How will the Flood Drivers of Rain-on-Snow Events Change Through the 21st Century?
The California-Nevada New Year’s flood event of 1997 (December 31st, 1996 to January 4th, 1997) flood drivers differentially respond to warming. The atmospheric river becomes more hazardous (left map plots), snowpack systematically decreases (top right violin plots), and short duration, high intensity rainfall increases, often exceeding theoretical precipitation scaling expectations (bottom right violin plots; black dots are theoretical scaling and gray dots are estimated from RRM-E3SM).
Image by the authors
We used a storyline approach to recreate California-Nevada’s infamous flood of record, the New Year's flood event of 1997 (December 31st, 1996 to January 4th, 1997), using the regionally refined capabilities of the Energy Exascale Earth System Model (RRM-E3SM). The 1997 flood was defined as a compound flood event that resulted from the interactions between extreme precipitation, high freezing levels, abrupt snowmelt, and saturated soils. The interaction between these flood drivers resulted in heavy runoff, widespread inundation, and high-water marks set throughout the Sacramento Valley, northern-to-central Sierra Nevada, and northern Nevada. We explored how much this flood event had already been influenced by climate change and how the flood event could respond to future warming levels.
The 1997 California-Nevada New Year's flood event caused over $1B in damages, remains a water management design storm, and is referred to by HyperFACETS stakeholders as the 1-in-100-year event. We test RRM-E3SM’s ability to hindcast and project the 1997 flood event in past, present, and future climates using the Betacast framework. This testing ensures that RRM-E3SM can provide value in anticipating and planning for future flood events, particularly given the changing climate and how flood drivers might differentially respond to warming. This study also highlights several take-home messages for water resource management and emergency planners in how the character of rain-on-snow events could change through the 21st century.
The 1997 flood event was notable not only because of its atmospheric flood drivers but also because of its land-atmosphere interactions (e.g., precipitation interacting with ripened snowpack and saturated soils). We identify that the peak flood hazard potential of a 1997 flood-like event occurs at warming levels of +1.7-2.5°C from pre-industrial conditions. At this warming level, snowpack can still exist throughout the Sierra Nevada, yet is more ripened and ready to produce meltwater, and antecedent soil conditions are wetter, particularly in the central-to-southern Sierra Nevada. Additionally, we show that although storm total precipitation does not significantly differ across warming levels, more intense short-duration rainfall occurs with more warming. We also estimate that the largest flood hazards will shift from the northern to southern portions of California-Nevada, along with enhanced leeside spillover.