The Seasonal-to-Multiyear Large Ensemble (SMYLE) prediction system using the Community Earth System Model version 2

Abstract

Abstract. The potential for multiyear prediction of impactful Earth system change remains relatively underexplored compared to shorter (subseasonal to seasonal) and longer (decadal) timescales. In this study, we introduce a new initialized prediction system using the Community Earth System Model version 2 (CESM2) that is specifically designed to probe potential and actual prediction skill at lead times ranging from 1 month out to 2 years. The Seasonal-to-Multiyear Large Ensemble (SMYLE) consists of a collection of 2-year-long hindcast simulations, with four initializations per year from 1970 to 2019 and an ensemble size of 20. A full suite of output is available for exploring near-term predictability of all Earth system components represented in CESM2. We show that SMYLE skill for El Niño–Southern Oscillation is competitive with other prominent seasonal prediction systems, with correlations exceeding 0.5 beyond a lead time of 12 months. A broad overview of prediction skill reveals varying degrees of potential for useful multiyear predictions of seasonal anomalies in the atmosphere, ocean, land, and sea ice. The SMYLE dataset, experimental design, model, initial conditions, and associated analysis tools are all publicly available, providing a foundation for research on multiyear prediction of environmental change by the wider community.

A multiyear tropical Pacific cooling response to recent Australian wildfires in CESM2

“…Cooling in the Niño3.4 region is simulated in both the AF and control ensembles as shown in Fig. 2A , where background drift has been removed (see Materials and Methods), suggesting that a component of the simulated cooling results from the initial state ( 14 ). By focusing on differences between the ensembles in subsequent analysis, the effects of this initial condition cooling are removed.…”

“…The control simulations for our experiments are from the Seasonal-to-Multiyear Large Ensemble (SMYLE), which prescribes SSP3-7.0 climatological fire emissions rather than variable emissions ( 14 ). The SMYLE experiment consists of 2-year-long hindcast simulations that span the period from 1970 to 2019, with seasonal initializations within each year.…”

“…2 , drift in Niño3.4 SST is estimated from the preceding set of August hindcast simulations from 1970 to 2018. Validation of the SMYLE ensemble has been conducted recently ( 14 ). In this work, additional simulations are performed to augment the SMYLE ensemble.…”

“…The ocean and sea ice initial conditions are based on the forced ocean-sea ice simulations using CESM2, constructed in line with the protocol for version 2 of the Ocean Model Intercomparison Project. SMYLE simulations are forced by smoothed biomass burning emissions, characterized by a mean climatological cycle of the satellite era, which are substituted for the default biomass burning emissions from the Coupled Model Intercomparison Project Phase 6 (CMIP6) due to associated nonlinear model responses (14,(36)(37)(38). Initial conditions in the ensemble are differentiated by the day of the month chosen at random to begin each simulation.…”

“…Like most climate models, it does not, however, include a plume model for estimating the vertical distribution of aerosols, leading to a possible underestimation of climate effects (see Results). Each ensemble consists of 30 members, initialized from highly similar, though not identical, atmospheric initial states within each ensemble ( 14 ). Corresponding members between the two ensembles use identical initial states, and there is therefore no initial state influence on the ensemble mean difference, which serves as an estimate of the climate system response to the Australian wildfires and hereafter is referred to simply as the wildfire response.…”