Optimally choosing small ensemble members to produce robust climate simulations

Evans, Jason P.; Ji, Fei; Abramowitz, Gab; Ekström, Marie

doi:10.1088/1748-9326/8/4/044050

Cited by 94 publications

(66 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Some studies have argued that certain GCMs may not be independent from one another because of shared code or parameterization schemes (Evans et al, 2013;Knutti et al, 2010). In an ensemble of opportunity like CMIP5, this dependence 5 may lead to high-density regions in climate variable space and hence influence the selection of models by methods like Kmeans clustering.…”

Section: Discussionmentioning

confidence: 99%

Transferability of climate simulation uncertainty to hydrological climate change impacts

Wang¹,

Chen²,

Cannon³

et al. 2018

Preprint

View full text Add to dashboard Cite

Increasing number of climate models are being produced to cover the uncertainty, which makes it infeasible to use all of them in climate change impact studies. In order to thoughtfully select subsets of climate simulations from a large 10 ensemble, several envelope-based methods have been proposed. The subsets are expected to cover a similar uncertainty envelope as the full ensemble in terms of climate variables. However, it is not a given that the uncertainty in hydrological impacts will be similarly well represented. Therefore, this study investigates the transferability of climate uncertainty related to the choice of climate simulations to hydrological impacts. Two envelope-based selection methods, K-means clustering and Katsavounidis-Kuo-Zhang (KKZ) method, are used to select subsets from an ensemble of 50 climate simulations over two 15 watersheds with very different climates using 31 precipitation and temperature variables. Transferability is evaluated by comparing uncertainty coverage between climate variables and 17 hydrological variables simulated by a hydrological model. The importance of properly choosing climate variables in selecting subsets is investigated by including and excluding temperature variables. Results show that KKZ performs better than K-means at selecting subsets of climate simulations for hydrological impacts, and the uncertainty coverage of climate variables is similar to that of hydrological variables. The subset 20 of first 10 simulations covers over 85% of total uncertainty. As expected, temperature variables are important for the snowrelated watershed, but less important for the rainfall-driven watershed. Overall, envelope-based selection of around 10 climate simulations, based on climate variables that characterize the physical processes controlling hydrology of the watershed, is recommended for hydrological impact studies. 25Hydrol. Earth Syst. Sci. Discuss., https://doi

show abstract

Section: Discussionmentioning

confidence: 99%

Transferability of climate simulation uncertainty to hydrological climate change impacts

Wang¹,

Chen²,

Cannon³

et al. 2018

Preprint

View full text Add to dashboard Cite

show abstract

“…Weighting based on correlations is supposed to remove much of the dependence between models, and thus make the sample more like a sample of independent models. Studies have shown that such a weighting scheme is superior to simple model averaging, with respect to both evaluating uncertainty and improving predictions (Bishop and Abramowitz 2013;Evans et al 2013).…”

Section: Assigning Different Weights To Each Model In a Multi-model Ementioning

confidence: 99%

Lessons from climate modeling on the design and use of ensembles for crop modeling

et al. 2016

View full text Add to dashboard Cite

Working with ensembles of crop models is a recent but important development in crop modeling which promises to lead to better uncertainty estimates for model projections and predictions, better predictions using the ensemble mean or median, and closer collaboration within the modeling community. There are numerous open questions about the best way to create and analyze such ensembles. Much can be learned from the field of climate modeling, given its much longer experience with ensembles. We draw on that experience to identify questions and make propositions that should help make ensemble modeling with crop models more rigorous and informative. The propositions include defining criteria for acceptance of models in a crop MME, exploring criteria for evaluating the degree of relatedness of models in a MME, studying the effect of number of models in the ensemble, development of a statistical model of model sampling, creation of a repository for MME results, studies of possible differential weighting of models in an ensemble, creation of single model ensembles based on sampling from the uncertainty distribution of parameter values or inputs specifically Climatic Change (2016) 139:551-564 DOI 10.1007/s10584-016-1803-1 Highlights ?•Ensembles of crop models can improve uncertainty estimates, impact projections, and collaboration.•The climate modeling community has identified and studied many of the questions related to design and analysis of ensembles.•Much of the climate experience could be adapted to crop modeling. oriented toward uncertainty estimation, the creation of super ensembles that sample more than one source of uncertainty, the analysis of super ensemble results to obtain information on total uncertainty and the separate contributions of different sources of uncertainty and finally further investigation of the use of the multi-model mean or median as a predictor.

show abstract

“…Given the ever-increasing collaborations of geophysical modelling communities in joint assessment studies, MM ensembles are becoming very popular and an opportunity to extend and generalize individual deterministic model results Solazzo and Galmarini, 2014;Galmarini et al, 2004;Vautard et al, 2012;Evans et al, 2013;Bishop and Abramowitz, 2013;and many others).…”

Section: Introductionmentioning

confidence: 99%

A science-based use of ensembles of opportunities for assessment and scenario studies

Solazzo

Galmarini

2015

Atmos. Chem. Phys.

View full text Add to dashboard Cite

Abstract. The multimodel ensemble exercise performed within the HTAP project context ) is used here as an example of how a pre-inspection, diagnosis and selection of an ensemble, can produce more reliable results. The procedure is contrasted with the often-used practice of simply averaging model simulations, assuming different models produce independent results, and using the diversity of simulation as an illusory estimate of model uncertainty. It is further and more importantly demonstrated how conclusions can drastically change when future emission scenarios are analysed using an un-inspected ensemble. The HTAP multimodel ensemble analysis is only taken as an example of a widespread and common practice in air quality modelling.

show abstract

Optimally choosing small ensemble members to produce robust climate simulations

Cited by 94 publications

References 28 publications

Transferability of climate simulation uncertainty to hydrological climate change impacts

Transferability of climate simulation uncertainty to hydrological climate change impacts

Lessons from climate modeling on the design and use of ensembles for crop modeling

A science-based use of ensembles of opportunities for assessment and scenario studies

Contact Info

Product

Resources

About