Every since the ARGO system has been in place the missing heat disappearing into the deep oceans has cropped up. The ocean heat content change is one of the most questionable of all climate data sets. The huge change in energy in 10^22 Joules is based on sparse measurements of vertical temperature, pressure and salinity over several various data acquisition types. Pretty much a hodge podge of methods spliced together with some serious statistical approaches.
I have said before that I am pretty impressed that the data looks as good as it does, but I am not all that convinced that the error ranges are all that accurate. Because of the "almost unbelievable" certainty, any adjustments will likely be small in order to not embarrass previous guestimators.