What is wind gust anyway



I’m having problem with wind gust is general. If you want to describe the wind in the last hour, you can have average wind speed and direction, but what about gust. Do you just mention that tiny little moment the wind had the highest speed, or are you going to take all gusts together in that period and average them out.
Wind gust depends on the sampling period, If I would sample wind once a minute, I could still calculate the maximum in the last hour and report that as the gust, but it would be different when sampled every second.

https://graphical.weather.gov/definitions/defineWindGust1hr.html defines gust as
Is a sudden, brief increase in speed of the wind. According to U.S. weather observing practice, gusts are reported when the peak wind speed reaches at least 16 knots and the variation in wind speed between the peaks and lulls is at least 9 knots. The duration of a gust is usually less than 20 seconds.

is that part of the gust report in weatherflow? I don’t think so.

anyway, it is not as simple as taking the max wind speed (and its direction) over some period.

Basically what I’m saying is that I don’t understand the current gust, unless someone defines it as simply being the max for practical reasons, in which case it would probably be better to call it just that, max wind speed.


According to the KNMI
maximum gust of wind (gust):
This is meant the maximum wind speed in a certain period of time.
Operationally it concerns the highest recorded 3 "average value
in a defined period of time, KNMI use 10 minutes. Values are in m/s


even that operation definition needs a bit more clarification. Is the 10 minute interval a sliding interval of the last 10 minutes or is the day simply divided into fixed segments of 10 minutes each.

What happens when weatherflow sky is in powersave and samples only every 15 seconds. I just realize that this is a question that should have a statistical answer. The 3 second sampling will have, on average, a higher max value than the 15 second sampling. Given the spread in 15 seconds sampling values and its max value, is there a statistical formula that tells how much higher the 3 second sampling max value is expected to be?


Well thats a good question about sliding intervals it seems to me that this is not used by the KNMI (In my opinion it should be a sliding time and not predifined intervals.

This is the pdf from the KNMI http://projects.knmi.nl/hawa/pdf/Handboek_H05.pdf for windobservation unfurtenally it’s in Dutch only and a lot is behind my knowledge


The highest vslue in previous rapid wind readings.

It it a not


interesting read, this KNMI stuff (official dutch weather agency).
interesting because it states that the wind direction is explicitly not a weighted average, but an average where each windvector is first scaled back to a unitvector, before being averaged. It is based on a mechanical construction, so at zero windspeed, the last recorded value is being used.

Also interesting, because all wind measurements are based on a 3 second sliding period consisting of 4 samples per second.

Most of the article is about how measurements are reported in abbreviated codes targeted for specific use, e.g. aerospace. They all have different ways of defining average windspeed and gust, which makes it hard to compare.


@sunny - I think you’re over-thinking this one.

WF calculates wind values by simple math using whatever rapid_wind measurements came in over the 60 second obs_sky period.

  • wind average is the simple average of those measurements
  • wind lull is the minimum of those measurements
  • wind gust is the max of those measurements

If you are in battery saving mode, you have 4 measurements to use. If you are in normal mode, you have 20 measurements to use…but regardless the math is the same.

Yes, running in battery saving mode means you’ll miss gusts that were in-between the every-15-second rapid_wind measurements and the math will come out how the math comes out.

This came up 3 weeks ago - you might want to read the bottom 4-5 posts of SKY: power save mode = 2yrs+ battery life


so the values of gust will be different when using powersave or not, which makes it hard to compare. The lower sampling rate could perhaps statistically be corrected to match the 3" one.
Actually I just found this. Formula 15) should give the answer.

But after reading the article from KNMI and others sources it is clear that there isn’t a clear answer. Both KNMI and linked article above do mention the 10 minute period for defining gust values.


There are many ways it is measured and calculated.

Pick the one you like.


As I am starting to discover from the parallel weewx UDP driver discussion on windGust and windGustDir, there is probably no wrong answer.

As the poor souls at WF can surely attest from the number of times their interpretations about calculated values are questioned, there is probably no correct answer either.

I believe that this is a classic 21st century example of the Kobayashi Maru…


Indeed, you might start feeling sorry for them. Creating a possibly better weather station, but being stuck with old and inconsistent ways of representing the data.


Stop and think about what are thinking.

First, DO NOT place the hardware and the current applications on the same level.

The hardware output is constantly being improved.

The applications are designed for the simplest of users.

The applications are NOT designed for those that want advanced weather information.


Preferring a little bit more precise display of the app is not what I intended to express here. What I was trying to say that it is a bit of a pitty (but understandable) that the hardware is capable to produce better results then what old and inconsistent rules of weather agencies specify. The rules are created for older types of instruments and human observations. Of course it somewhat makes sense, so one can compare with historic data.


Hi @sunny . Thanks for the productive discussion. First, make sure you are familiar with Data archive buckets explained. Then please let us know in detail how you propose the data to be displayed and we’ll take it into consideration. Thanks!


Some explanation on what we currently display:
SKY takes an instantaneous observation every 3sec (or 15sec in power save mode). Captures speed and direction.

1min Wind Average = sum of the obs over the last 60seconds divided by number of obs. Same logic for 5min, 30min, 3hr, 1day per Data archive buckets explained

1min Wind gust = the highest single ob over the last 60sec.

1min Wind lull = the lowest single ob over the last 60sec.

Curious to hear thoughts on labeling these values. “Gust / Lull” vs. “Max / Min” ? Remember, the default dashboard display data is purposefully simple intended for the average home owner, not the professional meteorologist. Home owner: “wow, it was super windy today. I wonder how windy it was to blow over my pool umbrella?” :wink:

For the professional meteorologists and serious weather gurus like you gals / guys … remember to keep the target audience in perspective. And should you desire custom derived data (like an alternative definition of gust), we absolutely encourage data manipulation by API or third party application!