Today on Facebook I saw once again a criticism of
meteorologist: "Chance of showers today, they said....."
Yeah, it's an old joke (sometimes appropriate, but many
times not).
In my environmental stats class we looked at how well some
weather models perform. In the past several decades we have seen much
improvement in weather forecasting capabilities (positive trend in graphs shown
below). There is an annual variability in the forecast skill. Forecasts tend to
be better in the winter months and worse in the summer months. We can also see that the GFS, EC, and UK models trend together with the EC model being the best overall. The CDAS is the worst of the four.
Below shows how well several models have performed at weather
forecasting in the last month. A value of 1 would be a perfect forecast. First,
notice that none of the models are perfect. Second, notice that no one model is
more right than others all the time. That's why it's important to consider all models when making a forecast. This graphic can be found here.
Granted, these errors are based on 500 millibar heights, which is perhaps the easiest variable to forecasts. The error trends between models for precipitation or wind speed would look quite different and likely have much less forecast skill.
No comments:
Post a Comment
Note: Only a member of this blog may post a comment.