[Bug] Stress period climate adjustment factors are not ingested/applied to data

[jvcarag]

Summary

When running ReEDS with climate impacts to hydropower or cooling water turned ON (i.e. GSw_ClimateHydro == 1 and/or GSwClimateWater == 1), climate adjustment factors for hydropower capacity factors and unappropriated freshwater seasonal distribution factors are not being read and applied for the stress periods.

Details:

The files not being read in and applied are the versions of climate_hydadjsea.csv and climate_UnappWaterSeaAnnDistr.csv found in the stress period folders of a ReEDS run (inputs_case/stress{model_year}i{iteration}). These files contain the stress-period values of the respective data they represent, and thus, should be read in alongside the representative period versions of the file, similar to cf_hyd and other similar parameters in 2_temporal_params.gms

ReEDS/reeds/core/solve/2_temporal_params.gms

Lines 462 to 494 in 62e6948

	* Written by reeds/input_processing/hourly_writetimeseries.py
	$onempty
	parameter cf_hyd(i,allszn,r,allt) "--fraction-- hydro capacity factors by season and year"
	/
	$offlisting
	$ondelim
	$include inputs_case%ds%%temporal_inputs%%ds%cf_hyd.csv
	$include inputs_case%ds%stress%stress_year%%ds%cf_hyd.csv
	$offdelim
	$onlisting
	/ ;
	$offempty
	$ifthen.climatehydro %GSw_ClimateHydro% == 1
	* Written by climateprep.py
	table climate_hydro_seasonal(r,allszn,allt) "annual/seasonal nondispatchable hydropower availability"
	$offlisting
	$ondelim
	$include inputs_case%ds%%temporal_inputs%%ds%climate_hydadjsea.csv
	$offdelim
	$onlisting
	;
	* adjust cf_hyd based on annual/seasonal climate multipliers
	* non-dispatchable hydro gets new seasonal profiles as well as annually-varying CF
	* dispatchable hydro keeps the original seasonal profiles; only annual CF changes. Reflects the assumption
	* that reservoirs will be utilized in the same seasonal pattern even if seasonal inflows change.
	cf_hyd(i,szn,r,t)$[hydro_nd(i)$(yeart(t)>=Sw_ClimateStartYear)] =
	sum{allt$att(allt,t), cf_hyd(i,szn,r,t) * climate_hydro_seasonal(r,szn,allt) } ;
	cf_hyd(i,szn,r,t)$[hydro_d(i)$(yeart(t)>=Sw_ClimateStartYear)] =
	sum{allt$att(allt,t), cf_hyd(i,szn,r,t) * climate_hydro_annual(r,allt) } ;

In the above example, without stress period data in the climate_hydro_seasonal parameter, Lines 490-491 calculates the non-dispatchable stress period data in cf_hyd as 0 for all years after the GSw_ClimateStartYear designation (see screenshot of inputs.gdx below):

Solution (ToDo):

• Adjust 2_temporal_params.gms to read in both versions of the climate_{hydadjsea|UnappWaterSeaAnnDistr}.csv from the inputs/{temporal_inputs|stress}/ folders.
  • Change climate_hydro_seasonal and watsa_climate from a GAMS table to parameter (required for multi-source parameter definition)
  • Fix reeds/input_processing/hourly_writetimeseries.py to output climate_hydadjsea.csv and climate_UnappWaterSeaAnnDistr.csv in long format with * appended to front of first column name