Anthropogenic climate change is one of the major challenges facing us today. The Department of Energy has recognized that many evident climate trends, if allowed to continue, would have major impacts on the energy sector (DOE, 2013). Many of these trends (such as storms and extreme precipitation) are inherently regional effects, and critically tied to resolution. Our primary tools for studying the impacts of climate change, global earth system models, are unable to resolve the scales important for modeling these processes explicitly and instead are dependent on uncertain parameterizations to capture their effects, often imperfectly. But resolving these processes in a global model takes enormous computational resources that are only now available, and a model capable of utilizing those resources efficiently. The Simple Cloud Resolving E3SM Atmosphere Model (SCREAM) was developed to address these challenges by delivering a computationally efficient, scalable global storm-resolving model for E3SM that can leverage our largest and fastest high performance computing resources. SCREAM has demonstrated incredible performance, scaling well to full utilization of the Frontier supercomputer at OLCF and winning the inaugural Gordon Bell Prize for Climate Modeling for this effort (Taylor et al., 2023).  

To date, the largest simulations possible with SCREAM have been on the order of a months to a year in duration (Caldwell et al. 2021, Donahue et al., in review). While these simulations have been useful at demonstrating performance of the model at capturing basic atmospheric phenomena, short simulations prevent robust evaluation of the mean state climate, and prevent aggregation of sufficient statistics to study extreme events (such as storms, heat waves, and droughts). We will be able to address these shortcomings for the first time with the completion of a multi-decadal simulation with SCREAM run under the well-established AMIP protocol to simulation present-day conditions.  

We present initial results from this ongoing simulation campaign, evaluating the mean state climate and beginning to aggregate statistics of extreme events. This simulation is also instrumented with an updated implementation of the CFMIP Observation Simulator Package and shows improved agreement with ISCCP-retrieved cloudiness relative to previous simulations with SCREAM, enabling more robust evaluation of modeled clouds.