Uncertainty in biogenic emission estimates and photochemical reaction rates can contribute significantly to modeling error in Photochemical Air Quality Simulation Models (PAQSMs). Uncertainties in isoprene emissions from biogenic sources, and isoprene atmospheric degradation rates have recently received considerable attention with respect to control strategy selection for the reduction of tropospheric ozone levels. This study addresses the effects of uncertainties in isoprene emissions and reaction rates on ambient ozone concentrations predicted by PAQSMs. Since PAQSMs are computationally intensive, propagation of uncertainty in reaction rate constants using traditional methods, such as Monte Carlo methods, is not computationally feasible. Here, a novel computationally efficient method of uncertainty analysis, called the `Stochastic Response Surface Method (SRSM)', is applied to propagate uncertainty in isoprene emissions and reaction rate parameters. Case studies include estimation of uncertainty in ozone concentrations predicted by (a) a box-model, (b) a plume trajectory model, the Reactive Plume Model (RPM), and (c) an urban-to-regional scale grid model, the Urban Airshed Model (UAM). The results of this analysis are used to characterize the relative importance of uncertainties in isoprene emissions and reaction rates on ozone levels for a wide range of conditions. Furthermore, this work demonstrates the applicability of the SRSM uncertainty propagation methodology to computationally intensive models such as the UAM.