PNNL

Abstract

Scenarios of future emissions and land use changes produced by integrated assessment or multi-sector dynamics models (IA-MSD models) have traditionally been developed without accounting for the possible feedbacks, both geophysical and socio-economic, that climate change impacts could produce on those emissions and land use. The omission of such feedbacks, the failure of ‘closing the loop’, risks skewing our assessments of the plausible range of future emission pathways and the associated range of changes in the Earth system. More broadly, a better integrated representation of human and Earth system changes and feedbacks would enable more accurate anticipation of the implications for human and natural systems of alternative socio-economic development pathways. In this study, we use the Global Change Analysis Model to investigate the potential magnitude of this effect by measuring the consequences of including specific impacts on its baseline scenario. Our intent is to determine if the implementation of these impacts has the potential to alter the course of human system dynamics enough to substantially alter emissions and climate. Such a result would call for an endogenous representation of these impacts when generating emissions scenarios. We implement impacts along several channels, both jointly and individually, as exogenous changes to water availability, crop and labor productivity, and energy demand and supply. We account for uncertainties in climate response using a climate model output emulator trained on two climate models producing different regional patterns for the same global warming levels. We then compare at global and regional scales the outcomes from the model runs that account for impacts and the model runs that do not. Our results indicate that the effects on global emissions lead to less than 0.1 C difference in warming by 2100, and therefore do not support endogenizing impacts. This conclusion is conditional on our modelling framework and the specific impact channels represented, but is consistent with other studies that have so far addressed the magnitude of feedbacks by implementing a two-way coupling. However, we do see regional impacts indicating that local economies and well-being measures may be affected significantly.