Quantifying wintertime O3 and NOx formation with relevance vector machines

This paper uses a machine learning model called a relevance vector machine (RVM) to quantify ozone (O₃) and nitrogen oxides (NO_x) formation under wintertime conditions. Field study measurements were based on previous work described by Olson et al. (2019), where continuous measurements were reported from a wintertime field study in Utah. RVMs were formulated using either O₃ or nitrogen dioxide (NO₂) as the output variable. Values of the correlation coefficient (r²) between predicted and measured concentrations were 0.944 for O₃ and 0.931 for NO₂. RVMs are constructed from the observed measurements and result in sparse model formulations, meaning that only a subset of the data is used to approximate the entire dataset. For this study, the RVM with O₃ as the output variable used only 20% of the measurement data while the RVM with NO₂ used 16%. RVMs were then used as a predictive model to assess the importance of individual precursors. Using O₃ as the output variable, increases in three species resulted in increased O₃ concentrations: hydrogen peroxide (H₂O₂), dinitrogen pentoxide (N₂O₅), and molecular chlorine (Cl₂). For the two termination products measured during the study, nitric acid (HNO₃) and formic acid (CH₂O₂), no change in O₃ concentration was observed. Using NO₂ as the output variable, only increases in N₂O₅ resulted in increased NO₂ concentrations.

The Role of Hydrates, Competing Chemical Constituents, and Surface Composition on ClNO2 Formation

Atomic chlorine (Cl^•) affects air quality and atmospheric oxidizing capacity. Nitryl chloride (ClNO₂) - a common Cl^• source-forms when chloride-containing aerosols react with dinitrogen pentoxide (N₂O₅). A recent study showed that saline lakebed (playa) dust is an inland source of particulate chloride (Cl^-) that generates high ClNO₂. However, the underlying physiochemical factors responsible for observed yields are poorly understood. To elucidate these controlling factors, we utilized single particle and bulk techniques to determine the chemical composition and mineralogy of playa sediment and dust samples from the southwest United States. Single particle analysis shows trace highly hygroscopic magnesium and calcium Cl-containing minerals are present and likely facilitate ClNO₂ formation at low humidity. Single particle and mineralogical analysis detected playa sediment organic matter that hinders N₂O₅ uptake as well as 10 Å-clay minerals (e.g., Illite) that compete with water and chloride for N₂O₅. Finally, we show that the composition of the aerosol surface, rather than the bulk, is critical in ClNO₂ formation. These findings underscore the importance of mixing state, competing reactions, and surface chemistry on N₂O₅ uptake and ClNO₂ yield for playa dusts and, likely, other aerosol systems. Therefore, consideration of particle surface composition is necessary to improve ClNO₂ and air quality modeling.

Time series analysis of wintertime O3 and NOx formation using vector autoregressions

Concentrations of 11 species are reported from continuous measurements taken during a wintertime field study in Utah. Time series data for measured species generally displayed strong diurnal patterns. Six species show a diurnal pattern of daytime maximums (NO, NOy, O3, H2O2, CH2O2, and Cl2), while five species show a diurnal pattern of night time maximums (NO2, HONO, ClNO2, HNO3, and N2O5). Vector autoregression analyses were completed to better understand important species influencing the formation of O3 and NOx. For the species studied, r2 values of predicted versus measured concentrations ranged from 0.82-0.99. Fitting parameters for the autoregressive matrix, Π, indicated the importance of species precursors. In addition, values of fitting parameters for Π were relatively insensitive to data size, with variations generally <10%. Variable causation was quantified using the Granger causation method. Assuming O3 and NOx behave as chemical products, reactants (in order of importance) are as follows: H2O2, N2O5, HONO, and ClNO2.