Integrating functional connectivity in designing networks of protected areas under climate change: A caribou case-study

Sarah Bauduin; Steven G. Cumming; Martin-Hugues St-Laurent; Eliot J. B. McIntire

doi:10.1371/journal.pone.0238821

Abstract

Land-use change and climate change are recognized as two main drivers of the current biodiversity decline. Protected areas help safeguard the landscape from additional anthropogenic disturbances and, when properly designed, can help species cope with climate change impacts. When designed to protect the regional biodiversity rather than to conserve focal species or landscape elements, protected areas need to cover a representative sample of the regional biodiversity and be functionally connected, facilitating individual movements among protected areas in a network to maximize their effectiveness. We developed a methodology to define effective protected areas to implement in a regional network using ecological representativeness and functional connectivity as criteria. We illustrated this methodology in the Gaspésie region of Québec, Canada. We simulated movements for the endangered Atlantic-Gaspésie caribou population (Rangifer tarandus caribou), using an individual-based model, to determine functional connectivity based on this large mammal. We created multiple protected areas network scenarios and evaluated their ecological representativeness and functional connectivity for the current and future conditions. We selected a subset of the most effective network scenarios and extracted the protected areas included in them. There was a tradeoff between ecological representativeness and functional connectivity for the created networks. Only a few protected areas among those available were repeatedly chosen in the most effective networks. Protected areas maximizing both ecological representativeness and functional connectivity represented suitable areas to implement in an effective protected areas network. These areas ensured that a representative sample of the regional biodiversity was covered by the network, as well as maximizing the movement over time between and inside the protected areas for the focal population.

Citation: Bauduin S, Cumming SG, St-Laurent M-H, McIntire EJB (2020) Integrating functional connectivity in designing networks of protected areas under climate change: A caribou case-study. PLoS ONE 15(9): e0238821. https://doi.org/10.1371/journal.pone.0238821

Editor: Laurentiu Rozylowicz, University of Bucharest, ROMANIA

Received: February 17, 2020; Accepted: August 24, 2020; Published: September 30, 2020

Copyright: © 2020 Bauduin et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: Data cannot be completely shared publicly because of the Endangered Status of the Atlantic-Gaspésie caribou population, so data that could help track and find these animals won't be made available publicly. However, the modeled data we used to quantify the balance between ecological representativeness and functional connectivity will be made available in a public repository (DRYAD #https://doi.org/10.5061/dryad.612jm641c) upon acceptance. However, the telemetry data collected on these endangered caribou population can be made available upon request by qualified researchers if they contact Prof. Martin-Hugues St-Laurent at Université du Québec à Rimouski (martin-hugues_st-laurent@uqar.ca) or services.clientele@mffp.gouv.qc.ca. Being co-owner of the data (along with the Québec Ministry of Forests, Wildlife and Parks), please note that I’ll then have to transfer the request to the authorities of the QMFWP to obtain their permission too.

Funding: Funding for this project was provided by: - the Québec Fonds Verts (S.G.C., E.J.B.M.) (URL: http://www.environnement.gouv.qc.ca/ministere/fonds-vert/) - the Ouranos Consortium (S.G.C., E.J.B.M.) (URL: https://www.ouranos.ca/) - the Centre d’étude de la forêt (S.B.) (URL: http://www.cef-cfr.ca/) - Natural Sciences and Engineering Research Council Discovery Grants (E.J.B.M., S.G.C.) (URL: https://www.nserc-crsng.gc.ca/NSERC-CRSNG/) - the Canada Research Chair’s program (E.J.B.M.) (URL: https://www.chairs-chaires.gc.ca) -the French National Research Agency (ANR‐16‐CE02‐0007, S.B.) (URL: https://anr.fr/en/).

Competing interests: The authors have declared that no competing interests exist.

Introduction

Habitat change is recognized as the main driver of the current declines of terrestrial species [1–3]. Securing habitats by creating or expanding protected areas networks is part of the solution to counter biodiversity loss [4]. A regional protected areas network could be considered ultimately effective insofar as it can sustain the region’s biodiversity into some reasonably foreseeable future. Such effectiveness is not guaranteed [5, 6] and could be limited by many factors [7]. We consider three such factors: ecological representativeness [8], functional connectivity among protected areas within the network [9] and resilience to climate change [10].

