Representation of species mixing in forest growth models. A review and perspective

Mixed-species forests can sometimes fulfil forest functions and services better than monocultures and are therefore receiving growing attention in forest science and practice. The productivity of mixtures and the interactions between species are influenced by the availability of different resources and climatic conditions, all of which change spatially and temporally. Models are a valuable tool for understanding and predicting how these interacting factors will influence the growth and other functions and services in mixed-species forests. However, concepts, models and theory are still mainly based on monocultures, which have dominated forest science since its beginnings. Recent empirical works report strong effects of mixing tree species on the environmental conditions within stands (e.g., vertical light profile, rooting space, humus layer), their functioning (e.g., photosynthetic rate, light use, growth), and tree and stand structure (e.g., crown and stem shape, shoot and root morphology). Process-based organ- or tree-level approaches reveal changed resource supply, flows and metabolic rates in high spatial and temporal resolution. Stand-level approaches usually provide predictions with a lower spatial and temporal resolution but can be more accurate in the longer-term. Many studies stress considerable effects of mixing tree species on growth dynamics, yield, resilience and stability, but these effects are still under-represented in existing models. We reviewed 54 forest growth models to show how they incorporate the variability of individual tree or species characteristics that occur in forests and how this variability influences the development and characteristics of the whole stand. While some organ- or tree-level models inherently integrate evident processes, stand-level models do not usually explicitly consider species interactions. Nevertheless, many processes that occur in mixtures can also be modelled at the stand level and are often included in stand-level monospecific models. Stand-level models are likely to offer a simpler alternative model structure for mixtures. We stress how both model approaches can be refined, benefit from mutual adjustment, and can be substantiated by further empirical research into mixed-species forests.