PDA

View Full Version : NOAA High for Today?



treant985
09-03-2007, 09:04 AM
The NOAA usually has a pretty good point forecast for Mt. Washington's temps, but for today they show:

"Labor Day: Mostly sunny, with a high near 80. West wind between 9 and 11 mph."

That's right--80 degrees. Here's the site, who knows how long it'll stay up like that: http://forecast.weather.gov/MapClick.php?CityName=Mount+Washington&state=NH&site=GYX

Their highs for the next few days seem right--around 50. But then they also show 75 for Friday.

Any idea why their predicted highs are so high??

Charlie
09-03-2007, 09:21 AM
thats in the valley?

Bill O
09-03-2007, 09:30 AM
thats in the valley?

Its totally messed up. Towards the end of the week it gets back to 75F on the summit. Even those mid-week numbers seem high for the summit. I wish they didn't have a "point" forecast for Mount Washington. Its never right and its never clear where the forecast is for. The summit, the valleys, the cirques?

Charlie
09-03-2007, 09:34 AM
what do they know ,they probably don't even know its a mountain

bclark
09-03-2007, 10:15 AM
These "point" forecasts for places like Mt. Washington are never good and should certainly never be trusted when making any sort of decision about hiking.

The reason they are so bad is because they aren't made by a human forecaster per se. Typically the forecasts you see on the web, where you can type in any city and get a forecast, are derived strictly from model data and almost never account for elevation. The forecast for "Your City, USA" may actually be the forecast for a nearby larger city that has had certain programed adjustments applied to it. There is sometimes some human intervention, but not on a scale that would make a forecast for Mt. Washington correct.

It is interesting to note however, that there is model data available specifically for Mt. Washington that is extremely helpful in writing the forecasts that we do for the higher summits. Since we provide observational data (on a contract basis) to the NWS, which in turn goes into the model runs, there is outputted data specifically for us. For instance, MOS data is available for Mt. Washington here (http://www.nws.noaa.gov/cgi-bin/mos/getall.pl?sta=KMWN). MOS data uses raw model output data, and combines that with climatological data to get a forecast. A little more info on MOS can be found on Wikipedia (http://en.wikipedia.org/wiki/Model_output_statistics).

It seems to me that with this sort of raw data available, websites that provide these point forecasts could fairly easily generate a much more accurate forecast for Mt. Washington using just the raw data. But then again I'm just a meteorologist, not a computer programmer. :cool:



Their highs for the next few days seem right--around 50. But then they also show 75 for Friday.

Actually, a forecast of temperatures around 50 is not really that close for the next few days. Our forecast for tomorrow (for the higher summits) is for temperatures to be in the low 40's.

You're better off just checking our forecast. (http://www.mountwashington.org/weather/forecast.php) :)

treant985
09-03-2007, 10:28 AM
While I agree that the forecast posted on this site is more reliable than on NOAA, the forecast on this site uses more vague numbers for predicted highs and lows. I use the NOAA point forecast to try to predict which spot will have the lowest temp in the country each day, so exact predictions are more interesting to me, even if they turn out to be a few degrees off. NOAA also has point forecasts for other unusual places like Death Valley, and they usually are not off by more than a couple degrees. For instance, the predicted high for DV yesterday was 118 I believe, and the ranger station predicted 120; the actual high was 122.

That NOAA page is the only mainstream page that usually has good forecasts for Mt. Washington, which is why I wondered why their forecasts right now are so off. I know that those kinds of pages usually don't provide good forecasts for extreme areas, but I know that their page usually has rather reliable temps for places like MWO.

Bill O
09-03-2007, 10:42 AM
We decided a few years ago there was no benefit to forecasting exact highs for the summit of Mount Washington. For example, hikers don't care if it is going to be 42F or 44F, they just want to know that it is likely going to be in the low 40's. Also, when you forecast down to a single degree you imply a precision that doesn't exist. A forecast in the low 40's indicates more variability and temps in the high 30's or high 40's wouldn't be a big surprise.

While the point forecasts for MWN are just terrible they seem to work much better in other mountain locations. For example, Alta, UT and the surrounding peaks often had very realistic forecasts.

bclark
09-03-2007, 11:19 AM
While the point forecasts for MWN are just terrible they seem to work much better in other mountain locations. For example, Alta, UT and the surrounding peaks often had very realistic forecasts.

That's interesting...I guess I haven't looked at that many other mountain sites. I look at the forecast for Big Bear Lake, CA (elevation ~7200 ft.) on a somewhat regular and it is usually pretty bad, although not quite as bad as Mt. Washington forecasts.

I think it has a big dependence on how good the derivation is that is used to get the point forecast. Some sites may have better derivations than others. Also some websites may have better derivations than others. This would also explain why a place like Death Valley (as treant985 mentioned) has a better forecast too. For instance:



The forecast for Mt. Washington according to accuweather.com:

Current Temp.: 68 F

Today: 81 F and mostly sunny

Tomorrow: 72 F and plenty of sunshine


The forecast for Mt. Washington according to weather.com

Current Temp.: 47 F

Today: 57 F and partly cloudy

Tomorrow: 42 F and partly cloudy

The actual temperature on the summit as I'm writing this is 46 F. Clearly the weather.com forecast is superior(in this case ;) ), and actually not too bad in general.

Bill O
09-03-2007, 11:29 AM
Maybe I am wrong. The forecasters in Salt Lake might just put more detail into the forecasts for the Wasatch mountains, which leads to better point forecasts. The Wasatch aren't that big and with the amount of people that recreate there its worth the extra effort.

I've noticed that weather.com puts out a very realistic forecast for MWN. I'm not sure if accurate is the right word but the numbers are generally reasonable. For example, you wont see an 80F forecast. Maybe they have a better algorithm that grabs MOS data more effectively.

Steve M
09-03-2007, 01:01 PM
I had it out not too long age with one of our local forecasters out of Tallahassee, FL over point forecasts for the surrounding area. They are regularly off on there temps but even farther off on there rain chance forecasts. He rudely informed me that the forecasts are very accurate and are regularly tested for accuracy. I think he's nuts.

JimS
09-04-2007, 04:01 PM
We decided a few years ago there was no benefit to forecasting exact highs for the summit of Mount Washington. For example, hikers don't care if it is going to be 42F or 44F, they just want to know that it is likely going to be in the low 40's. Also, when you forecast down to a single degree you imply a precision that doesn't exist.

Off Topic:
In my new life I was going over the difference between precision and accuracy with my students just today. I like this example!

Patrad Fischroy
09-04-2007, 08:13 PM
Continuing down the off- topic trail.
Jim,
Depending on the age of your charges and their ability to discern irony, this may also get the point across.
http://www.despair.com/consistency.html

JimS
09-05-2007, 06:58 AM
Continuing down the off- topic trail.
Jim,
Depending on the age of your charges and their ability to discern irony, this may also get the point across.
http://www.despair.com/consistency.html


I like it!

Mike D
09-05-2007, 10:13 PM
The Weather Service system makes assumptions that fail on Mt Washington, just like the METAR manual makes assumptions about observing weather that fail on Mt Wash. The Weather Service assumes that all zip codes are on regular terrain and can therefore get a cookie cutter forecast. The METAR manual assumes that all stations are at airports, or at least regular terrain, and therefore it must be impossible to have fog and a clear sky, for example.

The reason these one-size-fits-all solutions don't work on Mt Washington is that it's a unique locale with continuous surprises, which is why people type on forums like this one.