The Conditional Relationship Between Atmospheric River Moisture, Wind, and Precipitation in Satellite Observations

Atmospheric Rivers (ARs) are elongated and narrow filaments of water vapor transport in the atmosphere. Convergence of these water vapor plumes are often associated with heavy precipitation events (e.g., US West Coast). Our previous work has reproduced satellite-based AR climatology, variability, and associated precipitation following the integrated vapor transport (IVT) detection algorithm in Guan & Waliser (2015). Here we further quantify the relationship between an AR’s moisture, wind, IVT, and associated precipitation. Following Gonzales et al. (2020) who categorize US West Coast ARs as moisture (wet) versus wind-dominated (windy), we first develop a global, percentile-based classification of wet and windy ARs in satellite observations. It is found that wet ARs have a higher frequency near the equator whereas windy ARs have a higher frequency near polar regions, regardless of season.

Next, a binned scatter approach is used to better understand the sensitivity of mean AR precipitation to changes in other AR metrics (e.g., moisture, wind, and IVT). Binned scatter plots divide the independent variable into dynamically-sized bins and find the conditional means of a dependent variable, similar to the convective onset statistics for tropical convective precipitation against changes in water vapor (e.g., Peters & Neelin, 2006). Overall, we find a strong conditional relationship between mean AR IVT and mean AR precipitation, regardless of the ocean basin and season. For wet and windy ARs, the same relationship exists, but with varying sensitivity to AR IVT. Grid point linear regression suggests that regions with high AR precipitation sensitivity to AR moisture and wind correspond to areas with low frequency of wet and windy ARs, respectively. These conditional statistics in satellite observations may provide a benchmark for quantifying the relationship between ARs and precipitation in climate models.