Author Topic: Algorithm for Predicting Weather  (Read 4368 times)

0 Members and 1 Guest are viewing this topic.

Offline Inq

  • Member
  • *
  • Posts: 16
    • InqOnThat
Algorithm for Predicting Weather
« on: April 02, 2022, 08:43:47 AM »
Background
I'm teaching an introductory class at my local library and at the local community college about programming on a microprocessor.  I wanted a demonstration project more advanced than "hello world" or "blinking an LED" and I picked a simple weather station.  https://inqonthat.com/inqweather/  It has sensors, needs user input for adjusting, needs some kind of visual display, needs to store historical data and do some rudimentary calculations on data... IOW's sounded like the perfect project.  That it was useful when the student takes it home was just icing on the cake!  It only has sensors for temperature, humidity and pressure.   We have it doing all the simpleton things like:
  • Showing temperature (T)
  • Showing pressure (P)
  • Showing humidity (H)
  • Adjusting/normalizing pressure for location altitude
  • We have plans for easy things like showing high/low T,P,H in the last 24 hours, week, month... etc

Question/Problem
Seeing the basic weather stations like the one pictured below, I can see it must have the same set of sensors... T, P and H.  I would like to add some rudimentary weather prediction (showing an image of what's coming) like these units do.  The best chart I've been able to find so far is also pictured below.  Unfortunately, it is also dependent on wind direction which we don't have (and neither do these basic weather stations). 

Can anyone provide me with some algorithm or links to some information?  It doesn't have to be source code.   We can translate a chart into code.

Thanks.
VBR,
Inq
 [ You are not allowed to view attachments ]

P.S. Being my first post, I don't think the picture will upload.  I tried though.   :-(
« Last Edit: April 02, 2022, 08:45:34 AM by Inq »

Offline worachj

  • Forecaster
  • *****
  • Posts: 477
  • Twin Cities, Minnesota
Re: Algorithm for Predicting Weather
« Reply #1 on: April 02, 2022, 09:11:56 AM »
Your picture is of an AcuRite Display with forecasting. Here is what AcuRite says about their forecasting.

https://www.acurite.com/blog/precision-forecasting-white-paper.html

https://www.acurite.com/learn/glossary/learning-mode




AcuRite Atlas
KMNSAINT173

Offline Inq

  • Member
  • *
  • Posts: 16
    • InqOnThat
Re: Algorithm for Predicting Weather
« Reply #2 on: April 02, 2022, 09:45:28 AM »
Your picture is of an AcuRite Display with forecasting. Here is what AcuRite says about their forecasting.

https://www.acurite.com/blog/precision-forecasting-white-paper.html

https://www.acurite.com/learn/glossary/learning-mode

Thank you for responding.

I've just finished reading through those links and surfed from their links to their "white papers".  All talk about the marketing aspects that they do a "14 day learning", but no other real details.  I'm not looking to best them or even to get to the level of detail that they do my selecting 1 of 7 regions here in the US. 

Even before electronic units, a simple barometer had good/fair/poor.  I don't know, but I suspect there is something between seeing what the current pressure on an old barometer and those 14 day learning modes.  Something that can be a function of:
  • current temperature
  • temperature change rate
  • humidity
  • current pressure
  • change in pressure (rate) over any time period - I've read there is something special about a 3-hour pressure differential, but nothing further.

Over a great deal of time, I'm sure we could figure out a logic chart just by using our weather stations and making visual observations.  I'm just hoping to jump-start that process by asking a more learned group of people on this forum.

VBR,
Inq

Offline Storm017

  • Senior Contributor
  • ****
  • Posts: 172
Re: Algorithm for Predicting Weather
« Reply #3 on: April 02, 2022, 11:35:29 AM »

Offline Inq

  • Member
  • *
  • Posts: 16
    • InqOnThat
Re: Algorithm for Predicting Weather
« Reply #4 on: April 02, 2022, 12:06:14 PM »
Maybe this will help:

https://www.thoughtco.com/how-to-read-a-barometer-3444043

I've finished reading this page.  It's a good start.  It does show ranges, but it uses terms like rising/falling slowly and rapidly but doesn't define a range of rates for how fast pressure is changing.  I continue jumping from this page.

