Are niche-based species distribution models transferable in space?

Aim  To assess the geographical transferability of niche-based species distribution models fitted with two modelling techniques. Location  Two distinct geographical study areas in Switzerland and Austria, in the subalpine and alpine belts. Methods  Generalized linear and generalized additive models (GLM and GAM) with a binomial probability distribution and a logit link were fitted for 54 plant species, based on topoclimatic predictor variables. These models were then evaluated quantitatively and used for spatially explicit predictions within (internal evaluation and prediction) and between (external evaluation and prediction) the two regions. Comparisons of evaluations and spatial predictions between regions and models were conducted in order to test if species and methods meet the criteria of full transferability. By full transferability, we mean that: (1) the internal evaluation of models fitted in region A and B must be similar; (2) a model fitted in region A must at least retain a comparable external evaluation when projected into region B, and vice-versa; and (3) internal and external spatial predictions have to match within both regions. Results  The measures of model fit are, on average, 24% higher for GAMs than for GLMs in both regions. However, the differences between internal and external evaluations (AUC coefficient) are also higher for GAMs than for GLMs (a difference of 30% for models fitted in Switzerland and 54% for models fitted in Austria). Transferability, as measured with the AUC evaluation, fails for 68% of the species in Switzerland and 55% in Austria for GLMs (respectively for 67% and 53% of the species for GAMs). For both GAMs and GLMs, the agreement between internal and external predictions is rather weak on average (Kulczynski's coefficient in the range 0.3–0.4), but varies widely among individual species. The dominant pattern is an asymmetrical transferability between the two study regions (a mean decrease of 20% for the AUC coefficient when the models are transferred from Switzerland and 13% when they are transferred from Austria). Main conclusions  The large inter-specific variability observed among the 54 study species underlines the need to consider more than a few species to test properly the transferability of species distribution models. The pronounced asymmetry in transferability between the two study regions may be due to peculiarities of these regions, such as differences in the ranges of environmental predictors or the varied impact of land-use history, or to species-specific reasons like differential phenotypic plasticity, existence of ecotypes or varied dependence on biotic interactions that are not properly incorporated into niche-based models. The lower variation between internal and external evaluation of GLMs compared to GAMs further suggests that overfitting may reduce transferability. Overall, a limited geographical transferability calls for caution when projecting niche-based models for assessing the fate of species in future environments.

[1]  H. Pulliam.  On the relationship between niche and distribution , 2000 .

[2]  J. P. Grime,et al.  Ellenberg numbers revisited , 1993 .

[3]  J. C. van Houwelingen,et al.  Predictive value of statistical models , 1990 .

[4]  F. Kienast,et al.  Predicting the potential distribution of plant species in an alpine environment , 1998 .

[5]  Ronald I. Miller.  Predicting rare plant distribution patterns in the southern Appalachians of the south-eastern U.S.A. , 1986 .

[6]  R. Holt,et al.  Allee Effects, Invasion Pinning, and Species’ Borders , 2001, The American Naturalist.

[7]  Michel Loreau,et al.  Community Patterns in Source‐Sink Metacommunities , 2003, The American Naturalist.

[8]  G. C. Stevens,et al.  Spatial Variation in Abundance , 1995 .

[9]  Lalit Kumar,et al.  Modelling Topographic Variation in Solar Radiation in a GIS Environment , 1997, Int. J. Geogr. Inf. Sci..

[10]  Ari Huusko,et al.  Transferability of habitat suitability criteria of juvenile Atlantic salmon (Salmo salar) , 2002 .

[11]  M. Araújo,et al.  Climate warming and the decline of amphibians and reptiles in Europe , 2006 .

[12]  S. Lavorel,et al.  Biodiversity conservation: Uncertainty in predictions of extinction risk , 2004, Nature.

[13]  Ken D. Bovee,et al.  Application and testing of a procedure to evaluate transferability of habitat suitability criteria , 1993 .

[14]  R. Leemans,et al.  Assessing effects of forecasted climate change on the diversity and distribution of European higher plants for 2050 , 2002 .

[15]  D. Rogers,et al.  The effects of species’ range sizes on the accuracy of distribution models: ecological phenomenon or statistical artefact? , 2004 .

[16]  Joseph M. Culp,et al.  Transferability of Habitat Suitability Curves for a Benthic Minnow, Rhinichthys cataractae , 1997 .

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

[18]  Boris Schröder,et al.  Are habitat models transferable in space and time , 2000 .

[19]  T. Dawson,et al.  SPECIES: A Spatial Evaluation of Climate Impact on the Envelope of Species , 2002 .

[20]  W. Thuiller,et al.  Predicting species distribution: offering more than simple habitat models. , 2005, Ecology letters.

[21]  J. Michael Scott,et al.  Predicting Species Occurrences: Issues of Accuracy and Scale , 2002 .

[22]  Mike P. Austin,et al.  Continuum Concept, Ordination Methods, and Niche Theory , 1985 .

[23]  P. McCullagh,et al.  Generalized Linear Models, 2nd Edn. , 1990 .

[24]  M. Araújo,et al.  Reducing uncertainty in projections of extinction risk from climate change , 2005 .

[25]  David R. B. Stockwell,et al.  Future projections for Mexican faunas under global climate change scenarios , 2002, Nature.

[26]  A. Troumbis,et al.  Local adaptation enhances performance of common plant species , 2001 .

[27]  Gretchen G. Moisen,et al.  Comparing five modelling techniques for predicting forest characteristics , 2002 .

