Habitat loss influences the evolution of dispersal
In a recent paper published by myself, Ashif Shahid and Mark Fitzpatrick we highlighted the importance of considering the scale of habitat loss when assessing connectivity in the landscape. We investigated the impact of habitat loss on the evolution of dispersal in a population with two different dispersal strategies. In our paper we showed how strains of fruit flies with different dispersal strategies (less dispersive and more dispersive) dispersed equally well when habitat connectivity was high, but that under conditions of low connectivity dispersal was greatly reduced for flies with a less dispersive strategy. Although this is perhaps not so surprising it nonetheless suggests that habitat loss can have differential impact on a population if there is genetic variation underlying dispersal strategy. Our paper demonstrated that there were critical distances (thresholds) where connectivity simply broke down and that this threshold could change depending on natural allelic variation in a gene known to influence dispersal in flies.
So, why is this so important? The impact of habitat loss on biodiversity and communities is well studied in landscape ecology (and beyond) and readily deserves the increasing attention given how we humans continue to erode biodiversity on our planet. Dispersal is a key process by which organisms redistribute among habitat patches both at local and landscape scales. Dispersal is therefore the "glue" that connects and sustains community processes in such a way that under certain scenarios of habitat loss we may expect communities to appear "unaffected". As such, an important tool to manage our ecosystems is to promote connectivity in our landscapes. However, an important question is whether the impact of habitat loss on communities is gradual such that we can see a decline in connectivity that is proportional to the amount of habitat loss, or whether there is an invisible "threshold" beyond which habitat loss leads to a rapid decline or even collapse in connectivity (Fig. 1).
Fig. 1: The dispersal threshold is dynamic for a population with multiple dispersal strategies. For individuals with a more dispersive strategy (bold line) the influence of habitat loss (x-axis) on connectivity (y-axis) is less severe relative to individuals with a less dispersive strategy (dotted line). Adapted from Edelsparre et al. (2018).
This is important because the former indicates that we can quantify the impact of habitat loss on biodiversity and make management decisions that potentially reverse the effect whereas the latter indicate that the effect of habitat loss may go unnoticed (i.e. appear "unaffected") before we see a collapse in systems. Our findings using flies suggest that there indeed are thresholds (e.g. critical distances at which connectivity breaks down), but that the position of a given threshold depends on genetic variation underlying dispersal strategy. Our finding is thus a cautionary tale about the potential risk we accept by continuously letting our landscapes becoming fragmented without really understanding how this may impact landscape connectivity.
Landscape ecology is key to this understanding because it describes how structural connectivity (the distribution and pattern of habitat) influences functional connectivity (dispersal). Landscape ecology is therefore central to the understanding of how human altered landscapes impact ecosystems and over the last two decades this field have made tremendous theoretical and empirical advances that provide important insights into how we humans affect and create landscape patterns and how this process affect our ecosystems.
Generally, both theoretical and empirical studies in landscape ecology make two broad assumptions when investigating landscape connectivity (although there are exceptions). The first assumes that the number and sizes of habitat in the landscape influence functional connectivity. This assumption is intuitive because habitat provides key resources that organisms need. It is also convenient to model because habitat types can easily be quantified via remote sensing such as GIS systems etc. and so we can obtain fine scaled resolution of varying habitat types that appeal to a broad community of organisms. However, one potential caveat that arises from this assumption is that we ignore the structure of the gaps arising from habitat loss. The gap structure is essentially the "hostile" environment that individual organisms have to cross in order to reach habitat necessary to survive. The scale of these gaps can vary greatly even if the number and size of available habitat is constant. By quantifying the number and size of habitat in the landscapes researchers may therefore ignore the fact that it is the scale of habitat loss that affect connectivity and our paper highlights the importance of considering this caveat by demonstrating that it is the distance between habitat that affects connectivity and not the size of habitat.
A second assumption is that dispersers are a random subset of a population. In other words, that any individual from a population or species is weighed equally in terms of dispersal propensity. Making this assumption is understandable because in many cases we do not have information about multiple dispersal strategies in a given population/species and often we are trying to understand responses at the level of whole communities, so model complexity, by incorporating multiple dispersal strategies, can be daunting. But, recently the literature have been teaming with examples of how species consist of individuals with very different dispersal strategies. The idea that individuals consistently differ in intrinsic behavioral traits such as shy/boldness, aggression, sociability, novelty seeking etc. is not new, but only over the last couple of decades have researchers linked these behavioral differences with exploratory, dispersive and migratory movements. This has radically changed how researchers view the impact of dispersal on ecological and evolutionary processes. In fact, just recently Canestrelli et al. (2016) argued that such intrinsic behavioural differences among individuals have made significant contributions to shaping biogeographic patterns. The potential implication of individual behaviour acting as a "pacemaker" of evolutionary processes, as Canestrelli and colleagues (2016) put it, is interesting because it not only suggests that behaviour can take the lead in phenotypic evolution (e.g. by selecting on dispersal directly or indirectly via correlated behaviours) it also gives us an opportunity to understand the selective forces that maintain variation in behaviour. So both from an evolutionary and a ecological perspective it is therefore a major caveat to ignore variation in dispersal strategy since habitat loss likely will select for increased dispersal (as our paper suggests) and thus affect other important behaviours (if correlated with dispersal). This could greatly change the population dynamics, both within and between populations and species, if selection for increased dispersal leads to more aggressive and less social individuals, just to give an example.
Overcoming both assumptions will be a challenge, but remain an important next step as we go into the future. However, we have made tremendous steps in terms of understanding the behaviour and genetics behind organismal movement. In addition, our analytical abilities are constantly improving due to increasing computational power, even on conventional laptops. Bridging genotypic and phenotypic components of organismal movement with dynamic models will likely provide great insights into how habitat loss and landscape fragmentation will shape our species and communities.
Canestrelli, D., R. Bisconti, & C. Carere. 2016. Bolder takes all? The behavioural dimension of biogeography. Trends in Ecology & Evolution. 31: 35–43.
Edelsparre, A.H., S. Shahif, & M. Fitzpatrick. 2018. Habitat connectivity is determined by the scale of habitat loss and dispersal strategy. Ecology and Evolution. 8:5508–5514.