Thanks

Offline CamarilloWX

  • CamarilloWX
  • Senior Contributor
  • ****
  • Posts: 184
    • Camarillo Weather
Re: Algorithm for Predicting Weather
« Reply #5 on: April 02, 2022, 01:13:56 PM »
Search for Zambretti weather forecaster.
Eric

Offline Storm017

  • Senior Contributor
  • ****
  • Posts: 172
Re: Algorithm for Predicting Weather
« Reply #6 on: April 02, 2022, 01:50:01 PM »

Offline Inq

  • Member
  • *
  • Posts: 16
    • InqOnThat
Re: Algorithm for Predicting Weather
« Reply #7 on: April 02, 2022, 04:45:15 PM »
Search for Zambretti weather forecaster.

I definitely came to the right forum. - I first posted this request on the electronic forums I usually haunt.  Although very common among Arduino level electronic type projects, I did not receive even one reply over the last week.  I guess, they're so happy just reporting the current readings off their weather stations, they don't continue to the logical end and want to form some kind of predictive benefit.   Here... I get multiple great leads within hours.

This one especially... "Zambretti" alone opened a floodgate of Google hits including this one with actual JavaScript already written to take advantage of this 1915 technique.
http://www.beteljuice.co.uk/zambretti/forecast.html

Thank you all!

VBR,
Inq

Offline Mattk

  • Forecaster
  • *****
  • Posts: 2154
Re: Algorithm for Predicting Weather
« Reply #8 on: April 02, 2022, 07:27:51 PM »
Algorithm to Predict Weather! Well good luck with that one, most meteorological organisations don't get that right and especially when there are multiple models which never predict the same thing anyway. 

Offline Bashy

  • brecklandweather.com/meteo
  • Forecaster
  • *****
  • Posts: 1465
  • brecklandweather.com/index.php
    • Breckland Weather
Re: Algorithm for Predicting Weather
« Reply #9 on: April 02, 2022, 08:31:24 PM »
Search for Zambretti weather forecaster.

I definitely came to the right forum. - I first posted this request on the electronic forums I usually haunt.  Although very common among Arduino level electronic type projects, I did not receive even one reply over the last week.  I guess, they're so happy just reporting the current readings off their weather stations, they don't continue to the logical end and want to form some kind of predictive benefit.   Here... I get multiple great leads within hours.

This one especially... "Zambretti" alone opened a floodgate of Google hits including this one with actual JavaScript already written to take advantage of this 1915 technique.
http://www.beteljuice.co.uk/zambretti/forecast.html

Thank you all!

VBR,
Inq

A few of us use the Zambretti forumla on our sites, at least one weather data recording software uses it too. Its surprisingly accurate but it is only for the following 12 hrs.

Algorithm to Predict Weather! Well good luck with that one, most meteorological organisations don't get that right and especially when there are multiple models which never predict the same thing anyway. 

Thats why its called "prediction" i suppose, its only an etimate of the current data on hand, 3 days is normally close, 7 is stretching it and anything further out is just plain witchcraft if they get it right :D Here in the UK, i find that if the GFS and the ECMWF hi res match then its highly likely to happen but even then, no further than 3 days is trustworthy, heck, ive seen tomorrows change.
Kind regards
Bashy

Online gszlag

  • Forecaster
  • *****
  • Posts: 435
  • ..have you calibrated your barometer today?
    • Michael's Bay - Manitoulin Island weather
Re: Algorithm for Predicting Weather
« Reply #10 on: April 03, 2022, 07:25:57 AM »
Background
I'm teaching an introductory class at my local library and at the local community college about programming on a microprocessor.  I wanted a demonstration project more advanced than "hello world" or "blinking an LED" and I picked a simple weather station.  https://inqonthat.com/inqweather/  It has sensors, needs user input for adjusting, needs some kind of visual display, needs to store historical data and do some rudimentary calculations on data... IOW's sounded like the perfect project.  That it was useful when the student takes it home was just icing on the cake!  It only has sensors for temperature, humidity and pressure.   We have it doing all the simpleton things like:
  • Showing temperature (T)
  • Showing pressure (P)
  • Showing humidity (H)
  • Adjusting/normalizing pressure for location altitude
  • We have plans for easy things like showing high/low T,P,H in the last 24 hours, week, month... etc

Question/Problem
Seeing the basic weather stations like the one pictured below, I can see it must have the same set of sensors... T, P and H.  I would like to add some rudimentary weather prediction (showing an image of what's coming) like these units do.  The best chart I've been able to find so far is also pictured below.  Unfortunately, it is also dependent on wind direction which we don't have (and neither do these basic weather stations). 

