Assessing the Impact of Initialization on Decadal Prediction Skill

Sospedra‐Alfonso, Reinel; Boer, G. J.

doi:10.1029/2019gl086361

Cited by 15 publications

(11 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A recently developed approach ( 19 ) is applied to investigate the relative contributions of the external radiative forcing and initialization to the overall decadal prediction skill of summer ITP rainfall. First, the time series of the observation ( X ), the decadal prediction ( Y ), and the historical simulation ( U ) are decomposed as

where k represents the ensemble members; χ f , ψ f , and φ f represent the externally forced components; χ and ψ represent the predictable internally generated components; and x , y k , and u k represent the unpredictable noise components.…”

Section: Methodsmentioning

confidence: 99%

“…Decision-makers concerned with the climate adaptation and resilience could benefit greatly from the skillful decadal climate prediction. In general, both externally forced and internally generated climate variability contribute to decadal climate predictability ( 19 ). Decadal climate predictions can be enabled by initialized climate model simulations, which reproduce the historical externally forced variability by prescribing the external radiative forcing as in the historical climate simulations and forecast the internally generated components of the climate system by using data assimilation systems to preset the observed climate state at the beginning of the prediction.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Skillful prediction of summer rainfall in the Tibetan Plateau on multiyear time scales

Zhou

2021

Sci. Adv.

View full text Add to dashboard Cite

Skillful near-term climate predictions of rainfall over the Tibetan Plateau (TP), the Asian water tower, benefit billions of people. On the basis of the state-of-the-art decadal prediction models, we showed evidence that although the raw model outputs show low predicted ability for the summer Inner TP (ITP) rainfall due to low signal-to-noise ratios in models, we can produce realistic predictions by extracting the predictable signal from large ensemble predictions along with a postprocessing procedure of variance adjustment. The results indicate that the summer ITP rainfall is highly predictable on multiyear time scales. The predictability of ITP rainfall originates from the Silk Road pattern driven by sea surface temperature over the subpolar gyre region in North Atlantic. Real-time forecasts suggest that the ITP will become wetter, with 12.8% increase in rainfall during 2020–2027 relative to 1986–2005. Our results will help the water resources management in the surrounding regions.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Skillful prediction of summer rainfall in the Tibetan Plateau on multiyear time scales

Zhou

2021

Sci. Adv.

View full text Add to dashboard Cite

show abstract

“…Initialized predictions from CMIP6 show broad agreement on ACC skill for SPG SST, with high initial skill and skill degradation over time (Figure 7A). The only prominent outlier to this is the CanESM5 model, which displays a strong initialization shock until approximately lead year 7 due to issues in the North Atlantic region with the direct initialization from the ORAS5 ocean reanalysis (Sospedra-Alfonso and Boer, 2020;Tietsche et al, 2020). For this reason, we will discuss CanESM5 as a special case whenever appropriate.…”

Section: Toward Performance Based Weighting For Initialized Predictionsmentioning

confidence: 99%

Toward Consistent Observational Constraints in Climate Predictions and Projections

et al. 2021

View full text Add to dashboard Cite

Observations facilitate model evaluation and provide constraints that are relevant to future predictions and projections. Constraints for uninitialized projections are generally based on model performance in simulating climatology and climate change. For initialized predictions, skill scores over the hindcast period provide insight into the relative performance of models, and the value of initialization as compared to projections. Predictions and projections combined can, in principle, provide seamless decadal to multi-decadal climate information. For that, though, the role of observations in skill estimates and constraints needs to be understood in order to use both consistently across the prediction and projection time horizons. This paper discusses the challenges in doing so, illustrated by examples of state-of-the-art methods for predicting and projecting changes in European climate. It discusses constraints across prediction and projection methods, their interpretation, and the metrics that drive them such as process accuracy, accurate trends or high signal-to-noise ratio. We also discuss the potential to combine constraints to arrive at more reliable climate prediction systems from years to decades. To illustrate constraints on projections, we discuss their use in the UK's climate prediction system UKCP18, the case of model performance weights obtained from the Climate model Weighting by Independence and Performance (ClimWIP) method, and the estimated magnitude of the forced signal in observations from detection and attribution. For initialized predictions, skill scores are used to evaluate which models perform well, what might contribute to this performance, and how skill may vary over time. Skill estimates also vary with different phases of climate variability and climatic conditions, and are influenced by the presence of external forcing. This complicates the systematic use of observational constraints. Furthermore, we illustrate that sub-selecting simulations from large ensembles based on reproduction of the observed evolution of climate variations is a good testbed for combining projections and predictions. Finally, the methods described in this paper potentially add value to projections and predictions for users, but must be used with caution.

