Data were read from glider stored memory after recovery. The following table documents the specific steps applied to arrive at the Level 2 (L2) dataset. The details of the corrections and calibrations that were applied can be found in the Supplemental Files (Glider_L2_DissolvedOxygen_Corrections_20150817.pdf). Glider CTD dataset is described separately.

Regular sample interval	The CTD variables (temperature, salinity, and depth) are resampled at a 3 second sample interval.
Filter the CTD’s temperature and salinity	Filter the temperature and salinity (from the CTD), using cut-off frequency fc = 1/30 per second i.e. Tc = 30 seconds. Matlab filtfilt was used with a 4th order butterworth filter i.e. 4th order forward- and 4th order reverse-filtered.
Delay the CTD variables	Delay the CTD variables by 27 seconds.
Interpolate CTD data to the DO time base	The delayed CTD variables (filtered temperature, filtered salinity, and depth) are interpolated to the DO sensor’s time base, which has an irregular sample interval (median approximately 3.3 seconds).
Recalculate DO output variables	The TSP-corrected oxygen concentration (o2_tspcorr) is calculated using the modified CTD data: • Temperature filtered, delayed, and interpolated • Salinity filtered, delayed, and interpolated • Depth delayed and interpolated.
Apply Calibration Curve	The calibration curve (i.e. the slope and intercept described above in the “Calibration Fix” section) is applied. So, the calibrated TSP-corrected oxygen concentration (o2_sat) is: slope*o2_tspcorr + intercept.

BCO-DMO Processing:
-extracted data from MatLab .mat files
- added conventional header with dataset name, PI name, version date, reference information
- renamed parameters to BCO-DMO standard
- added yrday_utc and ISO_DateTime_UTC to served view
- reduced number of significant digits
- change NaN's to nd's

glider_DO.csv