Ecological representativeness measures the degree to which the various non-anthropogenic habitats or ecosystem types (sensu [11]) within a focal region are available within a protected areas network in proportion to their regional abundance [12, 13]. When protected area locations are skewed towards certain habitats [14, 15], usually for economic or social reasons, ecological representativeness will be low, and habitats considered to be of high-value for some species may be underrepresented [8]. Conversely, when ecological representativeness is high, it is reasonable to assume that the habitat requirements for many species will be satisfied within the protected areas network. This assumption is usual in conservation, and representativeness is one of the core concepts in systematic conservation planning [13, 16].

A high degree of ecological representativeness may be a necessary condition for an effective protected areas network, but it is not sufficient. There are species whose requirements are not automatically satisfied by ecological representative networks, such as endemic or threatened species [17]. Another exception, which we explore here, would be wide-ranging species with habitat requirements varying among seasons or life history stages. Functional connectivity is “the degree to which the landscape facilitates or impedes movement among resource patches” [18] or, in this case, among protected areas within a protected areas network. Functional connectivity is species- or population-specific [19, 20]. A protected areas network with high functional connectivity facilitates the movement between different protected areas for individuals of a given species, increasing their access to resources and, ultimately, the rates of recruitment or survivorship [21]. Increasing functional connectivity may thus increase population size and decrease extinction risk for the vulnerable species [21]. These effects would increase the effectiveness of a protected areas network.

Climate change is a major driver of ecosystem change and its negative impacts on biodiversity have increased rapidly over the past century [1, 3, 22, 23]. Climate change disrupts environmental patterns and species’ habitats globally [24, 25]. As a result, species distributions and individual movement patterns are impacted [26, 27]. Because of the “cost of waiting,” managers should proactively account for future climate change effects when implementing new protected areas networks [28] or expanding existing networks. The effectiveness of a fixed protected areas network designed for current conditions may decrease in the future if ecological representativeness or functional connectivity decline. Enhancing or maintaining functional connectivity inside protected areas networks is thus one approach to helping species cope with climate change [10, 29]; individuals would be better able to access resources available in distant protected areas, or even to migrate from less to more favorable areas. Therefore, the ability of a protected areas network to sustain functional connectivity under climate change is another dimension of effectiveness.

Many methods exist for designing protected areas networks to achieve ecological representativeness. In the systematic conservation planning literature [13], variations of the site selection problem are posed, where one seeks a subset of available sites that, in aggregate, achieve some measure of ecological representativeness at a near-minimum of total area or cost. Variations of these approaches exist that can also partially satisfy other conservation objectives such as topological connectivity, where selected sites are spatially aggregated to some degree [e.g. 30].

The aim of our paper is to present a new method to include functional connectivity in the design process, while also accounting for its persistence under climate change. Our method applies individual-based movement models to simulate maps of habitat use by a population, under present conditions and under hypothetical future conditions reflecting climate change scenarios. From these maps, we derive numerical indices of functional connectivity that allow alternate protected areas network designs to be compared. These are coupled with a novel variant of the site-selection algorithm that constructs protected areas networks of specified total area that can achieve a high degree of ecological representativeness while also satisfying secondary constraints on hydrological connectivity and intactness; the variant is detailed in Schmiegelow et al. [31] and an example of application can be consulted in Saucier [32].

Our goal was to identify protected areas network designs that simultaneously achieve high degrees of ecological representation and functional connectivity under present conditions and future climates. Achieving this is a multi-objective optimization problem [33, 34] that does not admit a unique solution. Instead, one may define a tradeoff surface by those points or “feasible solutions” where no single objective can be increased without decreasing at least one other. Such points are said to be Pareto optimal; points on the interior of this surface are suboptimal in that other solutions exist that are better in terms of all the dimensions considered [35]. Our design methodology used randomization to generate a large sample of feasible solutions for a given protected areas network design problem. Using these samples, we could approximately define the tradeoff-surface and identified a set of feasible, near-equivalent solutions in the vicinity of any specified point on that surface.

