“…First, these links were built up over a long period of time, and they have not only shaped a considerable diversity of geographical systems, but also constructed very powerful representations that still have a strong influence on debates on the efficiency of rail transport. Next, we have shown that the interaction between railways and demographic change is extremely dynamic, as the two have evolved together over decades, with periods of rapid readjustment and differences in both the chronology and nature of changes between one country or one region and another.Finally, knowledge about the future is vital for the rail sector, as it takes approximately 10 years to introduce new rolling stock and at least 20 to design and build a high speed line.In spite of its faults and weaknesses, we therefore still need forecasting (De Palma et al, 2010).However, it must be improved in two respects:first, greater transparency is required with regard to the construction and handling of the fundamental indicators and variables; andsecond, experts need to make greater use of geographic information systems in models that analyse the performance, vulnerability and resilience of the railway system (Tschopp et al, 2003) in order to promote a more geographicallybased approach.…”