WXforum.net

Weather Station Hardware => AcuRite Weather Stations => Topic started by: jakejake on October 19, 2011, 11:23:13 AM

Title: How accurate is "Precision Forecasting"?
Post by: jakejake on October 19, 2011, 11:23:13 AM
I'm giving strong consideration to Acurite's model 02010 (http://www.acurite.com/weather/weather-forecast-stations/acurite-professional-digital-weather-center-with-forecast-temperature-humidity-02010.html) because of the "Precision Forecasting" feature.

How well does it work?  Especially in Western Washington?

I like most things about the 02010, except I really wish I can see the daily highs/lows without having to push buttons (I need to monitor for freezing temperatures now that winter is here).  And I find the "indoor weather" to be totally unnecessary (especially since I have a furnace thermostat). 

So I'm trying to decide if Acurite's weather forecasting is worth it, or if getting a console that displays daily hi/low temperature is more advantageous.  If Acurite's "Precision Forecasting" is just as reliable as their competitors or Acurite's normal forecasting, then I'll give more weight to the daily hi/low choices, and my decision could become much simpler.

Thanks,
Jake

Title: Re: How accurate is "Precision Forecasting"?
Post by: Bushman on October 19, 2011, 01:10:06 PM
Rhetorical:  How accurate is ANY weather forecasting.  :)
Title: Re: How accurate is "Precision Forecasting"?
Post by: SLOweather on October 19, 2011, 02:06:46 PM
Rhetorical:  How accurate is ANY weather forecasting.  :)

Especially when you can get about 75% accuracy (at least around here) just by saying, "Tomorrow will be the same as today." It's that last 25% that is the art of forecasting.
Title: Re: How accurate is "Precision Forecasting"?
Post by: Sigdigit on October 19, 2011, 05:17:45 PM
If it's anything like the Davis unit, might as well use a Magic 8-ball.  I leave the forecasting to the experts at NWS.
Title: Re: How accurate is "Precision Forecasting"?
Post by: Bushman on October 19, 2011, 05:25:59 PM
Sounding harsh.... how accurate could "precision forecasting" be from a hundred dollar unit??
Title: Re: How accurate is "Precision Forecasting"?
Post by: Scalphunter on October 19, 2011, 08:03:24 PM
 Just how accuarite can an 60 dollar unit be when units costing 1000 dollars (VP2) only come in ballpark close. You can look out your window and guess better.

Title: Re: How accurate is "Precision Forecasting"?
Post by: Rhino on October 20, 2011, 10:18:27 AM
Of course you always have to keep in mind a forecast is just that- a forecast, not absolute truth. Even the weather man with radar and satellite and everything cannot ever be 100%- so don't expect better from ANY device.

We worked with atmospheric scientist Paul J. Roebber, Ph.D to develop a method of forecasting that takes more into consideration than pressure and temperature. We found it to be more accurate and much more detailed than what is typically found in the weather station industry. Our precision forecast will also calculate and display expected HI/LO as well as chance of precipitation. I know folks are EXTREMELY skeptical of the forecast features on these units, but this new precision forecasting has an awful lot of science and time put into it. I came up with the concept and regional data idea about 4 years ago-and we have been developing it since then! We have sold over 100K of them in canada and western U.S. with very positive feedback about the forecasting feature.

http://www.acurite.com/precision-forecasting
Title: Re: How accurate is "Precision Forecasting"?
Post by: Bushman on October 20, 2011, 11:01:59 AM
Since PF relies on derived altitude, can anyone with one of these units tell me how accurate that portion is?
Title: Re: How accurate is "Precision Forecasting"?
Post by: jakejake on October 20, 2011, 02:48:19 PM
"this new precision forecasting has an awful lot of science and time put into it."

I'm asking how the science compares to professional meteorologist and reality (validation).
For example, I can look at Weather Underground and get a hourly forecast, with probability, from the most expensive and high science/technology available.  I consider this the benchmark.
Anyone compare their Precision Forecast with their meteorologist reports?  Anyone compare the prediction with actual?  Anyone in PNW (where the weather patterns are difficult to predict)?

Title: Re: How accurate is "Precision Forecasting"?
Post by: SoMDWx on October 20, 2011, 03:12:05 PM
Get you BS in Meteorology and then you can forecast --- with SOME accuracy...
Title: Re: How accurate is "Precision Forecasting"?
Post by: jakejake on October 20, 2011, 05:46:57 PM
This is a lot easier to figure out, and doesn't need a college degree.
If I had the Precision Forecasting device right now, I'd take Acurite's modeled predictions and compare it to the hourly meteorologist modeled predictions.  Then compare to what's real.  I'd then score the results over a period of time, and that would be the answer.  Simple!

Has anyone done something like that for the PNW?
Or am I the only weather nerd that would do something like this?
Title: Re: How accurate is "Precision Forecasting"?
Post by: Scalphunter on October 20, 2011, 08:43:53 PM
 Try bliving in Alaska and firgure the weather out. You read the forecast, pireps and hope that things have not change  on the route your going to go. I have flown both in the PNW and Alaska and PNW is  an piece of cake compare to where I live now. So yes you can say that it has been compared. Have yet to see any of them even close. Acu-junk, davis or OS are all marginal when compare to what eye and sense can do. Here that 12 hour forcast hardly changes from Rain  or increasing clouds with rain when conditions come so fast  if you blink you can miss that break of sun.

