Empirical analysis of freeway flow-density relationships

Many researchers have reported on the occurence of gaps in freeway speed-density and flow-density data and have suggested that discontinuous functions are necessary to properly describe “observed” traffic behavior. This paper investigates the flow-occupancy (spot-density) relationship using an extensive data set collected on the Queen Elizabeth Way in Ontario. Daily time-traced plots of 5-minute average flow rates versus occupancy were analyzed. Results indicate that there is another interpretation of gaps in data, which does not imply a discontinuous function, but rather, an inverted V shape (continuous, but not continuously differentiable). Three conclusions were reached: (a) it is essential to provide details of data collection locations if one is to know whether a particular pattern in resulting data represents a “true” relationship, or just the specifics of a particular place; (b) there are clear advantages to examining daily time traces of traffic behavior, rather than relying on scatter diagrams of numerous days of accumulated data; and, (c) previously documented arguments for a discontinuous flow-occupancy relationship do not seem convincing, because knowledge of daily operations at a particular location could easily explain the occurence of gaps in the data.