An ensemble forecast is a weather or climate forecast made by running a numerical model many times instead of once. Each run, called a member, starts from slightly different initial conditions or uses a slightly altered version of the model. Because the atmosphere is chaotic, small differences grow, and the members diverge. Where they agree, forecasters have more confidence. Where they spread widely, the outcome is uncertain. The share of members showing an event gives an estimate of its probability.
The method turns a single prediction into a distribution, which is what financial users need. Energy traders watch changes in ensemble temperature outlooks because they shift expected demand, measured in each heating degree day and cooling degree day, and so move gas and power prices. Sellers of a weather derivative and insurers use ensembles to estimate an exceedance probability. Seasonal outlooks for the El Niño-Southern Oscillation are built the same way.