EOF analysis of three records of sea-ice concentration spanning the last 30 years

• Published in: Geophysical Research Letters
• Main Authors: Singarayer, JS, Bamber, JL
• Format: Article in Journal/Newspaper
• Language: English
• Published: 2003
• Subjects: Sea ice
• Online Access: https://hdl.handle.net/1983/17e2bf98-6765-4386-be14-6223fdeace7b; https://research-information.bris.ac.uk/en/publications/17e2bf98-6765-4386-be14-6223fdeace7b; https://doi.org/10.1029/2002GL016640; http://www.agu.org/pubs/crossref/2003/2002GL016640.shtml

Several continuous observational datasets of Artic sea-ice concentration are currently available that cover the period since the advent of routine satellite observations. We report on a comparison of three sea-ice concentration datasets. These are the National Ice Center charts, and two passive microwave radiometer datasets derived using different approaches: the NASA team and Bootstrap algorithms. Empirical orthogonal function (EOF) analyses were employed to compare modes of variability and their consistency between the datasets. The analysis was motivated by the need for a reliable, realistic sea ice climatology for use in climate model simulations, for which both the variability and absolute values of extent and concentration are important. We found that, while there are significant discrepancies in absolute concentrations, the major modes of variability derived from all records were essentially the same. Several continuous observational datasets of Artic sea-ice concentration are currently available that cover the period since the advent of routine satellite observations. We report on a comparison of three sea-ice concentration datasets. These are the National Ice Center charts, and two passive microwave radiometer datasets derived using different approaches: the NASA team and Bootstrap algorithms. Empirical orthogonal function (EOF) analyses were employed to compare modes of variability and their consistency between the datasets. The analysis was motivated by the need for a reliable, realistic sea ice climatology for use in climate model simulations, for which both the variability and absolute values of extent and concentration are important. We found that, while there are significant discrepancies in absolute concentrations, the major modes of variability derived from all records were essentially the same.