•A novel road description using stochastic models in a hierarchical structure.•A method for how to generate data possible to simulate in a dynamic model.•Hilliness and stop intensity are the most prominent properties in the chosen range.•The statistical method is compared with real log data and the results match well.•Ideas discussed on how to make realistic predictions of CO2 emissions from vehicles. We propose a novel statistical description of the physical properties of road transport operations by using stochastic models arranged in a hierarchical structure. The description includes speed signs, stops, speed bumps, curvature, topography, road roughness and ground type, with a road type introduced at the top of the hierarchy to group characteristics that are often connected. Methods are described how to generate data on a form (the operating cycle format) that can be used in dynamic simulations to estimate energy usage and CO2 emissions. To showcase the behaviour of the description, two examples are presented using a modular vehicle model for a heavy-duty truck: a sensitivity study on impacts from changes in the environment, and a comparison study on a real goods transport operation with respect to energy usage. It is found that the stop intensity and topography amplitude have the greatest impact in the sensitivity study (8.3% and 9.5% respectively), and the comparison study implies that the statistical description is capable of capturing properties of the road that are significant for vehicular energy usage. Moreover, it is discussed how the statistical description can be used in a vehicle design process, and how the mean CO2 emissions and its variation can be estimated for a vehicle specification.