Lyciasalamandra

Veith, Dennis Rödder Stefan Lötters Mehmed Öz Sergé Bogaerts Karolos Eleftherakos Michael, 2011, A novel method to calculate climatic niche similarity among species with restricted ranges-the case of terrestrial Lycian salamanders, Organisms Diversity & Evolution (New York, N. Y.) 11 (5), pp. 409-423 : 420

Lyciasalamandra —a special case?

Holt et al. (2005) hypothesized that isolated populations living in habitats close to their environmental tolerance limits may be most prone to evolving new physiological properties that result in a shift of their fundamental niche. Since population densities in Lycian salamanders are often remarkably high (5,000 –10,000 specimens per hectare; Veith et al. 2001), it can be expected that perhaps a high number of specimens from these source populations migrates to sink habitats just outside their climatic tolerance. This should theoretically enhance the chance of selection of novel adaptations (e.g. Holt et al. 2005; Jakob et al. 2010). However, this is contraindicated by our analysis, which showed that the realised niches of most species are quite similar ( Tables 1, 2), although populations live from 0 to 1,000 m above sea level ( Veith et al. 2001). One explanation may be that environmental conditions during summer times are too harsh to allow colonisation in the absence of the retreat possibilities provided by carstic limestone cavities. These harsh conditions may cause a stabilising selection on the species’ climate niches by restricting their range.

It should be noted that the detected niche constancy across species does not necessarily indicate that most Lyciasalamandra species exhibit a climatic niche that is close to that of a common ancestor, i.e. one that has not changed much during evolution. Alternatively, all species may have changed their niches in the same direction due to a pronounced similarity of climatic parameters of their respective habitats. Unfortunately, there is no a priori reason to decide between these two hypotheses. However, genetic variability, which is thought to facilitate niche shift through local level adaptation ( Holt and Gomulkiewicz 2004; Holt et al. 2005; Jakob et al. 2010) is rather low in Lyciasalamandra ( Veith et al. 2008) . This may advocate for a hampered evolutionary response to changing climate, and hence argues for niche conservatism over time in Lyciasalamandra .

Methodological implications and limitations

Some aspects need consideration when applying our novel jackknife approach. First of all, although Maxent has been shown to perform well when the available number of species records is rather limited ( Elith et al. 2006; Hernandez et al. 2006; Wisz et al. 2008), a great variety of SDM algorithms are available, and applying different algorithms may yield varying results (e.g. Araújo and New 2007). Selection of the appropriate algorithm may depend on the data at hand, i.e. the number of records and whether absence data is available or not ( Guisan and Zimmermann 2000). However, in the case of Lycian salamanders, Maxent is apparently the best choice since it frequently outperforms other methods when the sample size is low.

Our approach most easily identifies species that differ most from all others analysed. However, depending on the niche breadths and niche positions of the species at hand, one could imagine a variety of possible scenarios ( Fig. 2 ), which can be assessed in niche space using multivariate statistical methods. Imagine that some species are found in colder climates and a couple of species found in hotter environments. Pooling all species but one will perhaps result in a quite large overall niche breadth and, depending on the relative positions of the species omitted in niche space, niche similarity patterns may vary. A second scenario may describe a species that occupies a niche space nested in the middle of another species’ niches. Depending on the algorithm applied, the overall niche envelope may not change too much when removing it, even though it can occupy a niche space distinct from all other species.

Fortunately, these difficulties can be circumvented via comparisons of probability surfaces derived from SDMs. The relative occurrence probabilities at a given combination of environmental variables depend on the density of species records in niche space ( Fig. 2 ). As a consequence, multiple observations of different species and/or within single species in similar or identical parts of the ecological space lead to locally higher probabilities. Therefore, even in the unlikely case that two species occupy exactly the same part of the ecological space, removal of one of them will cause a local decrease in the probability, and a comparison of both probability distributions via Schoener’ s D will still identify a deviance. In order to detect such patterns, interpretation of results should always include inspection of the relative positions of the species in niche space spanning along the most important axes and the overall available niche space (e.g. as shown in Fig. 3 for a two-dimensional climate space). Note that in more complex cases, principal component plots may be used (e.g. as shown in Broennimann et al. 2007; 2011). These additional analyses become most important when assessing niche conservatism in a phylogenetic context. Here, we may consider jackknifing entire clades instead of single species in order to obtain a stronger signal.

Further application of our jackknife approach to a variety of taxa will need to show if the identification of species with deviant climatic niches and their subsequent omission from a pooled data set will help to derive robust SDM or, more accurately, clade distribution models (CDM), even in the absence of large sets of distribution data. Facing the ongoing habitat deterioration due to climate change urgently requires reliable tools for incorporation of predictive distribution analyses for conservation planning, especially in geographically restricted and thus potentially vulnerable species.

Figures

Fig. 2 Conceptual differences among proposed niche overlap analyses. Imagine a set of three species with different preferences along an environmental gradient. Using SDM projections, it is possible to derive for each species a probability distribution across the gradient. Comparing two species as proposed by Warren et al. (2008), the overlap (grey area in the middle panel) of their respective probability distributions in geographic space is computed (pair-wise niche overlap). In our jackknife approach, a probability distribution derived from all species records is compared to a probability distribution derived from all minus one species. The 1−the resulting overlap value reflects the relative contribution of the omitted species to the entire probability distribution (grey area in the lower panel) and can be used as measure to rank all species when omitting them iteratively. Note that these specific indices are comparable only across each method but not among the different approaches