I totally get how this could be a problem, especially in FL, and have had the same thing happen with a schedule being deferred when rain didn’t materialize.
On the other hand, a few days ago we had something like a 20% chance of rain. Two scheduled zones watered anyway under Flex, which appeared to be the correct decision at the time. An hour later that 20% chance of rain turned into over 1/2 inch, so it would have been better to skip the schedule.
This is what I struggle with. Any decision will help some while hindering others.
There are occasions where being conservative will have helped, and other times when it will not.
Will let you know if we come up with something in the middle.
Ok what about this.
If Probability of Precipitation (POP) is >= 60% we apply a weighting of 1 to the amount of precip forecasted (no different than what we do now).
If .25in is in forecast, we assume .25in for the day.
If Probability of Precipitation (POP) is < 60% we apply a weighting of 50% to the amount of precip forecasted.
If .25in is in forecast, we assume .125in for the day.
Thoughts?
@franz I think it may be hard to know intuitively whether this would be a positive or negative change overall. Ideally, it would be good to model some actual recent prediction data vs. actual recorded precip from that period. Based on that, you could run some experiments and start to see whether weightings like this would have had the desired outcome, i.e. skipping watering when it is likely to rain and not skipping when it isn’t.
I’m unclear on how the probability of precip and the inches of precip in the forecast are tied together. I’d assumed you were just using forecast inches to determine flex schedules now.
Unfortunately POP is somewhat nebulous since it is comprised of two different variables, area and confidence.
http://www.srh.noaa.gov/ffc/?n=pop
So, even with modeling internally, it is still almost flipping a coin since we don’t know area or confidence, just POP.
IMHO applying a variable weighting will get us a little closer to a common sense approach.
Thankfully we auto-correct each day by using observations, but I think we might be deferring too many waterings based on afternoon storms that never appear or provide much less precip than forecasted.
Interesting. Well, I’d come down on the side of watering unnecessarily rather than skipping a needed watering event for forecast precip that may not materialize.
Fwiw, the flex has generally been too conservative with watering my perennial beds and two of my grass zones. So I have some brown areas. Some of that was likely due to forecast precip that didn’t actually occur. Since I have mostly cool season grass in a hot area, letting it get dryer than intended even for a day can be bad.
My thought is this would help with those cases where we are too conservative. We’ll listen for feedback and continue to adjust.
It would be great to see current balance of each zone on the homepage zone info. Maybe the Heading “MY YARD” should Migrate in some way to ZONE STATUS. Think you have an opportunity to include some good info all on one screen. Keeps me from needing to click my way through your customer interface to see what is going on.
I believe some of this discussion could be linked to the topic “Flex schedule after unforecast rain” they are somewhat related.