Page 1 of 1
9am - 9am
Posted: Fri 15 Jul 2022 5:09 pm
by Mapantz
I have always used midnight to midnight for my weather data, but I never gave it any thought to use 9am to 9am. I didn't want to change it, for fear of putting my data out of kilter. That said, it makes sense when it comes to lowest maximums, and highest minimums, as I have missed out on many records. It got me thinking, rather than manually changing record data etc, is it possible for webtags to be implemented for certain data, just to show what the 9am - 9am figures are for lowest max temperature and highest min temperature?
Re: 9am - 9am
Posted: Sun 02 Oct 2022 8:47 pm
by OBSERWACJEMETEO
I would go one step further. I would give the possibility to manually select the end time of the measurement day, for example, at 6,7 or 8 a.m. Then one could be tempted to display maximum and minimum temperature values from 6 UTC to 18 UTC or precipitation day from 6 UTC to 6 UTC the following day. It would be extra if such functionality would be included in the new release of CumulusMX. I get up in the morning and know that the minimum temperature for the night was, for example, -9 degrees for the period from 18 UTC to 6 UTC the following day. I would also add the ability to choose the method of calculating the daily average air temperature, as it is in the NOAA reports in Cumulus MX. Following suit, I would add a formula for calculating pressure reduced to sea level. So that the non-reduced pressure would be recorded in the Cumulus MX report separately and the reduced to sea level separately. That is, CumuluxMx takes the pressure value from the weather station, reduces it to sea level and displays it in the main page. What do you guys think about this ?
Re: 9am - 9am
Posted: Sun 02 Oct 2022 9:06 pm
by freddie
That would be an awful lot of changes that would take quite a while to put in place, and would make Cumulus very complex.
What you could do is to use existing Cumulus functionality to populate a SQL database with the weather data, and create custom queries that return the data for the periods you are interested in.
Re: 9am - 9am
Posted: Tue 25 Jul 2023 5:38 am
by OBSERWACJEMETEO
I don't know enough about programming to do it. Here comes the learning of databases and php scripts I guess.
It would also be good if Cumulus had some hourly ranges to choose from, like 9-9, 8-8,7-7 or something like that. The program would get a new, in my opinion useful functionality. Keep in mind that the program is used by users in different countries. Each of them, I verify or compare their measurements to the official measurements of state weather stations. As you know, the measurement days of a meteorological day start at different times in different countries. Changing the time for a meteorological day, will allow users to adjust to the official measurements of their country. This would be a very odious change in cumulus.
Re: 9am - 9am
Posted: Tue 25 Jul 2023 7:06 am
by HansR
Ignoring the complexity for CMX (or any technique the user wishes to use), the idea of using free start and end times for the meteorological day is bad and destroys any form of comparability of averages between stations.
Meteo science - even for amateurs - does not profit of individual usage of statistical boundaries for the aggregate data. A lot of statistical work has gone into making old and new series - e.g. day averages of temperatures - comparable and create continuous time series. Why ignore that work and destroy the idea of continuity?
OBSERWACJEMETEO wrote: ↑Tue 25 Jul 2023 5:38 am
It would also be good if Cumulus had some hourly ranges to choose from, like 9-9, 8-8,7-7 or something like that. The program would get a new, in my opinion useful functionality. Keep in mind that the program is used by users in different countries. Each of them, I verify or compare their measurements to the official measurements of state weather stations. As you know, the measurement days of a meteorological day start at different times in different countries. Changing the time for a meteorological day, will allow users to adjust to the official measurements of their country. This would be a very odious change in cumulus.
Which countries use meteorological days other than midnight or 9 am? I tried to find the standardisation of the start of meteorological day at the WMO but could not find it.
Re: 9am - 9am
Posted: Tue 25 Jul 2023 7:08 am
by rogerthn
freddie wrote: ↑Sun 02 Oct 2022 9:06 pm
That would be an awful lot of changes that would take quite a while to put in place, and would make Cumulus very complex.
What you could do is to use existing Cumulus functionality to populate a SQL database with the weather data, and create custom queries that return the data for the periods you are interested in.
I do agree!
No promises but I might try to create a DB view from my "midnight" monthly table
Re: 9am - 9am
Posted: Wed 26 Jul 2023 5:03 pm
by rogerthn
If someone would like to "play".
Code: Select all
create or replace view LogRollover as select convert_tz(LogDateTime,'Europe/Stockholm','UTC') AS UTC,LogDateTime,if(120 / substring_index(timediff(LogDateTime,convert_tz(LogDateTime,'Europe/Stockholm','UTC')),':',1) + hour(LogDateTime) * 60 + minute(LogDateTime) between 0 and 659,substr(LogDateTime - interval 1 day,1,10),substr(LogDateTime,1,10)) AS LogRolloverDate,Temp,Humidity,Dewpoint,Windspeed,Windgust,Windbearing,RainRate,TodayRainSoFar,Pressure,Raincounter,InsideTemp,InsideHumidity,LatestWindGust,WindChill,HeatIndex,UVindex,SolarRad,Evapotrans,AnnualEvapTran,ApparentTemp,MaxSolarRad,HrsSunShine,CurrWindBearing,RG11rain,RainSinceMidnight,WindbearingSym,CurrWindBearingSym,FeelsLike,Humidex,hour(LogDateTime) * 60 + minute(LogDateTime) AS HMcnt,timediff(LogDateTime,convert_tz(LogDateTime,'Europe/Stockholm','UTC')) AS td,substring_index(timediff(LogDateTime,convert_tz(LogDateTime,'Europe/Stockholm','UTC')),':',1) AS TDH,120 / substring_index(timediff(LogDateTime,convert_tz(LogDateTime,'Europe/Stockholm','UTC')),':',1) + hour(LogDateTime) * 60 + minute(LogDateTime) AS TD2 from Monthly;
Re: 9am - 9am
Posted: Wed 26 Jul 2023 5:32 pm
by HansR
Re: 9am - 9am
Posted: Wed 26 Jul 2023 5:53 pm
by rogerthn
The S in SQL stands for structured not strange
