search

SuperDARN / rst

Radar Software Toolkit (RST)
https://superdarn.github.io/rst/
GNU General Public License v3.0
22 stars 18 forks source link

TDIFF values in FitACF 2.5 #493

Closed ecbland closed 2 years ago

ecbland commented 2 years ago

@egthomas , @ecbland , I was testing utilisation of the two tdiff values for calibrating non-stereo two-channel data and came across an apparent problem. It seems that there is a mix-up with applying tdiff values to stereo data by FITACF2.5. If I change tdiff[0], elevation values change in both channels but if I change tdiff[1], there are no changes in either of them. This problem does not exist for the data processed with FITACF3 when changes in tdiff[0] cause changes in channel 1 only, and changes in tdiff[1] lead to changes in channel 2 only. Please see attached a stack of PPT slides illustrating this issue.

TDIFF misfit for FITACF2.pptx

I used develop branch for this analysis.

I suspect that this issue may be caused by the way tdiff is declared in FitPrm structure (a single value instead of a two-element array), but I can easily be mistaken.

Originally posted by @pasha-ponomarenko in https://github.com/SuperDARN/rst/issues/455#issuecomment-1078672766

egthomas commented 2 years ago

@pasha-ponomarenko @ecbland thanks for identifying this issue. I think I have identified the problem - FitACFMake is called once by make_fit to populate the FitBlock structure (with a channel-dependent tdiff), but that function never gets called again so the tdiff values aren't updated when the record / channel changes. I'm working on a fix now and hope to have it ready for testing soon.