Problem: Loop Removal module of SBE data processing may not entirely remove short
periods of delay during CTD downcast, resulting in averaged values that produce slightly
lower density than the previous bin. The result is a non-monotonic increase in depth during
the downcast, despite a monotonic increase in measurement time. Since the data is sorted
by time (default), there is an apparent density inversion in the density profile. This inversion
disappears if the data is sorted by depth, rather than time. Example: Nov. 9, 2022, station
QU39.
Action (Jan. 24, 2023): Jessy will investigate if there are non-monotonic increases in depth in
the existing research ready dataset, that may have created false density inversions (about
500 casts of the ~ 4000+ casts already in the research ready dataset). Also: Manual QC of
the provisional dataset needs to be sorted by depth, rather than time, from now on.
False density inversion (Jan. 24, 2023)
Problem: Loop Removal module of SBE data processing may not entirely remove short periods of delay during CTD downcast, resulting in averaged values that produce slightly lower density than the previous bin. The result is a non-monotonic increase in depth during the downcast, despite a monotonic increase in measurement time. Since the data is sorted by time (default), there is an apparent density inversion in the density profile. This inversion disappears if the data is sorted by depth, rather than time. Example: Nov. 9, 2022, station QU39.
Action (Jan. 24, 2023): Jessy will investigate if there are non-monotonic increases in depth in
the existing research ready dataset, that may have created false density inversions (about 500 casts of the ~ 4000+ casts already in the research ready dataset). Also: Manual QC of the provisional dataset needs to be sorted by depth, rather than time, from now on.