(pretty much) all REMIND policy scenarios are misconfigured since three years

[0UmfHxcvx5J7JoaOhFSs5mncnisTJJ6q]

$ awk -F ';' '1 == NR { while ("path_gdx_ref" != $i) ++i; print $1, $i } $i ~ /SSP2EU-NDC/ { print $1, $i }' config/scenario_config_rev3p0p0.csv | column -t
title                    path_gdx_ref
SSP1-NDC                 SSP2EU-NDC
SSP1-NPi                 SSP2EU-NDC
SSP1-PkBudg500           SSP2EU-NDC
SSP1-PkBudg1150          SSP2EU-NDC
SSP2EU-NPi               SSP2EU-NDC
SSP2EU-PkBudg500         SSP2EU-NDC
SSP2EU-PkBudg1150        SSP2EU-NDC
SSP5-NDC                 SSP2EU-NDC
SSP5-NPi                 SSP2EU-NDC
SSP5-PkBudg500           SSP2EU-NDC
SSP5-PkBudg1150          SSP2EU-NDC
SDP-NDC                  SSP2EU-NDC
SDP-NPi                  SSP2EU-NDC
SDP-PkBudg500            SSP2EU-NDC
SDP-PkBudg1150           SSP2EU-NDC
SDP_EI-NDC               SSP2EU-NDC
SDP_EI-NPi               SSP2EU-NDC
SDP_EI-PkBudg500         SSP2EU-NDC
SDP_MC-NDC               SSP2EU-NDC
SDP_MC-NPi               SSP2EU-NDC
SDP_MC-PkBudg500         SSP2EU-NDC
SDP_RC-NDC               SSP2EU-NDC
SDP_RC-NPi               SSP2EU-NDC
SDP_RC-PkBudg500         SSP2EU-NDC
SSP2EU_lowEn-NDC         SSP2EU-NDC
SSP2EU_lowEn-NPi         SSP2EU-NDC
SSP2EU_lowEn-PkBudg500   SSP2EU-NDC
SSP2EU_lowEn-PkBudg1150  SSP2EU-NDC

but as is pointed out in the tutorials

• path_gdx_ref points to the run used for all t < cm_startyear, which can be used for example for delayed transition scenarios.

So pretty much every policy run not for SSP2EU that was not manually configured to do this right as been wrong since either three years or one and a half year since policy runs of the SSP1/SSP5/SDP varieties got fixed to the SSP2EU-NDC.
This in turn leads to infeasibilities with more specified input data (since three months).
Shout-out to @amerfort and @rauner who partially noticed something is fishy.

[LaviniaBaumstark]

well, that was what we have decided to do. We wanted to fix ALL scenarios to SSP2EU_NDC to have the same path independent of the SSP-scenario

[bs538]

(reading this by chance, still on parental leave, but got drawn in by the title). Agree with Lavinia, we discussed this during the 2.0 (I think) validation and decided to fix all scenarios to SSP2-NDC until 2020, as otherwise they would already diverge in 2020. It led to minor jumps in some variables from 2020 to 2025 (e.g. interest rates iirc), but we decided that we can live that.
(will unassign myself for now, if its still open at the beginning of 2023 I'm happy to re-join the discussion)

[0UmfHxcvx5J7JoaOhFSs5mncnisTJJ6q]

well, that was what we have decided to do. We wanted to fix ALL scenarios to SSP2EU_NDC to have the same path independent of the SSP-scenario

Well, then harmonize input data between SSPs until 2020, too. Until you do, SSP5-NDC is broken.

[orichters]

How could harmonizing the input data until 2020 be achieved? Does this mean a bigger rework of mrremind, mrcommons etc. or can this be done elsewhere?