Report of percentile QC malfunction

[BaptisteVandecrux]

We had a hint of this in GEUS-Glaciology-and-Climate/PROMICE-AWS-data-issues#140 but didn't take action.

First of all, let's look at a site where it did work as intended:

With the limitation that, within the malfunction period, there are still a lot of datapoints falling withing reasonable range which still need to be removed (manually?):

Now sites where good data were removed (pink dots):

FRE (no plot available)

Few good points of p, wspd, t removed at QAS_U, QAS_Uv3, NUK_K, NUK_L, KAN_B, KAN_L, KAN_M, KAN_U, TAS_L, TAS_U...