1 Introduction
Atmospheric Rivers (ARs) can generate extreme precipitation events, flooding, and contribute ~80% of low to mid latitude and ~30% of high latitude meridional fluxes of water and energy (Woods et al., 2013; Zhu & Newell, 1998). Their critical role in the atmosphere makes it important to understand their life cycle. Details of their precipitation are well studied (e.g., Lavers & Villarini, 2013; Mahoney et al., 2016; Waliser & Guan, 2017); however, less has been published regarding energy and moisture sources within ARs (e.g., Bao et al., 2006).
ARs could result from organized, local convergence associated with an extratropical cyclone (ETC) (e.g., Dacre et al., 2015), behave as ‘rivers’ representative of synoptic transports of heat and moisture (e.g., Shinoda et al., 2019; Sodemann & Stohl, 2013), or, on some occasions, represent a combination of both. Despite this question being examined from multiple observational and modeling perspectives, it remains not fully resolved (Eiras-Barca et al., 2017; Ramos et al., 2016; Payne et al., 2020). For ARs associated with extratropical cyclones (~82% of ARs in the North Pacific; Zhang et al., 2019), the primary energy source is the atmospheric response to baroclinic instability. However, they also found that only 45% of ETCs are associate with ARs, and this suggests that the moisture component to ARs may depend on unique circulation or surface fluxes. Dacre et al. (2019) posit that moisture transported into the AR via a feeder airstream is rained out before traveling too far poleward, based on their ETC-focused moisture budget using reanalysis data. If so, then one might expect subsequent moisture is added via surface fluxes. As such, the work herein is focused on the behavior of the ocean surface fluxes under ARs during the events.
One past constraint in determining AR energy sources was that observations of surface winds directly beneath ARs are obscured by precipitation (Bourassa et al., 2019). This impacted air-sea flux estimates that principally derive from bulk flux equations involving wind speed. The relatively recently (2017) launched NASA Cyclone Global Navigation Satellite System (CYGNSS) GeoPositioning Satellite (GPS) receiver constellation is unique for its ability to measure surface roughness to estimate wind speeds even under heavily precipitating clouds (Ruf et al., 2019), allowing air-sea fluxes to be estimated with higher fidelity (Crespo et al., 2019).
Using CYGNSS air-sea fluxes to analyze ETCs, Naud et al. (2021) show that the weakest fluxes out of the ocean occur in the storm warm sector where poleward advection of warm air causes surface air temperature to be as warm as the ocean below, limiting energy fluxes out of the ocean. Given that the warm sector of ETCs is often associated with an AR, the Naud et al. (2021) result suggests energy fluxed from ocean to atmosphere under AR regions may be small. To test this idea, we examine CYGNSS surface heat fluxes under ARs.
If not from directly below, where do ARs get their fuel? We supplement CYGNSS observational surface flux analysis using vapor source distribution tracers in the NASA GISS climate model to determine explicitly the provenance of moisture (heat) contained in ARs. Since the GISS model water tracers do not require a priori definition of evaporative region, they may be collated and examined in post-processing for any region of the world facilitating the comparison here to ~7500 ARs over the 2018-2022 time period. These tracers remain robust across the boundary layer and through time allowing for far-field moisture sources to be examined across greater distances than in some back-trajectory investigations. This tracer work allows us to quantify the relative extent to which AR moisture is provided by fluxes directly below by calculating climatological transport distances and sources of moisture in ARs and contrasting that to moisture sources during other times. Further, we can also analyze conditions during AR genesis specifically, to determine if there is a change in behavior when ARs initially form as compared to after formation. Importantly, our simulated results can be vetted against CYGNSS even when well-organized AR events rain heavily.