New rain rate calculation

[steve]

Currently, during normal running, that rain rate for Fine Offset stations as calculated by Cumulus is simply the amount of rain in the last hour. I have implemented an improved algorithm based on that used by Davis stations internally. I thought I'd explain how it works before I upload it.

The rate is calculated by measuring the time between tips of the rain bucket. After 15 minutes with no rain, the rate goes back to zero. When a reading indicates that one or more tips have occurred since the last reading, the time is noted. If further tips occur before the 15 minute limit, the rate is calculated based on the time between the tips, and the number of tips. Once the tips have stopped, the rate gradually falls back towards zero until the 15-minute limit is reached, when it is set to zero.

I explained that superbly, didn't I? Anyway, in theory it should be better than the current algorithm. Note that the new algorithm only operates during normal running, it's not really suitable for use during history processing. The algorithm for that will stay as it is - based on the amount of rain in each history entry.

I'll be uploading the new beta version shortly. If you decide to give it a try, back up your data. It may well produce silly values.

Steve

[robert]

Hi Steve
Thats sounds good, down here any silly readings will match the silly weather were having at the moment,
Looking forward to the next Download to try

It's all been pretty good so far

Robert

[steve]

Any feedback on how the new rain rate calculation is behaving?

There's a problem for some people at startup with 1.8.3, and I'd quite like to put the fix into the 1.8.4 beta rather than produce another version of 1.8.3 if possible. But if the beta itself has problems then obviously I can't do that.

Steve

[harrym1byt]

Sorry, I can't see any difference - the last rain here occured before you released the new beta. I'm not suffering any startup problems that I'm aware of.

[TrOjAn]

No rain here either... looks like you picked the wrong time to release a rain improvment, unless your releasing it made the sun come out

Cannot believe I want it to rain.. just so I can see it work, might get the hose out

TrOjAn

[dane]

Thank you, Steve, for this wonderful enhancement: haven't had a drop of rain since I installed 1.8.4
But maybe the attached forecast talks about possible rain: I don't know what the word "smallInt"" means

2008-11-22_152211.png

[steve]

A SmallInt is a 16-bit integer. But what it's doing in the forecast, I don't know

This new weather-controlling version of Cumulus is great - it's brought us a bit of snow, looks very nice.

I will investigate that strange forecast.

Edit: It should be 'short'. I should perhaps run a competition to see who can guess what happened

Edit2: By the way, did you turn on the solar display yourself, or did it come up that way? It shouldn't be on unless you turn it on.

Steve

[dane]

did you turn on the solar display yourself

I turned it on - I'm a curious nature....
I know it's meaningless for a fine Offset station, but gives a nice-looking display.
I shall probably turn it off soon.

Edit: re competition: you did a global change to change something from short(float) to integer

[steve]

I know it's meaningless for a fine Offset station, but gives a nice-looking display.

Yes, I know. I was a bit disappointed that I couldn't find a way to fit it in without spoiling the display

Edit: re competition: you did a global change to change something from short(float) to integer

That sort of thing, yes. The forecast code was in javascript, and I used an automated process to do the first pass at translating it to Delphi. A 'short' in javascript is a SmallInt in Delphi. Stupid translation program.

Steve

[Bob]

Steve,

I think there is a problem with this.....

today 2.1mm rain
rate now an all time record of 75.3mm/hr

is this what you wanted?

Bob

[harrym1byt]

Hi Steve,

Take a look at the graphs for yourself - http://myweb.tiscali.co.uk/m1byt/weather/trends.htm

Apparently both the wind speed anemometer and the rain buckets froze up during the early hours.

The anemometer (I think) froze at around 07:30 and started turing again at 09:45.

The rain gauge graph shows the point at which the snow sat on top of the rain gauge suddenly melted at around 09:50, to produce a record 2.54 inches per hour.

[steve]

today 2.1mm rain
rate now an all time record of 75.3mm/hr

If that amount of rain fell very quickly, that rate wouldn't be unreasonable. But, looking at your graphs, the figure does seem a bit high. The graphs are making the right sort of shapes, though. It's just that the peak seems rather high. I'll check the calculations.

Steve

[steve]

Harry,

The calculation is clearly coming out too high - you've just had another surge.

Can one of you send me your Nov08log.txt please?

Steve

[harrym1byt]

Nov08.zip attached.

[Bob]

and mine

Bob