# Community Support

Hi Pierre 🙂Â

You probably already heard of this indicator but I find it a good idea to have windrun as an accumulated value in Weather Station. Windrun is basically the distance wind travells at a given point over a given period of time. For instance, if the windspeed in the last hour was 20 kph in average the wind has travelled 20 kilometers. Thus, windrun is another (in my opinion more visual) way of describing the average windspeed.

At the moment I can give you some (unfortunatly not really reliable) reference, but I will provide more information as soon as I find some:

Windrun allows for a lot of interpretation and I would highly appreciate using it in the wonderful structures of Weather Station.

Chris

Â

I like the cold weather. It means you get work done.

Hi Chris!

I understand really well what is windburn (thanks to your description). I've just made some manual tests and, of course, if you choose to make it on a 10 minutes interval, hour interval, etc. I obtain (very) different values. Do you know the "recommended" way to do that? Do you have some links to point me out?

Thanks.

If you like Weather Station, please consider to make a review to help make it known. That would be the best way to thank me...

Hi Pierre!

I havn't really found anything useful on the web, however, I can provide you with an excerpt from "The Weather Observer's Handbook" by Stephen Burt. Lukily the section on windrun is publicly available: Link (Hope the Link works, otherwise I can provide a screenshot from my copy of the book).

As for the duration I mostly have seen windrun values accumulated over 24h. But since the value itself represents kind of an average also other timeframes are possible. For simplicity though, I'd go for one day 😉Â

Chris

Â

I like the cold weather. It means you get work done.

Ahaha. The page is not available, so if you can send me the excerpt by mail, that would be cool ðŸ˜‰

For the range, I don't speak of aggregation timerange. It seems 24 hours is the good one. I speak about the average value to take for the computation: if I take 10 minutes values, hourly values, etc. it gives me different wind run values (and yes, I compensate the sample length). So I hope your book will give me hints ðŸ˜‰

Â

If you like Weather Station, please consider to make a review to help make it known. That would be the best way to thank me...

Yes, I received the excerpt of the book (thank you for that) and, as you understand now, it doesn't help so much...

I will try to think about that by doing some tests on data I have. Just tell me when (if) you found something...Â

If you like Weather Station, please consider to make a review to help make it known. That would be the best way to thank me...

Ok, I put some thought into this, but before I continue, can you please clarify some things:

- When calculating the daily windrun a simple formula has to be fulfilled and can be stated as "golden rule":

windrun [km] = avg_windspeed [kph] * 24h

Right? - How is the daily average windspeed calculated in Weather Station (using data from each measurement or 10min averages etc.)?

I like the cold weather. It means you get work done.

I've try mainly with the average (hum, the arithmetic mean 😇 Â ) of 6Â successive records to obtain an hourly average, then the same formula as you.

I've made some tries too, with the median or mid-range (but that is worst in the 2 cases).

And some other things, but worst than worst ðŸ˜‰

Ok that's a bit weird, because theoretically the averaging period shouldn't change the outcome for windrun. Let's assume you get a wind speed value every minute (e.g. 1, 1, 2, 3, 5, 8, 13, 21, 34, 55 kph,...). For this dataset the windrun then gives *sum of all measurements * 1/60 = 2.38 km*. If I transfer this now to a two minute average the dataset is 1, 2.5, 6.5, 17, 44.5). Now I can apply again *sum of all measurements * 1/30 = 2.38 km*. And so on...

But if we're talking about moving averages it's an altogether different story and I'm currently thinking about this situation 🤔Â

P.S.: Mentioning the arithmetic mean is highly appreciated 🤗Â

I like the cold weather. It means you get work done.

Experimenting now with data from one of our stations I would go for a rather simple approach (though I'm not sure if this is possible in Weather Station): Everytime Weather Station fetches a new record, use the wind strength and multiply it with the difference between the timestamps of the last two records. Accumulate this value starting at 00:00 with windrun = 0.

Easy as this may sound (and I'm sorry if I was talking about the obvious so far) there is a major problem: The above assumes that weather data are gathered at rather small intervals (1-2 min). For Weather Station the intervals are usally higher and if you are only able to fetch the current wind reading (i.e. directly from the sensor without any major averaging), the error you get is quite significant. So we have to talk about the way stations are reporting wind readings. For file-based stations it's easier, because for instance the realtime-file includes an averaged wind reading over the last couple of minutes. However, for other stations you might get only a snapshot every 10 minutes or so and then you have to ask if it makes sence to calculate windrun and even how accurate the daily average values are...

The more I think about the implementation of windrun the more difficult it seems to me and I'm really starting to wonder if it makes sence for all kinds of stations.

I like the cold weather. It means you get work done.

Hi Chris.

You're following the same road as me: the more I think (and try), the more I understand the computation method depends on what the station is able to give as data aggregation. Not a system have the same way to do things as others...Â 🙄 Â That's really annoying (and I moderate my words)...*<digression>The absence of standardization in weather hardware protocols is absolutely stunning. That's not the first time I am confronted with this problem. We could almost believe that it's a competition to do things differently and inconsistently</digression>*

I've tried to do tests on my stations, and the best result so far is to take the daily median (the true median ðŸ˜‰ ) and multiply it by 24. But it converges really after 12-18 hours, not before. So doing it like that will allow to have it for each day (in historical data) but not for daily data (I had initially thought of something like "wind run so far..." in the current day).

I will try to investigate more on this, but I think it will be too short for 3.8. But, do we ever know ?

You're quite right and it's a bit frustrating searching for standards in weather station hardware (and software) 😐Â

I think using a 24 hour approach is sufficient, keeping in mind though, that for some stations the daily average (i.e. all types of averages 😀) windspeed values might not be as accurate as they suggest. But this is a problem of the stations and not of Weather Station.

Thanks again for your time and effort and I'm looking forward to it whenever it's ready 😉Â

I like the cold weather. It means you get work done.

I will try to make something "experimental" to see what happens. Maybe with parameters. I will try to do it for the 3.8...

Â