sparr: (Default)
Clarence "Sparr" Risher ([personal profile] sparr) wrote2009-03-12 06:33 pm
Entry tags:

Useless forecast

Pretty much every weather forecast site got today wrong, by around a hefty 10°F.

My question, then, is why leave that worthless information posted?  At some point around noon today it was 65° out, already well past the forecast high of 61°.  At that point, what possible need (other than poking fun at meteorologists) could I have for the existing forecast?  Why does [apparently] no one provide an updated forecast during the day, based on information from the earlier parts of the day?  They already produce a new forecast for Thursday on Wednesday night, and on Tuesday night, and a number of other days stretching back at least two weeks, with better certainty based on more data each time.  Why can that process not be extended another 6 to 12 hours?  With no advanced algorithms or additional data, I can tell you with better than 95% certainty that if it is 61° at 11AM then the high for today is going to be at least 65°.

At least WeatherBug has a "So Far Today" section.  Not what I want, but a small step in the right direction.

[identity profile] malachus.livejournal.com 2009-03-13 04:54 am (UTC)(link)
It's probably not worth the money. I'm guessing the usage of weather information has a couple of predictable spikes.

1) Early morning, when people are trying to figure out what the day will be like.
2) Before or during severe weather, when people are trying to find out how bad things are.

There may be some other, smaller spikes, like late afternoon as people check what things are like before they head home or out, but generally it probably doesn't matter nearly as much, since people don't tend to have the option to change how they are dressed/prepared at that point.

In other words, probably not enough people care to justify it.

[identity profile] sparr0.livejournal.com 2009-03-13 06:34 am (UTC)(link)
Define "early morning". At 6AM I would bet there was enough new data to produce a far more accurate forecast than the one we got stuck with at some time Wednesday night, and in plenty of time for the majority of people to see it.

[identity profile] malachus.livejournal.com 2009-03-13 12:23 pm (UTC)(link)
Yeah, but most people only care about the general pattern. Hot/not hot. Rain/Snow/Not rain or snow.

Basically, "Do I need my sweater/heavy coat or not?" and "Do I need my umbrella or not?" The updated forecast would probably rarely change the answer.

[identity profile] sparr0.livejournal.com 2009-03-13 03:40 pm (UTC)(link)
The forecast was 61/rain. The actual day was 75/dry. Rare, yes, but it happened.

[identity profile] malachus.livejournal.com 2009-03-14 02:50 am (UTC)(link)
Rare enough that the cost/benefit analysis doesn't justify spending the resources...