Include example of the "policy response process" described in paper about climatemargo.jl

[danielkoll]

Here is a minimal example. The example compares a 'delayed' policy scenario in which carbon capture only becomes available after 2100, versus an 'early' policy scenario in which carbon capture is available all along.

using ClimateMARGO
using ClimateMARGO.Models
using ClimateMARGO.Optimization
using ClimateMARGO.PolicyResponse
using PyPlot

## --------
## Define policy scenarios
params = deepcopy(ClimateMARGO.IO.included_configurations["default"])    # load default parameters

# First:
#  * solve no-carbon-capture baseline policy model
m_noCC = ClimateModel(params);
max_deployment_noCC = Dict("mitigate"=>1., "remove"=>0., "geoeng"=>1., "adapt"=>1.);  # disable carbon capture
@time optimize_controls!(m_noCC, max_deployment = max_deployment_noCC);

# Second:
#   * copy baseline
#   * step_forward to t=t0 so policy is preserved before t0
#   * make carbon capture available (= only effective after t0)
#   * finally, need to reset variable present_year; this doesn't seem to affect policy, but otherwise baseline CO2 emissions get messed up in plots
t0 = 80.  # [years]
m_withDelayedCC = deepcopy(m_noCC)
step_forward!(m_withDelayedCC,t0)
max_deployment_withCC = Dict("mitigate"=>1., "remove"=>1., "geoeng"=>1., "adapt"=>1.);
@time optimize_controls!(m_withDelayedCC, max_deployment = max_deployment_withCC);
m_withDelayedCC.domain.present_year = m_noCC.domain.present_year

# Third:
#   * compare to optimum policy if carbon capture had been available all along
m_withEarlyCC = deepcopy(m_noCC)
@time optimize_controls!(m_withEarlyCC, max_deployment = max_deployment_withCC);


## --------
## Make plots
fig, axes = ClimateMARGO.Plotting.plot_state(m_noCC);
fig, axes = ClimateMARGO.Plotting.plot_state(m_withDelayedCC);
fig, axes = ClimateMARGO.Plotting.plot_state(m_withEarlyCC);