Accuweather, the Pennsylvania-based private weather forecasting company, is no stranger to criticism from the meteorological community. Their website provides more sensationalism than forecasts and in general is held in very low regard by the profession.

They do some things quite well, however. They’re a money making machine and their forecasts tailored to some specific clients are about as good as they get. The radio forecasts you hear on WBZ in Boston or WCBS or WINS in New York are generally superb. Don’t expect to find a superb forecast on their website or from one of their sensational columnists, however.

In an effort to push the envelope and draw eye balls to their website they’ve begun to issue 25-day deterministic forecasts for locations across the country. Precipitation forecasts to within a hundreth of an inch and exact temperature forecasts as well!

Here’s the forecast for April 28th in Hartford.

Only 0.05″ of rain. Not bad! And not too breezy either – only WNW at 6 mph.

What a joke.

There’s a way to provide forecasts beyond day 10 to customers that are of some use. This is not the way.

Being able to express uncertainty in any forecast is very important, particularly when the forecast goes much past a day or two. Expressing uncertainty (both high and low confidence) is something that most meteorologists aren’t very good at. It’s difficult to communicate uncertainty whether it’s about a tornado warning or a 5-day forecast or a snow forecast.

Using a probabilistic over a deterministic forecast is one of the only ways to provide a forecast of any use beyond day 10 and especially beyond day 15.

Unfortunately this 25-day forecast is just a gimmick. It’s a stunt to grab publicity and eventually drive traffic to their website. Things like this make the profession look bad and are of absolutely no use to anyone.


10 thoughts on “AccuFail

  1. I sort of give them credit for putting this out there and potentailly taking heat oif their looooong range forecasts busts…if indeed anyone will remember. Maybe somebody can do some loose tracking of their prediction on days 15 -25?

    • I would give them credit as well if they were able to do it in a way that had any skill at all. Saying that there will be 0.02″ of precipitation 25 days from now is ridiculous. There’s virtually no skill and no use to that kind of forecast. Finding a creative way to display a probabilistic forecast for weeks 3 and 4 could be useful and I wouldn’t have a problem with it.

  2. The most important thing to remember about forecasting the weather is that all weather is caused by man-made global climate change. Before people invented the internal combustion engine, there was no weather. It was like living in San Diego year round.

Leave a Reply

Please log in using one of these methods to post your comment: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s