Can anyone provide me with some algorithm or links to some information?  It doesn't have to be source code.   We can translate a chart into code.

Thanks.
VBR,
Inq
 [ You are not allowed to view attachments ]

P.S. Being my first post, I don't think the picture will upload.  I tried though.   :-(

RE: Pressure

A good intro to Barometers in general.  I keep wanting to change the article to : "What are barometers good for?" Don't fall for all the folksy charm. The author is Dr. David Burch ,Phd Physics.
https://davidburchnavigation.blogspot.com/2013/04/modern-barometry-and-its-important-role.html

For an potential algorithm to develop, here is a something from David Burch to explore:
https://davidburchnavigation.blogspot.com/2014/11/barometer-for-wind-warning.html
Ambient Weather WS-2000
Ecowitt WS3900 console
Ecowitt GW1000/GW1100
Ecowitt WS68: Anemometer, UV/solar
Ecowitt WH40: Rain gauge
Ecowitt WH57 Lightning sensor
Ecowitt WH32E: Outside T & H sensor
Stratus Rain Gauge (manual)
Raspberry Pi 3B+ (WeeWX/CumulusMX)
Raspberry Pi Zero 2W (WeeWX/MQTT/Belchertown)
---
Barometer wiki: http://meshka.eu/Ecowitt/dokuwiki/doku.php?id=barometer#barometer
---
http://weather.glenns.ca (pwsdashboard - live)
http://weewx.glenns.ca
http://glenns.ca/cumulusmx2/index.htm
---
Uploading to: AWN, ecowitt.net, Weather Underground, PWSweather.com, AWEKAS, Windy.com, WOW

Offline Inq

  • Member
  • *
  • Posts: 16
    • InqOnThat
Re: Algorithm for Predicting Weather
« Reply #11 on: April 03, 2022, 07:11:20 PM »
Well... I converted the JavaScript version into a C function.  I left the wind component in, even if I don't have a gauge for that yet.  I was able to get rid of his long if/then logic tree and the results agree.  I still have one problem.  The JS page lets the user enter the trend (Falling, Steady, Rising).  I am having trouble getting a consensus on the range that is considered "Steady".  Somewhere, I recall seeing -1.6 mbar to 1.6 mbar, but can't find the reference again.  All other pages seem to gloss over the subject.  Any suggestions?

Thanks.
VBR
Inq

Code: [Select]
u8 zambretti(float P, float d3P, float minP, float maxP, bool north,
    float dirWind = NO_WIND)
{       
    float dLP = maxP - minP;
    if (!dLP)
        return 0;

    if (dirWind < 1000)     // NO_WIND = 1E6
    {
        if (!north)
            dirWind += 180;
        // This mess, just converts angle in degrees to 16 windrose cardinal
        // 0 = N to 15 = NNW
        u8 w = (u8)((dirWind + 11.25) / 22.5) % 16;

        const float wind_dir[]
            {6,5,5,2,-0.5,-2,-5,-8.5,-12,-10,-6,-4.5,-3,-0.5,1.5,3};
        P += dLP * wind_dir[w] / 100.0;
    }
   
    d3P = d3P > STEADY ? 1.0 : (d3P < -STEADY ? -1.0 : 0);
    if (month49 == north)   // Summer in each hemisphere
        P += d3P * 7.0 / 100.0 * dLP;
    u8 index = (u8)((P - minP) / dLP * 21.99F);
   
    // equivalents of Zambretti 'dial window' letters A - Z
    const u8 rise_options[]
        {25,25,25,24,24,19,16,12,11,9,8,6,5,2,1,1,0,0,0,0,0,0};
    const u8 steady_options[]
        {25,25,25,25,25,25,23,23,22,18,15,13,10,4,1,1,0,0,0,0,0,0};
    const u8 fall_options[]
        {25,25,25,25,25,25,25,25,23,23,21,20,17,14,7,3,1,1,1,0,0,0};
   
if (d3P > 0)
        return rise_options[index];
else if (d3P < 0)
return fall_options[index];
    return steady_options[index];
}

Offline Inq

  • Member
  • *
  • Posts: 16
    • InqOnThat
Re: Algorithm for Predicting Weather
« Reply #12 on: April 03, 2022, 07:24:48 PM »
I also wanted to comment.  The preset maximum range is 950 mbar to 1050 mbar.  Considering how much that range is used in calculating the equations, it seems rather arbitrary.  However, the worst I've seen here at the house over the last month is 1000 mbar to 1030 mbar corrected to sea-level.  And that included a severe rain / thunderstorm and a frigid Canadian Clipper high pressure.

