An improved calibration method for the drift of the conductivity sensor on autonomous CTD profiling floats by θ–S climatology

An improved method to estimate the time-varying drift of measured conductivity from autonomous CTD profiling floats has been developed. This procedure extends previous methods developed by Wong, Johnson and Owens [2003. Delayed-mode calibration of autonomous CTD profiling float salinity data by θ–S climatology. Journal of Atmospheric and Oceanic Technology, 20, 308–318] and Böhme and Send [2005. Objective analyses of hydrographic data for referencing profiling float salinities in highly variable environments. Deep-Sea Research Part II, 52, 651–664]. It uses climatological salinity interpolated to the float positions and observed θ surfaces and chooses 10 ‘best’ levels that are within well-mixed mode waters or deep homogeneous water masses. A piece-wise linear fit is used to estimate the temporally varying multiplicative adjustment to the float potential conductivities. An objective, statistical method is used to choose the breakpoints in the float time series where there are multiple drift trends. In the previous methods these breakpoints were chosen subjectively by manually splitting the time series into separate segments over which the fits were made. Our statistical procedure reduces the subjectivity by providing an automated way for doing the piece-wise linear fit. Uncertainties in this predicted adjustment are estimated using a Monte-Carlo simulation. Examples of this new procedure as applied to two Argo floats are presented.