Bioclimate envelope models: what they detect and what they hide — response to Hampe (2004)
Article first published online: 23 JUL 2004
Global Ecology and Biogeography
Volume 13, Issue 5, pages 471–473, September 2004
How to Cite
Pearson, R. G. and Dawson, T. P. (2004), Bioclimate envelope models: what they detect and what they hide — response to Hampe (2004). Global Ecology and Biogeography, 13: 471–473. doi: 10.1111/j.1466-822X.2004.00112.x
- Issue published online: 4 AUG 2004
- Article first published online: 23 JUL 2004
We welcome the informative contribution from Hampe (2004) to what is an ongoing and important debate. Indeed, the importance of assessing the usefulness of the BEM approach has been exemplified by a recent high-profile application of the approach by Thomas et al. (2004). Our review (Pearson & Dawson, 2003) aimed to present a balanced, although inevitably not exhaustive, appraisal of bioclimate envelope models (BEM) and we agree with a number of the additional points raised by Hampe (2004). Many of the matters raised are currently being investigated by a number of researchers working to test existing methods and develop new, improved approaches. We wish to make a few additional comments here.
We reiterate that BEM do indeed ignore the constraints that limited dispersal may have on future distributions (Pearson & Dawson, 2003; Hampe, 2004). Investigating the potential for species to disperse at sufficient rates to keep pace with their moving bioclimate envelopes is a lively area of research (e.g. Higgins & Richardson, 1999; Collingham & Huntley, 2000; Higgins et al., 2003a) and one which builds on the foundations of BEM. The models used in these studies are highly stochastic and thus address the nondeterministic nature of past range shifts as noted by Hampe (2004). A notable conclusion from such work is that rare long-distance dispersal (LDD) events, which are extremely difficult to predict, are likely to be the only mechanism by which species will be able to keep pace with future changes in the distribution of suitable climate space (Clark et al., 1998; Cain et al., 2003). Mechanisms by which LDD may be achieved are diverse and include the transportation of seeds in updrafts (Nathan et al., 2002) and accidental transportation by animals (e.g. in nest material or attached to fur; Ridley, 1930; Higgins et al., 2003b). The role of humans as vectors of LDD, both accidentally and deliberately through, for example, horticulture, is also important (Mack et al., 2000). Human-mediated dispersal may be of critical conservation value in the future as human encroachment into natural systems increases and artificial barriers to dispersal (e.g. large urban and agricultural areas; Collingham & Huntley, 2000) become increasingly insurmountable by natural processes. The utility of BEM for identifying sites that are likely to experience climatic regimes in the future that are similar to conditions where a species is currently found may thus be important for the managed relocation of species threatened by climate change.
In our review, we stressed the importance of applying BEM at an appropriate spatial scale, arguing that climate is most appropriately correlated with species distributions at coarse geographical scales, and conjecturing that climate may be one of a number of factors affecting species’ distributions in a scale-dependent hierarchical manner. Hampe (2004) notes a number of examples demonstrating that in some cases biotic interactions may also operate at broad spatial scales. Such examples confirm our original statement that a hierarchical viewpoint may be imperfect and over-simplified (Pearson & Dawson, 2003), yet do not contradict the central message that climate impacts on species distributions are best addressed at coarse geographical scales. The broad-scale coincidence between climate and realized niches has been demonstrated, at least for some species, using not only correlative models (as discussed in our review), but also physiologically based process models (e.g. Neilson, 1995; Chuine & Beaubien, 2001). Such broad-scale ‘equilibrium’ is likely to have been enhanced by the relative climate stability of the past few thousand years, as noted by Hampe (2004), yet we should not expect this equilibrium to hold over the next century under predicted rapid climate change. We note, in addition, that the potential usefulness of the multiscale hierarchical framework that we proposed (Pearson & Dawson, 2003) has been demonstrated in a study which applied the approach to uncouple the impacts of climate and land-cover on plant distributions in Britain (Pearson et al., 2004).
