CTD observations along the European shelf edge comparison of data processing algorithms

Abstract

As a part of their activities in the EU funded project SEFOS (Shelf Edge Fisheries and Oceanography Study). three laboratories decided to compare results of their CTD data processing routines, leaving aside calibralion procedures and lnstrument performances. The relevant modules of each laboratory data processing routines were applied to unprocessed (apart from sensor specific configuration) data. and the resulting series returned to the originator for analysis. Vertical structures larger than 10 dbar were revealed by all data processing schemes. One of the routines was applied to a self-recording instrument that slored data at a low sampling rate. Upon attempting to eliminate spiking by simple filtering during post processing. without correcting for sensor misalignment. this routine caused smearing of the vertical slructure and proved insufficient to eliminate density inversions. Routines designed to handle data acquired at high sampling rates gave rise to quite comparable compacted files. Typical standard deviations of the temperature and salinity dillerences, at constant pressure, between the series produced by two of the laboratories were around 0.004°C and 0.003, respectively. Standard deviations of up to nearly 0.03°C and 0.01 were found in stations possessing a great deal of structure, the largest individual differences occurring in regions of steep gradients. Different handling of the pressure series (pressure offset. pressure smoothing) in the dilferent routines was held responsible for the large temperature differences that were observed. Despite the differences encountered between the results of the three processing routines, the exercise proved encouraging and promoted significant progress in CTD data handling by the laboratories involved