Ensemble forecasting

Ensemble forecasting runs multiple simulations with varied initial conditions, model assumptions, or both. The spread and agreement among their forecasts help represent uncertainty.

Numerical weather prediction: The use of numerical models to predict the future state of the atmosphere. Ensemble forecasts repeatedly run weather-prediction models to sample possible atmospheric outcomes.
Numerical weather prediction ensemble: A collection of weather-model forecasts generated from varied initial states or model configurations. Operational weather centers use these collections to estimate forecast uncertainty.
Numerical model: A mathematical representation of a system that uses numerical computation to simulate its behavior. Each ensemble member is a numerical simulation of the system being forecast.
Deterministic forecasting: Forecasting that produces one predicted outcome for each specified set of inputs. A single deterministic run lacks the range of outcomes supplied by an ensemble.
Initial-condition perturbation: A deliberate small change to a model's starting state used to test sensitivity to initial conditions. Perturbing the starting atmosphere creates distinct ensemble members.
Hurricane track forecast: A prediction of the future path of a tropical cyclone. Differences among ensemble tracks show where a storm's path is uncertain.
Initial-value problem: A differential equation problem specified by the system's state at a starting time. Forecast trajectories depend on the initial state supplied to the model.
Single-model forecasting: Forecasting based on one model configuration rather than a collection of varied simulations. It does not directly sample uncertainty from varied initial states or assumptions.
Parameterization (climate modeling): A simplified representation of processes that a model cannot resolve directly. Changing parameter choices can produce members that reflect uncertainty in unresolved processes.
Flood forecasting: The prediction of river levels, flows, or inundation before or during a flood. Ensembles represent uncertainty in rainfall, runoff, and river-flow predictions.

Linked from 13 pages