Evaluating Uncertainty and Modes of Variability for Antarctic Atmospheric Rivers
Antarctic atmospheric rivers (ARs) are driven by their synoptic environments and lead to profound and varying impacts along the coastlines and over the continent. The definition and detection of ARs using ARDTs (Atmospheric River Detection Tools) over Antarctica accounts for large uncertainty in AR metrics, and consequently, AR impacts. Uncertainty quantification in the context of the AR-teleconnection relationship is necessary when analyzing the interplay between ARs and modes of variability (MOVs). Understanding ARs in the context of MOVs will ultimately lead to better predictability of these types of extreme events.
Antarctic-specific ARDTs better capture the interior AR footprint compared to global ARDTs, even those with polar constraints. We also find that MOVs hold greater influence over West Antarctica compared to East Antarctica and are generally robust across all ARDTs in sign, but not magnitude. Our study illustrates the importance of choosing an appropriate ARDT fit for purpose.
We find that Antarctic-specific detection tools consistently capture the AR footprint inland over ice sheets, whereas most global detection tools do not. Large-scale synoptic environments and associated ARs, however, are broadly consistent across detection tools. Using data from the Atmospheric River Tracking Method Intercomparison Project and global reanalyses, we quantify the uncertainty in Antarctic AR metrics and evaluate large-scale environments in the context of decadal and interannual modes of variability. The Antarctic western hemisphere has stronger connections to both decadal and interannual modes of variability compared to East Antarctica. Modes evaluated include the Southern Annular Mode (SAM), Pacific Decadal Oscillation (PDO), Pacific South American Mode -2 (PSA2), and the Indian Ocean Dipole (IOD) in and out of phase with El Nino Southern Oscillation (ENSO). The Indian Ocean Dipole’s influence on Antarctic ARs is notably stronger while in phase with ENSO.