Post by Teresa Vogl on Jan 14, 2020 8:37:59 GMT
Dear CloudnetPy team,
my colleagues and I recently discussed how cloud radar Doppler spectra could be implemented into CloudnetPy. We think that, by analogy with the moments, the first step would be to create "calibrated spectra files", which then can be further processed.
Here are some points we came up with, which should be discussed:
- how large can these calibrated spectra files be? If we go for the "classical" 24-hr files which are generated for the moments, file size could become too large. The threshold of maximum file size should be discussed, then we can test if it makes sense to use e.g. 12-hr or 6-hr files instead of the 24-hrs.
- which corrections are needed to obtain "calibrated" spectra? We collected the following items, but probably there are more points to consider: Removal of clutter, ghost echos and "specles" (single pixels of signal surrounded by pixels which do not contain signal), dealiasing if needed.
- Depending on the above point, it would be useful to create a 2D mask, which contains bits specifying which corrections were applied to which spectrum in time-height space. A similar concept is e.g. used in Dias Neto et al. (2019). This can be extended to include a check for turbulence, Gaussian shape of spectra, etc.
- Correction of reflectivity: Since this is a complicated topic, we suggest to not correct the reflectivity for attenuation. Absolute radar calibration however is another difficult topic open for discussion.
We would be happy to hear more thoughts on this!
Kind regards from Leipzig,
Teresa Vogl with Martin Radenz and Willi Schimmel