Tree neighbourhood matters - tree species composition drives diversity-productivity patterns in a near-natural beech forest

European beech forest with a variable admixture is one of the most important forest types in Central Europe. Growing evidence has demonstrated the positive effect of increased biodiversity on vital forest ecosystem functions and services such as productivity and nutrient cycling. Both complementarity in resource use and species identity are known to influence tree productivity but they have received relatively little attention in observational studies. Using a large dataset of repeat inventory trees in a near-natural deciduous forest in Central Germany we test whether tree diversity enhances tree productivity at the tree and the stand level, whilst accounting for tree size, tree vitality, local topography and the potentially confounding effects of spatial autocorrelation and negative growth estimates. Beech and hornbeam individual tree growth was sensitive to their neighbourhood diversity and composition whilst ash trees were only sensitive to the neighbourhood tree density. Neighbourhood complementarity effects were driven by differences in species' competitive strengths, whilst at the stand level productivity gains were primarily attributable to the density of ash and diversity effects were less prominent. We conclude that small-scale admixture with patches of different species promotes tree growth in European beech forest; congruent with current management plans for beech and hardwood forests.