Which weather site's best?

January 16, 2004 8:21 AM

Which of the major online weather sites are most reliable? (more inside)

I'm getting conflicting forecasts from Weather Underground and Weather.com (and fairly useless ones from the local media outlets' web sites). Is one typically more reliable than the other? is there a better place I should be checking? Sort of a burning issue because I'm driving from Mpls to Omaha tonight and want to figure out where there'll be an ice storm and where there won't be…

posted by COBRA! to science & nature (9 comments total)

If you live in the USA, I suggest sticking to Weather.com and NOAA
posted by riffola at 8:28 AM on January 16, 2004

If you REALLY want to know the weather, get onto the NOAA site and begin by reading the TEXT discussions in your area. Then, go to Unisys and watch the 9-days on a daily basis as well. Doing this, you will be far better at getting it right than either weather.com or your local weather news guy. Guaranteed.

/end geek
posted by BrodieShadeTree at 8:47 AM on January 16, 2004

I suggest NOAA too. However, I recently downloaded and installed a cool app for OS X (Meteo) that had the option of using Wunderground.com as a server, and I have found it to be most useful.
posted by terrapin at 8:51 AM on January 16, 2004

I Like Intellicast, the radar loops are accurate, in my experience.
posted by yoga at 9:58 AM on January 16, 2004

Weather.com is awful, generally.

Here's my credentials:

I have no credentials.

But, I took a meteorology 101 class, and part of our grade was forecasting the weather every day for 10 weeks. In a class of 400 (most of whom probably didn't care, granted) I came out 1st in accuracy. My method was this:

Take the NOAA forecast for the day for the nearest location and assume it will be correct.

Then go to a couple different sites, like say Unisys, Intellicast, your local paper, CNN.com, or even, yes, Weather.com. For temperature and pressure, take the average difference between these sites and NOAA, and apply it. For conditions, change NOAA's prediction slightly if there is a consensus among the other sites.

But, to answer your question, NOAA tends to be best, in my experience.
posted by Hildago at 10:31 AM on January 16, 2004

For temperature and pressure, take the average difference between these sites and NOAA, and apply it.

Do you mean you took the average difference between the others and NOAA and added that to NOAA? If so, then your result is just the average of the others (ignoring NOAA).

z+sum_i(x_i-z)/n = z+((sum x_i) -z*n)/n = z + (sum x_i)/n - z = sum (x_i)/n
posted by andrew cooke at 11:36 AM on January 16, 2004

Yeah, you caught me Andrew -- I didn't actually take the average, or do any math whatsoever -- I just sort of adjusted the NOAA score based on my perceived difference in credibility between the different sites, and how they differed with NOAA. No math is involved in my arbitrary, ridiculous system.
posted by Hildago at 1:15 PM on January 16, 2004

www.weatherbug.com
posted by Fupped Duck at 3:53 PM on January 16, 2004

Cool suggestions, y'all. I downloaded meteo, thanks terrapin!
posted by carter at 5:39 PM on January 16, 2004