Dis"GUST"ing wind!

Hi, there!

I have just made a minor discovery that strong wind gusts of short duration may not be recorded.

2020-09-18_100000 2020-09-18_100819

The above two graphs (DAVIS and OREGON) show a very strong single gust which occurred yesterday afternoon at 16:52. I do not know the actual peak value or duration but I would guess that it is over 20 m/s for a few hundred milliseconds, judging from damage to foliage. However it is very obvious that it was of very short duration.

The 3rd graph shows the same phenomenon, as recorded by my SKY/AIR/HUB combination, indicating that the gust was a tad over 3.0 m/s (NOT M/S on the graph). It is obvious that the system is unable to follow transient phenomena.


Interesting observation, and I think this highlights the big difference between spinning cups and sonic anemometers. You said that you think the gust was over 20 m/s for a few hundred milliseconds. This was long enough to get the Davis and Oregon cups spinning, and they then took a wind measurement as the cups were spinning down (i.e. the gust had already passed but the cups were still spinning because it takes time for them to slow down). Thus they record a signature of the gust even though it has already passed.

In contrast, the Sky has no moving parts so there is no inertia effect. It measures the wind speed almost instantaneously every three seconds, and therefore unless the gust coincided perfectly with a wind measurement, it would not be recorded by the Sky.


Thank you for your observations, Peter. Obviously, I agree with you and it shows that neither system may possibly be accurate under practical conditions. My knowledge of the effect of wind, under ordinary conditions, is looking at the way the pine trees are blowing 50 m downstream from where I am sitting in my office! Obviously ridiculous from a practical point of view, perhaps the theoretical ideal would be to take spot readings every millisecond and to follow them up with a bit of integral calculus!

For measuring wind speed, but not direction, hotwire anemometry could provide precise millisecond response times. It may even be conceivable to measure direction using the hotwire technique with an array of 8 or 16 sensors, with a thick wire giving a, say, one second or 5 second response time to partially average wind direction.

Obviously, what I am saying above is just my thinking out loud and not necessarily, in the least, practical!

Technically the Tempest could do more sampling for the wind but in how many cases will it bring really more versus, how long will the battery hold … as usual it is a balance between accuracy and longevity.

And more readings means more power usage.

1 Like