A Simplified Forest Model to Study Species Composition Along Climate Gradients

Forest models based on the gap dynamics hypothesis ( g ap models ) have gained an important role in forest ecology and have grown rather complex in the last 20 yr. They have been applied extensively to study the impacts of climatic change on ecosystems although they originally were not built for this purpose. The objectives of this study were (1) to develop a new forest gap model, ForClim, that includes only a minimum number of ecological assumptions but robust parameterizations of the effects of climate on plant population dynamics; (2) to test the realism of ForClim as compared to its predecessor model, FORECE; and (3) to examine the behavior of FORECE and ForClim systematically along climate gradients in Europe. ForClim is composed of three modular submodels: ForClim—P for plant population dynamics, ForClim—S for soil carbon/nitrogen turnover, and ForClim—E for providing reliable parameterizations of the abiotic environment. For the core model, ForClim—P, it was found that only four factors are sufficient to model tree growth, another four factors are required to model tree establishment, and only two factors are required to model tree mortality. The behavior of ForClim was tested at a large number of sites in the European Alps. The model yields tree species compositions that conform to field data and are very similar to those of the predecessor model. Based on this evaluation alone, it would not be possible to favor one of the models over the other. The behavior of both models then was examined systematically in a parameter space spanned by the annual mean temperature and the annual precipitation sum. From this exercise it became evident that both the pattern of aboveground biomass and the realized niches of the dominating tree species are simulated realistically by ForClim. Extremely steep gradients are characteristic of FORECE, and many ecotones are simulated to occur in the wrong places in FORECE. Thus, some of the current forest gap models can be simplified without reducing the realism of their behavior, and models other than FORECE should be scrutinized in this respect as well. The present study also suggests that the evaluation of model behavior at scattered sites is insufficient to show their validity for simulating forest dynamics along climate gradients. Further rigorous model comparisons and validation studies are required to increase the reliability of this promising class of models.