Previous - 3.5.2 Other Structural Choices Index Next - 3.5.4 Summary of Reconstruction Techniques and Structural Uncertainty
Yasunaka and Hanawa [2011] examined a range of climate indices based on seven different SST data sets. They found that the disagreement between data sets was marked before 1880, and that the trends in large scale averages and indices tend to diverge outside of the common climatology period. For the global average, the differences between analyses were around 0.2 K before 1920 and around 0.1-0.2 K in the modern period. Even for relatively well-observed events such as the 1925/26 El Niño, the detailed evolution of the SSTs in the tropical Pacific varied from analysis to analysis. The reasons for the discrepancies are not completely clear because each data set is based on a slightly different set of observations that have been quality controlled and processed in different ways, a problem that could be alleviated by running analyses on identical input data sets.
Combined with information about large-scale sampling uncertainties estimated in other ways, the spread between analyses suggests that the large-scale sampling uncertainty in global average SST anomaly is around 0.2 K in the late 19th century. For the large-scale sampling uncertainty of the global average to be much larger would require variability in the early record to have been different from variability in the modern period, which is a possibility. The resolution of such a question is most likely to be achieved via the digitisation of more observations from paper records.
Progress in assessing the differences between analysis techniques can also be made by studying the relative strengths and weaknesses of interpolation techniques on carefully prepared test data sets using synthetic data, or on withheld data from well observed regions. By running each analysis on the same carefully-defined subsets and tests, it should be possible to isolate reasons for the differences between the analyses and assess the reliability of analysis uncertainty estimates. The International Surface Temperature Initiative has been working on such benchmarking exercises for land surface air temperature data, building on work such as the COST ACTION project [Venema et al., 2012].
Previous - 3.5.2 Other Structural Choices Index Next - 3.5.4 Summary of Reconstruction Techniques and Structural Uncertainty