Simple climate emulators - fit for net-zero emission scenarios?

Simple climate models or climate emulators (CEs) are indispensable in the context of climate economics, where the lion share of compute resources is bound to the economic part of the problem. Associated applications traditionally focused on scenarios with strongly increasing carbon emissions. Here we ask to what degree CEs developed with such applications in mind are fit for purpose when it comes to strong reduction or negative emission scenarios, as in the context of the Paris agreement.

To this end, we present and discuss an augmented version of the CE used in the seminal DICE model. Augmentation consists of additional carbon reservoirs, notably a land biosphere and / or a middle ocean reservoir, on top of the three standard carbon reservoirs representing atmosphere, upper ocean, and lower ocean. The different CEs are calibrated and tested against output of large-scale Earth System Models from the Coupled Model Intercomparison Project, run on pre-defined emissions scenarios. For the calibration of the CE, we use the same strategy as in our previous work, i.e., two highly idealized test cases to separately calibrate the CE's carbon cycle and temperature response. In contrast to previous work, we strongly augment the set of test cases to which we expose the calibrated CE. Additional test cases address, in particular, strong mitigation as well as negative emission scenarios. As in our previous work, we demand that all test cases are well documented in the climate literature and can be easily implemented based on that literature. Based on these test cases, we discuss what added value the additional carbon reservoirs may offer with regard to the CE and associated econ studies.