I’ve noticed something that seems strange to me when I use EP (4.1) with fast ambient temperature data from a source other than the sonic. It seems that EP is treating these temperatures as sonic temperatures and applying the water vapor correction. The attached eddypro file produced the results in the attached Excel file. In there, the temperature listed in column BG (sonic_temperature) is the average fine-wire thermocouple temperature for each period and the temperature listed in column FT (ts_mean) is the average sonic temperature for each period, fine so far except for the mislabeling of BG. The strange part is that the temperature in column BH (air_temperature) is different from those in BG, seemingly ‘corrected’ (I have not included other temperature/biomet data in this case). Also, it seems that the BH temperatures are being used in calculations, rather than the BG ones–for example, the saturation vapor pressure is right for the temperatures in BH, but too low for those in BG (or FT).
you are actually right. In 4.1 (and most likely also in previous versions) the average “fast temperature” gets erroneously corrected for humidity. Note, however, that this error is not propagated to H fluxes and w’T’ covariances, and these are ok in the result files.
So in 4.1, when using a “fast temperature reading” , in the column labeled “sonic_temperature” you do find the actual average air temperature, while the value in column “air_temperature” is just an meaningless one. This is a bit awkward, but still all good results are available in the output file.
The fix will be available from the next release, but you can confidently use the flux results you got with 4.1.