Seasonal-to-Multiyear Large Ensemble (SMYLE) Experiment

The SMYLE set of initialized hindcasts using CESM2 that is specifically designed to explore Earth system predictability at forecast lead times ranging from 1 month out to 2 years. In addition to extensive hindcast output from all CESM2 component models, SMYLE includes historical reconstructions for the ocean, sea ice, and land component models; an experimental setup that can be replicated and/or modified; and python code for performing post-processing and skill assessment. Details are provided below:

SMYLE at a glance

  • CESM2 model: ocean (POP2, 1°, 60L); atmosphere (CAM6-FV, 1°, 32L); land (CLM5); sea ice (CICE5); ocean biogeochemistry (MARBL)
  • Hindcasts initialized quarterly (1st of February, May, August, November) from 1970 to 2019
  • 24-month simulations
  • 20-member ensembles
  • CAM6 initialization: JRA55 Reanalysis
  • POP2 initialization: JRA55-do forced ocean/sea-ice (FOSI) simulation
  • CICE5 initialization: JRA55-do forced ocean/sea-ice (FOSI) simulation
  • CLM5 initialization: CRU-JRAv2 forced land simulation
  • Data availability on Climate Data Gateway & DOI:
  • Data availability on NCAR’s campaign storage system (accessible from Casper):

Reproducing SMYLE

  • SMYLE was run in 2021 on the Cheyenne system at the NCAR-Wyoming Supercomputing Center. Bit-for-bit reproducibility may not be possible!
  • SMYLE was run using the following special tag of CESM2: cesm2.1.4-SMYLE. Code modifications for SMYLE have since been merged onto the main CESM2.1 trunk.
  • Restart files used for initializing SMYLE hindcasts are on NCAR’s campaign storage system (accessible from Casper):

The following paper documents the SMYLE experimental design and provides a broad overview of system skill:

Yeager, S. G., N. Rosenbloom, S. Glanville, X. Wu, I. Simpson, H. Li, M. J. Molina, K. Krumhardt, S. Mogen, K. Lindsay, D. Lombardozzi, W. Wieder, W. Kim, J. Richter, M. Long, G. Danabasoglu, D. Bailey, M. Holland, N. Lovenduski, and Warren G. Strand (2022): The Seasonal-to-Multiyear Large Ensemble (SMYLE) Prediction System using the Community Earth System Model Version 2, Geosci. Mod. Dev. Discuss., [preprint],


Acknowledging Use of SMYLE Data

We request that you cite both the data DOI ( as well as the dataset description paper in work that makes use of SMYLE. Please add your name to the SMYLE analysis registry to help us track who is using this resource. Finally, we welcome collaboration and information sharing, so please consider joining the CESM Earth System Prediction Working Group and presenting your work at upcoming ESPWG meetings.