We illustrated our method in the Gaspésie region (Québec, Canada) using the endangered Atlantic-Gaspésie population of woodland caribou (Rangifer tarandus caribou) as our focal population. This isolated herd, which has been identified as one of 12 Designatable Caribou Units considered irreplaceable components of Canada's biodiversity [36], has been declining since the late 19^th century and has reached a critical abundance level [37]. To measure functional connectivity with respect to this population, we applied an individual-based model of animal movement previously developed for this population [38]. The Québec government is currently engaged in expanding the existing network of protected areas in this region to attain a proportional area target of 12% [39]. To place our work in the context of this ongoing conservation planning exercise, we employed a variant of our design methodology that constructs protected areas networks by adding new protected areas to the existing network. We identified areas that were of high relative importance in achieving ecological representation in the Gaspésie region and functional connectivity for the Atlantic-Gaspésie caribou under present conditions and future climates.

Methods

Overview

We present a method to identify potential areas to prioritize to create an effective protected areas network in a given region, conserving both the regional biodiversity, represented by biophysical surrogates as well as focal, highly mobile endangered species, while also accounting for climate change. We illustrated this method in the Gaspésie region of Québec, Canada (Fig 1) using an individual-based model we developed previously [38] and for which we used telemetry data collected on several individuals. We had an Animal Welfare certificate (#52-13-112) for the capture and manipulation of these caribou, but for the current manuscript, no additional capture session was conducted.

Download:

Fig 1. The Gaspésie natural region with existing protected areas (black) and additions (grey) proposed by the Québec government [39].

The Gaspésie National Park (i.e. the black area designated by a white star) encompass most of the range and all the breeding habitats used by the Atlantic-Gaspésie caribou population. Right inset: Province of Québec (Canada) with study area outlined.

https://doi.org/10.1371/journal.pone.0238821.g001

We built candidate protected areas (hereafter referred to as CPAs) in the region and integrated random subsets of these with the existing protected areas (Fig 1) to create a large sample of candidate networks of the desired total area. Each candidate network was then evaluated for effectiveness. High priority conservation areas were identified as the CPAs that occur most frequently in the most effective networks. The complete workflow is summarized in Fig 2.

Download:

Fig 2.

Schematic of workflow for assembling candidate protected areas networks (bottom panel) from candidate protected areas (middle panel), which are constructed as hydrologically connected groups of intact hydrological catchments (top panel) rooted at the headwaters (shown in blue).

https://doi.org/10.1371/journal.pone.0238821.g002

Study area

The Gaspésie natural region (latitude extent: 47.98 to 49.20º N, longitude extent: 64.11 to 67.57º W; Fig 1) is a physiographically defined area of approximately 25,000 km² at the eastern end of the Gaspésie Peninsula, in eastern Québec [39]. Except for narrow coastal bands, it belongs to the balsam fir–white birch bioclimatic domain [40]. The climate is maritime with abundant precipitation. Wildfire is infrequent; the main natural disturbance is spruce budworm (Choristoneura fumiferana) outbreaks [40]. Approximately 90% of the region is covered by forests and 80% of these are on public lands [39]. Forest harvesting and the associated extensive road network are the main proximate human disturbances to date. Only 34% of the Gaspésie natural region is free of measurable human footprint [39].

An existing protected areas network covers 5.5% (1371 km²) of the region (Fig 1). This falls short of the target of an ecologically representative 12%, set by the government of Québec [41]. A scenario proposed by the Ministère du Développement Durable, de l’Environnement et de la Lutte aux Changements Climatiques (“the Ministry”, henceforth) defined 20 new protected areas that would increase the percentage of area protected to 12.3% (3080 km²) [39; Fig 1].

The Gaspésie National Park (802 km²) is the largest protected area in the region [42]. This park helps conserve 42 endangered and vulnerable species of plants and animals [39], including the Atlantic-Gaspésie caribou population. Because of its designated status as endangered [43], its capacity to move widely [44] and the vulnerability of its habitat [45], we chose this caribou population as our focal population to define functional connectivity. The population is designated because of its small size (currently ~70 individuals [37]) and the observed long-term decline [46]. These caribou rely on alpine tundra [47, 48] as a predator-free refuge against coyotes (Canis latrans) and black bears (Ursus americanus) [49], a habitat possibly threatened by climate change [50, 51]. Caribou also use mid-elevation, old fir forests during winter [47,48]. The fir forests around the park are affected by forest harvesting, and also possibly by climate change [52].

