Higher measured than modeled ozone production at increased NO _<i>x</i> levels in the Colorado Front Range

Abstract. Chemical models must accurately calculate the ozone formation rate, P(O3), to accurately predict ozone levels and test mitigation strategies. However, model chemical mechanisms can contain large uncertainties in P(O3) calculations, which can create uncertainties in ozone forecasts especially during the summertime when P(O3) can be high. One way to test mechanisms is to evaluate model P(O3) using direct measurements. During summer 2014, the Measurement of Ozone Production Sensor (MOPS) measured net P(O3) in Golden, CO, approximately 25 km west of Denver along the Colorado Front Range. Net P(O3) was compared to rates calculated by a photochemical box model using a lumped and a more explicit chemical mechanism. Observed P(O3) was up to a factor of two higher than that modeled during early morning hours when nitric oxide (NO) levels were high, contrary to traditional ozone chemistry theory. This disagreement may be due to model underestimation of hydroperoxyl (HO2) radicals relative to observations at high NO levels. These additional peroxyl radicals could come from the MOPS chamber chemistry or from missing volatile organic compounds co-emitted with NOx; additional cycling of OH into HO2 through reactions involving nitric oxide provides an alternate explanation for higher measured than modeled P(O3). Although the MOPS measurements are new, comparisons of observed and modeled P(O3) in NO space show a similar behavior to other comparisons between P(O3) derived from measurements and modeled P(O3). These comparisons can have implications for the sensitivity of P(O3) to nitrogen oxides and volatile organic compounds during morning hours, and can possibly affect ozone reduction strategies for the region surrounding Golden, CO in addition to other urban and suburban areas that are in non-attainment with national ozone regulations.