Prediction

Decadal Prediction Large Ensemble

Climate variations on interannual to decadal timescales can have enormous social, economic, and environmental impacts, making skillful predictions on these timescales a valuable tool for policymakers. As such, there is a growing interest in the scientific, operational, and applications communities in developing initialized forecasts to improve our foreknowledge of events such as floods, droughts, heat waves, and tropical cyclones. Such information requires skillful predictions of surface and upper ocean temperatures and precipitation among others as well as of their effects on the probability of extreme meteorological events. Additionally, deciphering the competing roles of internal and externally-forced variability is important at these timescales.

NCAR scientists recently completed a set of initialized prediction simulations using the Community Earth System Model (CESM). This CESM decadal predic­tion large ensemble (CESM-DPLE; Yeager et al. 2018) is composed of 40-member ensembles initialized each on 01 November between 1954 and 2015 (for a total of 62 start dates) and integrated for 122 months each. With this many ensemble members, CESM-DPLE is becoming a very valuable resource for evaluating decadal predictions of the Earth system in the large-ensemble limit. A unique aspect of the CESM-DPLE that, apart from its ensemble size, makes it unprecedented in the field of near-term climate prediction is the exitance of the complementary, unini­tialized large ensemble of historical simulations. Specifically, the CESM Large Ensemble (CESM-LE; Kay et al. 2015) is a highly successful community project that has accumulated a 40-member ensemble of historical and projection simulations spanning 1920–2100. The CESM-DPLE was generated using the same code base, component model configurations, and historical and projected radiative forcings as in the CESM-LE. Together, CESM-DPLE and CESM-LE offer a powerful means of disentangling the impacts of external forcing versus initialization on hindcast skill and of ascertaining how ensemble size (of both the initialized and uninitialized simulation sets) influences decadal prediction assessments. The CESM-DPLE simulations were performed on the NCAR’s Cheyenne supercomputer via a grant by the NCAR’s Computational and In­formation Systems Laboratory (CISL).

Initial results show that CESM-DPLE exhibits quite promising levels of skill for many different fields, across a broad range of forecast lead times up to decadal scales, both in the ocean and over land. Combined analysis of CESM-DPLE and CESM-LE is revealing significant and potentially useful skill at predicting low-frequency variations in hydroclimate over land, such as over Europe and Africa, that appears to highlight the role of the ocean in modulating decadal climate variations. Large en­sembles are needed to draw robust conclusions about the role of initialization in predictions of noisy atmo­spheric fields and to realize skill given what appears to be unrealistically small signal-to-noise ratios in the model. The inclusion of prognostic ocean biogeo­chemistry in CESM-DPLE opens up new prospects for predictability research that move beyond the physical climate system.

All the CESM-DPLE datasets are available to the community at:
www.cesm.ucar.edu/projects/community-projects/DPLE

Kay, J. E., C. Deser, A. Phillips, A. Mai, C. Hannay, G. Strand, J. Arblaster, S. Bates, G. Danabasoglu, J. Edwards, M. Holland, P. Kushner, J.-F. Lamarque, D. Lawrence, K. Lindsay, A. Middleton, E. Munoz, R. Neale, K. Oleson, L. Polvani, and M. Vertenstein, 2015: The Community Earth System Model (CESM) Large Ensemble project: A community resource for studying climate change in the presence of internal climate variability. Bull. Amer. Meteor. Soc., 96, 1333-1349, Doi: 10.1175/BAMS-D-13-00255.1
Yeager, S. G., G. Danabasoglu, N. A. Rosenbloom, W. Strand, S. C. Bates, G. A. Meehl, A. R. Karspeck, K. Lindsay, M. C. Long, H. Teng, and N. S. Lovenduski, 2018: Predicting near-term changes in the Earth System: A large ensemble of initialized decadal prediction simulations using the Community Earth System Model. Bull. Amer. Meteor. Soc., 99, 1867-1886, Doi: 10.1175/BAMS-D-17-0098.1