Building protected areas networks

We adapted the BEACONs (Boreal Ecosystems Analysis for Conservation Networks) approach to create protected areas networks by adding new protected areas to the existing ones [31, 32]. This approach uses mapped hydrological catchments as spatial units, rather than using the cells of an arbitrary grid or other tessellation (Fig 2). This allows the design of protected areas that simultaneously achieve terrestrial and hydrological connectivity [53, 54]. Although functional connectivity for aquatic species was not a major concern in our study area [39], it is increasingly recognized as a desirable goal in general [55]. The use of hydrological catchments does impose some limits on the shape and size of protected areas, due to the size of catchments and the method of stream network traversal; however, the size of the catchments was small, implying high spatial resolution to the design. CPAs are assembled as contiguous, hydrologically connected sets of catchments satisfying minimum size and intactness criteria (Fig 2). Using an initial “seed” catchment as starting point, a modified breadth-first traversal of the stream network is carried out. Catchments are added as they are encountered, provided they satisfy a catchment-level intactness criteria. The process continues until the target size is achieved, or all pathways are blocked by headwater or non-intact catchments.

We used a custom 1:50,000 catchment layer for Gaspésie [56] (Fig 2). The catchment’s average size was 2.3 km² (SD = 1.3 km²). We defined catchment intactness using existing 250 x 250m raster maps of the human footprint in Gaspésie [39]. These maps defined six disturbance types: forestry activities, roads and trails, agriculture, power-line rights-of-way, urban areas, and “other”. Only the first two of these disturbance types were widespread in our study area. From the six disturbance rasters, we derived a 250 x 250m raster of cell-level intactness, using a value of 0 for cells where no disturbance of any type was present, and a value of 1 otherwise. Catchment-level intactness was then calculated as the mean cell-level intactness over all cells within a catchment, or equivalently, as the proportion of non-intact cells. We defined our catchment-level intactness criteria as the median catchment intactness over all catchments on public lands, outside of existing protected areas (median value = 0.026) (Fig 2). In order to create feasible designs given land ownership in Gaspésie, we set the intactness of catchments completely overlapping private lands [39] to zero so that they would be excluded from the CPA construction (Fig 2).

Our measures of cell and catchment-level intactness give equal weight to each disturbance type. This was not applied, in general, at the species level. The response to different disturbance types is known to differ among species, as shown by Toews et al. ([57]) in Alberta for wolves (C. lupus), meso-carnivores and large ungulates, or more precisely within our study area for caribou, coyotes and bears that respond differently to recent harvesting, forest roads and hiking trails [58]. However catchment intactness as a design criteria was intended to be relevant to biodiversity more broadly, not to any particular species. Equal weighting of disturbance types reflects different responses to disturbances across species, and is thus a conservative assumption.

We prioritized the inclusion of headwater catchments in protected areas, because of their hydrological importance and to reduce the potential for upstream river contamination inside protected areas [59]. Accordingly, we used all intact headwater catchments as seeds. We used the mean size of the new protected areas proposed by the Ministry (Fig 1; mean = 85.5 km², SD = 61. 4 km²; [39]) as the target size. We then applied the construction process to each seed catchment, with target size and catchment intactness criteria as defined above. The procedure returns a list of the catchments selected and their total area. If this total area satisfied the target size, the result was accepted as a candidate protected area. The construction adds entire catchments one at a time, so target sizes are generally exceeded rather than satisfied exactly (Fig 2). Because the mean catchment area (2.3 km²) was small relative to the target size, we consider such discrepancies negligible. The minimum size of the Ministry’s proposed new protected areas was 31.2 km² [39], which is only 36.5% of our target size. To make our size criteria more comparable to theirs, we also accepted, as candidate protected areas, constructions smaller than the target size but larger than this minimum. Candidate protected areas satisfy, by construction, the criteria of size, intactness and hydrological connectivity [31]. They are not designed to satisfy ecological representativeness or functional connectivity criteria. These criteria pertain to networks of protected areas, not to individual protected areas.

