Rachio/Tempest weather

Recently installed and linked a Tempest to my Rachio gen 2.

Rachio is now using the Tempest for weather info and schedule control. Problem is, Rachio went ahead with watering scheduled for today despite noted anticipation of heavy thunderstorms. Seems like the networked weather intelligence did a better job of predicting and controlling than this hyper local PWS.

Heavy thunderstorms in the area, but does it show what the estimated precipitation is for these storms? Does that precipitation exceed the rain skip threshold (default is .25" I believe)?

From what I can see in the Tempest app, it doesn’t show anticipated rainfall amounts, only percent chance of rain. We actually ended up getting about .9”

Seems like the weather network did a better job than the linked Tempest. I do like the other data the PWS provides but doesn’t seem to be as useful for controlling rain skips.

I might have to defer to @franz for this one. I’m not sure that the Tempest is responsible for the forecasting data. I think it is more for live readings, and maybe forecasting is still handled by the same forecasting system for any/all WI.

That would make sense. Still odd that the schedule ran with the forecast as it was. Usually it is spot on.

That would make sense. The Tempest is not a forecasting device (although it presents a forecast in its own app display). Even when my (or your) Tempest is part of the Weather Underground network, it only supplies measured weather information that’s used as part of forecasting services that might be able to use it. Therefore, as a hyper-local weather station, Tempest is perfect for handling a LOCAL rain skip as it experiences it (like sticking our head out the window), but neither Tempest nor Rachio plans rain skips based on forecasts that I’m aware of (or experienced). In simple terms - if Rachio, just prior to the time a station is to be watered, can determine (from your Tempest or a near-local weather source or using Weather Intelligence) that sufficient water has already fallen on the station to skip the scheduled watering, it will do a “rain skip” if you’ve enable them. Otherwise, it won’t (including when watering the station is already underway).

I’m not a Rachio engineer, so the real device logic for skips might be different than I’ve described, but I’ve described exactly how it behaves in my configuration (Rachio, Tempest, Weather Underground, WI+, Flex Daily, flow metered).

1 Like

I think you have it pretty accurate. I’m also not a Rachio engineer or employee, so I might not be 100% in my statements above.

But, like you said, neither Rachio or Tempest have meteorologists on staff looking at everyone’s weather stations. They both pay for a service from somewhere…I’m just not 100% sure if Rachio uses the Tempest forecasting when it is linked, or stays with it’s own forecasting service. The live data is where the hyper-local benefit comes into play from the Tempest!

@Mikegoldnj Tempest DOES actually have their own forecasting network that they have built.

We do actually take forecasting into place and use that data it as much as observations. If the % of probability is > 80% we give the forecast a weight of 1 (.5 inches predicted is .5 inches assumed for data to skip).

If the % of probability is < 80% we use that as the multiplier. (i.e. .5 inches predicted at 70% is .5 inches x .7 used for data to determine if we skip)

@Mikegoldnj PM me the username you use for your Rachio account and I can have the engineering determine if their is a defect or not.



@franz, for the forecasting seen in Rachio, is that being pulled from the Tempest ecosystem (when linked) and bypassing Rachio’s normal forecasting source?

1 Like

Thanks! I knew that Rachio was using the weather network intelligence forecasts to initiate skips (and I’ve also seen instances where the skip is then reversed based on updated forecast). Just seemed like forecast wasn’t being taken into account in this last instance with my Tempest.

I will PM shortly


Yes correct. We wanted to use all of the Tempest weather data for customers with this first class integration.



@Mikegoldnj Weather review from the team.

Here is raw data, does look like we are using your weather flow station (which is good).

{"timestamp":1688440819,"scheduleId":"XXXXX","weatherStationId":"MJG 01","scheduleCriteria":{"scheduleType":"FIXED","startDate":"2023-06-10","endSunTime":"SUNRISE","rainDelayEnabled":true,"freezeDelayEnabled":true,"windDelayEnabled":false,"climateSkip":true,"seasonalShift":true,"smartCycle":true,"cycleSoak":false,"thriveModeEnabled":false,"soakTime":0,"cycleTime":0,"zoneDelayTime":0},"freezeSkipApplied":false,"climateSkipApplied":false,"precipSkipApplied":false,"noSkipApplied":true,"freezeThreshold":4.0,"precipThreshold":6.35,"temperature":22.0,"forecastedPrecip":3.348,"observedPrecip":0.0,"totalPrecip":3.348,"moistureLevelPredictedAtNextRunIfSkipped":-5.544154421257904,"runStartTime":"2023-07-04T00:11:55-04:00","runEndTime":"2023-07-04T05:32:00-04:00","runIntendedTime":"2023-07-04T05:32:00-04:00","sourceWeatherStations":"109549","weatherServiceFailure":false,"skipValidation":false}

The forecasted precip was 3.348 mm (0.13 inches). It also looks like you have rain skip set at .25 in which is actually a lot of rain. If you had it set at 0.125 this particular schedule would have skipped as over the threshold.

I can’t comment on Tempest’s forecasting engine (accuracy vs. our default weather service) since it probably performs better in certain areas of the country, as they all do.

The one guarantee is that any observations will be directly taken from your property which will definitely help our weather intelligence engine accuracy. Hope this helps.



Thanks! This is great and good to hear.

1 Like

Having installed a Tempest system last Fall and doing some minor tweaking of crop coefficients (now using the dynamic setting) and/or allowed depletion (due to soils), I would have to say that my Rachio system is performing noticeably better than it ever has. There is notably less over or underwatering this year. The “cadence” of watering seems nearly spot on.

1 Like

My Tempest integration definitely looks at potential rainfall, not just rain that has already fallen. In my area (SE FL) it is very accurate.

When i get a rain skip, I get an email that states the actual rainfall (if any so far), and anticipated rainfall (if any), and the sum of these two numbers. It seems to work exactly the same as the PWS weather network in regard to the forecast, and of course much more accurate in regard to the actual rainfall at my location.

However, I have no idea from where Tempest gets its forecast information from. Perhaps the National Weather Service?

If you want to look at mine, check the below link for the overall forecast:

and this link for the specific information from my weather station. If you click on an individual card, you can see the history:

1 Like

Do you have “Rain Check” enabled on your tempest by chance?

I’m not familiar with rain check. I do have rain start activated.

See Here

We found a noticeable difference in rain reporting with our Tempest with Nearcast Rain reporting turned on (the default for Tempest as delivered). NC Rain - WeatherFlow’s “neighborhood” rain reporting estimates - claims to report rain in 100 m x 100 m boxes based on multiple regional devices. We’re in an area in Texas with spurious, localized, and significant rain events, and disabling NS Rain and reporting actual single device results made an observable difference.

The setting in the Tempest app is at Settings / Stations / Your Station / Manage Devices / Your Device / Advanced / Show NC Rain (i)

Can’t easily tell if Rachio gets Tempest rainfall amounts as configured by Nearcast Rain setting or always from raw local device measurements, regardless of the setting. Our best Tempest-to-manual rain gauge comparisons seem to easily favor the local device measurements. My neighbors - all of whom are over 100 m. away) who use my Tempest as their “local” station might be misled by my reporting.

1 Like

Yes, I see that it is activated. I looked at the rain data and do not see where I can see the difference between the raw data and the corrected data. Last week I had two days with over 2 inches of rain each, and during the first 2+ inch rain event we had wind gusts over 90 mph, and sustained winds of about 35 mph.