Intercomparison of historical sea surface temperature datasets

• Published in: International Journal of Climatology
• Main Authors: Yasunaka, Sayaka, Hanawa, Kimio
• Format: Article in Journal/Newspaper
• Language: English
• Published: Wiley 2011
• Subjects: Sea ice
• Online Access: http://dx.doi.org/10.1002/joc.2104; https://api.wiley.com/onlinelibrary/tdm/v1/articles/10.1002%2Fjoc.2104; https://rmets.onlinelibrary.wiley.com/doi/pdf/10.1002/joc.2104

Abstract Seven historical sea surface temperature (SST) datasets are compared with each other: the Hadley Center sea ice and SST dataset (HadISST), version 1; the centennial in situ observation‐based estimate of SSTs (COBE); the extended reconstruction of global SST (ERSST), version 3; the optimal smoothing analysis by the Lamont‐Doherty Earth Observatory (LDEO); the monthly summaries of the International Comprehensive Ocean‐Atmosphere Data Set Release 2.1 (ICOADS); the second Hadley Center SST (HadSST); and SSTs by the authors at Tohoku University (TOHOKU). Differences in 30‐year climatologies and standard deviations of anomalies from the climatologies exist, especially in observation‐sparse areas and periods. Correlation among the datasets mainly depends on the number of observational data. Global means from all datasets are consistent with each other except for those from ICOADS. Signals of the El Niño/Southern Oscillation (ENSO) correlate highly with each other after 1880, although the durations and intensities of each event are different. Temporal variations of the Pacific Decadal Oscillation (PDO) correspond well after 1950, become gradually worse backward in time and are scattered before 1880. These differences are ascribed to the use of different interpolation methods for missing grids, treatments of satellite‐derived data, instrumental bias correction methods and other factors. Copyright © 2011 Royal Meteorological Society