The BIOCLIM algorithm provides a straightforward method to estimate the effects of climate change on the distribution of species. Estimating the core ranges of species from 5% and 95% quantiles of bioclimatic variables, the algorithm remains widely used even when more sophisticated methods of species distribution modelling have become popular. Where sufficient representative observations are available, I expect that BIOCLIM correctly identifies locations that would not be suitable in a future climate. To accommodate climate change investigations based on BIOCLIM for 48,129 tree species (a substantial subset of known tree species), I developed the TreeGOER (Tree Globally Observed Environmental Ranges) database, providing information on environmental ranges for 38 bioclimatic, 8 soil and 3 topographic variables. The database can be accessed from: https://doi.org/10.5281/zenodo.7922928. Statistics that include 5% and 95% quantiles were estimated for a cleaned and taxonomically standardized occurrence data set with different methods of outlier detection, with estimates for roughly 45% of species being based on 20 or more observation records. Inferred core bioclimatic ranges of species along global temperature and moisture index gradients and across continents follow the known global distribution of tree diversity such as its highest levels in moist tropical forests and the 'odd man out' pattern of lower levels in Africa. To demonstrate how global analyses for large numbers of tree species can easily be done in R with TreeGOER, here I present two case studies. The first case study investigated latitudinal trends of tree vulnerability and compared these with previous results obtained for urban trees. The second case study focused on tropical areas, compared trends in different longitudinal zones and investigated patterns for the moisture index. TreeGOER is expected to benefit researchers conducting biogeographical and climate change research for a wide range of tree species at a variety of spatial and temporal scales.