I’ve just collated ERSST v3b and ERSST v2 versions for the tropics and compared ERSST v3 and ERSST v2 versions. CA readers were undoubtedly ready for adjustments, but, using the technical terminology of the leading climate journals, the adjustments were “worse than we thought”. ERSST v3 lowered SSTs before 1880 by up to 0.3 deg C.
The reference article for ERSST v3 is here, which states in the abstract:
Improvements in the late nineteenth century are due to improved tuning of the analysis methods
Update: I’ve amended this article in light of Ryan O’s comments, with which, on further reading, I agree. The difference between ERSST2 and ERSST3 shown in the above graphic is related to the information in SR Figure 2 shown below. You can sort of see the 0.3 deg C shift in the 19th century in this figure, though it’s not explained. SR Figure 2.
What is a bit surprising, I guess, is that 19th century SST estimates are so volatile in respect to the parameters. The analysis reported uses a hurdle of 2 months per year (was 5); 2 years per LF (was 5); 25 deg area (was 15 deg)