Glossary

Design of experiment: Provides a framework for the extraction of all plausible information about the impact of each factor on the output of the numerical model

Exploratory modeling: Use of large ensembles of uncertain conditions to discover decision-relevant combinations of uncertain factors

Factor: Any model component that can affect model outputs: inputs, resolution levels, coupling relationships, model relationships and parameters. In models with acceptable model fidelity these factors may represent elements of the real-world system under study.

Factor mapping: A technique to identify which uncertain model factors lead to certain model behavior

Factor prioritization: A technique to identify the uncertain factors which, when fixed to their true value, would lead to the greatest reduction in output variability

Factor screening: A technique to identify model components that have a negligible effect or make no significant contributions to the variability of the outputs or metrics of interest

First-, second-, total-order effects: First-order effects indicate the percent of model output variance contributed by a factor individually. Second-order effects capture how interactions between a pair of parameter input variables can lead to change in model output. Total-order effects consider all the effects a factor has, individually and in interaction with other factors.

Hindcasting: A type of predictive check that uses the model to estimate output for past events to see how well the output matches the known results.

Pre-calibration: A hybrid uncertainty assessment method that involves identifying a plausible set of parameters using some prespecified screening criterion, such as the distance from the model results to the observations.

Prior: The best assessment of the probability of an event based on existing knowledge before a new experiment is conducted

Posterior: The revised or updated probability of an event after taking into account new information

Probabilistic inversion: Uses additional information, for instance, a probabilistic expert assessment or survey result, to update an existing prior distribution

Return level: A value that is expected to be equaled or exceeded on average once every interval of time (T) (with a probability of 1/T)

Return period: The estimated time interval between events of a similar size or intensity/

Sampling: The process of selecting model parameters or inputs that characterize the model uncertainty space.

Scenario discovery: Use of large ensembles of uncertain conditions to discover decision-relevant combinations of uncertain factors

Sensitivity analysis: Conducted to understand the factors and processes that most (or least) control a model’s outputs

Local sensitivity analysis: Model evaluation performed by varying uncertain factors around specific reference values

Global sensitivity analysis: Model evaluation performed by varying uncertain factors throughout their entire feasible value space

Uncertainty

Deep uncertainty: Refers to situations where expert opinions consulted on a decision do not know or cannot agree on system boundaries, or the outcomes of interest and their relative importance, or the prior probability Distribution for the various uncertain factors present

Epistemic uncertainty: Systematic uncertainty that comes about due to the lack of knowledge or data to choose the best model

Ontological uncertainty: Uncertainties due to processes, interactions, or futures, that are not contained within current conceptual models

Aleatory uncertainty: Uncertainty due to natural randomness in processes

Uncertainty characterization: Model evaluation under alternative factor hypotheses to explore their implications for model output uncertainty

Uncertainty quantification: Representation of model output uncertainty using probability distributions

Variance decomposition: A technique to partition how much of the variability in a model’s output is due to different explanatory variables.