John
Title: Re: How accurate is "Precision Forecasting"?
Post by: SoMDWx on October 21, 2011, 02:13:29 PM
This is a lot easier to figure out, and doesn't need a college degree.
If I had the Precision Forecasting device right now, I'd take Acurite's modeled predictions and compare it to the hourly meteorologist modeled predictions.  Then compare to what's real.  I'd then score the results over a period of time, and that would be the answer.  Simple!

Has anyone done something like that for the PNW?
Or am I the only weather nerd that would do something like this?


If it were that easy, then everyone could be a forecaster. From your statement, it sounds like you're very naive to what is involved in forecasting the weather.  A little LCD display on a weather station, ANY, weather station is not going to provide you with a forecast. What you see is a "forecast" based on current conditions. That little plastic box has NO clue of what is going on in 24,36, or 48 hours from now, from a mesoscale or synoptic perspective. So don't try to fool yourself into thinking this would work. Love when people think they can outdo a series of some of the world's most powerful computers that do nothing but suck in weather data, model, and then predict, 24/7... ](*,)
Title: Re: How accurate is "Precision Forecasting"?
Post by: Bushman on October 21, 2011, 02:30:18 PM
Here's something related to the relative performance question.  http://www.freakonomics.com/2008/04/21/how-valid-are-tv-weather-forecasts/
Title: Re: How accurate is "Precision Forecasting"?
Post by: jakejake on October 21, 2011, 02:45:43 PM
Don't make this more difficult than it needs to be.

This is how easy it is to validate...

1. Acurite provides a 4 hour forecast.
2. Weather Channel provides a 4 hour forecast (this is the benchmark).  
3. Record the forecast in spreadsheet.  
4. Note the differences between the forecast.
5. Wait 4 hours.
6. Compare forecasts to actual observation.
7. Collect the above information over a period of time for better comparison.

Then evaluate the data and answer the questions:
1.  How accurate is the Weather Channel forecast to actual?
2.  How accurate is Acurite forecast to actual?

Jake
Title: Re: How accurate is "Precision Forecasting"?
Post by: jakejake on October 21, 2011, 03:03:04 PM
@Bushman:  Thanks for the link. 

Meteorologists are forecasting ...3, 5, 7, 14... days out. 
Acurite is only forecasting 4, 4-8, 8-12 hours. 
And the forecasts provide probabilities, which are helpful in determining the forecast confidence. 





Title: Re: How accurate is "Precision Forecasting"?
Post by: SoMDWx on October 22, 2011, 08:56:26 AM
Spinning wheels.......
Title: Re: How accurate is "Precision Forecasting"?
Post by: Tawpie on November 25, 2011, 01:06:43 AM
[snip]I came up with the concept and regional data idea about 4 years ago [snip]

Curious... Now that I and others are uploading readings to my backyard weather, it seems like there is a wonderful opportunity to mine that data and 1) quantify the accuracy of the precision forecast and 2) continue refinements!  I don't remember if you asked for a lat-long when I registered the bridge, but if you did you'd have some very interesting check data!

FWIW, the station is also reporting to the weather underground and I gave them the location of the station.

To me, it would be wild and crazy cool if I could download my station's historicals along with the associated forecast in a .csv file. Might make for some nifty algorithmic brain teasers. I mean, if you can actually predict the weather isn't the next step predicting the stock market? The two present similar challenges but the stock market offers riches as an incentive.
Title: Re: How accurate is "Precision Forecasting"?
Post by: Weather Display on November 25, 2011, 04:44:25 PM
given enough real time data in an area, you could predict up wind of a station, using the data from down wind , to predict what is going to happen in the next time period...sort of thing...
would be usefull for if using solar power for example, knowing that there was a good period of sunshine or wind (for wind power) coming your way
so that you could then know you had lots of power coming and plan your daily power useage accordingly....sort of thing...
Title: Re: How accurate is "Precision Forecasting"?
Post by: xykotik on November 26, 2011, 10:55:28 AM
I don't know how I missed this thread earlier, but as a western-Washington resident, here are my two wet cents worth...

The TV weather forecasts here take up a good 10 minutes of the news, because they have to cover all of the microclimates.  For a 25-mile radius from Seattle, they will show spot-temp forecasts that vary by double digits (ie, Seattle, Tukwila, Sea-Tac, Renton, Kent, des Moines, Federal Way, Auburn), but it makes a difference depending on multiple factors and larger systems.  Some of them are fairly predicatable as compared to general and/or other local microclimate forecasts.  For a non-scientific example from my own observations, if there is a 50% chance of rain in Seattle, there is probably an 80% chance of rain in Lynnwood, 15 miles north.  If it is going to be extra windy in the Strait of Juan de Fuca, it may be only breezy in Seattle.

Now, the local TV meteos here do a decent job, and they are courteous enough to stagger their reports during the evening news so they can be compared (and they do differ).  The Weather Channel (.com) forecasts by zip-code are usually not even close and cast a wide net over a multi-zip area, and the Wunderground forcasts by zip aren't too bad for the areas I check, even the hourlys.  I just can't see how a fixed module in a budget device calling itself "precision" could be reliable at all in this region, so I wouldn't make a purchasing decision (for this region) based on that feature or personally spend any time doing accuracy research on it.