The spatial and temporal patterns of broad front bird migration are governed by the geospatial distributions of landmasses, mountain ranges, and weather conditions. These distributions interact with the birds' innate program during migration and are critical to successful migration. Hence, favourable environmental conditions for migration consequently lead to spatio-temporal concentrations, and the evolution of specific migratory flyways. Based on the aerodynamic properties of the ten most abundant nocturnal long-distance migrants, we developed a computational framework to simulate millions of individual trajectories across Europe using an agent-based simulation approach. We simulated a three-week period of autumn migration with a temporal resolution of 30 s. Departure conditions were derived from bird densities observed within the first two hours after sunset, which were extracted from the weather radar network. Individual itineraries were strongly influenced by the behavioural reactions to the environment such as the wind flow, coastlines and mountain ranges. Wind speed and direction are among the key factors that shape migration patterns such as reverse movements observed within some nights. Accumulation of migration was triggered by the combined effect of geographic barriers (coastlines, mountain ranges) and wind. The overall result of the simulation corresponds well with the large-scale pattern of bird migration intensities measured across the study area, for instance the high migration intensity observed between the Atlantic coast and the Pyrenees. Our model framework conjoins all important domains, such as the birds’ preconditions and behavioural scope, as well as the spatio-temporal dynamics of the environment. However, for the time being we cannot decide whether local discrepancies between model and data are due to environmental effects that are not yet captured by our simulation (i.e. effect of rain on migration intensity), or caused by differences in the sensitivity of the local weather radar systems. Due to the limited time period and lack of more accurate data for validation, our findings are preliminary. Further progress in data quality and hopefully better access of bird profiles from the continent-wide weather radar network will allow to improve model performance and implement a bird migration forecasting similar to weather forecasts. Apart from assessing conflicts between human activities and bird movements, this would greatly enhance our understanding of biomass flow on a large scale.