Weather Intelligence Plus

There are a few reasons why we can’t use ET PWS. The main two are that forecasted ET would be much different than observed PWS ET (simulated moisture graphs wouldn’t match reality) and most PWS users haven’t properly configured wind sensors which dramatically impacts ET calculations.


I understand it could implicate Rachio if my Ambient Weather Station isn’t properly installed, but that’s my problem as a PWS user. Our Solar Radiation, Temperature, and Wind Speed makes the largest impacts on my yard. Of couse Humidity matters too, but that’s very closely linked to Solar Radiation here in on the Front Range of Colorado.

If I select a PWS as my source, then shouldn’t all calculations be based on my local or selected PWS? Assuming I trust that PWS as a valid source of data?

Forecasting would of course be much larger scope than my single PWS.

Agreed that an excellently tuned PWS would provide better micro climate based ET (albeit a big ask for most PWS owners). The thing missing even if every PWS was properly setup is still the forecasted ET mismatch with the PWS observed ET, something we continue to think about how to solve.


What parameters of the PWS do you think need “calibration”?

I have compared my Rain Guage on the PWS to a conventional mechanical gauge, I trust the Ambient weather more than the “china calibration” of my conventional gauge, but they are for sure in the “Horse Shoes and Hand Grenades” level of agreement.

Temperature sensors are damn accurate anymore, even without an ICE POINT calibration method, at least close enough to consider the error irrelevant. I have yet to compare the solar radiation to something, but it sure seems to be close enough to the companies like Solar City and others to what they claim I could save if I contracted them to put solar panels on my house.

Wind is one of the more important variables for ET and sensor recommend be 33 feet above the ground which I think most people miss.


1 Like

My PWS is above roof-line or high enough that I consider the windspeed to be accurate enough for me (in compliance with the guideline), but yes @franz, when i had it installed conveniently, the only core aspect I loss was wind speed and direction.

I used 1.5" PVC, and 2" PVC to get it 20’ up. I still swags a bit in the wind, but not going to put a lightning rod up above the roofline, LOL. Even at say 6’ ground level, the temperature and solar radiation are still accurate.

I had my Rachio 2 fail and replaced with Rachio3 (for the water meter). I have a local MESOWEST weather station that is setup and maintained professionally, and much better for our microclimate than anything close by. The local station is available on my old Rachi2 screen (even though the device is offline) but it does not show up with Rachio3. Are you not using PWSWeather anymore?

The next nearest site does not work - we are in the hills and the santa-ana winds make temperatures,winds and humidity VERY different where I am than the nearest sensors.

If you’d like I can have the team remove that feature and get you back to PWS weather, just let us know.


Thanks for answering and for the offer!

With much help from Gene and his old code I’ve been able to cobble together something to get the data pushing to WUnderground as well as PWSWeather. Is there any way I can get the raw data that Rachio is using for my house so I can compare to real-world and my local observations? I know I can see basic info as a forecast and clicking on grayed out days in the schedule, but any way to get more detail?

Also, if they revert me to the PWSWeather, will it use the wind/temps and calculate ET from the PWSWeather site, or still use something else?

Thanks again!


If you get a personal API key from WU (I don’t know if that is still available since IBM acquisition) that will give you similar data, or might provide similar API access?

We still use national data to derive ET.


Thanks Franz,

So for calculating ET you use national wind data with local temperature data? Or for ET are you using both wind and temp data with a national system? Is there a way to find out what data (temp and wind, solar) were used to calculate ET at my location? I’d like to find out how close it is to reality…


This should help answer some of your questions.

Sorry I can not provide access to the data. If you can retrieve data from your closest national station that will be what we are using.


Got it. So there’s no way to go to the old system? The problem is my local station is reported on the NOAA site - it’s MESONET or MESOWEST RAWS (SBVC1), but I’m not sure if that is included in your “national data”. It wasn’t getting pulled into weather underground or PWSWeather. If it isn’t part of the forecast, I don’t see how I’ll be able to use the smart watering features…


@Lmc805 If you look on the weather intelligence map for the closest national station (starts with a K), that is the station we use to derive ET.


Am I understanding this correctly? All the local weather intelligence and personal weather station stuff is only used for precipitation?!

