Observing atmospheric rivers using multi-GNSS airborne radio occultation: system description and data evaluation

Abstract. Atmospheric Rivers (ARs) are narrow filaments of high moisture flux responsible for most of the horizontal transport of water vapor from the tropics to mid-latitudes. Improving forecasts of ARs through numerical weather prediction (NWP) is important for increasing the resilience of the western US to flooding and droughts. These NWP forecasts rely on the improved understanding of AR physics and dynamics from satellite, radar, aircraft, and in situ observations, and now airborne radio occultation (ARO) can contribute to those goals. The ARO technique is based on precise measurements of Global Navigation Satellite Systems (GNSS) signal delays collected from a receiver onboard an aircraft from setting or rising GNSS satellites. ARO inherits the advantages of high vertical resolution and all-weather capability of spaceborne RO observations and has the additional advantage of continuous and dense sampling of the targeted storm area. This work presents a comprehensive ARO dataset recovered from four years of AR Reconnaissance (AR Recon) missions over the eastern Pacific. The final dataset is comprised of ∼ 1700 ARO profiles from 39 flights (∼ 260 flight hours) from multiple GNSS constellations. Profiles extend from aircraft cruising altitude (13–14 km) down into the lower troposphere, with more than 50 % of the profiles extending below 4 km, below which the receiver loses or cannot initiate lock. The horizontal drift of the tangent points that comprise a given ARO profile greatly extends the area sampled from just underneath the aircraft to both sides of the flight track (up to ∼ 400 km). The estimated refractivity accuracy with respect to dropsondes is ∼ 1.2 %, in the upper troposphere where the sample points are closely collocated. For the lower troposphere, the agreement is within ∼ 7 % which is the level of consistency expected given the nature of atmospheric variations over the 300–700 km separation between the lowest point and the dropsonde.