20.48 inches of rainfall yesterday!

I wanted to post this yesterday, but it took me much longer to swim home than I anticipated.

I noticed rachio didn’t go off this morning as I expected it to so I checked out the moisture levels. Considering it didn’t rain a drop I was surprised to see 20.48 inches of rainfall!

I checked the weather station my rachio is associated with and this amount was indeed reported by the station.

https://www.pwsweather.com/archivewx.php?id=KTXBUDA16&d=20170615&t=

This is the first time this has happened, but I was hoping to be able help mitigate the effects of glitches like this.

A couple of ideas I had were to either throw out statistical outliers straight up, use nearby weathers stations to validate reading and throw out inconsistent readings and potentially alert of anomalies to allow the user to ignore a reading.

Right now I believe my only option is to reset water levels manually for all 12 zones.

1 Like

and it all happened in 5 minutes, maybe someone dumped a bucket on their pws? :grinning: But yes, that would be frustrating for sure

Was this a National Weather Station with the error?. I might use a PWS, but the glitch you experienced is worrisome.

@Slackeratx You make a good point with your suggestion. I will reach out to the API hosting site to see if that is something they’ve considered. Either way I think its something we could develop sometime down the line. What would you say the trigger point to consider a certain rainfall to be erroneous? Would you want to have control over this trigger, or do you think the Rachio should just do it for you by default?

I use a PWS, not mine but one relatively close by.

If I were implementing from an API it I would take the average of the national weather station and two other personal stations other than the one being requested. I would change the API to include the actual reading of the station being used, this average of 3 other stations and/or a percent difference from the reading being used.

Rachio could then use a threshold to flag readings as likely erroneous at say 25%. i.e if the current reading is 25% different (higher or lower) than mark it questionable. From rachio’s perspective I would want to control what the threshold is in cause you want to change it to 50% or 100% or whatever based on what you’re seeing in the data, but not have the end user play this this threshold, that might get ugly.

From the end users perspective I would want an option in my rachios settings for the app to;

1)silently ignore erroneous data (no notification, but the rainfall measurement is not used in calculations)
2)notify and ignore erroneous data. (notify you station reported 23 inches of rainfall yesterday we’re ignoring it)
3)notify and ask if measurement should be ignored.

I would probably use 2 or 3.

How about the nearest national weather service or airport. I use Wiley Post Airport. Just 2 miles from my house. Always accurate.

Yes an NWS should fix the core of the issue here (someone dumping water on their PWS or something), however I do like @Slackeratx idea on some sort of QA check for erroneous data. The thing most people don’t realize is that Rachio doesn’t have any say in the data it is presented from the weather API host, so having something like this would be helpful I would say.

@robertokc, you are lucky to have a NWS so close to you – my PWS is further away than your NWS!!! – my closest NWS is over 10 miles away, so the rain there is rarely what I get. So I’m not sure about going back to the NWS in case the PWS puts out bad data. I guess maybe it would be better than nothing???

1 Like

http://weatherflow.com/smart-home-weather-stations/ im waiting to see if these are worth the money. if they are, cha ching

1 Like

1 Like

If I had to choose in your situation, i would do the same. Periodically compare data from various weather stations, too.

I am sure this product is less expensive than a Campbell Scientific weather station. They are a major investment.

Yes, I periodically check all the PWS’s near me. But I think my birthday present for myself is going to be my own PWS this year!

2 Likes

I had a similar issue when I.set up my PWS. Rachio was no seeing my data when it rained due to a processing error by their weather provider. It took us several months to work out the bug but we did finally. During that time it would have been nice to be doing PWS data quality checks because my own data was invalid yet being used.

I know CWOP, Wundergrdoundand others do data quality checks and reject invalid data. Maybe Rachio could build off their historical quality checking algorithms rather than having to create new methods.

1 Like