The Regional Atmospheric Modeling System (RAMS), Comprehensive Air Quality Model with Extensions (CAMx), and a new emissions inventory were used to study O 3 formation downwind from the Israeli coast during summer 1997. While RAMS reproduced the observed diurnal variation of surface wind speeds and temperatures reasonably well, directions were somewhat less correlated. CAMx generally also reproduced the timing, location, and magnitude of measured surface and 300-m O 3 max over Jerusalem, but also showed an even larger concurrent max over western Jordan, an area without observations. Observed near-zero nighttime surface O 3 values, however, were over predicted because the CAMx grid was too large to accurately account for local emissions and titration effects. CAMx also generally accurately predicted another observed 300m local max near the Dead Sea, but could not account for a unique O 3 depletion mechanism from bromine monoxide released from its water surface. Factor-separation analysis showed that transportation emissions produced 66% and 29% of the Jerusalem and Jordan 300-m simulated maxima, respectively, while power plant emissions produced about 35% and 62%, respectively.