Habitat loss, degradation, and fragmentation are among the largest threats to biodiversity worldwide [1
]. Through a multitude of factors, land conversion as a result of anthropogenic activities plays a major role in the reduction of the abundance and distributions of many species [2
]. Habitat loss and degradation predominantly lead to the decline of local populations through the loss of available resources [1
]. The smaller the populations and the smaller the genetic variability, the greater the vulnerability to demographic and environmental stochasticity [3
]. Reduced connectivity between habitat exacerbates these threats by increasing the isolation of breeding populations, the likelihood of movement through inhospitable matrix, and the proportion of edge habitat, reducing successful dispersal between suitable habitat patches [4
]. With their relatively restricted movement capabilities and diverse yet specific habitat requirements, amphibians are among the most vulnerable groups of species to these threats [6
To understand species movement in a spatially explicit context, connectivity assessments are valuable contributions as the data-driven approach allows for verification of hypotheses of organismic movement across landscapes [9
]. Connectivity assessments often rely on modelling frameworks that identify the role of landscape structure in contributing to movement success of organisms [9
]. Connectivity modelling has been applied to a broad range of species and environments ranging from e.g., individual species [13
] to multi-species corridors [18
]. Multi-species approaches allow the generalization of findings across larger regions [21
], providing important baseline information for effective management measures and for implementation into green infrastructure [9
] and conservation management [24
In this study, we created functional and species-specific structural connectivity maps of five amphibian species in the Swiss lowlands and combined the results in order to produce multispecies connectivity maps that account for each species’ unique movement ecology. Using Circuitscape [25
], a standard connectivity modelling tool [27
], we provide a functional and species-specific analysis of connectivity with the goal of offering conservation managers a large-scale and comprehensive evaluation of functional landscape connectivity for amphibians for implementation e.g., in green infrastructure concepts.
Recently, claims to include connectivity assessments into green infrastructure concepts have been put forward as they may provide important baseline information to preserve and mitigate threats of habitat fragmentation [9
]. Here, we provided a regional-scale multispecies connectivity map to analyze the movement potential of amphibian species across a human-dominated landscape. The multispecies approach accounts for each species’ estimated or known dispersal ability and modulates the dispersal (“current”) strength based on the population size, thus allowing for a spatial source-sink dynamic. The results from this functional connectivity assessment are contrasted by an analysis relying solely on expert-based judgement on the potential connectivity given landscape structure only. The complementary aspects of the functional and structural connectivity maps provide a picture of realized and potential connectivity, together offering the insight required to preserve and mitigate threats to connectivity or improve and restore it. The sensitivity analysis showed that each species’ current map is highly sensitive to the choice of resistance scenarios.
There are various multispecies approaches in connectivity assessments, for example selecting ‘umbrella species’, assuming that a broad range of associated species would benefit from measures taken to preserve or restore connectivity for a particular species [44
]. However, identifying suitable surrogate species has remained a debated challenge [47
]. Alternatively, Koen et al. [37
] produced a regional map of potential functional connectivity for a generalized suite of forest-dwelling species that successfully predicted the movement corridors of a bird and several amphibian species. However, this approach lacks flexibility, as it is limited to groups of species that share a similar behavioral response to the landscape patterns. When the permeability of the landscape differs substantially among the focal species of a study, it would seem necessary to include separate resistance surfaces tailored to each species within the analysis. Beier et al. [47
] achieved this by overlaying the movement corridors predicted by individual species models in a multi-species least-cost path analysis of connectivity. Using Circuitscape [25
], however, insights are not limited to a few specific habitat patches or corridors, but are available in continuous, high-resolution, and large-scale maps across the study region indicating all potential movement routes. The species-specific approach used to generate the multispecies current maps offers the flexibility to include species with diverse movement ecologies and ensures that no species is potentially mismatched to the requirements of a single ‘umbrella’ species. The target species included in the analysis could even be expanded to include other groups of species, e.g., dragonflies or other wetland species that could synergistically benefit from conservation initiatives. While the multispecies current maps provide insights that can be used to improve landscape connectivity for some or all of these species, each individual species current map can also be used independently to prioritize regions of focus in conservation efforts and in the identification of integral landscape features to connectivity for a single target species.
