Satellite‐derived estimates of ultrafine particle concentrations over eastern North America

Abstract

High concentrations of ultrafine particles (UFP, i.e., particles with diameter < 100 nm) impact both human health and Earth's climate. Recent innovations in remote sensing technologies and data retrievals offer the potential for predicting UFP concentrations based on data from satellite‐borne instrumentation. Herein we present a physically based statistical algorithm to estimate UFP concentrations across eastern North America using remotely sensed aerosol optical depth, Ångstrom exponent, ultraviolet solar radiation flux, and ammonia and sulfur dioxide concentrations. The proposed algorithm is built and independently evaluated using an array of in situ observations. The algorithm is able to capture up to 60% of the variability in daily measured UFP number concentrations at a regionally representative reference site and is thus applied to generate seasonal UFP concentration estimates across eastern North America. The resulting UFP concentrations are cross‐evaluated with simulations from a global aerosol microphysics model. There is a negative bias in the model output relative to the satellite‐driven proxy, which is largest (up to 76%) in summer and may be due to overestimation of UFP from the satellite‐based algorithm derived herein, due to the higher availability of remote sensing data in clear‐sky conditions or uncertainty in the model simulation of new particle formation. Nevertheless, the model and algorithm indicate similar spatial and seasonal variability (spatial correlation coefficients of 0.10 to 0.56), indicating the value of the satellite‐based UFP proxy in global and regional model evaluation exercises and in efforts to identify regions where future in situ data collection should be prioritized.