If the old dataset is outside the error range, it seems there is an issue with estimating error and if it's not, then what changed? (realclimate.org)
Von Schuckmann & Le Traon (2011) also estimate the errors in global trends from the period analysed, and also future error uncertainty. (skepticalscience.com)
Next, we use the observed characteristics of present - day interannual NAO variability to estimate the error on future NAO SLP trends and associated SAT and P trends. (link.springer.com)