search

chrisroadmap / fair-calibrate

Calibration package for fair (>=2.1)
MIT License
3 stars 0 forks source link

separate treatments for stochastic and deterministic EBMs #36

Open chrisroadmap opened 2 years ago

chrisroadmap commented 2 years ago

I was too hasty to generalise: the stochastic EBM of Cummins et al. (2020) is a slightly different model structurally to Leach et al. (2021) FaIR 2.0. The latter can be approximated by the former by choosing a large value for gamma and setting sigma_xi and sigma_eta to zero. But I don't like approximations.

chrisroadmap commented 2 years ago

This seems pretty difficult to have two slightly different EBMs side to side. When running deterministically, gamma = 1000 is visually indistinguishable from deterministic, so this has been made the default value initialised in EnergyBalanceModel. Issue remains open