The following pages tell you everything you need to know about Argo data, how to access it and where to go to learn more about it.
Advice on Pressure Biases in the Argo Data Set
Wednesday 07 April 2010
A part of the global Argo data is subject to biases in reported pressures. These biases are usually less than 5db, but occasionally can be larger (> 20db). These bias errors are being steadily removed by the reprocessing of historical Argo data. We expect that by the end of 2010 these errors will be removed from the global Argo data set in both the delayed-mode and real-time data. Adjusted pressure data are stored in the PRES_ADJUSTED variable, while the raw biased data remains in the PRES variable.
Normally, Argo pressure data have an accuracy near 2.4 db after bias correction. A subset of Argo floats cannot be corrected as the pressure bias was not transmitted by the floats. Within this subset, some will have a high probability of developing large biases. These floats are identified in the delayed-mode processing of Argo data and are flagged with higher pressure errors (20db) in the PRES_ADJUSTED_ERROR variable.
For science applications sensitive to small pressure biases (e.g. calculations of global ocean heat content or mixed layer depth) we recommend that:
1. Users utilise the quality flags in the Argo data files and data labelled with QC = 1
2. Only delayed-mode data are used
3. Only ADJUSTED data be used
4. PRES_ADJUSTED_ERROR be checked and where values are ≥ 20db, these data be rejected
Pressure sensor drifts in Argo and their impacts are described in a paper by Paul Barker and co-authors.