Author Topic: Am I the only one?  (Read 4249 times)

0 Members and 1 Guest are viewing this topic.

Offline WeatherHost

  • Forecaster
  • *****
  • Posts: 3649
Re: Am I the only one?
« Reply #25 on: September 28, 2016, 10:46:00 AM »
But my question is more why  ...

... people use WU at all.  Or most other web or national sources.  Refer to your local NWS Forecast office and/or local media services.

You want local data and interpretations for local weather.


Offline WheatonRon

  • Forecaster
  • *****
  • Posts: 1237
    • WUnderground
Re: Am I the only one?
« Reply #26 on: September 28, 2016, 10:50:14 AM »
But my question is more why  ...

... people use WU at all.  Or most other web or national sources.  Refer to your local NWS Forecast office and/or local media services.

You want local data and interpretations for local weather.

For what it is worth, the local forecasts the past 7 days have been absolutely off the mark. Rain was forecast 3 of the past 5 days and we got none. They were as wrong as WU!
Davis VP2 with SHT31 (3 complete VP2 systems—2 with a daytime fan and 1 that has a 24 hour fan); CWOP--CW5020, FW3075 and FW4350; WU--KILWHEAT17, KILWHEAT36 and KILWHEAT39; WeatherCloud.net; CoCoRaHS--IL-DP-132; and Weatherlink 2.0

Offline Jáchym

  • Meteotemplate Developer
  • Forecaster
  • *****
  • Posts: 8605
    • Meteotemplate
Re: Am I the only one?
« Reply #27 on: September 28, 2016, 10:59:33 AM »
This is not to to disrespect any mets out there but the problem is there isn't enough data. They are making forecasts based on model data and local knowledge. That isn't enough so they are left with making an educated guess as to how much of a chance there is for rain, and most times it's wrong.
Twice I have seen a 100% chance of rain for my area and we didn't get a drop. One of those times there wasn't any rain for 50 miles. Another time they were so sure of a wide spread tornado event most of the business's closed early to send people home. Nothing happened.
Unless we have airborne nano sensors every 1000ft and spaced out a half mile apart mets don't truly know whats going on up there. That's why a rain forecast is just a best guess.
So what we need is a VP2 every 1000'! Davis will be thrilled and may give them the push to release a VP3!

Well... unfortunately no :D What you need in order to get 100% accurate forecast is overcoming the laws of physics :) When you get to the very basic principles of how the atmosphere behaves there are random motion effects etc. which cannot be by any means predicted. And as you know even these small nuances can lead to major differences - the butterfly effect. So even if you had a hypothetical supercomputer that had data from every single place on Earth, with all the atmosphere parameters, terrain data, soil types etc etc... you would probably be very good and very accurate, but not always. :)

Offline gwwilk

  • Southeast Lincoln Weather
  • Forecaster
  • *****
  • Posts: 2578
    • SouthEast Lincoln, NE Weather
Re: Am I the only one?
« Reply #28 on: September 28, 2016, 01:36:19 PM »
This is not to to disrespect any mets out there but the problem is there isn't enough data. They are making forecasts based on model data and local knowledge. That isn't enough so they are left with making an educated guess as to how much of a chance there is for rain, and most times it's wrong.
Twice I have seen a 100% chance of rain for my area and we didn't get a drop. One of those times there wasn't any rain for 50 miles. Another time they were so sure of a wide spread tornado event most of the business's closed early to send people home. Nothing happened.
Unless we have airborne nano sensors every 1000ft and spaced out a half mile apart mets don't truly know whats going on up there. That's why a rain forecast is just a best guess.
So what we need is a VP2 every 1000'! Davis will be thrilled and may give them the push to release a VP3!

Well... unfortunately no :D What you need in order to get 100% accurate forecast is overcoming the laws of physics :) When you get to the very basic principles of how the atmosphere behaves there are random motion effects etc. which cannot be by any means predicted. And as you know even these small nuances can lead to major differences - the butterfly effect. So even if you had a hypothetical supercomputer that had data from every single place on Earth, with all the atmosphere parameters, terrain data, soil types etc etc... you would probably be very good and very accurate, but not always. :)
Well put, Jachym.  The only 100% reliable way to know the future is to experience it because the only 100% accurate model is the one we're living in from moment to moment.
Regards, Jerry Wilkins
gwwilk@gmail.com

Offline SLOweather

  • Global Moderator
  • Forecaster
  • *****
  • Posts: 3456
    • Weatherelement Moline IL
Re: Am I the only one?
« Reply #29 on: September 28, 2016, 01:37:45 PM »
One site that I find interesting is http://www.forecastadvisor.com/.
They compare the various forecasts for accuracy

Here the one for my zip code:

http://www.forecastadvisor.com/California/SanLuisObispo/93405/

Once you get past the individual forecast service accuracy numbers, check out the Persistence figure. Persistence is basically saying that the weather today will be the same as yesterday. For our area, Persistence always seems to be about 65%. Some forecasts are barely better than that at 70-72%.

