Applications of island biogeography theory beyond
islands
Islands provide a powerful tool to address ecological questions because
they represent discrete environments offering multiple replicates of
comparable eco-evolutionary processes. Furthermore, there is a
well-established theory that postulates a connection between the
macroecological patterns observed in a set of islands and the identity
and interactions between the species that integrate their species’
assemblages (Wilson 1969). In its initial formulations, however, island
biogeographical models were essentially neutral, insofar as they only
relied on species identity and endemism, treating all species as
functionally equivalent. This is analogous to say that the position of
each species on the niche space is random and does not affect the
properties of the overall island biodiversity. We now know this is a
strong assumption, which can be relaxed by incorporating a functional
description of the species occurring in each island (Hubbell 2011;
Rosindell et al. 2011). This way, the position of each species is
defined in the n-dimensional overall trait space of each island,
allowing us to make further assumptions on how the presence of
different, non-random combinations of traits affect the macroecological
status of the islands: a crucial step to connect community scale
patterns with island-level observations.
Indeed, the results of null-modelling confirms that the combination of
traits in each island is not random and can be linked to the predictions
of the island biogeography model. In other words, our analysis suggests
a causal relationship among the island age, the changes in the niche
availability and the overall island’s biodiversity. Our results indicate
how in young islands functional richness and evenness approach the
values expected by change for the same species richness, illustrating a
multidirectional expansion of the trait space as a consequence of the
species radiation and ecological opportunity. However, as an island
matures, the traits space becomes smaller and more ordered when compared
to the null-model, reflecting how the existence of a more connected
network of species interactions optimizes the use of the available niche
space. Subsequently, as an island is eroded and its niche space reduced,
species will start to be extirpated. An initial stage of this phase can
be seen in Gran Canaria, where the null-modelling reveals that the trait
space functional richness is larger than expected by chance, whereas
evenness approaches the randomly predicted values. Indeed Gran Canaria,
despite being younger than Fuerteventura, has already undergone a long
erosive period, characterized by a lack of volcanism and massive
landslides dramatically affecting its geomorphology. The trait space of
Fuerteventura resembles the values obtained by null-modelling,
illustrating the collapse of the island’s ecological network during its
final ontogenetic stage, just before it will disappear under the ocean.