Advancing Atmospheric River Forecasts into Subseasonal-to-Seasonal Time Scales

Atmospheric rivers are elongated plumes of intense moisture transport that are capable of producing extreme and impactful weather. Along the west coast of North America, they occasionally cause considerable mayhem – delivering flooding rains during periods of heightened activity and desiccating droughts during periods of reduced activity. The intrinsic chaos of the atmosphere makes the prediction of atmospheric rivers at subseasonal-to-seasonal time scales (S2S; ~3 to 5 weeks) an inherently difficult task. Using reanalysis data, our work shows the potential exists to advance forecast lead times of atmospheric river activity into S2S time scales through knowledge of two of the atmosphere’s most prominent oscillations: the Madden-Julian oscillation (MJO) and the quasi-biennial oscillation (QBO). To harness this potential predictability, we develop an empirical prediction scheme of anomalous atmospheric river activity using the current state of the MJO and QBO as predictors. We demonstrate skillful subseasonal “forecasts of opportunity” with lead times extending into S2S time scales. In contrast, state-of-the-art numerical weather prediction (NWP) models from the S2S Prediction Project database lack skill in forecasting AR activity at S2S lead times. To understand the NWP models’ deficiencies, we perform a process-oriented evaluation of the biases in location and frequency of atmospheric river activity along the west coast of North America. Biases in atmospheric river activity are directly related to biases in the models’ depiction of the 500-hPa geopotential heights over the North Pacific. In turn, these biases in 500-hPa geopotential heights are linked to biases in the models’ ability to accurately simulate the MJO and QBO, the subtropical jet, their interaction as a Rossby wave source, and the subsequent development of downstream teleconnections and blocking. In light of the wide-ranging impacts associated with landfalling atmospheric rivers, even modest gains in their subseasonal prediction have great worth in supporting early action decision making and thereby benefit numerous sectors of society.