Traffic is more turbulent around motorway ramps due to route choice-related lane changes and anticipatory or cooperative manoeuvres. These manoeuvres result in changes in speed and headways and have a negative influence on traffic safety and capacity. However, the distance upstream where turbulence starts and the distance downstream where it dissolves are yet unknown. In this paper, we propose a new method for detecting the start and end distances of turbulence. This method relies on the analysis of large quantities of empirical loop detector data from multiple on-ramps and off-ramps at different sites. By comparing the traffic operations near ramps to those on a regular motorway section, the length of the turbulence influence area can be estimated. The scope of the research is limited to three-lane motorways in the Netherlands, and shows that the distribution of traffic over the motorway lanes is a useful indicator for turbulence.