Global change may induce shifts in plant community distributions at multiple spatial scales. At the ecosystem scale, such shifts may result in movement of ecotones or vegetation boundaries. Most indicators for ecosystem change require timeseries data, but here a new method is proposed enabling inference of vegetation boundary movement from one ‘snapshot’ (e.g. an aerial photograph or satellite image) in time. The method compares the average spatial position of frontrunners of both communities along the vegetation boundary. Mathematical analyses and simulation modeling show that the average frontrunner position of retreating communities is always farther away from a so-called optimal vegetation boundary as compared to that of the expanding community. This feature does not depend on assumptions about plant dispersal or competition characteristics. The method is tested with snapshot data of a northern hardwood-boreal forest mountain ecotone in Vermont, a forest-mire ecotone in New Zealand and a subalpine treeline-tundra ecotone in Montana. The direction of vegetation boundary movement is accurately predicted for these case studies, but we also discuss potential caveats. With the availability of snapshot data rapidly increasing, the method may provide an easy tool to assess vegetation boundary movement and hence ecosystem responses to changing environmental conditions.