Like all Conductivity-Temperature-Depth (CTD) instruments, the CTDs on Argo floats all have sensor response errors that must be corrected as best as possible. The Sea-Bird Electronics model SBE-41 and SBE-41CP CTDs that are widely used on profiling floats are no exception. Work on sensor response corrections by PMEL and other scientists is illustrated in the figures below, using data from a segment of a profile from a SBE-41CP CTD mounted on an Ice-Tethered Profiler (ITP) ascending through a portion of the water column in the Arctic Ocean containing well-defined thermohaline staircases. Thermohaline staircases are maintained by a diffusive instability, and are characterized by regions of vertically homogenous temperature and salinity separated by sharp gradients. CTD data collected in thermohaline staircases are often excellent for quantifying sensor response errors.

Figure 1. Raw (black) and corrected (cyan) temperature plotted versus pressure from ascending cycle 175 of ITP2. Each + denotes a measurement from the 1-Hz time-series. The correction for the thermal inertia of the thermistor results in sharper gradients between thicker homogenous regions of the thermohaline staircase.

Figure 2. Salinity (black) derived from uncorrected, or raw temperature and conductivity measurements has spikes while ascending through strong temperature (green) gradients and is not homogenous in regions of homogenous temperature. Use of corrected temperature to derive salinity (magenta) eliminates much of the spiking. Further applying a small time-shift between conductivity and temperature before estimating salinity (red) has little effect. Finally, further correcting for conductivity cell thermal mass error before estimating salinity (blue line) results in salinity and temperature profiles with well-matched high gradient and homogenous regions, as expected for a thermohaline staircase.

Correcting these sensor response errors in SBE-41 and SBE-41CP CTD data, especially the conductivity cell thermal mass error, is important to producing the highest-quality data possible. These corrections can eliminate erroneous fresh salinity biases reported by these CTDs within strong thermoclines, and reduce erroneous fresh spikes reported at the bottoms of surface mixed layers above strong thermoclines.