Any time series can also be called a “realization.” Each realization itself is composed of smaller realizations. Whenever a larger realization is broken into smaller pieces, those smaller pieces together become an “ensemble.” Any time series, such as hydrographs, temperature or precipitation records, or even forecast data, can be turned into an ensemble. One key point to remember is that each ensemble member is specific to a certain time period, such as a particular year.
In this animation, the large puzzle piece can represent a time-series dataset. Since any time series can be represented by smaller parts, we can show the larger puzzle piece as being composed of smaller puzzle pieces. It is these smaller puzzle pieces that are the ensemble members of the original puzzle piece.