show abstract

“…The expectation is that by taking advantage of predictable slowly-varying internally generated fluctuations of the climate system, including those originating from the ocean, climate phenomena such as multi-year atmospheric circulation changes (Smith et al, 2010;Sutton and Dong, 2012;Monerie et al, 2018), their impact on near-surface temperature and rainfall (Zhang and Delworth, 2006;Boer et al, 2013;McKinnon et al, 2016;Sheen et al, 2017), and the frequency of extreme weather events (Eade et al, 2012;Ruprich-Robert et al, 2018), can be predicted a year or more in advance. Furthermore, initialization affects the model response to external forcing and can potentially correct unrealistic simulated trends (Sospedra-Alfonso and Boer, 2020).…”

Section: Introductionmentioning

confidence: 99%

“…Assessing the added value of decadal climate predictions over climate simulations can, however, be difficult. In the presence of a strong long-term climate response to external forcing, as is the case for near-surface air temperature for instance, the externally forced component carries most of the predictable variance and can mask the contribution of relatively weaker internal variations to the skill (Smith et al, 2019;Sospedra-Alfonso and Boer, 2020). Underestimation of the predictable signal can also degrade decadal forecast skill (Sienz et al, 2016), motivating the use of large ensembles to better extract the predictable component of the forecast (Scaife and Smith, 2018;Yeager et al, 2018;Smith et al, 2019Smith et al, , 2020Deser et al, 2020).…”

Section: Introductionmentioning

confidence: 99%

Decadal climate predictions with the Canadian Earth System Model version 5 (CanESM5)

Sospedra‐Alfonso

Merryfield

Boer

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Abstract. The Canadian Earth System Model version 5 (CanESM5) developed at Environment and Climate Change Canada's Canadian Centre for Climate Modelling and Analysis (CCCma) is participating in phase 6 of the Coupled Model Intercomparison Project (CMIP6). A 40-member ensemble of CanESM5 historical decadal forecasts (or hindcasts) is integrated for ten years from realistic initial states during 1961 to present using prescribed external forcing. The results are part of CCCma's contribution to the Decadal Climate Prediction Project (DCPP) component of CMIP6. This paper evaluates CanESM5 large ensemble decadal hindcasts against observational benchmarks and against historical climate simulations initialized from preindustrial control run states. The focus is on the evaluation of the potential predictability and actual skill of annual and multi-year averages of key oceanic and atmospheric fields at regional and global scales. The impact of initialization on prediction skill is quantified from the hindcasts decomposition into uninitialized and initialized components. The dependence of potential and actual skill on ensemble size is examined. CanESM5 decadal hindcasts skilfully predict upper-ocean states and surface climate with a significant impact from initialization that depend on climate variable, forecast range, and geographic location. Deficiencies in the skill of North Atlantic surface climate are identified and potential causes discussed. The inclusion of biogeochemical modules in CanESM5 enables the prediction of carbon cycle variables which are shown to be skilful on decadal time scales, with a strong long-lasting impact from initialization on skill in the ocean and a moderate short-lived impact on land.

show abstract

Assessing the Impact of Initialization on Decadal Prediction Skill

Cited by 15 publications

References 30 publications

Skillful prediction of summer rainfall in the Tibetan Plateau on multiyear time scales

Skillful prediction of summer rainfall in the Tibetan Plateau on multiyear time scales

Toward Consistent Observational Constraints in Climate Predictions and Projections

Decadal climate predictions with the Canadian Earth System Model version 5 (CanESM5)

Contact Info

Product

Resources

About