Issue #1: No (Co)Spectra for CO2 and H2O
In EddyPro 6.2.0 (also seen for 6.2.2) there is an issue where (co)spectra are not created for CO2 and H2O. It was possible to obtain the (co)spectra in a later version of EddyPro (7.0.0 rc8). Since there were some (smaller) updates done after 6.2.0, I assume that this issue has been addressed in versions > 6.2.2.
Without the (co)spectra, the spectral correction after Fratini et al. (2012) cannot be applied, and EddyPro switches to the correction by Moncrieff et al. (1997) instead.
This issue was observed for the data CH-OE2: 2017_2, 2017_3 and 2018, when only the IRGA72 was running. In 2016 and then in 2017_1, when the IRGA72 was installed alongside the IRGA75, this problem did not occur. I am currently not sure why this discrepancy between time periods was the case.
Since in EP 6.2.0 no (co)spectra were output, I tried different settings to check whether a specific setting caused this bug, but even minimal settings did not work. The same was observed for EP 6.2.2. I then created a new .eddypro and .metadata file from scratch in the EP 6.2.2 GUI and tested it with these files, but no success. There was no way to get the (co)spectra files.
I then tested the .eddypro and .metadata files that I created in EP 6.2.2 in a new version of EP (EddyPro 7.0.0 rc8). In this version, the CO2 and H2O (co)spectra were succesfully written to the files, but there was a new, unknown error that caused the execution to crash on Windows (on Mac it worked). This new error was probably due to an error caused in the (co)spectra sorting algorithm within EP. GF was able to pin down the error and at the same time speed up EP significantly (percentiles were calculated more often than needed, hence the speed increase). I received new executables (bin) from GF, pasted these files over the files in the EP installation folder (bin) and ran 6-hour test data (from CH-Oe2 2017_2) again from the GUI. Everything executed as expected. I assume the error is therefore resolved.
– LH, 4 Apr 2019