search

openmodels / META-2021

The Model for Economic Tipping (point) Analysis
GNU General Public License v3.0
14 stars 11 forks source link

SAF cannot be switched off in Monte Carlo mode #17

Closed pwaidelich closed 10 months ago

pwaidelich commented 11 months ago

Currently, runsim() in montecarlo.jl has no switch to deactivate SAF and always runs the following line of code:

update_param!(inst, :SAFModel_ECS, draws.TemperatureModel_fair_ECS[ii]) # SAFModel.ECS

Unless I'm overlooking something here, this seems to implicitly assume that the model provided by full_model() features the SAF component and hence breaks for runs that feature some tipping points but not SAF, such as the following:

model = full_model(rcp="RCP4.5", ssp="SSP2", saf = false) draws = getsim(50, "Fit of Hope and Schaefer (2016)", "Cai et al. central value", "Nordhaus central value", "Distribution", "none", false, false, false) results = runsim(model, draws, true, true, true, "Cai et al. central value", "Distribution")

16 provides a potential suggestion for adding an SAF switch in Monte Carlo mode

jrising commented 10 months ago

Thanks for this! But I want to check why you want to turn it off. My understanding is that the SAF model is a definite improvement over the default climate dynamics in PAGE (or FAIR for that matter). I'm told that SAF doesn't really have tipping point dynamics, so it doesn't really go with the other components, and using a constant feedback effect is just inaccurate.

pwaidelich commented 10 months ago

I want to turn it on/off in MC mode to isolate the effect of the SAF component on model results (e.g., GDP loss) so I can attribute the overall change with tipping points vis-a-vis the base model without tipping points to the different components. But thanks for the additional context; that's helpful to know!

jrising commented 10 months ago

Got it. Okay, you should be good to go! I'll close this.