Weather forecasting question...

[EsoCyn]

Why do the Weather Channel's forecasts and the National Weather Service forecasts vary so much? Like, anytime there's a major event, the NWS and the Weather Channel seem to forecast completely different. Is it a difference in methodology (if there is any) or guidelines?

For example, today the NWS is forecasting for Austin, a 20% chance of storms the entire week with little to no possibility of severe weather. But the Weather Channel, on the other hand, is forecasting a 40% chance of storms with a possibility of some being severe today, and a 30% tonight and tommorow. Likewise, the Storm Prediction Center isn't saying anything about even a slight risk of severe storms in my area until, at least, tommorow (which then, is like a 10% risk.)

So, if I may, what the fuck?

[Grommitt]

The National Weather Service is required by Congress to adhere to specific accuracy parameters. Perhaps they hedge their bets to stay within the mandated numbers? According to this site (http://online.wsj.com/public/article/SB115816640287562053.html), NWS has the least accurate forecast of those sampled.

the rankings for overall accuracy for this year thus far: The Weather Channel (right 73% of the time), followed by MyForecast (72.3%), AccuWeather (71.4%) and Intellicast (71.4%), and the National Weather Service (71%). But accuracy in your area may vary.
Only 2% difference between the most and least accurate.

[David B]

The National Weather Service is required by Congress to adhere to specific accuracy parameters. Perhaps they hedge their bets to stay within the mandated numbers? According to this site (http://online.wsj.com/public/article/SB115816640287562053.html), NWS has the least accurate forecast of those sampled.

Only 2% difference between the most and least accurate.

And all of them way better than 50-50.

David B

[GrandpaMithras]

Weather channel comes out with their own forecasts, so does accuweather. They have the fortunate ability to look at various models and NOAA forecasts and then make their own. I've been in this game for some time, and I can tell you that site might be right about say temperature forecasts, but when it comes to severe weather trust the NWS 100% of the time. They know what they are doing, and this is what they concentrate on. Plus there are no commercials.

The Weather Channel also hypes things.

[GrandpaMithras]

Oh! And coverage areas. When they're forecasting rain its not 20% chance that it rains, its 20% of coverage area expected to get rain. Otherwise how could you verify the forecast? So, if the weather channel has a different geographical area they're talking about than the NWS, that could explain the science behind the different results.

[Silent Acorns]

Oh! And coverage areas. When they're forecasting rain its not 20% chance that it rains, its 20% of coverage area expected to get rain. Otherwise how could you verify the forecast?

This is not what "20% chance of rain" means.

One key to an accurate forecast is accurate knowledge of the current conditions. Unfortunately, we never has as much knowledge about the current conditions as we would like, especially when it comes to predicting highly sensitive things like rainfall. So what can forecaster's do?

One technique called "ensemble forecasting". Basically, you randomly nudge your imperfect current conditions data and run the computer forecast several times (a different random nudge for each run). If you make 10 runs and 2 of them predict rainfall, then the forcaster says that there is a 20% chance of rain.

Edit: although, perhaps ensemble forecasting isn't as widespread as I think. Upon further reflection, your explanation is reasonable too.

[GrandpaMithras]

I was incorrect.

http://www.weathernotebook.org/transcripts/1998/09/25.html

Thanks Silent Acorns!

[Breath]

For example, today the NWS is forecasting for Austin, a 20% chance of storms the entire week with little to no possibility of severe weather. But the Weather Channel, on the other hand, is forecasting a 40% chance of storms with a possibility of some being severe today, and a 30% tonight and tommorow. Likewise, the Storm Prediction Center isn't saying anything about even a slight risk of severe storms in my area until, at least, tommorow (which then, is like a 10% risk.)

I'm just wondering what practical value there is in predicting a single event with a 20% or 40% chance of it happening. Either a storm occurs, or it doesn't.

[Silent Acorns]

I'm just wondering what practical value there is in predicting a single event with a 20% or 40% chance of it happening. Either a storm occurs, or it doesn't.

But it could easily miss or overshoot your location. One of the most difficult things to predict is the exact moment a storm cloud will actually start to produce rain because the mechanisms that trigger rainfall are functions of internal cloud fluctuations that are smaller then our instruments are capable of measuring.

Without perfect information, you can't make a perfect forecast. All we can do is make a probabalistic forecast, especially for things as unpredictable as rain.

[hclincha]

Rain isn't that unpredictable. If you were to ask me what weather was going to be like here tomorrow, and I said rain, I'd be right 2 times out of 3.