We used the set of selected candidate protected areas to generate a sample of 500,000 candidate protected areas networks (Fig 2). Each candidate network was constructed by adding a random sequence of candidate protected areas to the existing network, until the summed area of unique catchments exceeded the Ministry’s area target of 3080 km² (Fig 2). The candidate protected areas included within a given candidate network may overlap with each other or with elements of the existing network, but this does not affect the candidate network area. It is easy to show that the properties of hydrological connectivity and intactness are conserved by aggregating overlapping candidate protected areas. Thus, the only material consequence of such overlap is that spatially disjunct network elements may exceed the minimum size criteria, and may be fewer than expected based on the number of candidate protected areas that were added. However, other things being equal, larger protected areas are preferred [60], and the number of spatially separated components is not one of our network design criteria, so we do not consider this overlap of relevance to the present study.

Network ecological representativeness

Ecological representativeness is measured at the candidate network level [31], in terms of four environmental attributes: elevation, surficial deposit, drainage class and potential vegetation type [39] (Fig 2). Potential vegetation was defined by the Ministère des Forêts, de la Faune et des Parcs du Québec (MFFP) as the vegetation present on a site or potentially present, in the absence of disturbance [41]. Surficial deposit, drainage class and potential vegetation type had been used for a gap analysis conducted to inform the proposed expansion of the existing protected areas network [39]. The four attributes define habitats independent of human activities such as harvesting history. All four attributes were available for the entire study and were provided by the Ministry. Catchment-level attributes were calculated by intersecting a shapefile of the catchment layers with the various raster grids and taking means for continuous attributes and frequency tables for categorical attributes.

We measured candidate network representativeness using nonparametric, two-sample univariate dissimilarity measures. For each attribute, we obtained its distributions over all catchments within the candidate network, and over all other catchments within the study region. For continuous attributes (e.g., elevation), we calculated a two-sample Kolmogorov-Smirnov statistic. For categorical attributes (e.g. surficial deposit, drainage class, and potential vegetation type), we calculated the Bray-Curtis statistic. These statistics ranged from 0 to 1, measuring a candidate network’s deviation from perfect representation with respect to an attribute, and were equal to 0 only when the two distributions are identical. The univariate statistics for multiple attributes were combined into a univariate distance metric by calculating a Euclidean norm [31, 32]. We took the inverse of this distance metric as the ecological representativeness score for each candidate network, so that a larger score indicates a greater degree of representativeness.

We also calculated the representativeness score of the Ministry’s proposed network. To do this, the protected areas within this network were approximated to the resolution of hydrological catchments. Because these catchments were relatively small, we assumed approximation errors to be negligible. We note that this measure of ecological representativeness does not depend on the specific choice of ecological attributes or on the specifics of our design methodology; it can be applied to any existing protected areas network using whatever ecological attributes are of interest.

Network functional connectivity

We defined candidate network functional connectivity for the Atlantic-Gaspésie caribou population using a spatially explicit individual-based movement model (IBM) adapted from Bauduin et al. [38, 61] (Fig 2). This model simulates caribou movement as a two-state behavioral model process with a habitat-mediated random walk in habitats of high quality and a foray loop movement in habitats of low quality. The IBM includes the attraction of individuals’ mating areas during mating season. Habitat quality is modeled using a Resource Selection Function (RSF; [62, 63]) model developed for this specific population of caribou [58] and based on four habitat types relevant to this population (i.e., alpine tundra, mature fir stands, regenerating stands and stands of other tree species, primarily broad-leaved species) as well as three classes of transportation routes (paved roads, gravel/secondary roads, and hiking trails). These landscape components represent resources or barriers for the caribou (e.g., alpine tundra and mature fir stands for food resources and shelter from predators, paved roads as movement barriers) or proxies for the presence of their predators (e.g., regenerating stands where bears and coyotes frequently occur).