Should I let the sensors just accumulate "my" local min and max or use the defaults which were for somewhere in England???? 

Online gszlag

  • Forecaster
  • *****
  • Posts: 435
  • ..have you calibrated your barometer today?
    • Michael's Bay - Manitoulin Island weather
Re: Algorithm for Predicting Weather
« Reply #13 on: April 03, 2022, 07:47:14 PM »
Quote
...am having trouble getting a consensus on the range that is considered "Steady".  Somewhere, I recall seeing -1.6 mbar to 1.6 mbar, but can't find the reference again.  All other pages seem to gloss over the subject.  Any suggestions?

The afore-mentioned Dr. Burch defined steady pressure as < 1mb in a standard 3 hr time period.
Ambient Weather WS-2000
Ecowitt WS3900 console
Ecowitt GW1000/GW1100
Ecowitt WS68: Anemometer, UV/solar
Ecowitt WH40: Rain gauge
Ecowitt WH57 Lightning sensor
Ecowitt WH32E: Outside T & H sensor
Stratus Rain Gauge (manual)
Raspberry Pi 3B+ (WeeWX/CumulusMX)
Raspberry Pi Zero 2W (WeeWX/MQTT/Belchertown)
---
Barometer wiki: http://meshka.eu/Ecowitt/dokuwiki/doku.php?id=barometer#barometer
---
http://weather.glenns.ca (pwsdashboard - live)
http://weewx.glenns.ca
http://glenns.ca/cumulusmx2/index.htm
---
Uploading to: AWN, ecowitt.net, Weather Underground, PWSweather.com, AWEKAS, Windy.com, WOW

Offline zoomx

  • Senior Contributor
  • ****
  • Posts: 192
Re: Algorithm for Predicting Weather
« Reply #14 on: April 04, 2022, 07:26:11 AM »
Here you can find Zambretti algoritm for Arduino insithe the main sketch
https://github.com/fandonov/weatherstation
Here another version for ESP32
https://github.com/G6EJD/ESP32_Weather_Forecaster_TN061

Online gszlag

  • Forecaster
  • *****
  • Posts: 435
  • ..have you calibrated your barometer today?
    • Michael's Bay - Manitoulin Island weather
Re: Algorithm for Predicting Weather
« Reply #15 on: April 04, 2022, 07:41:48 AM »
I also wanted to comment.  The preset maximum range is 950 mbar to 1050 mbar.  Considering how much that range is used in calculating the equations, it seems rather arbitrary.  However, the worst I've seen here at the house over the last month is 1000 mbar to 1030 mbar corrected to sea-level.  And that included a severe rain / thunderstorm and a frigid Canadian Clipper high pressure.

Should I let the sensors just accumulate "my" local min and max or use the defaults which were for somewhere in England????

1000mb to 1030mb is way too small a range. Up here, we had 979 SLP just last week. 950 to 1050 is a reasonable range. Does your local weather service have extreme high/low records for barometric pressure?
Ambient Weather WS-2000
Ecowitt WS3900 console
Ecowitt GW1000/GW1100
Ecowitt WS68: Anemometer, UV/solar
Ecowitt WH40: Rain gauge
Ecowitt WH57 Lightning sensor
Ecowitt WH32E: Outside T & H sensor
Stratus Rain Gauge (manual)
Raspberry Pi 3B+ (WeeWX/CumulusMX)
Raspberry Pi Zero 2W (WeeWX/MQTT/Belchertown)
---
Barometer wiki: http://meshka.eu/Ecowitt/dokuwiki/doku.php?id=barometer#barometer
---
http://weather.glenns.ca (pwsdashboard - live)
http://weewx.glenns.ca
http://glenns.ca/cumulusmx2/index.htm
---
Uploading to: AWN, ecowitt.net, Weather Underground, PWSweather.com, AWEKAS, Windy.com, WOW

Offline Inq

  • Member
  • *
  • Posts: 16
    • InqOnThat
Re: Algorithm for Predicting Weather
« Reply #16 on: April 04, 2022, 08:54:36 AM »
Algorithm to Predict Weather! Well good luck with that one, most meteorological organisations don't get that right and especially when there are multiple models which never predict the same thing anyway.

