Precipitation addition time zone dependent?


#1

I am in Brisbane Australia, UTC+10, and I seem to be having issues with schedules vs recorded precipitation.

I am wondering whether there is a time zone issue in that my recorded precipitation seems to lag by one day, and, after heavy rainfall on the 4th of April, 15mm, my lawns still irrigated for 30 mins per zone even though the ground was already waterlogged.

Have I missed a time zone setting?

Reading the flex schedule info I learned that schedules are recalculated 12 hours and 1 hour before planned start time so I thought that the accumulated precipitation, being greater than the 3.2mm threshold, would have skipped today’s watering. Today we had another 4.5mm.

ps. I am using Weather Intelligence and my own weather station for actual MET data.


#2

@Ghawgood

I will have the engineering team look into this today. Thanks for bringing this to our attention!

:cheers:


#3

@Ghawgood

We looked at our data and it seems to be correctly identifying, taking time zone into account. I noticed that you did water on 4/5 and based on the chart precipitation it looks like you should not have. Doing more research we did make the correct decision based on the forecasted precipitation, which you can not see in the graph below (it only shows observed precipitation).

Ironically, I did seem to find an issue with the PWS “reported” precipitation and what our weather service provider is sending us. We received .07 inches for 4/4 and station is reporting .64 inches for 4/4. I opened a support ticket with them and they are working on a solution.

:cheers:

Franz,

Based on the API query, it used the EOD24h (end of day 24 hour value), but looking at the data for this station:
https://www.pwsweather.com/obs/archive/2018/04/04/IQNSWEST4.html

It starts the day at 0.01", then goes to zero, then works its way to 0.63" before falling back to 0.07". It seems like the 24-hour precip value is reporting a bit differently than most stations.

I have contacted our data team to look at this stations data a bit more to determine if its something specific to the station / software combination or some other potential issue.

I will provide an update as the team researches.


#4

Thank you guys, looks like you are on to it.

We are in for a very welcome string of fine days now so expect to see
the watering schedule increase.


#5

@Ghawgood

Relaying the response from our weather service provider.

Our team has investigated this station, found the cause and pushed out a resolution.

Specifically, This station is not sending a true max 24-hour precip value nor what seems to be a running 24-hour precip value. Our daily summary platform has support to handle this type of station, but a couple of unique events happened which cause the daily total not to be calculated properly.

The station sent a value from the previous day 0.01 for the first 3-4 minutes of the new day. This can be due to time differences on the station or other hardware between. Normally a station will send a maximum of 1 observation with previous the previous day’s data, which was being handled properly. Unfortunately, with this station on April 4th, the three first observations had precip from the previous day, then the Maximum 24-hour precip occurred in the middle of the day and reset to 0.07 which occurred at the end of the day. This caused an issue with our ruleset, and the system fell back to using the final 24hour precip value of the day, which was 0.07.

This can be seen here: https://www.pwsweather.com/archivewx.php?id=IQNSWEST4&d=20180404&t=day

On the following day, the station only sent the .07 on the first observation, though we noticed the station switched from updates per minute to every few minutes.

All said the team has updated the ruleset to offer improved handling of this type of scenario and the station is now reporting properly for the daily total.

We have already updated the complete history for this station and running an update to catch any other such scenarios in our archived data.

I hope this helps explain things and the user should find improved results going forward.