence is parameterized is another consideration. The degree of predictability is very much related to the scale of the phenomena you wish to predict. Each of these limitations are involved in evaluating the degree of accuracy of the forecast model. This is an area where there is continual meteorological research.

Answer 2

When the model doesn't initialize. For example, lets say the model puts a cold front right smack dab through the middle of Des Moines, Iowa. As a forecaster you check the real time data and find that it's actually 30 mile farther west. That means that (depending on how fast the system is moving) the 12-24-48 hour forecasts are going to be incorrect, probably exponentially. So you throw that model out the widow and look at other tools.

Answer 3

We don't use just one model, we use several and compare them with each other. If three models give similar results and one gives a different one, it's likely that we will base the forecast on the result that comes from three models, rather than one.

We also keep track on what results the models have given over the last few days - if they give different results every day, that is a sign that the weather situation is difficult for the models, which means that in the end the forecasts are less reliable than usual.