For those of us in Southern California (where precipitation is negligible most of the year, and is less often micro-climate driven) that means the “intelligence” is inversely proportional to your distance from the “National” weather station. All the Weather underground stations start with a “K” so I don’t know which one is used for ET data, but here Evotranspiration and irrigation is effectively all there is. On the other hand, in many cases the coastal weather (where the bigger weather stations are) is dramatically different from the weather a few miles away from the coast. I would strongly suggest giving customers the ability to override the ET calculation source. Even if the forecast is off, that is less of a problem than the actual ET being WAY off right?

From what I can tell the closest national station is 9 miles away. The temperature where I am at 10PM is 14 degrees hotter and the humidity is 23% instead of 66%. How can I correct for that?

1 Like

I agree with you in that it’s not a perfect solution.

There are many more variables than just using immediate ET values from PWS stations for determining day-to-day flex schedule runs. The system needs to use forecasted ET values which we have to use to show a general sense of what watering will happen in the next two weeks (moisture graph) to give customers a very close understanding of how their schedule will behave. Our forecasts from our weather provider uses a blend of national, radar, and satellite data which usually is very close to national station real time data. It’s not perfect but given our data and constraints its the best solution the team has been able to create. If we could get forecasted data based exclusively from PWS stations so expectations would match reality than it would be more of a possibility to use PWS stations for ET (only if the user opts-in).

At one point we were using PWS stations for ET (if the user had selected a PWS station) but that created other problems. The forecasted ET values took into account much higher wind values than localized PWS stations that generally don’t. The soil moisture graph never matched the actual watering which created a lot of confusion.


I’ve been reading these posts concerning ET with interest. So, my understanding is that Rachio uses “national” weather stations (4 letter identifiers) to forecast ET rather than PWS because of more reliable wind values. I’m trying to understand how a “national” station that may be 11 miles away (and normally located at airports with lots of open areas with little vegetation) is going to forecast ET any better than the PWS in my backyard. The wind at most PWSs is going to be different than the wind values at an airport, plus the vegetation mix is also going to be different at both which would affect ET. Wouldn’t you want the forecasted weather values to match the area at the Rachio controller rather than sampling from a different landscape 11 miles away? My Davis Vantage Pro2 PWS calculates ET hourly using air temp, rel. humidity, average wind speed and solar radiation.
I suppose if you obtain more data from “national” stations than you do from PWSs, this would make a difference.

I can relay my experience with ET: my weather station is marginally good at temperature and useless at wind (a big factor for ET here with persistent north winds in the summer). I’ve no idea how good the solar radiation sensor is. My Ambient station will not go over 15 MPH, while a correctly set-up and maintained NOAA station 35 miles north is clearly reading 25+ MPH or gusts above 25 MPH. And that’d be what I’m actually experiencing at my location. I would not want ET to be calculated using my weather station… since my trees would get cumatively underwatered big time. I’m just too lazy to deal with the wind instrument issues. But I am absolutely interested in my rain gage working since that is critical information for irrigating my trees, especially in the spring when transitioning to the hot, windy dry season.

I think at one time a bug caused my PWS station to be used for ET. It was a big, meaningful difference when comparing to a reliable State of California CIMIS station about 30 miles south of here. The incorrect ET also didn’t meet my subjective perception of the weather. We nuked that bug.

Anyway, if a PWS is to be used for ET, it need be “pro” hardware, well setup & calibrated instruments used for calculating ET, and regularly maintained. I’m happy to let NOAA do that :slight_smile: but I’m sure there are a few great PWS’s out there with excellent, knowledgeable people maintaining them. This user is not one of them lol. I just make sure my rain gage is functional.


I just installed my 3.0 a couple of weeks ago and put it on a flex schedule. We are expected about 6 inches of rain in my area over the next 2 days, and I note that the program is telling me it’s not going to run again until May 5, so I assume it’s decision is based on the predicted weather. I’ll have to wait and see if the May 5 date changes if we don’t get as much as predicted.

@marked It looks like flex daily is pushing out the front yard past 5/13 based on precipitation forecasts. If they don’t materialize it will correct for that as well.