Identifying Better Indicators of Aerosol Wet Scavenging During Long-Range Transport

Abstract. As the dominant sink of aerosol particles, wet scavenging greatly influences aerosol lifetime and interactions with clouds, precipitation, and radiation. However, wet scavenging remains highly uncertain in models, hindering accurate predictions of aerosol spatiotemporal distributions and downstream interactions. In this study, we present a flexible, computationally inexpensive method to identify meteorological variables relevant to estimating wet scavenging using a combination of aircraft, satellite, and reanalysis data augmented by trajectory modeling to account for air mass history. Treating the enhancement (Δ) ratio of black carbon and carbon monoxide (ΔBC/ΔCO) measured by aircraft as an in situ proxy for wet scavenging, we assess the capabilities of an array of meteorological variables to predict ΔBC/ΔCO using regression statistics derived from curve-fitting and k-fold cross-validation. We find that accumulated precipitation along trajectories (APT) – treated as a wet scavenging indicator across multiple studies – is unable to accurately capture ΔBC/ΔCO trends, suggesting that APT is not a good indicator of wet scavenging effects. In contrast, the frequencies of precipitation or high relative humidity along trajectories better predict ΔBC/ΔCO trends and magnitudes, suggesting that these types of meteorological variables are better than APT for estimating the degree of wet scavenging in an air mass. Precipitation characteristics (e.g., intensity, frequency) from satellite retrievals are better indicators of ΔBC/ΔCO than those calculated from reanalysis, supporting previous studies that demonstrated reanalysis to be less reliable than satellite retrievals in terms of precipitation. Finally, top quantiles (e.g., 90^th) of relative humidity are able to consistently capture the behavior of ΔBC/ΔCO and may also be a more suitable indicator of wet scavenging than APT. Future studies can use the best-performing meteorological variables identified in our study to estimate wet scavenging. Furthermore, this method can be repeated for different regions to identify region-specific factors influencing wet scavenging, and our findings may be useful for informing scavenging parametrization schemes in models.