Modelling and mapping the suitability of European forest formations at 1-km resolution

Proactive forest conservation planning requires spatially accurate information about the potential distribution of tree species. The most cost-efficient way to obtain this information is habitat suitability modelling i.e. predicting the potential distribution of biota as a function of environmental factors. Here, we used the bootstrap-aggregating machine-learning ensemble classifier Random Forest (RF) to derive a 1-km resolution European forest formation suitability map. The statistical model use as inputs more than 6,000 field data forest inventory plots and a large set of environmental variables. The field data plots were classified into different forest formations using the forest category classification scheme of the European Environmental Agency. The ten most dominant forest categories excluding plantations were chosen for the analysis. Model results have an overall accuracy of 76%. Between categories scores were unbalanced and Mesophitic deciduous forests were found to be the least correctly classified forest category. The model’s variable ranking scores are used to discuss relationship between forest category/environmental factors and to gain insight into the model’s limits and strengths for map applicability. The European forest suitability map is now available for further applications in forest conservation and climate change issues.

[1]  Ramón Díaz-Uriarte,et al.  GeneSrF and varSelRF: a web-based tool and R package for gene selection and classification using random forest , 2007, BMC Bioinformatics.

[2]  A. Townsend Peterson,et al.  Novel methods improve prediction of species' distributions from occurrence data , 2006 .

[3]  G. Moisen,et al.  Further advances in predicting species distributions , 2006 .

[4]  Atte Moilanen,et al.  Replacement cost: A practical measure of site value for cost-effective reserve planning , 2006 .

[5]  Brian J. McGill,et al.  Can niche-based distribution models outperform spatial interpolation? , 2007 .

[6]  M. Araújo,et al.  The importance of biotic interactions for modelling species distributions under climate change , 2007 .

[7]  N. Zimmermann,et al.  Predictive mapping of alpine grasslands in Switzerland: Species versus community approach , 1999 .

[8]  T. Simons,et al.  Spatial autocorrelation and autoregressive models in ecology , 2002 .

[9]  M. Austin,et al.  An Ecological Perspective on Biodiversity Investigations: Examples from Australian Eucalypt Forests , 1998 .

[10]  Desiré Paelinckx,et al.  A concept for biological valuation in the marine environment , 2007 .

[11]  G. Kenk,et al.  Management of transformation in central Europe , 2001 .

[12]  W. Oberhuber.  Influence of climate on radial growth of Pinus cembra within the alpine timberline ecotone. , 2004, Tree physiology.

[13]  A. Bolte,et al.  The north-eastern distribution range of European beech—a review , 2007 .

[14]  Andy Liaw,et al.  Classification and Regression by randomForest , 2007 .

[15]  Jingyun Fang,et al.  Climatic limits for the present distribution of beech (Fagus L.) species in the world , 2006 .

[16]  Antoine Guisan,et al.  Predictive habitat distribution models in ecology , 2000 .

[17]  T. M. Smith,et al.  A new model for the continuum concept , 1989 .

[18]  M. Araújo,et al.  Validation of species–climate impact models under climate change , 2005 .

[19]  Antoine Guisan,et al.  Spatial modelling of biodiversity at the community level , 2006 .

[20]  Wilfried Thuiller,et al.  Using niche‐based modelling to assess the impact of climate change on tree functional diversity in Europe , 2006 .

[21]  A. Lotter,et al.  Central European vegetation response to abrupt climate change at 8.2 ka , 2001 .

[22]  M. Austin.  Species distribution models and ecological theory: A critical assessment and some possible new approaches , 2007 .

[23]  David A. Seal,et al.  The Shuttle Radar Topography Mission , 2007 .

[24]  Trevor Hastie,et al.  Making better biogeographical predictions of species’ distributions , 2006 .

[25]  J. Gégout,et al.  The derivation of species response curves with Gaussian logistic regression is sensitive to sampling intensity and curve characteristics , 2006 .

[26]  Piermaria Corona,et al.  European forest types: Categories and types for sustainable forest management reporting and policy , 2006 .

[27]  J. Svenning,et al.  Limited filling of the potential range in European tree species , 2004 .

[28]  A. Deslauriers,et al.  Evidence of threshold temperatures for xylogenesis in conifers at high altitudes , 2007, Oecologia.

[29]  A. Prasad,et al.  Newer Classification and Regression Tree Techniques: Bagging and Random Forests for Ecological Prediction , 2006, Ecosystems.

[30]  Achim Zeileis,et al.  Conditional variable importance for random forests , 2008, BMC Bioinformatics.

[31]  Achim Zeileis,et al.  Bias in random forest variable importance measures: Illustrations, sources and a solution , 2007, BMC Bioinformatics.

[32]  Boris Schröder,et al.  Habitat models and their transfer for single and multi species groups: a case study of carabids in an alluvial forest , 2001 .

[33]  M. Willig,et al.  Alternative Configurations of Conservation Reserves for Paraguayan Bats: Considerations of Spatial Scale , 2002 .

[34]  William N. Venables,et al.  Modern Applied Statistics with S , 2010 .

[35]  Leo Breiman,et al.  Bagging Predictors , 1996, Machine Learning.

[36]  A. Hampe.  Bioclimate envelope models: what they detect and what they hide , 2004 .

[37]  Emmanuel S. Gritti,et al.  Quantifying components of risk for European woody species under climate change , 2006 .

[38]  Š. Šmelko,et al.  Relations between selected geomorphology features and tree species diversity of forest ecosystems and interpolation on a regional level , 2004, European Journal of Forest Research.

[39]  Leo Breiman,et al.  Random Forests , 2001, Machine Learning.

[40]  Ronald P. Neilson,et al.  A rule-based vegetation formation model for Canada , 1993 .

[41]  J. L. Parra,et al.  Very high resolution interpolated climate surfaces for global land areas , 2005 .

[42]  Christina Gloeckner,et al.  Modern Applied Statistics With S , 2003 .

[43]  M. Araújo,et al.  Consequences of spatial autocorrelation for niche‐based models , 2006 .

[44]  Flemming Skov,et al.  Could the tree diversity pattern in Europe be generated by postglacial dispersal limitation? , 2007, Ecology letters.

[45]  A. Lehmann,et al.  Improving generalized regression analysis for the spatial prediction of forest communities , 2006 .

[46]  A. Guisan,et al.  Equilibrium modeling of alpine plant distribution: how far can we go? , 2000 .

[47]  Carolin Strobl,et al.  The behaviour of random forest permutation-based variable importance measures under predictor correlation , 2010, BMC Bioinformatics.

[48]  Edzer J. Pebesma,et al.  Multivariable geostatistics in S: the gstat package , 2004, Comput. Geosci..

[49]  W. Cramer,et al.  A global biome model based on plant physiology and dominance, soil properties and climate , 1992 .

[50]  Felix Kienast,et al.  Comparing models for tree distributions : concept, structures, and behavior , 2000 .

[51]  C. Dormann.  Promising the future? Global change projections of species distributions , 2007 .