"Prob' bly rain" - Eeyore

You might have noticed that the main purpose of the question was as an electronics teaching project... not to be the end-all prediction tool.  In that regard, this forum has been invaluable in helping me.  With a more pessimistic mindset, I did once read that if you simply predict tomorrow's weather to be the same as today's weather, you'll get a higher percentage than most weather services.  Also have read... they (news/weather stations) heavily weight their forecasting to the negative.  It's increases viewing/hits and thus revenue. 

I have no delusions or aspirations to turn the metrological society on its ear with this project.  But for less than $6 worth of parts, my introductory level IoT students can have a graphical weather station that can rival any of the simple weather stations like the one pictured in my OP.  Do any of those (<$100) units have the ability to display the results on you smart phone, tablet, desktop from anywhere else in the world?  Not the last time I checked.

The students (and I) will be able to set the background picture to any of the 26 conditions by simply dropping the files into a folder.  I plan to take pictures out my front door for the conditions predicted. 

VBR,
Inq

 [ You are not allowed to view attachments ]

Offline Inq

  • Member
  • *
  • Posts: 16
    • InqOnThat
Re: Algorithm for Predicting Weather
« Reply #17 on: April 04, 2022, 09:15:09 AM »
Here you can find Zambretti algoritm for Arduino insithe the main sketch
https://github.com/fandonov/weatherstation
Here another version for ESP32
https://github.com/G6EJD/ESP32_Weather_Forecaster_TN061
If I ran across one of these first, I probably would have used them.  Both of these projects are far more cluttered (and expensive) on the hardware side... needing to display the results using a connected LED/LCD type displays.  The first one uses a huge if/then tree to solve Zambretti and doesn't take into account wind direction.  I prefer the cleaner, array look-up solution, we've ended up with.  Although our project doesn't have a way of detecting wind, a student (or I) might get a wild-hair and add it later.  The algorithm will be ready if we do. 

VBR,
Inq

P.S.  Like Monty Python!  [tup]

Offline Inq

  • Member
  • *
  • Posts: 16
    • InqOnThat
Re: Algorithm for Predicting Weather
« Reply #18 on: April 04, 2022, 09:57:35 AM »
I also wanted to comment.  The preset maximum range is 950 mbar to 1050 mbar.  Considering how much that range is used in calculating the equations, it seems rather arbitrary.  However, the worst I've seen here at the house over the last month is 1000 mbar to 1030 mbar corrected to sea-level.  And that included a severe rain / thunderstorm and a frigid Canadian Clipper high pressure.

Should I let the sensors just accumulate "my" local min and max or use the defaults which were for somewhere in England????

1000mb to 1030mb is way too small a range. Up here, we had 979 SLP just last week. 950 to 1050 is a reasonable range. Does your local weather service have extreme high/low records for barometric pressure?

I never thought to look... Here in the mountains they're always wrong, so I've never looked to see what other information I could drill-down into.  Just now looking at my closest (big city) Asheville, NC...  The min/max for the last 12 months is 996 to 1036.

I guess what I am asking here...   

I'm looking at the Math being done in the algorithm and the amount of adjustment made is proportional to the 950 to 1050 defaults.  We experience very bad (for us) thunder storms, high winds and rain of 4"/hour and at the other extreme bright/dry... not a cloud to be seen.  My logic tells me that our actual range being so much smaller (40% of the defaults) the Zambretti results are not being "adjusted" as much and thus... might not ever show a 'Z', "Stormy, much rain" even though we certainly experience them here.

My "logic" is telling me that I should maybe start out with defaults (you clued me into [tup]) of our 996 to 1036 and let the sensors expand this range as necessary when they experience a larger extreme.  This way, I would get 100% scaling "tuned" to my area.  I'm wondering if my logic makes meteorological sense???

