hdrake
commented
4 years ago This was a really easy change to the code. Just had to change the mitigate_cost from units of [%GWP] to [trillion USD / year / GtCO2] and change the mitigation cost function from
CM = E * mitigate_cost * f(M)
to
CM = q * mitigate_cost * f(M)
in both the cost function of the Diagnostics submodule and the optimize_controls function of the Optimization submodule.
Implemented in https://github.com/hdrake/ClimateMARGO.jl/pull/27
The mitigation costs (as a fraction of instantaneous GWP) are currently fixed for a given value of the fractional mitigation
M. This is fine if the baseline emissions do not change much over time or increase at a rate similar to the GWP, but it doesn't make any sense when emissions are falling while GWP increases. For example, with the current default emissions pathway, mitigation costs (per ton of CO2 mitigated) increase towards infinity as baseline emissions go to zero between 2100 and 2150. As a result, in almost all of the optimization runs, controlled emissions decrease as expected from 2020 to 2100 but then unexpectedly increase again between 2100 and 2150.