Understanding and mitigating the impacts of habitat fragmentation due to anthropogenic causes (e.g. logging) on wildlife populations consists of three major components: (1) understanding both historical and current population dynamics, structure and behavior of the species of interest; (2) identifying the role and extent that various human activities and changes in land-use patterns play in causing the resulting population structure; (3) understanding how best to mitigate negative anthropogenic impacts on a given landscape. As pointed out by van Oort, McLellan & Serrouya (2011), in many instances the assumption is made that where humans have fragmented once contiguous habitats, wildlife will inevitably establish a new metapopulation structure for the species. If land and wildlife managers and conservation practitioners begin with the premise that a metapopulation structure is in place before confirming this with data, and if these managers proceed to attempt to mitigate our impacts to putatively ensure and enhance habitat connectivity, landscape permeability and genetic and demographic connectivity, our conservation efforts and funds may be misguided.
As a hypothetical example, species X is perceived to have been historically distributed fairly evenly across an intact landscape as a relatively panmictic population. At some point in time a road network is constructed within the once contiguous habitat. As time proceeds, the population begins to decline and the management and conservation community decides that these anthropogenic structures (i.e. roads) have severed habitat connectivity for species X forcing the species into a new metapopulation structure. As a result, the decision is made to spend funds to reduce habitat fragmentation through best management practices that reduce road densities and/or via some mitigating structure(s) (e.g. road crossing structures) to enhance both genetic and demographic connectivity (i.e. rescue effect; Simberloff, 1988; Hanski & Gilpin, 1991; Hanski et al., 1995). In this hypothetical example, behavior studies are even conducted demonstrating that species X would utilize patches of minimum size Y and crossing structures with design parameters A, B and C. Once the identified best management practices are used to reduce fragmentation levels and correctly designed crossing structures constructed, species X still continues in decline ultimately going locally extinct. In this hypothetical example, what was clearly missing was an empirically based, baseline understanding of population structure and behavior before fragmentation of the landscape.
In the case of mountain caribou Rangifer tarandus caribou, van Oort et al. (2011) claim that spatial severing of habitat by human activities did not lead to the creation of a metapopulation structure in which natal dispersal (or even breeding dispersal) links subpopulations, because the weak ‘innate’ ability to disperse continues to prevent metapopulation structuring and associated rescue effects of these small, isolated subpopulations. Did caribou in this study area demonstrate this lack of ‘innate ability’ to disperse historically for long periods of time or is the inability and/or unwillingness to disperse a function of current social dynamics? In other words, caribou may likely not disperse in the current situation in British Columbia due to population sizes now falling below some critical threshold level of size/density that would drive both natal and breeding dispersal (Dobson, 1982; Travis, Murrell & Dytham, 1999). This idea is touched on by van Oort and colleagues in their discussion of population eruption and expansion by other caribou populations (e.g. the Mulchatna caribou herd). In the woodland caribou case study it appears that human activities (logging and roads) have fragmented the habitat, but that the lack of dispersal and connectivity among populations may not be due to habitat fragmentation and barriers per se and instead the proximate cause of the lack of connectivity is the low-density populations' unwillingness to disperse. Disentangling between social/behavior-caused and habitat fragmentation-caused lack of dispersal has huge implications for how best to recover populations or mitigate the impacts of anthropogenic-caused habitat fragmentation. For example, if the lack of demographic connectivity is simply a spatial habitat fragmentation issue, then mitigation measures such as logging practices that keep habitat patches above some critical size-threshold and/or road crossing structures in critical locations [e.g. across Trans-Canada Highway (1)] may be successful strategies to enhance/restore demographic connectivity and functional populations. In contrast, if lack of dispersal and resulting ‘extreme non-equilibrium metapopulation’ structure is simply due to behavioral responses to low population numbers, then population augmentation measures such as the ones van Oort et al. (2011) briefly mention may be the only effective mitigation measures to restore genetic and demographic connectivity for mountain caribou.
In their discussion, van Oort and colleagues emphasize that ‘incorporating the metapopulation paradigm in conservation is likely appropriate for species with a naturally fragmented distribution but should be used more cautiously with species that were recently more uniformly distributed’. While I concur, it also is worth noting that simply because a species has recently been more uniformly distributed across the landscape does not mean that some form of subpopulation structure did not exist historically. If anything, the van Oort and colleagues paper clearly points out the need for understanding the historical population dynamics, structure, and behavior of the species before landscape fragmentation. It is important to use caution and validate assumptions that no population substructuring existed historically in populations that were at one time perceived to be more uniformly distributed. It is possible a species that appears uniformly distributed across a landscape can exhibit subpopulation structure due to lack of dispersal even in the absence of human-caused fragmentation of the landscape (e.g. through density-dependent dispersal processes or other behavioral population regulation mechanisms). In the case of mountain caribou, the lack of dispersal (natal and breeding) and the resulting summer/fall composite range (SFCR) and subpopulation structure identified by van Oort et al. (2011) may have been in existence before any fragmentation of the landscape by humans (as evidenced by low levels of movement even among near-neighboring composite summer ranges described by van Oort and colleagues that are clearly within the species' movement capacity). It would be interesting to know which of the SFCRs and subpopulations identified by van Oort et al. (2011) as not being connected actually were spatially separated by a human-caused barrier or fragmentation (e.g. a road or region of human-altered habitat) between them. Although historical lack of demographic connectivity seems unlikely for mountain caribou given long-distance movement capabilities of some caribou populations (particularly barren-ground caribou), the fact that the Mulchatna caribou herd expanded its geographical distribution during population eruption events, suggests that this population of mountain caribou may behave similarly and only expand via dispersers during such infrequent population eruptions (see Hinkes et al., 2005). If this is indeed how this population historically expanded and hence was demographically connected, then human-induced fragmentation may not be the direct cause of the population structuring, but more likely plays a secondary and exacerbating role (e.g. through increased predation due to habitat change; Wittmer, Sinclair & McLellan, 2005; Wittmer et al., 2007) by reducing the overall population below a critical threshold where such eruptions and expansion are no longer possible – an idea acknowledged by van Oort and colleagues. The proximate cause of a lack of demographic connectivity and metapopulation structure may have nothing to do with anthropogenic habitat fragmentation, but may simply be the result of historical behavioral patterns displayed by woodland caribou in this region. Given the lack of juveniles collared and studied in the past for this population the data to address this question may not be available from these sites, but may be available from other caribou studies.
The take home message is that mitigating impacts of human-induced habitat fragmentation on wildlife populations depends on a clear understanding of the historical, underlying population structure that can only come from baseline data (both behavioral and demographic) collected before the fragmentation of the landscape. In the absence of this level of understanding that only comes from empirical, field-based data, we are often left guessing what came first, the habitat fragmentation or the population structure.