Hampe (2004) refers to the importance of spatial autocorrelation in the context of BEM and highlights the overestimation of model fit statistics that can result from pseudoreplication. A number of approaches that address spatial autocorrelation within the BEM framework have been investigated, including the use of autocovariate terms in logistic regression and generalized additive models (Smith, 1994; Segurado & Araujo, 2004). However, we note that the implications of spatial autocorrelation for studies of large-scale systems that cannot be ‘replicated’ in a classical experimental context remain keenly debated (e.g. Hargrove & Pickering, 1992; Oksanen, 2001; Cottenie & De Meester, 2003). A thorough analysis of the role of spatial autocorrelation within niche-based modelling, that embraces spatial autocorrelation as an investigative tool (Diniz-Filho et al., 2003), quantifies bias in model validation statistics (Hampe, 2004) and incorporates statistical treatments to deal with spatial autocorrelation (e.g. Keitt et al., 2002), may prove insightful. However, the principal aim of characterizing species-climate relationships at an appropriate spatial scale should have priority over the issue of replication.
We note, in addition, that the complexity of the question as to climate change impacts on species’ distributions is exacerbated when considering the potential direct impacts of increased concentrations of atmospheric CO2 on species’ physiology (Houghton et al., 2001). Elevated levels of CO2 can be expected to alter species’ distributions, and the composition of ecological communities, by affecting how species respond to their physical and biological environment. For example, Catovsky & Bazzaz (1999) report varying influences of increased CO2 on seedling regeneration patterns for two species of birch on sites with differing soil moisture content in temperate forests. Similarly for grassland and savanna ecosystems, Bond et al. (2003) discuss the importance of CO2 concentrations for determining postburn recovery rates which affect postdisturbance succession and therefore the boundaries between fire tolerant and fire intolerant vegetation. Such effects are not incorporated within the BEM framework and our limited understanding of likely impacts (see for example the ambiguous evidence relating to well-studied cereal aphids; Newman, 2003) makes their inclusion within BEM predictions currently problematic. The potential impacts of elevated CO2 may be best addressed using dynamic mechanistic models of species’ response (e.g. Cramer et al., 2001; Sitch et al., 2003) yet the application of such approaches to individual species and specific regions is restricted by problems including substantial data requirements and the inclusion of poorly understood processes (Woodward & Beerling, 1997).
In conclusion, we agree with Hampe that future conservation strategies would benefit from models that incorporate more biological realism than BEM can so far provide. New less-deterministic modelling approaches that address the complex and nonlinear nature of the biosphere (Levin, 1999; Sole & Goodwin, 2000) are required. However, the capacity to build more realistic models is hindered by limitations in our current understanding of complex ecological systems and by the limited data available regarding such factors as genetic heterogeneity. Despite the limitations of BEM, a number of recent studies have demonstrated the success of the approach for predicting unknown species distributions (Raxworthy et al., 2003) and for simulating past distribution changes (Martinez-Meyer et al., 2004). We thus maintain that BEM can provide a useful first approach for understanding the potential impacts of climate change on species distributions, although the models must be carefully applied with due consideration for the limitations involved. Highlighting these limitations was the central aim of our review paper (Pearson & Dawson, 2003) and addressing these issues, and those raised by Hampe (2004), should form the basis for future research. BEM, and the discussion they generate, should at least hone our understanding of the critical questions as to climate change impacts on natural systems, and should help us to formulate new and more insightful questions for investigation.
We thank Miguel Araújo and Canran Liu for useful discussion.
