Create principal components without heat flux

[rplzzz]

We recently added the ocean heat flux variable to the concentration and ESM principal components, but it turns out that there are a lot of the CMIP5 models that don't have useable heat flux data, so we're going to have to revert to the old version. However, the versions that include the heat flux are still occasionally useful, so we're going to keep them.

This is all simple enough, but the one thing we have to figure out is what to call everything. I'd like to reserve the original names we've been using all along for the ones that we are actually using in the Bayesian Monte Carlo. Unfortunately, this means doing a bit of renaming. The new names will be:

• pc_conc Concentration-driven principal components (temperature only)
• pc_emiss Emissions-driven principal components (temperature and CO₂)
• pc_conc_hflux Concentration-driven PCs with temperature and ocean heat flux
• pc_emiss_hflux Emissions-driven PCs with temperature, CO₂ and heat flux.

The tests for the principal components functions currently use the data with heat flux, and I think I'm going to leave it that way, since that provides a more comprehensive test.

[rplzzz]

Ok, this is done. I was thinking I would need to rerun the projections onto the changed components, but since we never updated the projections to the principal components with the fluxes, this turns out to be a no-op. (I reprojected them anyhow, and the new projections match the old ones, which serves as an additional check on the revised components.)