Scaling of binned and full (co)spectra outputs

Environmental Forum Forums EddyPro Scaling of binned and full (co)spectra outputs

This topic contains 2 replies, has 3 voices, and was last updated by  gerardo 2 years, 3 months ago.

Viewing 3 posts - 1 through 3 (of 3 total)
  • Author
    Posts
  • #7851

    swchan
    Participant

    Hi EddyPro developers,

    I suspect that the column headers for the full (co)spectra output files are incorrectly labeled. In EP versions 5.2.1 and 6.1, they are labeled as f_nat*spec() or f_nat*cospec(). The labels do not indicate that the full (co)spectra are normalized by the corresponding variances or covariances.

    This is in sharp contrast to the binned (co)spectra output files which are normalized by the (co)variances. I have tried comparing the full and binned (co)spectral outputs and could only get them to align if I assumed that the full (co)spectra were normalized.

    The source code (v6.1) appears to support this. Lines 770 and 781 in file spectral_analysis.f90 appear to scale the (co)spectra by the corresponding (co)variances.

    Can you confirm and update the headers in future releases? Thanks again for all of the continued development efforts!

    #7854

    IsraelB
    Keymaster

    Hi swchan,
    Thanks for bringing this up. I did a quick check and it does look like it is normalized. We will need to update the headers in the future.

    Thanks

    #7902

    gerardo
    Participant

    Hi again swchan! 🙂
    It is incredibly useful to have EddyPro users contribute with such detailed investigations. We would not be able to spot all such issues on our own. So, thanks a lot again!
    I checked and can confirm that we need to change the labels, as full length co-spectra are indeed normalized with both the frequency and the co-variance. We will make the change available in the next release.
    Ciao,
    Gerardo

Viewing 3 posts - 1 through 3 (of 3 total)

You must be logged in to reply to this topic.

Comments are closed.