Wind direction in UDP messages

Developers
,
1

For the Tempest hub’s ‘obs_st’ messages, is the wind direction parameter just a moment in time reading, similar to the ‘rapid_wind’ messages, or is it some kind of average of the direction over the reporting interval?

Thanks!

2

Everything in an observation (not an event) is an average over that observation’s time period, as far as I’m aware.

2 Likes
3

From what David has indicated, the device takes a wind reading every three seconds and records it. The one minute reading is the calculated average of the 20 readings.

1 Like
4

So I averaged the direction degree values from the rapid_wind readings over the 1 minute interval and they aren’t matching up to the direction degree value in the obs_st readings… I was just doing a simple average though so I am not sure if that works on degrees… I did notice that the degrees in the last rapid_wind reading before a obs_st reading seem to be very close or often match, so it seems to me that its not an average, just the most recently measured direction reading…

5

Unfortunately a simple average does not work. You have to do a vector average. To illustrate, if the wind is blowing with a direction of 350 degrees for 3 seconds, and then it blows with a direction of 10 degrees for 3 seconds, we can instinctively see that the average direction is exactly 0 degrees. However, a simple arithmetic mean (350+10)/2 will give 180 degrees. This is not correct.

Similarly, the arithmetic mean of 0 and 360 degrees is also 180 degrees, despite the fact that 0 and 360 degrees are the same direction.

You can read more here and see the equation required to calculate the mean of circular quantities: https://en.wikipedia.org/wiki/Mean_of_circular_quantities

6

sure - but to answer the original question, it is an average value as calculated and reported by WF. Whether it’s a simple average or a weighted one doesn’t matter re: the original question in this thread - it’s an average value, not an instantaneous value.

7

Yes, I completely agree. I was just responding to this:

explaining why a simple average of the rapid_wind messages won’t match the 1 minute obs. The mis-match doesn’t mean that WF are taking an instantaneous value.