ENSO updates

[forsyth2]

Resolves #292.

• Update derivations for NET_FLUX_SRF and LHFLX.
• Change code to use TS if SST is not available.
• Update start_yr and end_yr parameter handling.

[forsyth2]

@chengzhuzhang I still need to run/test this code, but are these changes what you had in mind?

[chengzhuzhang]

Looking good, the system tests need to be updated for parameters specifying year range.

[forsyth2]

@chengzhuzhang Ready for review. I tagged you on a few questions I had.

[chengzhuzhang]

 'NET_FLUX_SRF': OrderedDict([
        (('FSNS','FLNS','LHFLX','SHFLX'), lambda fsns, flns, lhflx, shflx: netflux4(fsns, flns, lhflx, shflx)),
        (('rsds','rsus','rlds','rlus','hfls','hfss'), lambda rsds, rsus, rlds, rlus, hfls, hfss: netflux6(rsds, rsus, rlds, rlus, hfls, hfss)),

I meant to add an entry
(('FSNS','FLNS','QFLX','SHFLX'), lambda fsns, flns, lhflx, shflx: netflux4(fsns, flns, lhflx, shflx)),

with this netflux4 function needs to take in account the qflx to lflx conversion. Or we can have a new netflux4 function just for take care of this.

[forsyth2]

@chengzhuzhang See the second commit for my changes. Please test again and see if that works for you. If so, I'll fixup the commits and merge.

I spent some time trying to add a test case, but that ran into complications. I wanted to add a test where the start years are unequal. However, it looks like we only have two years of data for TREFHT in the tests directory, and if you set one of the start years to equal the end year, then the code fails (this is true on master as well). I'm not sure if that's expected behavior or not -- in any case, that would be outside the scope of this pull request to address.

[chengzhuzhang]

https://compy-dtn.pnl.gov/zhan429/enso_reg_diags_alpha22_to_obs/viewer/enso_diags/index.html

The regression map has succesfully generated, but scatter plots are missing. due to the AttributeError: 'EnsoDiagsParameter' object has no attribute 'start_yr' error..