Carbon cycle tuning

[znicholls]

write a method, tune_carbon_cycle, that:

• takes an emissions timeseries and forcing timeseries (can extend to concs later)
• info on how long each timestep is (default to yrs)
• defines a function which takes emissions, 3 a values (4th constrained by fact that sum over a = 1), r0 and beta (to scale rT and rC together), runs fair and returns radiative forcing
• uses scipy.optimize.curve_fit to optimise this function to the target radiative forcing
• prints the optimised parameters

example of using scipy.optimize.curve_fit

from scipy.optimize import curve_fit
a = curve_fit(PD100_atune,
yearsad,
fit,
p0 = [0.2173,0.2240,0.2824,0.2763,1.134083,0.523307],
bounds = ([0.0,0.0,0.0,0.0,0.1,0.1],[1.0,1.0,1.0,1.0,10.0,10.0]))[0]

MAGICC's forcing formula in the Fortran code to confirm that it's the same as FAIR's

DAT_CO2_RF%DATGLOBE(NEXT_YEAR_IDX) = CORE_DELQ2XCO2 /DLOG(2.0D0) * & DLOG(DAT_CO2_CONC%DATGLOBE(NEXT_YEAR_IDX) / DAT_CO2_CONC%PREIND_DATGLOBE)

MAGICC Default RCP values C_0: 279.51ppmv F_2X: 3.71 W/m^2

[znicholls]

Slight complication is that if you want to get the feedbacks right you need to define what everything else is doing. For now I think there are a few options:

• prescribe all non-CO2 forcings into FAIR (without tuning its temperature response). This is what I'm doing for now
• prescribe a temperature timeseries to FAIR (i.e. bypass the temperature response)
• tune FAIR's temperature response before tuning the carbon cycle (essentially prescribing temperatures but in a slightly more coherent way as we probably won't match temperatures exactly even after tuning)

[znicholls]

@Myles @richardcode @njleach This is now written and working I believe. Have a look at the last commit of the oo-develop branch