- 2003) Long-distance dispersal. Ecology, 84, 1943–1944. , & (
- 1999) Elevated CO2 influences the response of two birch species to soil moisture: implications for forest community structure. Global Change Biology, 5, 507–518. & (
- 2001) Phenology is a major determinant of tree species range. Ecology Letters, 4, 500–510. & (
- 1998) Reid's paradox of rapid plant migration: dispersal theory and interpretation of paleoecological records. Bioscience, 48, 13–24. , , , , , , , , , , , & (
- 2000) Impacts of habitat fragmentation and patch size upon migration rates. Ecological Applications, 10, 131–144. & (
- 2003) Comment on Oksanen (2001): reconciling Oksanen (2001) and Hurlbert (1984). Oikos, 100, 394–396. & (
- 2001) Global response of terrestrial ecosystem structure and function to CO2 and climate change: results from six dynamic global vegetation models. Global Change Biology, 7, 357–373. , , , , , , , , , , , , , , , & (
- 2003) Spatial autocorrelation and red herrings in geographical ecology. Global Ecology and Biogeography, 12, 53–64. , & (
- 2004) Bioclimate envelope models: what they detect and what they hide. Global Ecology and Biogeography, DOI: 10.1111/j.1466-822x.2004.00090.x. (
- 1992) Pseudoreplication: a sine qua non for regional ecology. Landscape Ecology, 6, 251–258. & (
- 2003a) Estimating plant migration rates under habitat loss and fragmentation. Oikos, 101, 354–366. , & (
- 2003b) Are long-distance dispersal events in plants usually caused by nonstandard means of dispersal. Ecology, 84, 1945–1956. , & (
- 1999) Predicting plant migration rates in a changing world: the role of long-distance dispersal. The American Naturalist, 153, 464–475. & (
- 2001) Climate Change 2001: the scientific basis. Contributions of Working Group I to the Third Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, UK. , , , , , , & (
- 2002) Accounting for spatial pattern when modelling organism–environment interactions. Ecography, 25, 616–625. , , & (
- 1999) Fragile dominion. Princeton University Press, Princeton. (
- 2000) Biotic invasions: causes, epidemiology, global consequences, and control. Ecological Applications, 10, 689–710. , , , , & (
- 2004) Ecological niches as stable distributional constraints on mammal species, with implications for Pleistocene extinctions and climate change projections for biodiversity. Global Ecology and Biogeography, 13, 305–314. , & (
- 2002) Mechanisms of long-distance dispersal of seeds by wind. Nature, 418, 409–413. , , , , , , & (
- 1995) A model for predicting continental-scale vegetation distribution and water balance. Ecological Monographs, 5, 362–385. (
- 2003) Climate change and cereal aphids: the relative effects of increasing CO2 and temperature on aphic population dynamics. Global Change Biology, 10, 5–15. (
- 2001) Logic of experiments in ecology: is pseudoreplication a pseudoissue? Oikos, 94, 27–38. (
- 2003) Predicting the impacts of climate change on the distribution of species: are bioclimate envelope models useful? Global Ecology and Biogeography, 12, 361–371. & (
- 2004) Modelling species distributions in Britain: a hierarchical integration of climate and land-cover data. Ecography, in press. , & (
- 2003) Predicting distributions of known and unknown reptile species in Madagascar. Nature, 426, 837–841. , , , , , & (
- 1930) The dispersal of plants throughout the world. Reeve, Ashford, UK. (
- 2004) An evaluation of methods for modelling species distributions. Journal of Biogeography, in press. & (
- 2003) Evaluation of ecosystem dynamics, plant geography and terrestrial carbon cycling in the LPJ dynamic global vegetation model. Global Change Biology, 9, 161–185. , , , , , , , , , , & (
- 1994) Autocorrelation in logistic regression modelling of species’ distributions. Global Ecology and Biogeography Letters, 4, 47–61. (
- 2000) Signs of life: how complexity pervades biology. 1st edn. Basic Books, New York. & (
- 2004) Extinction risk from climate change. Nature, 427, 145–148. , , , , , , , , , , , , , , , , , & (
- 1997) The dynamics of vegetation change: health warnings for equilibrium ‘dodo’ models. Global Ecology and Biogeography Letters, 6, 413–418. & (