VBR,
Inq




Online gszlag

  • Forecaster
  • *****
  • Posts: 435
  • ..have you calibrated your barometer today?
    • Michael's Bay - Manitoulin Island weather
Re: Algorithm for Predicting Weather
« Reply #19 on: April 04, 2022, 11:16:33 AM »
Quote
My "logic" is telling me that I should maybe start out with defaults (you clued me into [tup]) of our 996 to 1036 and let the sensors expand this range as necessary when they experience a larger extreme.  This way, I would get 100% scaling "tuned" to my area.  I'm wondering if my logic makes meteorological sense???

Sure. If your sensors can expand as necessary - go for it.
Ambient Weather WS-2000
Ecowitt WS3900 console
Ecowitt GW1000/GW1100
Ecowitt WS68: Anemometer, UV/solar
Ecowitt WH40: Rain gauge
Ecowitt WH57 Lightning sensor
Ecowitt WH32E: Outside T & H sensor
Stratus Rain Gauge (manual)
Raspberry Pi 3B+ (WeeWX/CumulusMX)
Raspberry Pi Zero 2W (WeeWX/MQTT/Belchertown)
---
Barometer wiki: http://meshka.eu/Ecowitt/dokuwiki/doku.php?id=barometer#barometer
---
http://weather.glenns.ca (pwsdashboard - live)
http://weewx.glenns.ca
http://glenns.ca/cumulusmx2/index.htm
---
Uploading to: AWN, ecowitt.net, Weather Underground, PWSweather.com, AWEKAS, Windy.com, WOW

Offline Inq

  • Member
  • *
  • Posts: 16
    • InqOnThat
Re: Algorithm for Predicting Weather
« Reply #20 on: April 07, 2022, 06:40:54 PM »
Quote
My "logic" is telling me that I should maybe start out with defaults (you clued me into [tup]) of our 996 to 1036 and let the sensors expand this range as necessary when they experience a larger extreme.  This way, I would get 100% scaling "tuned" to my area.  I'm wondering if my logic makes meteorological sense???

Sure. If your sensors can expand as necessary - go for it.
I wanted to thank the forum again.

I also wanted to give feedback on the little experiment with adjusting the Zambretti
"local weather range".  I don't know if anyone is interested, but I like to be thorough even if there is only one person I can help. 

I don't know which is the original site/author, but I found this site first http://www.beteljuice.co.uk/zambretti/forecast.html.  Later I found the site copied at several other URL's.  Most sites had the defaults of local weather range set to 950 to 1050 mbars.  However, one site had them set to some numbers that were in in-Hg, but the label for the main entry was still in mbar.  Putting in a legitimate mbar number always gave a bad answer.  This caused me to look more closely at the mathematics.  And that's when I asked the question above about it making meteorological sense.

I have several units worth of components so I made two units and set one using the defaults of 950 to 1050 and a second one using the range I found were my local extremes for all of last year being 996 to 1036.
 [ You are not allowed to view attachments ]
I had both sitting next to each other outside and ran them for two days during which a storm came through.  It lasted about four hours, with thunder and lightning, but only dropped about 1" of rain.  I was working on my computer all day and had the two visual dashboards on my desktop.  Being quite different images between good and bad weather when the Zambretti value changed, it easily caught my eye any time either changed.
 [ You are not allowed to view attachments ]  [ You are not allowed to view attachments ]
It turns out changing the range did significantly change the predicting ability as I would have guessed from looking at the Math.  Unfortunately... maybe too much.  The stock numbers were way too optimistic... it said relatively clear weather until only a couple hours before the storm when it maxed out about 'X'.  The one with my local extremes was way too manic and went above X about 12 hours before and in the last couple of hours went to Z.  I think over time, if I get these tuned to my area it will have very good predicting ability.

Here is the graphed data for one of the units. 
 [ You are not allowed to view attachments ]

Offline davidefa

  • Forecaster
  • *****
  • Posts: 436
Re: Algorithm for Predicting Weather
« Reply #21 on: April 08, 2022, 07:03:10 AM »
It would be nice if you could share a link to the project ( once your students will complete it )

Offline Inq

  • Member
  • *
  • Posts: 16
    • InqOnThat
Re: Algorithm for Predicting Weather
« Reply #22 on: April 08, 2022, 07:12:32 AM »
It would be nice if you could share a link to the project ( once your students will complete it )

Was in the O.P.
https://inqonthat.com/inqweather/

Online gszlag

  • Forecaster
  • *****
  • Posts: 435
  • ..have you calibrated your barometer today?
    • Michael's Bay - Manitoulin Island weather
Re: Algorithm for Predicting Weather
« Reply #23 on: April 08, 2022, 07:31:41 AM »
Quote
I also wanted to give feedback on the little experiment with adjusting the Zambretti
"local weather range".  I don't know if anyone is interested, but I like to be thorough even if there is only one person I can help. 

