Ozone production efficiency and removal of NOx in power plants - a comparison of results from models and measurements

Models and measurements are used in combination to investigate ozone production efficiency per NOx (OPE) and photochemical removal of NOx in large power plants. Large power plants account for approximately 35% of NOx emissions in the U.S. and are believed to contributed significantly to O3, especially during regional pollution events. Ryerson et al. (1998) estimated power plant OPE directly from measurements in Tennesee and concluded that OPE in large power plants was much lower than in small power plants or in urban plumes. Here, the measurements from Ryerson et al. are compared with model results and reinterpreted. The initial model calculation yields 33%-60% higher NOx and SO2 then measured in the power plant plumes, even at distances less than 2 hours downwind. This suggests either that emission estimates are erroneous or that a significant portion of emissions were exported above the daytime convective mixed layer during the event. Models with high initial NOx also underestimate NOx removal rates, leading to large errors vs. measured NOx in downwind plume transects. When the model is modified to match the initial measured NOx and SO2 in the plume, results for downwind NOx concentrations and removal rates show much better agreement with measurements. OPE is derived from measured fluxes of SO2, NOx and O3 in plumes, using modified techniques. These result in estimated OPE 1.5-2, compared to 2-3 in models, for 9 hours downwind from the plume source. Model calculations predict that plume OPE increases to 3-4 at distances 18 hours downwind, a value that is comparable to OPE from urban NOx emissions. This higher downwind OPE is linked to model assumptions of greater horizontal dispersion in plumes that were transported overnight. Estimates of power plant OPE based on correlations between O3 and inert tracers (SO2) were found to underestimate the true OPE in models.