Advancing Stochastic Streamflow Prediction and Projection at Single and Multiple Sites by using Large Scale Climate Indices

Abstract:

Alleviating the impacts of global warming in water resources and freshwater availability requires reliable short- and long-term streamflow scenarios. Many stochastic methods provide ensemble streamflow under current and changing conditions, and yet little attention has been given on how large scale climate indices (LSCIs) affect interannual and seasonal variability. We address this gap by proposing a copula-based methodology that incorporates LSCIs (such as Pacific Decadal Oscillation, El Niño Southern Oscillation, Arctic Oscillation, etc.) in ensemble streamflow generation at single and multiple site. We diagnose the most impactful indices at the basin scale and then use Vine copulas to condition streamflow to these indices. We demonstrate our method in three headwater reaches in southern Alberta, Canada, where large-scale climate indices markedly impact streamflow variability. Results show that integration of LSCIs improves streamflow predictions and projections. Our method is general and can be coupled with different streamflow generators and applied to other regions. We deem this tool can support improved water resources management in the era of climate change.