v2 LR Description

The E3SMv2 lower resolution model is the physical coupled model consisting of atmosphere, land, river transport, ocean, and sea-ice components. E3SMv2 represents a significant evolution from E3SMv1 with substantial updates to its numerical core and improvements in the simulated climate.

The lower resolution configuration of E3SMv2 includes a grid spacing of 110 km in the atmosphere, 165 km in the land, 0.5° in the river routing model, and a grid spacing that varies between 60 km in the mid-latitudes and 30 km at the equator and poles in the ocean and sea ice components.

E3SMv2 is approximately twice as fast (or efficient if measured in terms of power) compared to E3SMv1. The efficiency gains are achieved in the atmosphere and ocean components. In the atmosphere, they arise from a new semi-Lagrangian tracer transport method and a new grid for physics calculations. The gain in the ocean is due to a longer allowable timestep.

The atmospheric physics, while based on the same set of parameterizations as v1, underwent significant retuning in v2 to improve clouds and precipitation (Figure 1).

Additionally, a new convective trigger function for the deep convection (Xie et al., 2019) significantly improves the phase of the diurnal cycle of precipitation, but the amplitude remains weaker than observed (Fig. 2).

The equilibrium climate sensitivity (ECS) is a common metric to quantify the warming of a model in response to an increase in greenhouse gases. ECS is defined as the equilibrium surface temperature change resulting from a doubling in CO₂ concentrations. Because it is not practical to run a model to equilibrium (it would take thousands of years), ECS is approximated by linear regression of top-of-atmosphere radiation vs. surface temperature in a 150 year simulation where CO₂ is abruptly quadrupled. In E3SMv2, ECS is significantly reduced compared to v1 (4.0 K vs 5.3 K) which is mostly attributable to a smaller cloud feedback. An ECS value of 4.0 K is much more plausible than 5.3 K according to a recent assessment (Sherwood et al., 2020). The transient climate response (TCR) measures a model’s response to CO₂ on shorter time scales. TCR is defined as the change in surface temperature averaged for a 20 year period around the time of CO₂ doubling from a simulation in which CO2 concentration is increased by 1% every year. In E3SMv2, TCR is more realistic with a magnitude of 2.4 K compared to 2.9 K in v1. The effective aerosol forcing which measures the radiative impact of aerosol and their interaction with clouds remains essentially unchanged in v2 (ERF aer = −1.5 W m⁻² ). While relatively large, this value is within the likely range assessed by Bellouin et al. (2020).

Unfortunately, the combination of a weaker climate sensitivity without significant reduction in aerosol forcing leads to a worse representation of the global mean surface temperature during the second half of the 20th century (Figure 3). An analysis of single-forcing simulations indicates that correcting the historical record in E3SMv2 would require a substantial reduction in the magnitude of aerosol forcing (∼75%), and a moderate reduction in the TCR (∼25%) (gold line in Figure 3).