fix on reference run: landuse in REMIND standalone runs + MAGICC6

[orichters]

Purpose of this PR

• Hopefully fixes https://github.com/remindmodel/development_issues/issues/221#issuecomment-1953940386
• for Costs|Biomass|Price integral presolve: For cm_startyear > 2005, load the value from reference run and overwrite only for t, not ttot. As this parameter is used only in this single equation for ttot > cm_startyear, this should have no effect on REMIND, but fixes the reporting here in reportCosts.
• for Net Taxes|GHG emissions|w/o CO2 LUC and Net Taxes|GHG emissions|CO2 LUC: The problem was that this equation returned different results if cm_multigasscen differed between scenarios. Load pm_taxrevCO2LUC0 from input_ref.gdx and overwrite it only for t, not ttot
• for Net Taxes|SO2. The effect was very small, caused by the fact that pm_gdp_gdx can slightly change between iterations. In the reference run, the value for p21_tau_so2_tax was not recalculated in postsolve, but in the next runs in the cascade it was indeed calculated based on the adjusted value. As this was a rough approximation anyhow, use pm_gdp which does not change. @strefler is fine with that.
• simply fix MAGICC6 variables in post-processing. We will not (be able to) fix that at the source, and the error is on the order of 0.1%, so I think that is the easiest (say: only) option. @gabriel-abrahao: maybe you can have a short look?

Type of change

• Bug fix

Checklist:

• My code follows the coding etiquette
• I performed a self-review of my own code
• I explained my changes within the PR, particularly in hard-to-understand areas
• All automated model tests pass (FAIL 0 in the output of make test): [ FAIL 0 | WARN 0 | SKIP 6 | PASS 81 ]
• The changelog CHANGELOG.md has been updated correctly

Further information (optional):

• Test run that fixes the errors as intended, see also this development issue: /p/tmp/oliverr/remind-smallfix/output/SSP2EU-PkBudg1050_2024-02-21_09.47.34

[dklein-pik]

The solution for the correct fixing (overwrite only for t and not ttot) look good. Thank you!