Limit use of execute_load

[tonnrueter]

Purpose of this PR

Using execute_load to load values from GDX risks replacing symbol data entirely. Use execute_load only iff you want a clean slate for all attributes of a variable. In all other cases use execute_loadpoint to merge new data from a GDX into existing symbols while preserving it's attributes (i.e. bounds, levels, ..).

• Switch from execute_load to execute_loadpoint except for
• Usage in core/datainput.gms (lines 1627-1632) since there are no pre-existing values to preserve

Overview of the changes made:

File Path	Lines	Action	Reason
core/datainput.gms	1627-1632	Keep execute_load	Fresh initialization of reference parameters (p_prodSeReference, pm_prodFEReference, p_prodUeReference, p_co2CCSReference, pm_macCost) from scratch for comparison purposes. No pre-existing values to preserve.
core/preloop.gms	170	Switch to Execute_Loadpoint	Loading historical prices (t < cm_startyear) for pm_PEPrice, pm_SEPrice, pm_FEPrice. Future prices may already exist and should be preserved.
modules/21_tax/on/datainput.gms	153-154	Switch to Execute_Loadpoint	Loading baseline tax revenue parameters (pm_taxrevCO2LUC0, pm_taxrevGHG0). Should merge with existing baseline data.
modules/24_trade/se_trade/datainput.gms	189-190	Switch to Execute_Loadpoint	Loading reference demand values (p24_demFeForEsReference, p24_demFeIndSubReference). Should merge with existing reference data.
modules/29_CES_parameters/calibrate/datainput.gms	198	Switch to Execute_Loadpoint	Loading CES calibration data p29_cesdata_load from input.gdx. Should merge with existing calibration data.
modules/29_CES_parameters/calibrate/preloop.gms	29, 38	Switch to Execute_Loadpoint	Loading set mappings (ces2_29, regi_29_load) from input.gdx. Should merge with existing sets.
modules/30_biomass/magpie_40/datainput.gms	79	Switch to Execute_Loadpoint	Loading biomass cost data pm_pebiolc_costs_emu_preloop. Should merge with existing data.
modules/36_buildings/simple/datainput.gms	61	Switch to Execute_Loadpoint	Loading reference building demand p36_uedemand_build. Likely partial data that should be merged with existing data.
modules/37_industry/subsectors/datainput.gms	651	Switch to Execute_Loadpoint	Loading variable vm_demFeSector_afterTax. Variable bounds (.lo, .hi) should be preserved.
modules/37_industry/subsectors/datainput.gms	872	Switch to Execute_Loadpoint	Loading only .l attribute of v37_plasticWaste. Variable bounds should be preserved.

Please note that this might affect scenario results in unforeseen ways. Preliminary test runs:

• No changes, baseline run: /p/tmp/tonnru/clean_remind/output/SMIPv08-M-SSP2-NPi2025-def_2025-11-04_14.37.55
• Blanket replacement of execute_load, ✨ converged ✨: /p/tmp/tonnru/clean_remind/output/SMIPv08-M-SSP2-NPi2025-def_2025-11-04_18.02.35
• Partial replacement. Converged, but had INFES: /p/tmp/tonnru/clean_remind/output/SMIPv08-M-SSP2-NPi2025-def_2025-11-05_17.57.57
• Partial replacement, other scenario, ✨ converged ✨: /p/tmp/tonnru/clean_remind/output/SMIPv08-ML-SSP2-EcPrice300-def_2025-11-05_22.51.01

Please suggest further scenarios to be tested!

Type of change

Indicate the items relevant for your PR by replacing ◻️ with ☑️.
Do not delete any lines. This makes it easier to understand which areas are affected by your changes and which are not.

Parts concerned

• ☑️ GAMS Code
• ◻️ R-scripts
• ◻️ Documentation (GAMS incode documentation, comments, tutorials)
• ◻️ Input data / CES parameters
• ◻️ Tests, CI/CD (continuous integration/deployment)
• ◻️ Configuration (switches in main.gms, default.cfg, and scenario_config*.csv files)
• ◻️ Other (please give a description)

Impact

• ◻️ Bug fix
• ☑️ Refactoring
• ◻️ New feature
• ◻️ Change of parameter values or input data (including CES parameters)
• ◻️ Minor change (default scenarios show only small differences)
• ◻️ Fundamental change of results of default scenarios

Checklist

Do not delete any line. Leave unfinished elements unchecked so others know how far along you are.
In the end all checkboxes must be ticked before you can merge.

• I executed the automated model tests (make test) after my final commit and all tests pass (FAIL 0)
• I adjusted the reporting in remind2 if and where it was needed
• I adjusted the madrat packages (mrremind and other packages involved) for input data generation if and where it was needed
• My code follows the coding etiquette
• I explained my changes within the PR, particularly in hard-to-understand areas
• I checked that the in-code documentation is up-to-date
• I adjusted forbiddenColumnNames in readCheckScenarioConfig.R in case the PR leads to deprecated switches
• I updated the CHANGELOG.md correctly (added, changed, fixed, removed, input data/calibration)

Further information (optional)

• Runs with these changes are here:
• Comparison of results (what changes by this PR?):

[LaviniaBaumstark]

looks good. Did I get it right, that not all testruns did work? Could you provide a link to a compareScenarios for before and after your changes?

[tonnrueter]

looks good. Did I get it right, that not all testruns did work? Could you provide a link to a compareScenarios for before and after your changes?

Please find attached the compareScenario summary. Some observations
Uploading compScen-compScen-base_blanket_partial--2025-11-10_12.23.29-summary.pdf…

• Some differences in the FE pCap variable after 2060 for the different runs (p.24ff)
• Price|Carbon (p. 50) seem identical

[LaviniaBaumstark]

sorry, where is the compareScenario.pdf file?

[tonnrueter]

I've uploaded the compareScenario*.pdf here. Three SSP2-NPi2025 scenarios (base, blanket, partial) correspond to varying degrees of execute_load[point] replacement. Some plots show small differences (e.g. FE|Industry|Solids, Carbon Management|Materials|Plastics, Emi|CO2|w/o Bunkers|Energy and Industrial Processes, CDR|Rocks weathering), however as @LaviniaBaumstark told me, these can be assigned to solver behavior. @JakobBD, @amerfort could you please have a look and let me know if this is alright with you/your teams?

[amerfort]

Hi all,
The only region that shows differences between your test runs is USA and there in the last time steps. It looks to me like the main differences are coming from the energy supply side (PE and SE mixes change). My guess is, that this also affects Industry and Carbon Management variables, but my gut expects the root in the Energy supply side after scanning the compare scenarios.
I can't really track where these differences in PE and SE stem from, and I don't know if this is possible. @robertpietzcker Do you think this is because the optimum is quite flat for a carbon price of 30$/t in USA at the end of the century? Or would you say, this is a problem for this PR and we should investigate more?
One example of this is that somehow only in your "blanket" scenario USA decides to produce lignocellulosic bioenergy crops (PE|Production|Biomass in EJ):

From a purely "Carbon Management and CDR" point of view, the differences are fine.