A Monte Carlo study of upper tropospheric reactive nitrogen during the Pacific Exploratory Mission in the Western Pacific Ocean (PEM‐West B)

Abstract

A subset of Pacific Exploratory Mission in the Western Pacific Ocean (PEM‐West B) data (northwestern Pacific, March 1994), selected to represent upper troposphere (UT) midlatitude conditions, is analyzed to answer the following questions: (1) Is there a shortfall in total reactive nitrogen (NOy) as measured during PEM‐West B in this region? (2) If so, what are the likely contributions of interfering constituents like HCN or of other reactive nitrogen species? For question 1 our analyses show that 87% of total reactive nitrogen measured in the UT is accounted for by NOx + HNO3 + PAN (similar to Kondo et al. [this issue (a, b)], Talbot et al. [this issue], and Singh et al. [1997a, b]). For question 2 we find that less than 20 pptv (<5% of mean NOy) is possibly HCN. A one‐dimensional model that simulates mean mixing ratios of this PEM‐West B data is used with a Monte Carlo approach to explore other candidates for unmeasured nitrogen species and NOy partitioning. Using standard gas‐phase reactions with varying rate coefficients [Thompson and Stewart, 1991; Stewart and Thompson, 1996], it is found, on average, that NOx + HNO3 + PAN = 399 pptv (observed mean equal to 432±97 pptv). A more complete inventory for total reactive nitrogen (Σ NOi = NOx + HNO3 + PAN + HNO4 + CH3O2NO2 + alkyl nitrates + C2H5O2NO2) is 494±91 pptv, with 19% consisting of HNO4 + CH3O2NO2 + alkyl nitrates. Thus, whether unmeasured forms of reactive nitrogen are present or not, total reactive nitrogen as measured at midlatitude UT during PEM‐West B is accounted for within the measurement uncertainty. The greatest kinetics uncertainties are in thermolytic losses for HNO4, PAN, and CH3O2NO2 (120–150% by the method of Stewart and Thompson [1996]). Nonetheless, comparison with PEM‐West B data shows that panel‐recommended kinetics expressions [Demore et al., 1994] can explain reactive nitrogen observations without invoking extreme rates or heterogeneous processes. In summary, large concentrations of unmeasured reactive nitrogen species were not prevalent during midlatitude UT PEM‐West B sampling although the observed shortfall (13%) can be explained by HNO4 + alkyl nitrates + CH3O2NO2. Agreement between theory and observations may also reflect improved instrument capabilities for measuring reactive nitrogen.