Friday, March 25, 2011

Discrepancy in Computer Models Leaves us Guessing

You rarely hear me say on the air "I have no idea what the weather will do this weekend or next week." After all, that is my job, right? To forecast the weather is to know the weather and how all the forces in the atmosphere influence weather features and weather patterns.

When meteorologists like myself make forecasts, we often come away with two modes of thought. Either we are fairly confident on how the forecast will play out as predicited, or not very confident at all.

You can be confident in your forecast, but that does not guarantee that you will be accurate. The other scenario is true, however. We may make a forecast based on our instinct, or gut feeling, and could be dead on, even though we did not have high confidence from the start.

So where does our confidence in making a forecast come from? The answer is simple: the computer models that we look at on a daily basis.
A computer model is exactly what is says, a model. There are always errors in computer models due to the bad data that goes into them from the start. We also have several models to view and they each give us different outcomes.
The problem we run into is when the different computer models we look at don't agree with each other. When that happens, we say there is uncertainty in the forecast.
This is the problem we have for our weather next week, where the different models are giving us different solutions on the timing of cold fronts, and the likelihood of precipitation.
To add to the issue, the computer models I've looked at the past couple of days have been flip flopping. In other words, If model A shows solution A and model B shows solution B, those are different solutions. But when they flip flop, model A may show a solution similar to solution B and vice versa.
Alright, I'm sure you are confused by now. What I'm getting at is the discrepancy in the computer models makes a forecast very challenging. That's where going on experience and instincts takes over. It's what also makes our job more challenging.