We describe the diversity and heterogeneity of urban CO₂ levels observed using the BErkeley Atmospheric CO₂ Observation Network, a distributed instrument of > 50 CO₂ sensors stationed every ~ 2 km across the San Francisco Bay Area. We demonstrate that relatively simple mathematical techniques, applied to these observations, can be used to detect the small changes in highway CO₂ emissions expected to result from upcoming fuel economy regulations, affirming the policy relevance of low-cost sensors.