I don't know which is the original site/author, but I found this site first http://www.beteljuice.co.uk/zambretti/forecast.html.  Later I found the site copied at several other URL's.  Most sites had the defaults of local weather range set to 950 to 1050 mbars.  However, one site had them set to some numbers that were in in-Hg, but the label for the main entry was still in mbar.  Putting in a legitimate mbar number always gave a bad answer.  This caused me to look more closely at the mathematics.  And that's when I asked the question above about it making meteorological sense.

I have several units worth of components so I made two units and set one using the defaults of 950 to 1050 and a second one using the range I found were my local extremes for all of last year being 996 to 1036.

I had both sitting next to each other outside and ran them for two days during which a storm came through.  It lasted about four hours, with thunder and lightning, but only dropped about 1" of rain.  I was working on my computer all day and had the two visual dashboards on my desktop.  Being quite different images between good and bad weather when the Zambretti value changed, it easily caught my eye any time either changed.

It turns out changing the range did significantly change the predicting ability as I would have guessed from looking at the Math.  Unfortunately... maybe too much.  The stock numbers were way too optimistic... it said relatively clear weather until only a couple hours before the storm when it maxed out about 'X'.  The one with my local extremes was way too manic and went above X about 12 hours before and in the last couple of hours went to Z.  I think over time, if I get these tuned to my area it will have very good predicting ability.

Have you seen Dr. KFS research on the Zambretti algorithm? I did a very quick read-thru. It appears the baro range changes depending on whether or not pressure is steady falling or rising. Unless I missed it, I did not see adjustments for latitude. Interesting reading.
https://integritext.net/DrKFS/zambretti.htm

 [ You are not allowed to view attachments ]

« Last Edit: April 08, 2022, 07:33:29 AM by gszlag »
Ambient Weather WS-2000
Ecowitt WS3900 console
Ecowitt GW1000/GW1100
Ecowitt WS68: Anemometer, UV/solar
Ecowitt WH40: Rain gauge
Ecowitt WH57 Lightning sensor
Ecowitt WH32E: Outside T & H sensor
Stratus Rain Gauge (manual)
Raspberry Pi 3B+ (WeeWX/CumulusMX)
Raspberry Pi Zero 2W (WeeWX/MQTT/Belchertown)
---
Barometer wiki: http://meshka.eu/Ecowitt/dokuwiki/doku.php?id=barometer#barometer
---
http://weather.glenns.ca (pwsdashboard - live)
http://weewx.glenns.ca
http://glenns.ca/cumulusmx2/index.htm
---
Uploading to: AWN, ecowitt.net, Weather Underground, PWSweather.com, AWEKAS, Windy.com, WOW

Online gszlag

  • Forecaster
  • *****
  • Posts: 435
  • ..have you calibrated your barometer today?
    • Michael's Bay - Manitoulin Island weather
Re: Algorithm for Predicting Weather
« Reply #24 on: April 08, 2022, 08:01:29 AM »
..found a code snippet based on Dr.KFS research
https://w4krl.com/zambretti-forecaster/
Ambient Weather WS-2000
Ecowitt WS3900 console
Ecowitt GW1000/GW1100
Ecowitt WS68: Anemometer, UV/solar
Ecowitt WH40: Rain gauge
Ecowitt WH57 Lightning sensor
Ecowitt WH32E: Outside T & H sensor
Stratus Rain Gauge (manual)
Raspberry Pi 3B+ (WeeWX/CumulusMX)
Raspberry Pi Zero 2W (WeeWX/MQTT/Belchertown)
---
Barometer wiki: http://meshka.eu/Ecowitt/dokuwiki/doku.php?id=barometer#barometer
---
http://weather.glenns.ca (pwsdashboard - live)
http://weewx.glenns.ca
http://glenns.ca/cumulusmx2/index.htm
---
Uploading to: AWN, ecowitt.net, Weather Underground, PWSweather.com, AWEKAS, Windy.com, WOW

 

anything