||A multiple-box model was designed to determine how anthropogenic, biological, and meteorological processes combine to produce diel cycles of carbon dioxide (CO2) concentrations within the urban Salt Lake Valley (uSLV). The model was forced by an anthropogenic CO2 emissions inventory, observed winds, sounding-derived mixing depths, and net biological flux estimates based on temperature, solar radiation, day of year, and ecosystem type. The model was validated using hourly CO2 data from a network of sensors around the uSLV for years 2005-2009. The model accounted for 53% of the observations on an hourly basis and accounted for 90-94% of the mean diel cycle of the observations depending on the season. Salt Lake Valley suffers from prolonged temperature inversions during the winter that trap pollutants and gases at the surface. The CO2 network (co2.utah.edu) was compared with the CO2 multiple-box model to determine whether the model could capture the main drivers of CO2 variability during the Persistent Cold Air Pool Study (PCAPS). Time-height analyses were performed to facilitate investigation and explanation of CO2 variability during PCAPS intensive observation periods (IOPs). The analyzed data included atmospheric soundings, CO2 network data, quasivertical CO2 profiles collected ascending by foot or vehicle, and laser-ceiliometer data.