The 2017 paper by Moore et al., used temperature, CO2, and precipitation in its regression, however, only the temperature part of the regression is implemented in MimiGIVE. Is this correct?
This question may be best posed directly to the specific model developers on the MimiGIVE.jl or MimiMooreEtAlAgricultureImpacts.jl models and related publications, but in brief the calculations of social cost of agriculture in MimiGIVE follow the methods used in Moore et al. (2017) to calculate the social cost of carbon via damage functions. From Moore et al. (2017)'s Methods section:
Calculating the social cost of carbon
Results of the economic modeling are used to create damage functions that relate changes in economic welfare (measured as percentage of the value of agricultural sector output) with temperature change. GTAP results are aggregated from the country level to the 16 FUND regions. Damage functions are based on a linear interpolation between the point-estimates of welfare changes at 1, 2, and 3 °C of warming and then a linear extrapolation beyond 3 °C (results reported in main text) or on a quadratic fitted through the point estimates (Supplementary Table [5]).