Aim Knowledge about suitable ecological conditions for tree species remains largely incomplete. We aim to quantify and compare suitable environmental conditions for some of the most common tree species in Europe using a modelling approach and a large array of predictors that describe the different dimensions of the environmental niche. Location France. Methods Using a large database of 46 768 plots distributed across France (550 000 km2), we compared the performance of 61 physiologically relevant predictors that describe temperature, soil nutrient supply, soil water availability and aeration at fine spatial resolution to model the distribution of 32 tree species. To allow some comparisons between species along the same gradients, the most efficient predictors were determined and used to characterize the species requirements. Results Mean annual temperature, soil water deficit, soil pH, C/N and temporary and permanent waterlogging gave the best performances. The efficiency of species distribution models based on these six variables was comparable to models calibrated with 61 variables. For the first time, those models provide a unique quantitative framework of homogeneous information concerning the requirements for a large number of tree species. Conclusions Our results highlight the importance of soil descriptors to shape plant distribution and allow each species to delineate areas inside the climate envelope with favourable nutritional and moisture conditions. These results emphasize a gap between existing studies on species distribution, mainly based on large-scale climatic predictors, and empirical knowledge or existing practices where soil is recognized as a main driver.