NOTE: This story was originally written Friday.
With our new wind-flow technology we're able to illustrate why hurricane forecasting can be very tricky beyond a few days out.
Check out the two images below. The first one is a model comparison between the GFS and European model at 3 p.m Friday, looking ahead to Sept. 8. The second one is another model comparison, about 24 hours ago, again looking ahead to Sept. 8.
Notice the difference.
There are other computer models we use and multiple variations of these models also. This is why it's important not to concentrate on any one particular model, but rather get a complete story behind what's happening.
All computer models have at least some error, which in the short term usually isn't a huge deal. But when you translate those errors a week into the future, they can be off by hundreds of miles and the intensity can be wildly incorrect also.
Here at Bay News 9 we firmly believe in what is referred to as an 'ensemble model' approach. We look at all the iterations of the GFS and European models and well as other models and we come to a general consensus. It's like taking an average of all the models put together. This kind of approach has better results than just randomly picking one single model.
This is why it’s important for meteorologists to add experience and knowledge as we understand the biases in these computer models and their tendencies.
We will continue showing the comparison each day so you can see at home what we’re seeing. But keep in mind that the results will likely change 20 times between now and the end of next week when it really matters.
There won’t be any tropical threats near us this holiday weekend, but by the late part of next week we could have a hurricane close to our state.
It’s way too early to speculate on that so we’ll just calmly deliver ‘no-hype’ forecasts with data that we’re seeing on a daily basis.