It strikes me that the error bar issue is a bit of distraction, because
error bars suggest there is some reliable causal relationship working here.
The statistical significance is, however, still limited as the large
error bars suggest.
Not exact matches
From 1992 to 2003, the decadal ocean heat content changes (blue), along with the contributions from melting glaciers, ice sheets, and sea ice and small contributions from land and atmosphere warming,
suggest a total warming (red) for the planet of 0.6 ± 0.2 W / m2 (95 %
error bars).
What it would cost them is certainty during the post 1970 period e.g bigger
error bars on the gradient, strangely they seem to have bigger
error bars in the summary than the AOGCMS
suggest.
From 1992 to 2003, the decadal ocean heat content changes (blue), along with the contributions from melting glaciers, ice sheets, and sea ice and small contributions from land and atmosphere warming,
suggest a total warming (red) for the planet of 0.6 ± 0.2 W / m2 (95 %
error bars).
Now I'm as fond of
error bars and confidence limits and parameter tests as the next guy, nut let me play the contrarian a moment and
suggest that they don't add much to the argument here.
The early scientific reviews
suggest a couple of reasons: firstly, that modelling the climate as an AR (1) process with a single timescale is an over-simplification; secondly, that a similar analysis in a GCM with a known sensitivity would likely give incorrect results, and finally, that his estimate of the
error bars on his calculation are very optimistic.
Even in the ARGO era (2003 --RRB-, the
error bars and uncertainty ranges for our educated guesses (that's what they are) about deep ocean heat are 10 times greater (and more) than the
suggested temperature changes (hundredths of a degree) themselves.
From 1992 to 2003, the decadal OHC changes (blue) along with the contributions from melting glaciers, ice caps, Greenland, Antarctica, and Arctic sea ice plus small contributions from land to atmosphere warming (red)
suggest a total warming for the planet of 0.6 ± 0.2 W m − 2 (95 %
error bars).
This confidence interval does not overlap with the
error bars associated with distances less than 2 Mb
suggesting the existence of systematic dependencies between methylation differences at distances up to 2 Mb.