Plausible future long-term streamflow time series are essential for evaluating different policies and management strategies in river basins and testing the operation of water resource systems. Relying solely on stationary historical data may not be sufficient in a changing climate. Alternatively, uncertainty in the wide range of streamflow projections from General Circulation Models calls into question their direct use in water resources planning. There is thus a need for an intermediate approach to identify ensembles of streamflow time series based on assumptions that provide a rationale for plausible future hydrologic conditions. We developed storylines of plausible future conditions that describe such a rationale by quantitatively defining the associated assumptions and then identifying matching streamflow ensembles. These representative ensembles provide inputs needed for running models to support planning that may need a scenario with specific characteristics or studies that rely on a wide range of scenarios, such as in Decision Making under Deep Uncertainty. Applying this approach in the Colorado River Basin, we worked to identify representative ensembles for each storyline. While three storylines were well matched among existing ensembles there was not a good match for the plausible storyline of warming-driven declining streamflow with increasing variability. To address this gap, we developed a general approach to create new streamflow ensembles using a stochastic nonparametric approach that combines observed and paleo-reconstructed flows and adjusts the marginal distribution of the streamflow time series to incorporate the estimated decline and increasing variability in future flow.

Stochastic hydrology produces ensembles of time series that represent plausible future streamflow to simulate and test the operation of water resource systems. A premise of stochastic hydrology is that ensembles should be statistically representative of what may occur in the future. In the past, the application of this premise has involved producing ensembles that are statistically equivalent to the observed or historical streamflow sequence. This requires a number of metrics or statistics that can be used to test statistical similarity. However, with climate change, the past may no longer be representative of the future. Ensembles to test future systems operations should recognize non-stationarity, and include time series representing expected changes. This poses challenges for their testing and validation. In this paper, we suggest an evidence-based analysis in which streamflow ensembles, whether statistically similar to and representative of the past or a changing future, should be characterized and assessed using an extensive set of statistical metrics. We have assembled a broad set of metrics and applied them to annual streamflow in the Colorado River at Lees Ferry to illustrate the approach. We have also developed a tree-based classification approach to categorize both ensembles and metrics. This approach provides a way to visualize and interpret differences between streamflow ensembles. The metrics presented and their classification provide an analytical framework for characterizing and assessing the suitability of future streamflow ensembles, recognizing the presence of non-stationarity. This contributes to better planning in large river basins, such as the Colorado, facing water supply shortages.