Tempest Rain Inaccuracy (Not Birds)

I have seen a lot of threads discussing false rain readings on the Tempest due to things like vibration, wind or birds but I don’t seem to have any of that, yet my Tempest after a rain event reads about twice the amount as nearby stations or even my old fashioned rain gauge. During the almost 3 weeks since I have had the station installed, I have never had a false reading where it showed it was raining when it was not. Today, we had a gentle rain that lasted several hours. My gauge shows around 0.25" of rain yet my Tempest shows 0.65" of rain. The previous rain event was also exaggerated by 2X. That’s a huge discrepancy. My Tempest is located above the roof of our house with no nearby obstructions like trees.

Will this accuracy improve over time due to RainCheck corrections or will I always have to rely on RainCheck to tell me the following day how much rain actually fell?

Your mount also effects the sensitivity. Weatherflow can adjust your sensitivity to match your other gauges. If you fill in this form weatherflow can adjust the sensitivity to be more accurate. I fill it in regularly trying to help.

But it will not always match because there are so many variables that they are always trying to improve.
I have Skys which are similar but Tempest has an improved sensor mount. Sometimes I think my Sky is amazingly accurate then the next rain event it might be 50% different so I never rely on it when discussing how much rain I have had.
cheers Ian :slight_smile:

Thanks, I’ll do that and see if it improves. FYI… RainCheck updated my amount to 0.29" which is now more or less what my rain gauge says.

1 Like

Just had the remains of Sally come through NC and this is the first time I’ve seen the gauge be this far off. I have 2 rain gauges in the back yard that fill up the a tube, both measured 2.2x amount but tempest came in with a total of just 1.08 for the day. Looking at reports from around the county I’m in, the 2.2x looks about spot on.

It’s a shame to have to post for the very last time on tempest giving low amounts each and every time it rains. It’s a take it or leave it kind of thing and having tempest up for nearly three months has convinced me I will not see much or any improvement in this area. I would love to get feedback from as many tempest owners as possible to hear if you are experiencing the same thing under reporting rain rather then over. If any one is happy with there rain amounts then at least I would know it’s not a complete loss. Today is another classic example of tempest falling way behind on the order of 0.07 to 0.10 range at least. I’m glad many of the other readings are much better just a disappointment about the rain issue.

1 Like

I have issues with my Tempest under-reporting rain totals. I’ve come to accept it’s a shortcoming. So I don’t bother looking at rain totals; just rain notifications.

1 Like

Well, my expectation for the haptic sensor is that it will never be as good as a conventional rain gauge. I’m happy to know that it started raining, how long it rained, and a rough measure of intensity. I take the amount of rain reported as more of a “we got a little rain” vs. “we got a lot of rain” indication instead of a value.

Having said that, we had a short downpour about half an hour ago. I looked at what my Tempest reported - 0.04" of rain. I looked at stations around mine (within a 5-10 mile radius) and they all were in the same ballpark - 0.04 - 0.06" with a couple of others further away at 0.01" and 0.12". So, I’m satisfied with the haptic sensor readings I’m getting, given it’s inherent limitations.

Yeah hate to hear that is what we are up against and that is unacceptable when you shell out nearly 300 or so for better results. They should have put way more thought into this rain part of tempest especially if people are using this as there soul source.

All sensors have their inherent strengths and weaknesses. A haptic rain sensor is excellent at recording the onset of rainfall, how long it has been raining, and an instantaneous measure of the relative intensity of that rain. It’s weakness is an accurate measure of the accumulation. The accuracy of the accumulation expected by WeatherFlow is +/- 10%. In contrast, a tipping bucket is great for accumulation, but weak for onset, duration, and intensity.

It’s very rare to find a sensor that is all around perfect, so trade-offs have to be expected. If accumulation is particularly important for you, then you might need to consider buying a stand along rain gauge, or wait for a tipping bucket to be developed by WeatherFlow

1 Like

Yeah if they come up with a tipping bucket wouldn’t that be something, but I would be surprised if it comes to that.

I’m currently having the same issue. My collocated bucket and the weather station about 1/4 mile away are in sync with each other and the tempest has generally been 40-50% less in its measurements. Today my sensor is at .14 and the tempest is at .06. The sensor down a bit down the road is at .14. I have been providing the results to the google sheet report they have set up. That hasn’t resulted in anything. Will be submitting a ticket to see if it can be addressed.

Great to hear from you, it sure needs addressed and more voices to be heard may help.


Just turn off rain check. Their algo struggles but the sensor self calibrates with a few showers.

does it? if it does, it does it by using raincheck values or if you send the the expected values thru the special form. (https://docs.google.com/forms/d/e/1FAIpQLScIVy0zwUhQfMAN48_N-VZwzyQDx5QB-VXtnmGmlOKW2KbEXA/viewform )

I’ve had my unit for about 45 days and turned off “rain check” on Day 1. I’m not so sure about the rain sensor self calibrating. Yesterday, I had rain that varied from very heavy to light. My Tempest only recorded 0.29 inches while my Atlas reported 0.80 inches. This is not a one-off event but typical where the Tempest is under-reporting rain totals with a delta of at least 0.4 inches.

Bottom line, a rain sensor that uses today’s technology for a haptic sensor will be inferior to a tipping mechanism for total measurements. WF could use software algorithms to compensate for the difference but the new argument will be “actual” data vs “synthetic” data. In other words, did WF create data to make up for the rain total difference.

So I’ve been sending data thru the form and have seen no positive correction to my actual rain amounts compared with my local co-located sensor and one a short distance from my house. Is there an expectation on how that calibration happens?

I had it on but it kept correcting values incorrectly. Once I turned it off our values have been dead in line with our tube gauge. It at the start was way off but kept getting better.

Do you remember the rain intensity levels? The reason I ask is a lot of end-users have noticed the Tempest falls short when the intensity is heavy to extreme.

Rain-check does not run in my area due to terrain. From day one I bought a cheap physical gauge. put it 20ft from the tempest and have been recording its amounts every rain.

Out of the box my tempest was easily 30-50% off and not getting better due to no rain check. I used the google forms to report my co-located gauge and after 10-11 good rain events I emailed support which calibrated my tempest based on those reports.

Since then my tempest has been absolutely spot on with the physical. In fact, its common for the tempest to end up a little over the physical.

I cant vouch for everyone’s situation but it seems if you would record precipitation with a physical guage and have it calibrated on those amounts it ends up spot on.

I am very pleased with the results of recent rainfall measurements. I was very skeptical at first but am convinced now that my tempest will produce accurate rainfall amounts even without rain check


Very intense to extreme. It was bad I’m the beginning but great now