Something Wrong With The Anemometer?

Today, the station I maintain at Raleigh’s Place began reporting high wind gusts along with a NE to E direction when the rest of the reporting stations that are within 10mi are reporting S & SE winds. The highest wind gust so far from the other reporting stations is 29mph. The Tempest recorded 47mph. I figured it could be some water collecting near the wind gap that could be causing the speed issue. Not sure why it’s reporting winds from the NE though. Tempest is oriented 0° True N.

23872 is the station ID

@KJ.Davis I remember you telling me that you had some issues before with water collecting just outside of the wind gap.

Here is an update: The anemometer appears to be back reporting wind data that is similar to the nearby stations…

Yes, droplets that were suppose to continue falling down the side of the top would stop at the bottom and hang. Those droplets would cause wind speeds to skyrocket. No droplets would be present on sonic transducer plates, only where I drew the arrow in the photo below. Once I dried off the droplets, the wind speeds returned to a normal level.

1 Like

I wonder if it would also affect wind direction reading?

It did on my Tempest. Direction and speed were affected by the hanging droplets.

1 Like

That’s what happened on the Raleigh’s Place Tempest. My tempest down in PCB probably had the same issue when Sally was around.

The CL algorithm seems to ignore this wind speed & direction anomaly and just let them be recorded. It has happened to me on several occasions.

I don’t think CL looks through the wind data. Just QCs UV, Rain, Pressure, & RH.

@dsj @tony.mcgee @WFsupport

I apologize for the mistake in saying CL algorithm. It’s more of a filter algorithm that the Hub is suppose to check for this kind of issue before sending the data on to the servers. Basically, I still see it as a form of CL algorithm since it should learn by looking and filtering out the issues even though WeatherFlow may not directly classify it as CL.

1 Like

No problem. I’m far from being an expert on WF’s system. The stuff you just told me I didn’t even know about.

1 Like

But how would the hub recognize it? The only reason I know is that I compared the data to quality sources.

As far as I understand it, SKY and Tempest run the wind RAW data through a quality control filter that is processed in real-time before being sent to the HUB. Next, the HUB runs that data through another quality control filter before the data is finally sent to the WeatherFlow servers. SKY and Tempest were designed to get smarter as time goes on from processing this data to stop this type of problem. The key is “get smarter” which tells me it is basically like a form of Continuous Learning Algorithm.

Sometimes this bad data gets through and we see it in the graphs. All we can do right now is try and catch it ourselves and get the water droplets off the affected areas to help fix the bad data. Hopefully, WeatherFlow can one day have these filters block it entirely. This is one reason why all three of my weather stations are easy to reach. You never know when you might have to climb the ladder.

Today after a feeder band of rain from Delta passed by, I had to wipe the droplets off around the wind gap area. The winds were reading NW when they were actually from the SE plus wind gust were higher. After wiping the droplets away the speed and direction returned to normal. See the graph below where you can clearly see when I wiped the droplets away.


Fantastic mounting job there!

Thanks! It’s a little more steadier than the previous mount.

Mine was from the NE and winds was 25-40mph even though it was lighter. Do you think WF might need to improve the Tempest design to keep this issue from happening?

It may be hard to keep every little variable from skewing the data from sonic anemometers, but I’m not an engineer so leave that for the WeatherFlow team. What I know from what I have experienced is that issue may need to be addressed. Next time it happens to me, I will take a photo of the droplets on the sides while the speed and direction shows erroneous data.

WeatherFlow knows what they are doing and it’s hard to account for everything. As long as we share our experiences to help them see and fix potential problems, it’s a win for everyone. Even if it takes time in these challenging everyday circumstances.


I know that the ultrasonic anemometer that Davis uses is terrible when it comes to rain. Not sure how well the ambient weather ultrasonic works during misting rain either. That coating on the outside would help except it would be rubbed off due to handling. Wished there was something else out there that they could use to get rid of the hanging droplets.

1 Like

I wonder a viable solution is to put rain-x on the outside area of the wind tunnel. I’m thinking if I put rain-x on the unit, it would help keep the droplets from sticking. Thoughts?