Tree species composition in European pristine forests: Comparison of standdata to model predictions

Citation
Fw. Badeck et al., Tree species composition in European pristine forests: Comparison of standdata to model predictions, CLIM CHANGE, 51(3-4), 2001, pp. 307-347
Citations number
64
Categorie Soggetti
Environment/Ecology,"Earth Sciences
Journal title
CLIMATIC CHANGE
ISSN journal
01650009 → ACNP
Volume
51
Issue
3-4
Year of publication
2001
Pages
307 - 347
Database
ISI
SICI code
0165-0009(200112)51:3-4<307:TSCIEP>2.0.ZU;2-I
Abstract
The degree of general applicability across Europe currently achieved with s everal forest succession models is assessed, data needs and steps for furth er model development are identified and the role physiology based models ca n play in this process is evaluated. To this end, six forest succession mod els (DISCFORM, ForClim, FORSKA-M, GUESS, PICUS v1.2, SIERRA) are applied to simulate stand structure and species composition at 5 European pristine fo rest sites in different climatic regions. The models are initialized with s ite-specific soil information and driven with climate data from nearby weat her stations. Predicted species composition and stand structure are compare d to inventory data. Similarity and dissimilarity in the model results unde r current climatic conditions as well as the predicted responses to six cli mate change scenarios are discussed. All models produce good results in the prediction of the right tree functional types. In about half the cases, th e dominating species are predicted correctly under the current climate. Whe re deviations occur, they often represent a shift of the species spectrum t owards more drought tolerant species. Results for climate change scenarios indicate temperature driven changes in the alpine elevational vegetation be lts at humid sites and a high sensitivity of forest composition and biomass of boreal and temperate deciduous forests to changes in precipitation as m ediated by summer drought. Restricted generality of the models is found ins ofar as models originally developed for alpine conditions clearly perform b etter at alpine sites than at boreal sites, and vice versa. We conclude tha t both the models and the input data need to be improved before the models can be used for a robust evaluation of forest dynamics under climate change scenarios across Europe. Recommendations for model improvements, further m odel testing and the use of physiology based succession models are made.