Characterizing Storm Types that Drive Heavy Precipitation at the Watershed Scale

Heavy precipitation is associated with a multitude of negative impacts at the watershed scale including flooding, landslides and debris flows, and damage to water management infrastructure such as dams. A comprehensive understanding of the meteorological drivers of heavy precipitation can improve our process-level knowledge while helping forecast interpretation for better anticipation and mitigation of the associated risks. However, for any given watershed, a range of meteorological patterns and processes may be associated with heavy precipitation. Here we demonstrate an approach to develop a comprehensive climatology of the range of storm types that are associated with heavy precipitation at the watershed scale, focusing on three western US watersheds: the Upper Yuba of California, the Willamette of Oregon, and the Green of Washington. We employ a variant on the self-organizing map (SOM) approach to cluster spatial patterns of integrated water vapor transport for a five-day sequence of days ending on a heavy precipitation day. The resulting SOMs provide a detailed depiction of the range of storm types that drive heavy precipitation over each watershed and provide a platform for further meteorological diagnosis of the processes involved in each storm type. For all three watersheds, atmospheric rivers are a key feature for the days with the heaviest precipitation, while the proportion of heavy precipitation days not associated with an atmospheric river varies by watershed. The methods developed here are extensible to any watershed and can be customized to focus on the season(s) and meteorological variables of interest.