How do I get accurate rainfall?

I disabled rain check. It always under reports. So now my rainfall totals are WAY over! For example, on 01/01/2021 I recorded 1.13" of rain (actual) yet TEMPEST shows 2.82" Raincheck trims that to a mere 0.81".

I would rather have the actual rainfall within +/- 0.10" of certified gauges at my location. Particularly in the summer where we can get 5" here and just 1 mile away gets zip!

How can this be achieved?

Mount it on a very short pole in the ground sheltered from wind next to a manual funnel rain gauge to compare and submit accumulation reports to weatherflow to calibrate it.
Cheers Ian :slight_smile:

Ok so what happens when it goes back on the tower where it’s supposed to be to get accurate readings?

1 Like

The rain calibrations change when you change the mount.
If you can not remove ALL the vibrations from the mount on the tower you will get false rain in high wind.
It is difficult to get accurate rain readings when it is on a pole with vibrations but not impossible.
I did build a vibration resistant mount as a challenge to see if I could solve the vibration problem and I believe I was successful.

And the Tempest is less sensitive to vibrations than the Sky due to more cushioning within the Tempest.
I have now removed my vibration resistant mount because I keep experimenting with different designs and it perished in my extreme UV. If I was building another one I would paint the rubber.
I do still use a second Sky in my garden for accurate rainfall.
cheers Ian :slight_smile:

I’ve had strong winds with clear skies and do not see any precip shown. This is not a vibration issue. This is an issue with the sensor itself recording excessive rainfall numbers. I have a friend across town and his is exhibiting the same behavior as well. His is mounted on a mast maybe six feet above the ground.

I think it is just a known limitation of the device. If you want accurate rain measurements you should get a proper rain gauge (analog) and if you are really interested… .join the COCORAHS network. Not just for weather geeks :-).


I don’t think that it is a vibration issue in my situation, either. It is the combination of Rain Check and the sensor reports being far apart.

My Sky is mounted on a heavy-duty satellite dish mount onto the gable end of our house. The pole extends less than 3 feet up above the peak of the roof.
Nevertheless, I had so much problem with Rain Check undercutting the rain amounts that I had to turn off Rain Check several months ago.

The Rain Check would often report totals significantly less with both the official amount for our city (Weather Service) AND other reporting stations that were geographically close (a mile or so) from my house.

I’ve given up paying any attention to the rain amounts from my Weatherflow Sky and I’m not going to go to the expense of purchasing a different rain gauge just to calibrate something in one location and then get someone to have it moved back to the permanent location (I cannot go onto roofs or climb ladders). So for me, the need to get the Sky unit unmounted from the permanent mount, move it to a temporary mount, install a new gauge nearby, send in reports, have Weatherflow do something, unmount the Sky unit from the temporary mount, have it moved back to the permanent location, move the new gauge, etc. – I can’t do all of that.

These things were supposedly calibrated before shipping – and now we hear that to get them to work accurately (maybe) each one needs a separate calibration after shipping. This sounds like either a major quality control issue or a design issue and is disappointing. I"ve had my Weatherflow Sky (and Air) installed for just under 12 months. Most of the sensors seem to be working great – just not the rain sensor.

1 Like

May I suggest you leave your device where it is and send in a few rain reports comparing nearby trusted sources. Simply register your trusted source, ideally one that can have the data viewed publicly by WF, and then send in a few accumulation reports from your Sky. Some of my devices needed it, some did not. They are pretty sensitive to mounting methods.

1 Like

Agree with Carl completely - the reason for getting this was to improve the accuracy of our reporting. Finding out that it is inaccurate is quite disappointing.

I recently inquired about getting cumulative totals - another key reason to have an onsite rain sensor. It appears that was not originally designed into the system - I do hope that you guys are working on this. As a grape grower, we really care about cumulative (and instantaneous) rainfall amounts.

How can we help as engaged users?


I’d suggest spending $35 US for a CoCoRAHS manual rain gauge as a trusted reference of ‘true’ rain and comparing the information daily with your station.

Unless you have a colocated measuring device you can trust is accurate, you’ll never know if the WF gear is reading high or low nor how well (or badly) it is measuring precipitation.

