Coppied from: http://davidappell.blogspot.com/2012/01/trenberth-response-to-todays-loeb-et-al.html

"It seems to have its main point that our earlier paper was wrong. I will certainly disagree with that. In our earlier papers (refs 2,8 and 27) we pointed out that there were major discrepancies between the inventory of energy uptake in the climate system, mainly in the oceans, and the top-of-atmosphere observed changes. The period of "missing energy" was 2008-2009 which was a La Nina period and their Fig 2 shows indeed that there was a large input of energy into the climate system at that time. But 2008 was the coldest year this century and so where did the heat go?

We did not make a big deal about the uncertainties in the observations which are highlighted in this paper. But we were well aware of them. The main point of our paper was that yes, perhaps the observations are consistent within the error bars but if so, the error bars (uncertainties) are so large as to make the values useless. A key purpose of our paper was to challenge both the ocean heat content community and the CERES (atmospheric radiation) communities to do better. Both have responded and the situation has improved somewhat. The latest CERES data as reported here has corrected their data and found about 20% of the problem. In addition the OHC communities have improved their estimates and some of the problem has gone away from that standpoint too. But there remain some major problems. As they note on p 3: the correlation with two of the OHC data sets is only 0.05, and they choose to use the one that is correlated 0.46. Even that is not very good and is not significant for so few values.

Moreover, the uncertainties computed by Loeb et al for CERES appear to be wrong. They included the systematic error in the interannual error bar, so the real error bar on the change is less than shown and in fact it seems likely the agreement is not within the uncertainty.

So while their conclusions may be valid: yes there is no evidence of a discrepancy, given their uncertainties, and yes there is no "statistically significant" decline in OHC rates of change, but the uncertainties are so large that neither dataset is useful to know what is really going on, and that is the key point. The discrepancies among OHC data sets remain huge. We MUST do better. So the key point in their title is "within uncertainty". It should add: "but the uncertainty is too large."