Historical Atmospheric River Optimization to Estimate Maximum Precipitation through a Total Storm Approach

Probable maximum precipitation is a common tool to assess the risk to existing infrastructure under worst case scenario storm events. In the American River Watershed, probable maximum precipitation estimates have been computed to ensure the reservoir can safely operate during extreme events. A typical probable maximum precipitation estimate is limited to a 72hr duration maximum. However, this does not account for the role of long duration storms or the possibility of events in quick succession which threaten reservoir operations if cumulative inflow is too large to manage. A maximum precipitation methodology is presented to maximize historical atmospheric rivers for their total storm period. Six historical events are maximized through Atmospheric Boundary Condition Shifting and Relative Humidity Optimization through dynamical downscaling of atmospheric reanalysis data and physical models. Results indicate the importance of storm duration contributing to the cumulative precipitation depth as well as the definition used to define an individual storm period. Of the six events considered, five were substantially intensified by the maximization procedure, highlighting the importance of considering historical atmospheric rivers that did not impact the watershed directly. The one event that did not increase considerably by maximization was already near optimal in terms of landfall location and atmospheric moisture so only slight alterations were needed. By considering the total storm period in the maximum precipitation estimate, as well as by using physically based models, the atmospheric and land interactions that contribute to extreme flooding due to severe precipitation can be fully incorporated into future risk assessments.