Whenever I speak before a group about weather, I like to mention Persistence, and how much effort is put into gaining just a few percentage points of accuracy.

Offline gwwilk

  • Southeast Lincoln Weather
  • Forecaster
  • *****
  • Posts: 2578
    • SouthEast Lincoln, NE Weather
Re: Am I the only one?
« Reply #30 on: September 28, 2016, 01:46:04 PM »
One site that I find interesting is http://www.forecastadvisor.com/.
They compare the various forecasts for accuracy

Here the one for my zip code:

http://www.forecastadvisor.com/California/SanLuisObispo/93405/

Once you get past the individual forecast service accuracy numbers, check out the Persistence figure. Persistence is basically saying that the weather today will be the same as yesterday. For our area, Persistence always seems to be about 65%. Some forecasts are barely better than that at 70-72%.

Whenever I speak before a group about weather, I like to mention Persistence, and how much effort is put into gaining just a few percentage points of accuracy.
If you live in an arid or tropical climate, the persistence effect is probably a fair guess.  More temperate climates such as mine give persistence fits.  Here is the Forecast Adviser evaluation of various local forecasts for the past month and the past year.  Persistence is at the bottom by a huge margin.
Regards, Jerry Wilkins
gwwilk@gmail.com

Offline Jáchym

  • Meteotemplate Developer
  • Forecaster
  • *****
  • Posts: 8605
    • Meteotemplate
Re: Am I the only one?
« Reply #31 on: September 28, 2016, 02:12:56 PM »
Well I cant help but to mention my Forecast Comparison plugin for Meteotemplate which I made about a yr ago. What it does is it takes forecasts from many providers (yr.no, WU, forecastIO, Yahoo etc.) and it compares them side by side. In addition, you can make a "snapshot" which saves the current forecast and then week later you can load that forecast (which has of course been updated since then, but this script loads exactly what it showed when you saved it) and it adds to the stats and graphs a line which shows what your station really measured on those days.
This way you can at least see in long term which service is the best for your particular location.

One way of getting 100% accurate forecast would be time travel or if we talk about the concept of Matrix.... well you probably didnt want such answer did you :D

Offline WheatonRon

  • Forecaster
  • *****
  • Posts: 1237
    • WUnderground
Re: Am I the only one?
« Reply #32 on: September 28, 2016, 02:48:23 PM »
One site that I find interesting is http://www.forecastadvisor.com/.
They compare the various forecasts for accuracy

Here the one for my zip code:

http://www.forecastadvisor.com/California/SanLuisObispo/93405/

Once you get past the individual forecast service accuracy numbers, check out the Persistence figure. Persistence is basically saying that the weather today will be the same as yesterday. For our area, Persistence always seems to be about 65%. Some forecasts are barely better than that at 70-72%.

Whenever I speak before a group about weather, I like to mention Persistence, and how much effort is put into gaining just a few percentage points of accuracy.

Interesting summary by forecast advisor. My summary weather is here:
http://www.forecastadvisor.com/Illinois/Wheaton/60187/

Interesting to see the difference in Persistence in my area vs. California vs. Nebraska. I guess the take away is California has gradual changes, less so for Illinois and Nebraska.
« Last Edit: September 28, 2016, 03:23:01 PM by WheatonRon »
Davis VP2 with SHT31 (3 complete VP2 systems—2 with a daytime fan and 1 that has a 24 hour fan); CWOP--CW5020, FW3075 and FW4350; WU--KILWHEAT17, KILWHEAT36 and KILWHEAT39; WeatherCloud.net; CoCoRaHS--IL-DP-132; and Weatherlink 2.0

Offline sundevil01010101

  • KAZSURPR41/54/57
  • Senior Contributor
  • ****
  • Posts: 211
  • Eye on the sky...
    • Surprise AZ Weather
Re: Am I the only one?
« Reply #33 on: September 28, 2016, 03:35:38 PM »
Cool site WheatonRon, didn't know about that one.

http://www.forecastadvisor.com/Arizona/Surprise/85379/ is mine.    Not numbers they should be very proud of.