However, lacking independent movement data to validate the connectivity models and the resistance maps they are derived from, it is difficult to assess just how accurate these maps are. Despite reasonably high overlap of high current regions in the sensitivity analysis, the breadth of landscape resistance scenarios captured by the analysis is by no means all-inclusive. It is also not likely that landscape categories can be discretely divided into categories of resistance rankings, or that the number of categories would be equal across species. In addition, only land cover/land use was considered. Other factors, such as pollution and habitat quality, weather conditions, traffic volumes, or microhabitats may additionally impact movement patterns of amphibians [31
Therefore, options to improve connectivity models are manifold. It is well established that connectivity analyses based on expert knowledge generally perform worse than and are best applied only as a complement to more data-driven approaches [11
]. Empirically derived field-data that accurately relates the movement of dispersers to the landscape would improve the spatial aspects of dispersal [12
]. There are a number of methods available to landscape ecologists for this purpose. Mark-recapture and telemetry studies can quantify the movement rates, distances, and paths of individuals in order to identify the behavioral responses of a species to its environment. Besides being quite resource-intensive, the challenge with these methods is capturing the movements of an actual disperser. Many amphibians have a high fidelity to their natal ponds, and dispersal rates between breeding populations can be very low [51
]. Movement patterns of an amphibian within its terrestrial home range likely differ to that of a dispersing individual exposed to various qualities of matrix habitat over much greater distances [50
]. Alternatively, the use of gene flow as a measure to delineate connectivity among populations has been widely applied [53
], also in a conservation context [56
]. Through the comparison of the genetic characteristics between breeding ponds, it is possible generate estimates of gene flow that can then be used to estimate the resistance values of landscape features that separate them after accounting for the multigenerational processes that determine genetic structures [1
It is also important to note that the current maps only indirectly describe the actual quality of the landscape with respect to supporting connectivity. High current regions within the functional connectivity maps are predominantly determined by the size and number of connected breeding populations; landscape resistance simply shapes the flow of current. Generally, current that flows across regions of poor permeability to movement will be highly concentrated through the few landscape features that promote movement, like streams. Alternatively, in areas where the terrain is broadly favourable to movement, current flow unrestricted in wider swaths. Furthermore, if absolutely no movement corridors exist through poor terrain, the flow of current will be indistinguishable from that of a uniformly high quality region of connectivity. Distinguishing such scenarios from these current maps requires a keen eye and familiarity with the terrain. Sinsch’s et al. [58
] review of the literature suggests that this may actually be an accurate portrayal of the effect of landscape resistance on amphibian movements. However, it is also likely that the high cost of dispersal over a poorer matrix habitat has a negative effect on the likelihood of an immigrant’s reproduction success, which should be factored into predictions of functional connectivity [1
]. Circuitscape calculates an ‘effective resistance’ metric between each node pair as a function of the cumulative cost-distance and redundancy of paths between nodes [26
]. Further weighting of the amount of current flow between nodes by effective resistance may allow a better representation of functional connectivity within the region that takes into account landscape quality.
Regardless of the limitations of the models within this study, we believe that the methods outlined here to generate functional and species-specific structural connectivity maps could be very valuable to conservation efforts for these species and, if scaled up accordingly, for even more species and regions. Both sets of current maps, functional and structural, provide a great deal of information concerning landscape connectivity and could seemingly be used in a multitude of conservation-related applications. The functional connectivity current maps act as excellent visual aids that are easily accessible and intuitively allow a user to locate areas within each species’ distribution where connectivity is high or low on a cantonal scale. Focusing in on specific areas elucidates the importance of fine-scale features that can inform decisions in development and land use change, suggest locations along highways where amphibian tunnels may be needed, or identify sensitive movement corridors that could benefit from reinforcing protective measures. The structural connectivity maps can guide restoration efforts, highlighting suitable locations for the creation of stepping stone ponds between isolated clusters of breeding populations. In addition, such maps could provide decision-makers with the insight needed to mitigate the impact of development on biodiversity and identify regions where landscape connectivity can be improved. This feature is particularly relevant in the context of designing green infrastructure [9
], a concept in spatial planning policy included in the EU2020 Biodiversity Strategy that involves the strategic planning of development and land use to ensure the long-term persistence of biodiversity and ecosystem services [9
]; EEA (https://www.eea.europa.eu/themes/sustainability-transitions/urban-environment/urban-green-infrastructure/what-is-green-infrastructure
Additionally, approaches to overcome the computer-intensive analysis with Circuitscape may allow large scale maps [23
], allowing for large-scale multi-species connectivity maps capable of guiding the needs of conservation efforts.