Identifying Eastern US Atmospheric River Types and Evaluating Historical Trends

Abstract

An atmospheric river (AR) is the primary moisture transport forcing in the Western United States, making ARs the predominant producer of extreme precipitation events in this region. A growing body of evidence suggests similar impacts for the Central and Eastern US. This study determines the most prominent types of ARs in the Central and Eastern US study domain through the implementation of a machine learning methodology. Self-organizing maps (SOMs) are leveraged to determine what “flavors” of ARs exist in the study domain. Four atmospheric river detection criteria are utilized to investigate the variability AR types. Mann-Kendall trend analyses on AR strength and size are produced to evaluate changes over the study period. The results confirm extratropical cyclones as the most common driver of ARs, however, limited kinematic forcing can also instigate the development of AR events. Results show coastal cyclones and lee-side cyclones are responsible for producing the strongest ARs. The trend analysis results suggest that ARs associated with Nor'easters and ARs originating in the Gulf of Mexico are exhibiting increasing trends in intensity and/or size. Increasing moisture transport by mature cyclones across the Central and Eastern US have important implications for flooding in highly populated corridors. Areas of concern include the Northeast and Southeast US, while localized enhancement of rainfall is seen along the eastern and southern slopes of the Appalachian Mountains. In addition to the physical findings, this research highlights the importance and sensitivity of statistically significant results to the specific atmospheric river detection criteria that was leveraged.

Description
Keywords
Atmospheric river, Self-organizing maps, Machine learning, Mid-latitude cyclone
Citation
Ramseyer, C. A., Stanfield, T. J., Van Tol, Z., Gingrich, T., Henry, P., Forister, P., et al. (2022). Identifying Eastern US atmospheric river types and evaluating historical trends. Journal of Geophysical Research: Atmospheres, 127, e2021JD036198. https://doi.org/10.1029/2021JD036198