Daily Average Temperature Query

[quinny]

Hi. Just a general question, but what is the agreed measure for determining an average daily temperature? Is it the average from midnight to midnight over 24 hours; a given 12 hour window (say 7am to 7pm); or is it between sunrise and sunset? I've pinged the Met Office that question, but they've not gotten back to me.

I only ask as I've recently got a new Ecowitt station, and looking at the software it calculates the average daily temp over a 2 hour period. But when meteorological services talk about average or mean daily temperatures, do they calculate it that way?

Hope my question makes sense. Thanks, Liam

[HansR]

As far as I understand - no doubt others will jump in - the following is true:

1. The MetOffice calculates the average daily temperature as (Max+Min)/2; You can opt to have CMX calculate it this way in the NOAA report
2. CMX calculates the average over 24 hrs standard as SUM(all samples)/NumberOfSamples; This is standard use where daily average is used.
3. The difference between the two can be seen in a chart by CUtils (Graphs=>Misc Graphs=>Select: Clash of Averages)

Tbh: I have no idea about your 2 hr calculation of daily average on an Ecowitt, the CMX samples the value as it is at a certain moment. Could you explain where you get this from?

[quinny]

As far as I understand - no doubt others will jump in - the following is true:

1. The MetOffice calculates the average daily temperature as (Max+Min)/2; You can opt to have CMX calculate it this way in the NOAA report
2. CMX calculates the average over 24 hrs standard as SUM(all samples)/NumberOfSamples; This is standard use where daily average is used.
3. The difference between the two can be seen in a chart by CUtils (Graphs=>Misc Graphs=>Select: Clash of Averages)

Tbh: I have no idea about your 2 hr calculation of daily average on an Ecowitt, the CMX samples the value as it is at a certain moment. Could you explain where you get this from?

Thanks for the reply: that's very useful to know. As for the 2hr calculation bit: that's a fat finger issue on my part - I meant to type 24 hour ... but missed out the '4'...

[mcrossley]

If you mean agreed world-wide, I don't think there is such a thing.

On their daily high/lows page the UK met office say...

Max temp, lowest max temp are 0900 to 2100 on the date shown
Min temp, most rainfall, sunniest are 2100 previous day to 0900 on the date shown

Which doesn't help with average, but as the met office uses a 0900 - 0900 met day, I imagine the average will be over that period.

(the rainfall statement seems odd, as it would mean they only counted 12 hours per day - I think that must be a typo)

[freddie]

Yes, rainfall is 0900-0900 and attributed to the date on which that period started. With 0900-0900 temperatures, the max is attributed to the date on which the period starts - same as the rainfall. With min temperature, it is attributed to the date on which the period ends.

As a general rule, the rain and the max temp are attributed to the date on which the period starts, and the min temp to the day on which the period ends. This applies to 0900-0900, 0900-2100, 2100-0900 and midnight to midnight (where midnight is defined as the start of the day (23:59). There is no midnight at the end of the day (i.e. 24:00) as the clock goes from 23:59:59 (on day 1) to 00:00:00 (on day 2).