Atmospheric River (AR) that often takes place over the eastern North Pacific Ocean is a type of intense storm characterized with a narrow region of strong horizontal water vapor flux associated with polar cold fronts. Many studies show that AR is a critical contributor to extreme precipitation and flooding events of the west coast of the United States. Except for integrated water vapor (IWV) observations gathered from special sensor microwave/imager (SSM/I) and some specially designed field experiments, the ARs remain poorly observed by the existing global observing system. With the launch of the Constellation Observing System for Meteorology, Ionosphere and Climate (COSMIC) mission in April 2006, ~2,000 GPS RO near-real-time soundings per day uniformly distributed around the globe in all weather are now routinely available. Such GPS RO observations will provide valuable observations for AR events.

In this paper, we examine the impact of GPS RO observations on an intense landfalling atmospheric river that devastated portions of the Pacific Northwest with torrential rains and severe flooding on 6-7 November 2006. We first conducted a one-week-long data assimilation cycling experiment from 3 to 9 November 2006 using NCEP GSI analysis system coupled with WRF forecast model. GPS RO soundings from COSMIC and CHAMP missions, in addition to all other observations routinely used by NCEP, were assimilated. The assimilation of GPS RO profiles made use of a local refractivity operator as well as an advanced non-local excess phase operator which considers the effects of atmospheric horizontal gradients. The results show that the assimilation of GPS RO soundings improved the moisture analysis and precipitation forecast. The use of nonlocal excess phase observation operator produced larger and more robust analysis increments, and led to further improvement in precipitation forecasts.