Don’t trust surrounding “trusted” sources because you can’t trust they are any good either.

Instead, colocate a manual gauge and ‘be’ a trusted source for your location.


I find the haptic sensor to be horribly inaccurate in anything other than medium or heavy rain. Light rain and mist simply doesn’t register. For example, we have had heavy mist and light rain all day today. My CoCoRahs rain gauge that is located right next to my Tempest has registered 0.30” today. My Ambient Weather which is located about 20’ away has registered 0.33”. My tempest is reporting a whopping 0.03”. The Tempest up the road from me is registering 0.05”. I am reporting my CoCoRahs amounts to weather flow via the form, but the problem is a severe accuracy problem inherent in the design. The sensor has to “hear” rain…it can’t “hear” mist or light rain so it doesn’t get registered. Not sure any amount of calibration will fix that. It’s the only part of the sensor I don’t like and don’t really rely on at this point. Wish there was a way to have the rain sensor as a traditional bucket type or similar rather than a haptic sensor.


Well I guess it beats the weather rock. :wink:

The Sky or Tempest mount and location effects its calibration.
If rain accuracy is the only important reason for owning it (some people might use it to control garden watering and do not care about the wind accuracy) Then to achieve the most accurate rain calibration it works better on a very short pole in the ground.
When the same Sky or Tempest is moved to a different mount its sensitivity in different types of rain and wind changes so it is not possible to adjust its sensitivity on one mount and expect it to maintain its accuracy when it is moved to a tower.
Weatherflow can change many of the rain sensitivity settings and must be a complicated process to adjust the calibrations for the many types of rain in different winds and then on different mounts.
I have 2 manual funnel gauges on opposite sides of the house and they sometimes record 10% different.
To have your calibrations improved requires you to submit accurate local readings.
After every rain event over a day or so I fill out this form:

Sometimes my readings are spot on.
Many times they are not accurate so each time I submit my report is adding to their data to improve the calibration in different circumstances.
The Tempest is more accurate than the Sky due to extra cushioning and less vertical height near the haptic sensor.
When I look at the history of my past 70 reports for the past 11 months the total in the funnels was 1404mm and the total from the Sky was 1186mm edit and the average accuracy was 94%.
cheers Ian :slight_smile:

Actual it’s 85% of actual, which is horrid. A manual gauge is +/- one or two percent depending on how precisely you read it.

What we need is a way to manually calibrate the WF rain accumulation (sensor).

You can’t calibrate wrong when it’s hard rain, missing when it’s light rain, and wrong at other times. Best you can do is turn raincheck off and cross your fingers.

I live in a suburb surrounded by trees and buildings. My two funnel gauges are some times 50% different to each other because I am exposed to strong winds. I watch rain bouncing off the roof of the neighbours house in strong wind which causes double the amount of rain where it lands, and less where it is sheltered.
The amounts of rain falling on my roof top is very different to what falls in the garden.
It is only since recording the rain from 2 funnel gauges that I have learned how different the amounts can be within 40m of each other.
My figures change over time with calibrations so giving my overall totals is not an accurate indication, but is just to get a rough idea.
I do have a Tempest and I have noticed that it is much better than my Sky at recording accurately but I have changed its mount so I can not show a result.
cheers Ian :slight_smile:

I’ve pretty much given up on getting reliable rain readings from my Tempest. For a month now I’ve struggled with the mount and trying different things. I now have a setup that is vibration-free along the pole, but I still get false rain reading when the wind picks up. My Tempest is located in an area of the Texas Hill Country that is perpetually windy. I get gusts over 30mph almost every day with sustained winds easily in the 8-12mph range most of the time.

Just now I was getting false (but tiny) rain readings, so I went over to the Tempest to see what was happening. There were no vibrations on the pole that I could feel, but I did notice that when the wind blew strong the Tempest itself would whistle. That seemed to coincide with a false rain reading. So, I now think the problem is that the Tempest sets itself off in strong winds due to it’s own wind-whistling.



Hi Brian,
Wow that is a great find. It makes me think about installing a microphone to listen to what is happening in my Tempest…
Cheers Ian :slight_smile: