Ozone production efficiency is a parameter used for evaluating the effect o
f anthropogenic NOx on tropospheric ozone. For this work, zero-dimensional
simulations were run to compare the differences in ozone production efficie
ncy when air masses of different origin were separated and when they were m
erged. The purpose of the simulations was to estimate whether coarsely reso
lved models might under or overpredict ozone production due to their blendi
ng of air masses of different origin. Cases were run for several combinatio
ns of air mass origin, different latitudes, times of year, air mass dilutio
n ratios and initial gas concentrations. The main result of this study is t
hat integrated ozone production may be overpredicted by as much as 60% in c
oarse-model grid cells exposed to different air masses. Under certain condi
tions, such as in rather finery resolved urban airshed models, ozone produc
tion may actually be underpredicted by about 20% in mid-latitudes during su
mmer. The results imply that large-scale global models may have a difficult
time correctly predicting ozone concentrations near, for instance, urban/f
ree tropospheric boundaries, a conclusion supported by other studies examin
ing parameters other than ozone production efficiency. (C) 2000 Elsevier Sc
ience Ltd. All rights reserved.