I see my IRO watered today from my web dashboard Moisture levels, but I do not see it in the history panel. Also not getting alerts like I used to. Any ideas?
Any rain forecast for today?
No doesn’t look like it.
You are showing a 20% chance of precipitation. Click on “more details” for this zone and post that here. We should be able to tell what is happening from that. Most likely the forecast changed and the hour before it was due to water enough precip was predicted that it didn’t need to water. Then the forecast changed and it thought it was going to water (but now the window of opportunity has passed). This is a display oddity that happens sometimes for “today”. Currently the whole of “today” is handled as a forecast in moisture levels even though in actualality it is partially history and partially future. It should correct for tomorrow’s watering.
If you give it 24 hours, my guess is that this will update to show the corrected/actual history for today’s date (24th), where there was no watering. It’s a little non-intuitive, but the last entry to the right of the graph is a forecasted entry. Even though it’s today’s date, it’s not reporting what happened this morning. Only what the Flex schedule formula thinks will happen this morning. Actual observations can cause this to deviate from the forecast. I’ve also seen this when playing with the empty/fill settings.
I have mentioned in the past that this particular way of reporting today’s activity on the graph is confusing. My suggestion was to delineate actual events from forecasted events by using solid lines for actuals and dashed lines for forecasts.
edit: @Linn beat me to it…
Ohhh I see - I thought it did water before I woke up and I just didn’t get alerted. So you are saying it planned to water, but then it rained enough to not water so the 24th will look different tomorrow (prob not as high) as it looks now? Also, based on my weather station’s data here, it looks like only .02 inches of rain have fell in the past 24 hours. This would not be enough to drop below the allowed depletion line for tomorrow. However, it doesn’t look like the flex schedule adjusted to schedule a watering for tomorrow to compensate? I am confused. lol
Here are more details:
Right, this is an example where the MAD watering algorithm is doing the correct thing, but the presentation of what the algorithm has concluded is lagging a bit. Tomorrow, the 24th will move out of forecasted status and you’ll see what actually happened.
For my own knowledge, does the water droplet mean water fell not sprinklers watered. So a droplet on a past date means either it rained or grass was watered by sprinkers, or a droplet today means either sprinkers are expected to run or water is expected to fall?
A droplet only indicates days that watered, or days that will water (if the day is a forecasted day).
Precipitation is not explicitly reflected in the graph - meaning, there is no icon in the graph that represents that it rained or that rain is forecasted.
I do believe, now that FLEX is back, that this is an opportunity for Rachio to make the moisture graphs better to clear up confusion.
So the fact that there is a water droplet on my graph today means the system intended to water but my lack of notification or watering history means it actulay didn’t? And it didn’t why?
My guess is the 0.05 of anticipated precipitation forecasted for today was enough to keep your zone’s moisture balance above your allowed depletion threshold. As to why the moisture graph still predicts a watering today when one actually didn’t happen, you’d have to ask the Rachio folks.
@plainsane and I went back and forth on this in an epic thread that went nowhere, so you’re not alone.
Same thing happened to me a couple of weeks ago and I have a good theory for it. IRO adjusts the schedule in real time as the weather forecast changes. The watering was scheduled based on yesterday’s forecast. Then before the schedule was suppose to run this morning, the precipitation forecast changed and it was more than 0.05 inch (say 0.1 inch). That was enough to keep the moisture balance above 0 and the schedule was skipped. Then the forecast changed again (precipitation down to 0.05 inch) and watering is needed. However, the scheduled run time has passed so it won’t run. Had you set the start time to 5pm, it would run. In short, when IRO adjusts the same day schedule, it doesn’t take into account whether the start time has passed. Don’t use the manual fill/empty. IRO will correct itself and run the schedule tomorrow morning when it sees that no watering happened today.
@goblue you described this very well. @franz, I am in agreement with this, wanted to write it up, but won’t as this is a very good description. If something could be changed in how “today” shows in the moisture levels, it will save a whole lot of confusion. I realize that it really is just perception as we the users see it, since it corrects itself when it gets ready to water, but it sure would be nice if “today” was able to show the mix of what already has happened, and what could still happen. In other words, once the opportunity to set the watering off has passed, the scheduled part gets set to zero, but the prediction of the precipitation amount still shows. Yes, just cosmetic, but still very helpful.
The proble. Is that flex knows more about what to do than almost anybody and to surface thy data to the normal user will only create more confusion.
Having beta tested flex I have gone through this then and I still do now.
It’s on opaque box where data goes into one side and fucking magic comes out the other in the form of a sprinkler opera.
The graph prediction does not update realtime and @franz will have to correct me but that graph is not updated when the schedule is determine to skip.
I’m pretty sure the algorithm and graph are fairly disjointed
Right, but the hard and most important part (the algorithm) is done, and accurate. Just need to bring the graph in line which can probably be done with a few more conditionals and some color/pattern changes to indicate past/forecast:
Correct, this is something we can make more in-sync. The most confusing thing for folks is always today’s moisture levels since it is a simulator, and isn’t necessarily using the data we use 1 hour before the flex schedule is supposed to run, hopefully using the most up-to-date, real time weather data.
We are literally re-building large parts of our cloud infrastructure systems from the ground-up to accommodate things like doing predictions, skips, etc. up to 24 hours in advance for flex, fixed, etc. schedules. Over time these systems will continue to evolve, become more intuitive, and eventually be running large government facilities. Well that last part probably is not true, but you get what I mean. Our development team size is definitely moving in the right direction
I hope you guys are building a cep engine…it’s a perfect fit for this use case.
Looks like you guys were right. This morning, yesterday correctly shows no watering, today does, and it looks ok. New Question, why didn’t flex adjust the watering time to bring the level closer to 100%?
Short answer: Read up on Managed Allowed Depletion (MAD)
Long answer: Watering times are based on your zone characteristics (Nozzle PR, Sun, Slope, Crop Type, etc). The goal of flex isn’t to saturate the zone to field capacity. That’s just an upper limit (technically 120% is the upper upper limit). The goal of flex is to make sure that soil moisture never drops below the allowed depletion level. This concept is called Managed Allowed Depletion (MAD). Flex has determined that your zone needs exactly 0.36 in to saturate your soil type at your crops’ optimal root depth. If you kept your zone at 100% capacity, you would be overwatering, likely experiencing run-off, and not giving your crops/grass and chance to root deeper.
Super helpful. Thanks! This is really the best product community I’ve seen.