Wet Antecedent Soil Moisture Increases Atmospheric River Streamflow Magnitudes Nonlinearly

Abstract Atmospheric rivers (ARs) drive most riverine floods on the U.S. West Coast. However, estimating flood risk based solely on AR intensity and duration is challenging because precipitation phase, antecedent conditions, and physical watershed characteristics (e.g., slope and soil depth) can influence the magnitude of floods. Here, we analyze how antecedent soil moisture (ASM) conditions contribute to variability in streamflow during AR events and how that changes across climatic regimes and physiography in 122 U.S. West Coast watersheds. We identify a robust nonlinear relationship between streamflow and ASM during ARs in 89% of watersheds. The inflection point in this relationship represents a watershed-specific critical ASM threshold, above which event maximum streamflow is, on average, 2–4.5 times larger. Wet ASM conditions amplify the hydrologic impacts of more frequent but weaker, lower moisture transport AR events, while dry ASM conditions attenuate the hydrologic impacts that stronger, higher moisture transport AR events could otherwise cause. Our research shows that watersheds prone to ASM-amplified streamflows have higher evaporation ratios, lower cold-season precipitation, lower snow-to-rain ratios, and shallower, clay-rich soils. Higher evaporation and lower precipitation lead to greater ASM variability during the cold season, increasing streamflow during wet periods and buffering streamflow during dry periods. Lower snow fraction and shallower soils limit the antecedent water storage capacity of a watershed, contributing to greater sensitivity of streamflow peaks to ASM variability. Incorporating ASM thresholds into hydrologic models in these regions prone to AR-amplified streamflow could improve forecasts and decrease uncertainty. Significance Statement Atmospheric river (AR) storm strength and duration are useful indicators of potential flooding, but because the on-the-ground conditions when a storm arrives can influence the timing and amount of precipitation reaching river channels, prestorm soil wetness should be considered. In this study, we identify important watershed-specific values of prestorm soil moisture wetness, above which, high streamflows are more likely. We find that with wet prestorm conditions, streamflow is typically 2–4.5 times larger than streamflow with dry prestorm conditions. We also establish the types of regions where prestorm soil wetness conditions have the largest influence on streamflow and suggest that flood prediction in these regions could be enhanced by considering more detailed information on the prestorm hydrologic states.