Paper No. 135-2
Presentation Time: 1:55 PM
TWENTIETH CENTURY EXTREME PRECIPITATION DETECTED IN A HIGH-RESOLUTION, COASTAL LAKE SEDIMENT RECORD FROM CALIFORNIA (Invited Presentation)
California faces increasing economic and societal risks from extreme precipitation and flooding associated with atmospheric rivers (ARs) under projected twenty-first century climate warming. Lake sediments can retain signals of past extreme precipitation events, allowing reconstructions beyond the period of instrumental records. Here, we provide a pre-instrumental record of AR-related extreme precipitation from the last century using a natural archive of lake sediments. Our site – Wildcat Lake, Point Reyes National Seashore – is in the latitudinal zone of the highest frequency landfall for modern ARs in California. First, we cored Wildcat Lake in 2021 and estimated the sediment ages using a Bayesian-based 210Pb, 137Cs, and 226Ra model (± 4 years in the calibration period) and obtained high-resolution imagery, elemental data from XRF, and grain size data to identify detrital layers in the sediment core. Excursions in erosional proxy data (Ti/Al) are positively and significantly correlated (rmedian = 0.45, pmedian = 0.04) with modern records of integrated vapor transport (kg m-1 s-1), a key metric of AR intensity, using correlations that incorporate age model uncertainty. Signals of extreme precipitation – Ti/Al and K/Al excursions – coincide with every extreme AR year (n=23) that occurred since CE 1948. Water years with short or less intense ARs were not captured by this record, suggesting a site-specific threshold for energetic sediment delivery to the core site or the efficiency of the sediment runoff from nearby slopes. Despite the land-use change near the study site, the data suggest intense and long-lasting AR storms are identifiable in this sedimentary record. These results demonstrate the potential of lacustrine archives to provide records of extreme AR activity and provide context for modern variability in AR frequency and magnitude.