models can provide an estimation of the environmental impact of road traffic.
However, decision makers need to be confident in these assessments in order to
implement reduction strategies. The key issue at stake, especially in dense urban
zone, is to describe accurately the traffic dynamic and particularly the congestion
periods. The proper definition of the link mean speed is the ratio of total travelled
distance and total time spent during a given period. This spatial speed description
can be easily obtained from a dynamic traffic simulation. However, in operational
conditions, it is often deduced from observed speeds on loop detectors or speed
limit, which inevitably implies a bias on related emissions to be quantified.
For this study we focused on vehicle trajectories in the morning peak for a
typical weekday in a 3km2 urban network. These detailed traffic data represent
a considerable amount of data, but allows us to operate any spatiotemporal
aggregation used for emission assessment sake. The emission calculations were
made at link level each 6 minutes, combining the various traffic indicators and
either the Copert emission factors database or Phem model. The related fuel
consumption and NOx emissions are compared.