I recently scanned through lot's of CloudNet quicklooks from several sites and noted here and there some more or less severe issues in the data and/or the processed products. Maybe it could be a good idea that users just note down in this forum what looks weird to them and then somebody can check and maybe even re-process the data. I think this could greatly help to increase the overall data quality.
Of course, alternative suggestions how to deal with this issue are very welcome!
Lindenberg: 2016-11-18, Very high/weird Doppler velocities
Palaiseau: 2012, Nov/Dec, Very weird looking classification, probably caused by lidar data Palaiseau: 2011, Oct-Dec, Lidar artefacts, also radar data have kind of ground clutter and other issues Palaiseau: 2015, June 19-21, interrupted radar data with very weird values
Mace Head: 2009, April 25-26, arteficial line at 4-5 km in radar
Munich: 2019-2020, radar disturbances during daytime (see for example 2019-01-16, 2020-01-14), partly filtered in classification
ARM SGP: 2006, very weird radar 'ghost' echos, for example: 2006-03-01, throughout entire year but would need to check exact time period ARM SGP: 2006, also issues with the lidar data, for example see classification on 2006-10-06 (similar issues also 2007)
Post by simotukiainen on Jan 24, 2020 12:56:51 GMT
Thanks for spotting these anomalies, I will notice Ewan about this. He has processed these data..
Some of the legacy data indeed contain dubious / incorrect values. I guess there have not been huge amount of quality control in the legacy data. This will change soon(ish) when we change from Matlab to CloundetPy which contains better routines to monitor quality. The idea is also have a separate NRT data flow and a "fully" quality controlled data set released later.
Anyway, I think that reporting data issues on this forum is a good idea in general!