Volatile organic compounds (VOC) play an important role in air quality and climate by acting as precursors for tropospheric ozone and secondary organic aerosols (Logan 1985). Forest ecosystems represent an important source of VOC to the atmosphere, contributing ten times more to the global VOC budget than anthropogenic sources (Fuentes et al. 2000). Modeling studies suggest that VOC oxidation in rural environments is not well understood. Specifically, first-generation VOC oxidation products are overestimated (Ganzeveld et al. 2008) and oxidant levels are underestimated (Lelieveld et al. 2008). Recent work attributes these model biases to inadequacies in model treatments of gas-phase chemistry and dynamic processes in the forest canopy.One major source of uncertainty stems from the inability of global- and regional-scale models to capture the complexities of foliar emissions. Emission rates and VOC composition vary by plant species, temperature, and, often, light availability (Guenther et al. 1995). Isoprene (C5H8) is a VOC emitted in large quantities from broadleaf vegetation at a rate dependent on light and temperature. At UMBS, many species emit isoprene, though light attenuation in the forest canopy can reduce understory emissions. Mixed forests like UMBS contain distinctly different understory and overstory species and are thus ideal environments for studying how complexities in foliar emissions influence local chemistry. To test the sensitivity of biogenic gas-phase chemistry to vertical heterogeneity in foliar emissions, further detail about the vertical canopy foliage distribution and speciation is needed. Here I propose to collect canopy data at UMBS to constrain model simulations. I will perform two simulations using a 1D forest canopy model in which the composition of BVOC emissions per model vertical layer is (1) constant (e.g., the current model configuration; Bryan et al. 2012) and (2) variable as a function of species-specific foliage structure. Unique distributions for the dominant species at UMBS—bigtooth aspen, red maple, quaking aspen, etc. (Garrity et al. 2012)—will be constructed using the observed lower and upper limits of foliage (i.e. trunk and crown heights), and used to determine the composition of emissions for each model layer individually. The outcome of the two simulations will demonstrate the degree to which complexities in foliage structure influence biogenic chemistry. Results from this study will be used to explain the model biases in biogenic chemistry, highlighting the importance of capturing the complexities in foliage structure in atmospheric models. An understanding of the sensitivity of biogenic chemistry to vertical heterogeneities in foliar emissions will guide improvements to emissions parameterizations in atmospheric chemistry models.