From DiNezio et al. (2018): "We used CESM1, a model that simulates realistic IPWP climate and is sensitive to changes in the configuration of land masses over the Maritime Continent—a response that is important for simulating glacial climates. CESM1 consists of coupled general circulation models of the atmosphere and ocean, as well as sea ice and land models. Other components of the Earth system, such as the carbon cycle and marine ecosystems, can also be simulated using CESM1; however, we kept them inactive because our focus here is on the climate response to glacial boundary conditions. The climate at the LGM was simulated by prescribing the following boundary conditions: (i) reduced GHG concentrations, (ii) insolation changes due to the orbital configuration at 21 ka, (iii) orography changes due to ice sheets and corresponding roughness length, (iv) surface albedo changes due to ice sheets, and (v) changes in the land-sea distribution and altitude due to lower sea level. A simulation of preindustrial climate was used as control. A series of simulations forced with individual glacial boundary conditions, also known as “single forcing” runs, were used to isolate the climate responses to different glacial drivers. The climate responses and single forcing simulations used to compute them are listed in Table 1. Full details on the implementation of the LGM boundary conditions and the single forcing simulations are available in the Supplementary Materials. The ensemble was augmented by simulations in which the CESM1 atmosphere and land models (CAM5 and CLM4) were coupled to ocean models of varying complexity. These simulations allowed us to explore the importance of ocean-atmosphere coupling in response to ice sheets. First, we replaced the fully dynamical ocean (POP2) with a model of the ocean mixed layer. In this model, the effect of vertical mixing, entrainment, and horizontal currents is prescribed as a seasonally varying heat source or Q-flux. Changes in SSTs computed by this “slab” ocean model can only be influenced by energy exchanges with the atmosphere, such as changes in evaporation or clouds. In the second case, the ocean model consists of climatological SSTs and sea ice extent from the preindustrial control. In this configuration, the air-sea heat fluxes are computed but cannot change the prescribed climatological SST and sea ice extent. Therefore, climate changes simulated in this configuration are due to “uncoupled” atmosphere or land changes. “Thermally coupled” and uncoupled responses were computed from each configuration as in the full-coupled cases by differencing the simulations as specified in Table 1."

See DiNezio et al. (2018) Supplementary Materials for a complete model description.