JC112 Originator CTD Data Processing
Sampling Strategy
A total of 87 CTD casts were deployed during OSCAR cruise JC112, of which one was a test cast and another to complete a previous cast. In addition, 20 of the 87 casts were tow-yo casts below 3000 m.
Data Processing
CTD data processing procedures were as follows:
- Processing was carried out using Sea-Bird Electronics data processing software (Version Seasave V 7.23.1).
- The raw .hex files were converted to .cnv and .ros files.
- The pressure data were filtered by a low-pass filter of 0.15 seconds.
- The conductivity data were filtered to remove conductivity cell thermal mass effects (alpha=0.03, 1/beta=7).
- Bad scans marked by pressure slowdowns (slower than the fixed minimal velocity of 0.25m/s) or pressure reversal, were given a flag value of -9.990e-29. All of the data were then averaged over 1 dbar bins.
- One of each temperature, salinity, and conductivity channels were selected.
- Oxygen concentration values were converted to umol/L.
- Sound velocity was derived using the UNESCO algorithm.
- Data was split into up and down cast. Only down cast datafiles were kept. Tow-yo data were split into files, one for each down cast.
- Ascii files created from the .cnv files.
CTD Conductivity Calibration
The two conductivity sensors were calibrated using water bottle samples that were taken at each cast at discrete intervals: around 100 m apart for the shallower levels, and 250 m to 500 m apart at the deeper levels. In total, approximately 1560 water samples were analysed in the lab using Autosalinograph for conductivity comparison with CTD sensor measurements. Comparing the two CTD measured conductivities and the calibrated bottle conductivity estimates, the outliers were identified as values that differed more than 0.01 mS/cm. The final CTD calibration used 1545 "good" data points, after removing outliers. The linear regression of autosal calibrated data versus CTD uncalibrated data revealed: C calibr = 0.9996 x C insitu1 + 0.01170 mS/cm for CTD sensor 1, and C calibr = 1.0003 x C insitu2 - 0.01027 mS/cm for CTD sensor 2. Above 2000 m depth, the water column is much less stable leading to large differences between autosal and in-situ conductivity measurements, more than 0.005 mS/cm. Thus, higher precision of CTD conductivity calibration is achieved sampling waters below about 2000 meters, i.e. below water masses of shallow overturning circulation. In addition, the difference between the two CTD conductivity sensors was tested in a profile with 20-minute stops at different depths, designed to calibrate MicroCATs. The difference was less than 0.001 mS/cm most of the time in stable waters, however, in the presence of turbulence, the difference could reach almost 0.005 mS/cm. In the tested profile, sensor 1 was more stable than sensor 2.
CTD Oxygen Calibration
The CTD oxygen sensors were calibrated using water bottle samples taken at 62 stations. The higher the oxygen concentration, the higher the mismatch is between in-situ and corresponding lab oxygen values. The range of oxygen concentrations up to 100 umol/L is better sampled and less noisy than the upper range of oxygen concentrations, therefore two separate linear regression fits to the data were performed, one below and one above 100 umol/L. The CTD oxygen sensor calibration results are: Ox calibr = 1.085 x Ox insitu + 0.161 umol/L for low oxygen concentrations (<100 umol/L), and Ox calibr = 1.000 x Ox insitu + 8.805 umol/L for high oxygen concentrations (>100 umol/L).