[28]  John Bell,et al.  A review of methods for the assessment of prediction errors in conservation presence/absence models , 1997, Environmental Conservation.

[29]  H. Akaike.  Maximum likelihood identification of Gaussian autoregressive moving average models , 1973 .

[30]  Brian Huntley,et al.  Plant species' response to climate change: implications for the conservation of European birds , 2008 .

[31]  Mike P. Austin.  Searching for a Model for Use in Vegetation Analysis , 1980 .

[32]  M. Austin.  Spatial prediction of species distribution: an interface between ecological theory and statistical modelling , 2002 .

[33]  C. Jeffree,et al.  Temperature and the Biogeographical Distributions of Species , 1994 .

[34]  M. Sykes,et al.  Climate change threats to plant diversity in Europe. , 2005, Proceedings of the National Academy of Sciences of the United States of America.

[35]  M. Hill,et al.  Extending Ellenberg's indicator values to a new area: an algorithmic approach , 2000 .

[36]  M. Groom,et al.  Allee Effects Limit Population Viability of an Annual Plant , 1998, The American Naturalist.

[37]  P. C. Dias,et al.  Sources and sinks in population biology. , 1996, Trends in ecology & evolution.

[38]  R. Brandl,et al.  Geographic variability of ecological niches of plant species: are competition and stress relevant? , 2002 .

[39]  Stefan Dullinger,et al.  Habitat distribution models, spatial autocorrelation, functional traits and dispersal capacity of alpine plant species , 2004 .

[40]  M. Boyce,et al.  Evaluating resource selection functions , 2002 .

[41]  Michael Kleyer,et al.  Validation of plant functional types across two contrasting landscapes , 2002 .

[42]  Jörg Schulla,et al.  Hydrologische Modellierung von Flussgebieten zur Abschätzung der Folgen von Klimaänderungen , 1997 .

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

[44]  J. R. Landis,et al.  The measurement of observer agreement for categorical data. , 1977, Biometrics.

[45]  A. Peterson,et al.  Geographic analysis of conservation priority: endemic birds and mammals in Veracruz, Mexico , 2000 .

[46]  I. C. Prentice,et al.  Boreal forest futures: Modelling the controls on tree species range limits and transient responses to climate change , 1995 .

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

[48]  A. Fielding,et al.  Testing the Generality of Bird‐Habitat Models , 1995 .

[49]  M. Zobel,et al.  The relative of species pools in determining plant species richness: an alternative explanation of species coexistence? , 1997, Trends in ecology & evolution.

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

[51]  A. Peterson,et al.  Predicting the potential invasive distributions of four alien plant species in North America , 2003, Weed Science.

[52]  William J. Mitsch,et al.  A spatial habitat model for the marsh-breeding red-winged blackbird (Agelaius phoeniceus L.) in coastal Lake Erie wetlands , 1997 .

[53]  Tomas Vitvar,et al.  A comparative study in modelling runoff and its components in two mountainous catchments , 2003 .

[54]  Michel Loreau,et al.  Coexistence in Metacommunities: The Regional Similarity Hypothesis , 2002, The American Naturalist.

[55]  R. Tibshirani,et al.  Generalized additive models for medical research , 1986, Statistical methods in medical research.

[56]  Trevor Hastie,et al.  Generalized linear and generalized additive models in studies of species distributions: setting the scene , 2002 .

[57]  P. D. Körner.  Alpine Plant Life , 1999, Springer Berlin Heidelberg.

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

[59]  J A Swets,et al.  Measuring the accuracy of diagnostic systems. , 1988, Science.

[60]  Michael Gottfried,et al.  Prediction of Vegetation Patterns at the Limits of Plant Life: A New View of the Alpine-Nival Ecotone , 1998 .

[61]  H. Birks,et al.  The altitudinal gradient of vascular plant richness in Aurland, western Norway , 1999 .

[62]  A. Lehmann,et al.  Regression models for spatial prediction: their role for biodiversity and conservation , 2002, Biodiversity & Conservation.

[63]  M. Freeman,et al.  Transferability of Habitat Suitability Criteria for Fishes in Warmwater Streams , 1997 .

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

[65]  A. Guisan,et al.  An improved approach for predicting the distribution of rare and endangered species from occurrence and pseudo-absence data , 2004 .

[66]  A. Guisan,et al.  Assessing alpine plant vulnerability to climate change: a modeling perspective , 2000 .

[67]  W. Thuiller.  Patterns and uncertainties of species' range shifts under climate change , 2004 .

[68]  T. Dawson,et al.  Model‐based uncertainty in species range prediction , 2006 .

[69]  P. Carey,et al.  The Use of GIS to Identify Sites that will become Suitable for a Rare Orchid, Himantoglossum hircinum L., in a Future Changed Climate , 1994 .

[70]  M. Austin,et al.  New approaches to direct gradient analysis using environmental scalars and statistical curve-fitting procedures , 1984 .

[71]  A. Peterson,et al.  Preliminary distributional analysis of US endangered bird species , 2000, Biodiversity & Conservation.

[72]  A. Baltensweiler,et al.  Spatially distributed hydrotope-based modelling of evapotranspiration and runoff in mountainous basins , 1999 .

[73]  E. Weber,et al.  Current and Potential Ranges of Three Exotic Goldenrods (Solidago) in Europe , 2001 .

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

[75]  J. Franklin.  Predictive vegetation mapping: geographic modelling of biospatial patterns in relation to environmental gradients , 1995 .

[76]  Stefan Dullinger,et al.  A regional impact assessment of climate and land‐use change on alpine vegetation , 2003 .