Heat/Cool Degree Days calculation.

[aldofer]

Today while I was comparing mean temperatures between nov-dec 2013 and nov-dec 2012,I noticed an anomaly in the values of Heat/Cool Degree Days of NOAA reports.
Calculated values differ from expected values, here some examples:
Yesterday
mean temp     1.7°C
Heat Deg Days    29.9
expected value    16.6

02 August
mean temp     28.5°C
Cool Deg Days    18.4
expected value    10.2

heating and cooling base temperatures are 18.3°C (65 °F)
target_unit = METRIC

In short, the relationship between the expected values and the calculated values is always 0.55 which is the ratio Fahrenheit/Celsius.
Is that correct? and if so, how can I fix it ?

Hi Aldo

(My NOAA reports of 2012 and 2011 were generated by Cumulus and calculated values are the same as expected values)

[Thomas Keffer]

Hello, Aldo

I am away from my development machine right now so I can't investigate it thoroughly, but the problem could be that converting the difference between two temperatures is different than the difference between two converted temperatures.

Try changing the heatbase and coolbase options in skin.conf to:

    [[DegreeDays]]

        # This section sets the base temperatures used for the calculation

        # of heating and cooling degree-days.

                

        # Base temperature for heating days, with unit:

        heating_base = 18.3, degree_C

        # Base temperature for cooling days, with unit:

        cooling_base = 18.3, degree_C

-tk

--
You received this message because you are subscribed to the Google Groups "Weewx user's group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to weewx-user+...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

[aldofer]

Hi, Tom
OK it works !!
I checked before the dec-2013 report, then when I was sure, I deleted the other "wrong" reports and they have been regenerated correctly.
Do you think that this is the final solution, or just a temporary fix ?

Ciao Aldo
and thank you very much.

[Thomas Keffer]

Glad it worked.

There is a similar problem with the "trend" tag ($trend), which can also involve taking the difference of two temperatures. 

A "final" solution involves making temperature differences a first-class unit type (e.g., "delta_C" and "delta_F"). It's on my list of things to fix.

-tk

--