Obviously weather prediction is extremely complex with variables upon variables upon variables. 

A quantum computer doing weather forecasting will fix this since it can calculate all possible variations of any future instant's weather instantly.  :twisted:
« Last Edit: September 28, 2016, 03:37:09 PM by sundevil01010101 »
Visit https://www.surpriseazweather.com/ !!!
Bucket List - 1,000,000 visitors  :twisted: Lol...

Offline hankster

  • Forecaster
  • *****
  • Posts: 482
    • NFM Weather
Re: Am I the only one?
« Reply #34 on: September 28, 2016, 05:20:47 PM »
One site that I find interesting is http://www.forecastadvisor.com/.

Here is what they say for forecasts in my area. Appears WU is doing pretty good.

Weather forecast accuracy for Fort Myers, Florida

Last Month
Weather Underground     83.33%
The Weather Channel     81.25%
NWS Digital Forecast     79.76%
Dark Sky (forecast.io)     79.76%
AccuWeather     78.87%
Foreca     75.60%
MeteoGroup     74.40%
Persistence     70.05%
World Weather Online     65.48%

Last Year
MeteoGroup     84.83%
Weather Underground     84.13%
The Weather Channel     83.96%
AccuWeather     81.62%
Foreca     79.32%
NWS Digital Forecast     76.77%
Dark Sky (forecast.io)     71.53%
World Weather Online     70.73%
Persistence     63.03%

Offline WheatonRon

  • Forecaster
  • *****
  • Posts: 1237
    • WUnderground
Re: Am I the only one?
« Reply #35 on: September 28, 2016, 05:30:34 PM »
One site that I find interesting is http://www.forecastadvisor.com/.

Here is what they say for forecasts in my area. Appears WU is doing pretty good.

Weather forecast accuracy for Fort Myers, Florida

Last Month
Weather Underground     83.33%
The Weather Channel     81.25%
NWS Digital Forecast     79.76%
Dark Sky (forecast.io)     79.76%
AccuWeather     78.87%
Foreca     75.60%
MeteoGroup     74.40%
Persistence     70.05%
World Weather Online     65.48%

Last Year
MeteoGroup     84.83%
Weather Underground     84.13%
The Weather Channel     83.96%
AccuWeather     81.62%
Foreca     79.32%
NWS Digital Forecast     76.77%
Dark Sky (forecast.io)     71.53%
World Weather Online     70.73%
Persistence     63.03%

All very interesting. But isn't WU now owned by The Weather Channel? If so, you would think they would be the same,no?
Davis VP2 with SHT31 (3 complete VP2 systems—2 with a daytime fan and 1 that has a 24 hour fan); CWOP--CW5020, FW3075 and FW4350; WU--KILWHEAT17, KILWHEAT36 and KILWHEAT39; WeatherCloud.net; CoCoRaHS--IL-DP-132; and Weatherlink 2.0

Offline BigOkie

  • Forecaster
  • *****
  • Posts: 1367
  • Tulsa, OK
    • KOKTULSA13
Re: Am I the only one?
« Reply #36 on: September 28, 2016, 05:37:27 PM »
One site that I find interesting is http://www.forecastadvisor.com/.

Here is what they say for forecasts in my area. Appears WU is doing pretty good.

Weather forecast accuracy for Fort Myers, Florida

Last Month
Weather Underground     83.33%
The Weather Channel     81.25%
NWS Digital Forecast     79.76%
Dark Sky (forecast.io)     79.76%
AccuWeather     78.87%
Foreca     75.60%
MeteoGroup     74.40%
Persistence     70.05%
World Weather Online     65.48%

Last Year
MeteoGroup     84.83%
Weather Underground     84.13%
The Weather Channel     83.96%
AccuWeather     81.62%
Foreca     79.32%
NWS Digital Forecast     76.77%
Dark Sky (forecast.io)     71.53%
World Weather Online     70.73%
Persistence     63.03%

All very interesting. But isn't WU now owned by The Weather Channel? If so, you would think they would be the same,no?

Not any more.  I'd heard IBM had bought it..

