A one-dimensional Lagrangian model for atmospheric transport and photochemistry has been developed and used to interpret measurements made at Pellston, Michigan, during the summer of 1998. The model represents a moving vertical column of air with vertical resolution of 25 m near the ground. Calculations have been performed for a series of trajectories, with representation of emissions, vertical mixing, and photochemistry for a 3-day period ending with the arrival of the air column at Pellston. Results have been used to identify causes of the observed decrease in isoprene at night, to investigate causes of high nighttime OH. Significant OH can be generated at night by terpenes if it is assumed that some fast-reacting monoterpenes are emitted at rates comparable to inventory emissions for terpenes. However, this nighttime OH is confined to a shallow surface layer (0--25 m) and has little impact on nighttime chemistry. The observed decrease in isoprene at night can be reproduced in models with low OH, and is attributed primarily to vertical dilution. There is also evidence that transport from Lake Michigan contributes to low nighttime isoprene at Pellston. Model results compare well with measured isoprene, NOx, and with isoprene vertical profiles. Significant model-measurement discrepancies are found for OH, HO2, methylvinylketone, and formaldehyde.