We are still running into issues with solving for S and diff. Steve suggested that Magicc might not have had this problem because of the multiple temperature boxes and we thought that since we could easily test that capability with the recent developments in Hector!!
Hmm this is what we saw when we did a similar exercise but only git to global temperature.
So the pink global temp is the CESM1-CAM5 temperature data that we compared with in step 1. The blue is the Hector output we fit to in step 3. The black lines are the different Hector fits that range in S from 2 to 19 and diff from 0.5 to 20. I would like to stress that while we did not fit to heat flux or global temp in part 3 I wanted to see what was going on so I included them in this plot.
We do not see much of a different in the 21st century temperature time series (global, ocean, land) even though those runs use a wide range of S and diff combinations. But we do see large differences in the 21st century heat flux data!