Hourly measurements of O3, NO, NO2, PAN, HNO3 and NOy concentrations, and eddy-covariance fluxes of O3 and NOy over a temperate deciduous forest from June to November, 2000 were used to evaluate the dry deposition velocities (Vd) estimated by the WRF-Chem dry deposition module (WDDM), which adopted Wesely (1989) scheme for surface resistance (Rc), and the Noah land surface model coupled with a photosynthesis-based Gas-exchange Evapotranspiration Model (Noah-GEM). Noah-GEM produced better Vd(O3) variations due to its more realistically simulated stomatal resistance (Rs) than WDDM. Vd(O3) is very sensitive to the minimum canopy stomatal resistance (Ri) which is specified for each seasonal category assigned in WDDM. Treating Sep-Oct as autumn in WDDM for this deciduous forest site caused a large underprediction of Vd(O3) due to the leafless assumption in 'autumn' seasonal category for which an infinite Ri was assigned. Reducing Ri to a value of 70sm-1, the same as the default value for the summer season category, the modeled and measured Vd(O3) agreed reasonably well. HNO3 was found to dominate the NOy flux during the measurement period; thus the modeled Vd(NOy) was mainly controlled by the aerodynamic and quasi-laminar sublayer resistances (Ra and Rb), both being sensitive to the surface roughness length (z0). Using an appropriate value for z0 (10% of canopy height), WDDM and Noah-GEM agreed well with the observed daytime Vd(NOy). The differences in Vd(HNO3) between WDDM and Noah-GEM were small due to the small differences in the calculated Ra and Rb between the two models; however, the differences in Rc of NO2 and PAN between the two models reached a factor of 1.1-1.5, which in turn caused a factor of 1.1-1.3 differences for Vd. Combining the measured concentrations and modeled Vd, NOx, PAN and HNO3 accounted for 19%, 4%, and 70% of the measured NOy fluxes, respectively. © 2011 Elsevier Ltd.