Land Temperature and Ocean Heat Content and Baseline Selection

This is one of those in progress posts.  Since there was such a good correlation between North Atlantic Sea Surface temperatures and "Global" land surface temperatures, I was curious how well the Ocean Heat Content (OHC) data agreed with the land surface temperature data.  The data is the Climate Research Unit CRUtemp4 data (global) and the 0-700 meter OHC data from NOAA.

The data is massaged a bit with a common baseline anomaly from 1955 to 2012 to match the OHC data period, then normalized by dividing by the standard deviation of the baseline period. As you can see there is a very good fit between the CRUtemp4 data and the North Atlantic OHC data.  Near the end, about 2005, there appears to be a divergence between CRU4 and Global OHC, but that is a pretty short period and it includes the ARGO data.  Time will tell if that is real or not.

Pretty spiffy eh?  Well, since I used the full length of the OHC data for the anomaly period, that pretty much has to fit.  I have beat the data into submission.

Here I have added the CMAR Sea Level Rise data using the same baseline, almost, 1955 to 2006 so you can see how well I can whip data.  If you extent the trend lines for each of the series the OHC and GMSL will have about the same trend while the CRU4 series will have a much different trend.  That appears to make Global Mean Land Surface Temperature the odd data out using this anomaly baseline. 

But is it Land Surface Temperature or just the Hemispheres since most of the land surface is in the Northern Hemisphere and has a lot of influence from the North Atlantic and the Gulf Stream?

Here is the Hadley Center SST version 3 with the GMSL with the same massaging using 1880 to 2006 for the baseline period.  Using the GMSL as a reference, in this case it appears that the Northern Hemisphere has much larger variations around the mean sea level "mean". 

So let's replace the NH SST with the "Global Land Surface Temperature".  This is using the same 1880 to 2006 baseline for anomaly with the same normalization.  Temperature of the Southern Oceans, Sea Level Rise and Global land surface temperature seem to make some sense with the MSL "mean".   Base on this plot, temperatures and sea level have been rising together since at least 1880 with some "wandering" of both the SST and GLST.  I could select a different baseline period and get radically different results.

Finding the right "reference" to avoid chasing wild geese should be priority one in paleo climate reconstructions.