Environmental Engineering Reference
In-Depth Information
because we do not know
why
a species is absent in some sites, we are in the dark
about its niche. The discussion of what a niche is, and what we are modelling, has
sparked several interesting and not always compatible publications (e.g. Kearney
2006; Sober
´
n 2007). Hence, Ara
´
jo and Guisan (2006) have named the “clarifica-
tion of the niche concept” the first of five challenges for SDMs. While we cannot
resolve this issue here, it is important to realize that the “niche” based on the
correlation between geographic distributions and environmental conditions is quite
a bit more vague than the niche discussed in evolutionary ecology, where resources
and other causal drivers are envisaged (see, e.g. Losos 2008).
More to the point in this context is the challenge to quantify how much of the
fundamental niche is actually covered by the realized niche as extracted from
SDMs. If, on one extreme, the realized niche is pretty much also the fundamental
niche (i.e. there are no biotic interactions alike to constrain the distribution at the
scale we are analyzing), then we can merrily predict future distributions of this
species (e.g. under climate or land-use change). At worst, we are overestimating the
future, “potential” distribution (if in the future biotic interactions may become
limiting or if species do not reach the sites). At the other extreme, if the fundamental
niche is considerably wider than what we model, any projection can be fundamen-
tally flawed (Dormann et al. 2010). I am not aware of any study assessing the
overlap of realized and fundamental niche for geographic distributions (see also
Nogu´s-Bravo 2009). It could require transplant experiments into areas beyond the
current range and the manipulation of biotic interactions there. The few studies
going into this direction point at a large discrepancy between fundamental and
realized niche. Battisti et al. (2006), for example report on a range shift after a
particularly warm summer, which was not reverted afterwards, indicating that it
was dispersal limitation that prevented a filling of the niche. Similarly, several
studies point at the importance of dispersal limitation (Nekola 1999; Ozinga et al.
2005; Samu et al. 1999; Svenning and Skov 2004), leading to both a bias in the
modeled environment-occurrence relationship as well as the width of the niche
itself. There is, as yet, no standard way to wed SDMs and dispersal (for attempts
see, Johst et al. 2002; King and With 2002; Lavorel et al. 2000; Lischke et al. 2006;
Midgley et al. 2006; Schurr et al. 2007; Thuiller 2004).
Niche Evolution
Another important and fast developing field related to species distribution model-
ling is the study of niche evolution. I shall use this term very loosely, as is often
done, to also include micro-evolutionary changes, genetic (and ecological) drift
within species and genotypic plasticity (Pfenninger et al. 2007). Climate change
projections using SDMs rely on the assumption that species are not able to adapt
significantly to altering environmental conditions. This assumption is implicit in the
extrapolation of the fitted niche: if a species was able to adapt rapidly, then the
present niche would not be related to its future niche.