Yep...a while back actually.  Almost a year now:

https://www.wunderground.com/blog/JeffMasters/weather-underground-bought-by-ibm
Current setup: Davis Vantage Pro 2 Plus Wireless
Weather radios:
Reecom R-1650
Sangean CL-100
Uniden Home Patrol I

Offline Jáchym

  • Meteotemplate Developer
  • Forecaster
  • *****
  • Posts: 8605
    • Meteotemplate
Re: Am I the only one?
« Reply #37 on: September 28, 2016, 05:59:44 PM »
Cool site WheatonRon, didn't know about that one.

http://www.forecastadvisor.com/Arizona/Surprise/85379/ is mine.    Not numbers they should be very proud of.

Obviously weather prediction is extremely complex with variables upon variables upon variables. 

A quantum computer doing weather forecasting will fix this since it can calculate all possible variations of any future instant's weather instantly.  :twisted:

Terry not even quantum computer would help you. It is like trying to predict radioactive decay - impossible. These are the phenomena that really are "random" (unlike the pseudo-randomness we use most of the time).

So a quantum computer could predict all the possible scenarios, but it wont be able to tell you, which one is going to happen :D And we can pretty much do that today already :D I can tell you that it will be somewhere between -50 and 150 F, humidity somewhere between 0 and 100%, pressure somewhere between 25 and 35 inHg and it will be 0 to 100% cloudy. Oh, and it will be dark at night :D. Do you want to bet with me I am correct? :D

Offline WheatonRon

  • Forecaster
  • *****
  • Posts: 1237
    • WUnderground
Re: Am I the only one?
« Reply #38 on: September 28, 2016, 06:22:14 PM »
One site that I find interesting is http://www.forecastadvisor.com/.

Here is what they say for forecasts in my area. Appears WU is doing pretty good.

Weather forecast accuracy for Fort Myers, Florida

Last Month
Weather Underground     83.33%
The Weather Channel     81.25%
NWS Digital Forecast     79.76%
Dark Sky (forecast.io)     79.76%
AccuWeather     78.87%
Foreca     75.60%
MeteoGroup     74.40%
Persistence     70.05%
World Weather Online     65.48%

Last Year
MeteoGroup     84.83%
Weather Underground     84.13%
The Weather Channel     83.96%
AccuWeather     81.62%
Foreca     79.32%
NWS Digital Forecast     76.77%
Dark Sky (forecast.io)     71.53%
World Weather Online     70.73%
Persistence     63.03%

All very interesting. But isn't WU now owned by The Weather Channel? If so, you would think they would be the same,no?

Not any more.  I'd heard IBM had bought it..

Yep...a while back actually.  Almost a year now:

https://www.wunderground.com/blog/JeffMasters/weather-underground-bought-by-ibm

Not entirely correct. IBM bought The Weather Company, Weather Underground, The Weather Company brand, etc. The TV segment was not acquired by IBM, but is licensing weather data and analytics from IBM under a long-term contract. While announced in 2015, the transaction did not close until January 29, 2016. This verbiage I took from IBM's SEC filings. The long and short of this, the data that WU and the Weather Channel use are both owned by IBM, but I understand why IBM would want to keep the data separate.
« Last Edit: September 28, 2016, 11:27:39 PM by WheatonRon »
Davis VP2 with SHT31 (3 complete VP2 systems—2 with a daytime fan and 1 that has a 24 hour fan); CWOP--CW5020, FW3075 and FW4350; WU--KILWHEAT17, KILWHEAT36 and KILWHEAT39; WeatherCloud.net; CoCoRaHS--IL-DP-132; and Weatherlink 2.0

Offline cospringswx

  • Forecaster
  • *****
  • Posts: 4136
    • Colorado Springs Weather
Re: Am I the only one?
« Reply #39 on: September 29, 2016, 12:31:02 AM »
Found this year's ago.


Why is it Difficult to Forecast?

Consider a rotating spherical envelope of a mixture of gases…occasionally murky and always somewhat viscous.  Place  it around an astronomical object nearly 8,000 miles in diameter. Tilt the whole system back and forth with respect to its source of heat and light. Freeze it at the poles of its axis of rotation and intensely heat it in the middle. Cover most of the surface of the sphere with a liquid that continually feeds moisture into the atmosphere.  Subject the whole to tidal forces induced by the sun and a captive satellite. Then try to predict the conditions of one small portion of that atmosphere for a period of one to several days in advance.
 




Ryan 

Colorado Springs, CO
www.cospringsweather.com
Davis Vantage Vue
Weather Display Software
Amcrest HD IP Camera

 

anything