Delayed-mode quality control of dissolved oxygen concentration measured by ARGO floats in the North- Atlantic with LOCODOX - Floats deployed in 2012, 2014, 2015, 2016 and 2017

As part of the french Equipex NAOS and European H2020 AtlantOS projects, the LOPS is in charge of (1) Implementing and maintaining an array of Argo floats equipped with oxygen sensors in the North-Atlantic Ocean; (2) Performing the quality control of the oxygen data acquired by those floats; (3) Transmitting the qualified data to the Coriolis data center To achieve the last two objectives, the LOPS has developed a Matlab tool (LOCODOX) to correct Argo Oxygen data based on reference datasets. The qualification procedure is done in different steps: (1) Correct the real-time Quality Control flag (QC) in order to put a QC=4 (bad data) to the outliers and "hooks" at the base of the profiles (generally over the first 50m of the ascending profile); (2) When necessary, add an additional pressure correction on DOXY in the form DOXYcorrpres=DOXY*( 1+coef*PRES/1000); (3) Apply the correction estimated by LOCODOX This document summarizes the quality control preformed on the floats deployed by the LOPS in the North-atlantic over the period 2012-2017 and focusses on the LOCODOX corrections. There is one document per year of deployment (2012, 2014, 2015, 2016 and 2017