wind direction average issue + idea

[Kevin Crivelli]

I kinda want to try and tackle the wind direction average problem that we often see when we go to longer range charts such as months & years. When the wind isn't blowing it tends to take a reading of "north" and that skews the averages data.

My idea is to set a min in the [[MinMax]] section for wind to 1 instead of 0

and maybe set something in corrections like I do for lightning something like

windDir = windDir if  windSpeed < 1 else None

has anyone tried anything like this and if so how were the results? did it mess up your data or help reign it in better like I hope to do

[michael.k...@gmx.at]

The "average" direction is done by calculating the resulting vector. So if the average is tending to north, something in the calculation is not right. Zero wind has no direction. Even if the hardware is showing a direction with zero wind (for instance: the direction the vane is pointing) there cannot be an effect on the "average" direction, because the current vector has a zero length and cannot influence the result of the "average" direction in any way.

Since I never have seen tending the "average" direction in any direction in calm winds: is this really a weewx issue? Or something else (hardware, driver, skin, ...)?

[Cameron D]

When I look at my year chart, winter months tend to westerly, while summer clusters SE or NE.

It sounds as if you might have a problem in your database.

For no wind, the speed value should be zero and the direction NULL.

Is it possible that you are recording a small positive speed instead of zero speed and this would allow a zero for direction instead of null?

Can you examine the database values?