Welcome to the Cumulus Support forum.

Latest Cumulus MX release v3.2.4 build 3060 - 10 December 2019.
    Legacy Cumulus 1 release v1.9.4 (build 1099) - 28 November 2014

    See the Wiki for downloading either version.

    Lack of accuracy of weather station and Cumulus forecasts

    Discussion of version 3 of Cumulus, which runs on Windows, Linux, and OS X. All Cumulus MX queries in here, please.

    Moderator: mcrossley

    Post Reply
    martinu
    Posts: 51
    Joined: Wed 01 Jan 2014 7:54 pm
    Weather Station: Davis Vantage Vue
    Operating System: Raspberry Pi: Raspbian Stretch
    Location: Rudston, North Yorkshire

    Lack of accuracy of weather station and Cumulus forecasts

    Post by martinu » Sat 29 Jun 2019 2:43 pm

    I appreciate that weather stations and Cumulus use fairly low-tech algorithms to work out a weather forecast, but has anyone ever had a forecast that bears any resemblence to the current or near-future weather?

    I've set the lat/long/altitude to my location. The wind, rainfall, temperature, pressure and humidity readings seem sensible - and changing over time. The weather station (Davis Vantage Vue) is sited about 1.5 m above ground on a pole which is on a large area of grass, so the temperature is not affected by heat reflected off a hot patio, and there's probably about 10 m all round that is clear of bushes etc. (When I work out the best place, I'd like to mount the station higher up, clear of many of the bushes and yet hopefully not shielded too much by the gable-end of the roof or subject to air that has been heated by roof tiles.)

    And yet the station and Cumulus forecasts seem to be fairly unchanging and generally pessimistic, no matter what the weather is doing.

    As an example, it is very hot and sunny at the moment, and MetOffice and The Weather Outlook predict this to continue (full sun, temp gradually falling from 20-17 deg C between 1500 and 2200, then cloudy overnight and hazy sun tomorrow from 0500 to 1200; chance of rain < 5% throughout). But the weather station and Cumulus predict "Increasing clouds with little temperature change. Precipitation possible and windy within 6 hours" and "Occasional rain, worsening" respectively.

    Is there some factor that I've set wrongly which causes *such* incorrect forecasts, and forecasts which vary very little despite the real weather?

    Weather station results are at http://goosebears.co.uk/weather/ Ignore the static temp and humidity between about 1400 and 1500 today - the console lost comms with the remote sensor and had to be rebooted. The spike in inside temp at around 0800-1000 happens some mornings - I think low-angle early-morning sun is falling on the console: I'll have to try moving it a bit.

    User avatar
    beteljuice
    Posts: 2792
    Joined: Tue 09 Dec 2008 1:37 pm
    Weather Station: None !
    Operating System: ex XP SP3 - now W10 laptop grr
    Location: Dudley, West Midlands, UK

    Re: Lack of accuracy of weather station and Cumulus forecasts

    Post by beteljuice » Sat 29 Jun 2019 8:49 pm

    ??? There's actually very little difference (apart from a few billion £ worth of investment)

    .... cloudy overnight and hazy sun tomorrow from 0500 to 1200; chance of rain < 5% throughout
    "Increasing clouds with little temperature change. Precipitation possible and windy within 6 hours"
    "Occasional rain, worsening"

    Saying the same thing surely ?

    Actually for your Cumulus / Zambretti forecast, the pressure trend is 'on the cusp' and if it were a little bit lower the forecast would become "Fine, possible showers".
    Image
    ......................Imagine, what you will KNOW tomorrow !

    freddie
    Posts: 521
    Joined: Wed 08 Jun 2011 11:19 am
    Weather Station: Davis Vantage Pro 2
    Operating System: GNU/Linux (Ubuntu 18.04)
    Location: Dorrington, Shropshire
    Contact:

    Re: Lack of accuracy of weather station and Cumulus forecasts

    Post by freddie » Sun 30 Jun 2019 11:17 am

    You should bear in mind that the forecasts are based on algorithms that highlight the weather that typically occurs with a given set of parameters. The troposphere is a very complex fluid in motion and you should not expect too much from an algorithm that has just a set of readings from a single point in that fluid as its input parameters.
    Freddie
    Image

    Post Reply