GCWerks CRDS calculates the std stdev from all std C(nonlin corrected) values within approximately +/- 10 days of a given point
However, it appears to do this calculation once per day, starting at the time of the first point data gap (even if flagged/rejected/filtered), then recalculating every 24 hrs and 1 minute after that
So for example on the RGL Picarro there is a gap before the point at 120307.1208, so that the std_stdev gets updated at 120308.1209, 120309.1210, 120310.1211 etc.
It needs at least 3 std C(nonlin corrected) values in the calculation, otherwise it returns nan
What do we want to do with the nan points? These probably have poorer precision than usual, because there are few std cylinder measurements around these times. Maybe we could set mf_repeatability for these points to equal that max mf_repeatability within that calendar year, or something similar?
What do we want to do with the nan points? These probably have poorer precision than usual, because there are few std cylinder measurements around these times. Maybe we could set mf_repeatability for these points to equal that max mf_repeatability within that calendar year, or something similar?