Cumulus Support

CumulusMX seems to produce some really strange standard forecasts. For instance webtag #forecastenc is currently saying "Occasional rain, worsening"!! Where did it get that from when there is not a cloud in the sky and humidity is at 51%?

The Sager forecast from BT's script is currently saying "Fair; No important change. Some tendency for slight increase in winds during day, diminishing in evening. East or Southeast winds.", which is much more accurate.

Cumulus standard forecasts will not often be correct.

TL;DR
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.

water01 wrote: ↑Fri 24 Apr 2020 11:46 am Where did it get that from when there is not a cloud in the sky and humidity is at 51%?

Adding to what Freddie says (although I recognise his intended brevity), the Cumulus forecast only uses current pressure, whether that is rising or falling, and current wind direction as inputs, as far as I recall it does not take account of rate of change of pressure parameter or whether wind direction is variable. It certainly does not take inputs about cloud cover or humidity, so what Freddie calls a set of readings is a tiny proportion of what is available about that single point!

freddie wrote: ↑Fri 24 Apr 2020 12:28 pm 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.

Calculations based on 3 values are never going to give more that a guess at the forecast. You would need to read up about the 1915 Negretti and Zambra forecaster to learn more about the available combinations.

I would hazard a guess that Steve Loft would say he added the feature, in version 1.8.3 if it matters, as a bit of fun, not as a serious forecast, in fact I'm pretty sure he wrote somewhere you should remember the Met Office uses high computer power and lots of modelling to attempt a good forecast. I have not found that, but I did find viewtopic.php?f=4&t=112&p=644&hilit=forecast#p649.

water01 wrote: ↑Fri 24 Apr 2020 11:46 am The Sager forecast from BT's script

You can read in his thread how much more complicated the processing is for this feature. From memory, he uses a METAR report as source that does include cloud cover, humidity and much more in the way of inputs.

OK thanks both that helps me understand. Think I will have to try harder with the changes that betaljuice pointed me at to get the BT Sager forecast appearing in the SteelSeries gauges scrolling bar as it much more correct.

One other thing about the forecast, you should set the min/max pressure values used by Cumulus in the settings to values applicable to your location.

The forecast algorithm was really intended to be used in Western Europe, is application elsewhere is even more suspect!

It should be noted that both Zambretti and Sager forecasts should really have 'observational' data.
ie. The wind direction should really be at altitude (cloud movement) and compared to ground level (veer or not). None of this data is available - so second guessing just from ground level.

BTJ implementation of Sager 'cheats' in getting some data from a METAR source, but still is not altitude winds.
Might as well get the TAF if there is one to be honest

Station forecast (which ever one) either works for your locale or not - simple as.
.. but should give good indications on rapid change etc.