The original model was re-estimated from newly available GPS locations [44, 64]. Model movement behavior is partially driven by spatial variation in habitats as described by the RSF. Following Bauduin et al. [38, 61], we generated five different habitat maps to represent the current landscape and four landscapes for 2080 forecast under different climate change scenarios. The current landscape was derived from the ecoforestry maps from Québec’s 4^th decennial forest inventory program (source: MFFP). The future landscapes represented a range of potential climate change impacts (none: CC₀; minimum: CC_Min; moderate: CC_Med; high: CC_High) on vegetation dynamics and natural disturbances, combined with predicted cumulative timber management for 2080. Primary impacts were reduced areas of tundra [50, 65] and of fir forest [52], and a decreased severity of spruce budworm outbreaks [66, 67] leading to less area with young fir stands. A complete description of the four climate scenarios and of the construction of the corresponding 2080 landscapes is given in Bauduin et al. [61] (and in the S1 Material). A fifth scenario represented current conditions.

For each landscape, we created 20 individuals in each of the three caribou subpopulations (see [38]; current population estimates are 20, 35 and 15, [37]) and ran 10,000 replicates of four-year model simulations. For each scenario and replicate, we calculated and mapped the number of caribou visits per 1-ha landscape cell (see S2 Material). For each candidate network, under each scenario, we measured network functional connectivity by taking the mean, over replicates, of the number of caribou visits in cells within the network’s protected areas. This represented the simulated movement patterns within and between the protected areas. The differences in results between climate scenarios were relatively small. We considered evaluating differences between them to be unimportant relative to our core message concerning the inclusion of functional connectivity in a protected areas network design. Accordingly, and because no climate change scenario was defined as more likely than another, we used the mean functional connectivity over the four climate change scenarios as an index of network connectivity under future conditions (S2 Material). We calculated the functional connectivity of the Ministry’s network under current and future conditions in the same way.

Identifying priority conservation areas

By plotting the distributions of indicators (i.e., ecological representativeness, current and future functional connectivity), the outlines of the tradeoff surface may be visualized. To identify one specific location on this surface, we incrementally decreased a quantile threshold Q from 1 to 0 by steps of 0.001 to find the maximum Q for which at least 500 networks had indicator values above the Q-th quantile for each of the three indicators. This identified a subsample of 0.1% of the candidate networks that were highly ranked under all three criteria. This represents a level of tradeoff which, informally, values all three attributes equally. We then identified the candidate protected areas included in the network subsample, calculated their selection frequencies, and mapped their locations color-coded by selection frequency. These selection frequencies may be interpreted as priorities for including candidate protected areas within a new or expanded protected areas network [30]. Similarly, spatial clusters of high priority areas may highlight regions of high importance relative to network conservation goals. We set a selection frequency threshold for high priority catchments by plotting the selection frequencies in decreasing rank order, and identifying an inflection point on this curve. We then added the locations of high priority catchments to the above-mentioned map. To evaluate the sensitivity of the priority conservation areas selection to the choice of conservation goals, we applied the preceding analysis to two alternate rankings of the same set of 500,000 candidate networks. In the first instance, we ranked the networks by ecological representation alone. Then, we ranked them according to their joint functional connectivity under current and future conditions.

Results

Protected areas networks

We constructed 690 unique candidate protected areas satisfying our size and intactness criteria. Their mean size was 86.8 km² (SD = 4.0 km²), only very slightly larger than the target size of 85.5 km². The 500,000 candidate protected areas networks (candidate networks) included a mean of 25 (SD = 2.2) candidate protected areas added to the 62 existing protected areas. The mean candidate network area was 3138.5 km² (SD = 24.6 km²), again slightly larger than the target of 3080 km². Network ecological representativeness scores ranged from 2.76 to 7.63 (mean = 4.71, SD = 0.64) (Fig 3). Network functional connectivity ranged from 4.21e⁺⁰⁵ to 4.79e⁺⁰⁵ (mean = 4.38e⁺⁰⁵, SD = 9.42e⁺⁰³) under current conditions (Fig 3A), and from 4.41e⁺⁰⁵ to 5.00e⁺⁰⁵ (mean = 4.58e⁺⁰⁵, SD = 9.51e⁺⁰³) under mean future conditions (Fig 3B). There was a negative relationship between ecological representativeness and functional connectivity, under both current (slope = -3922, p < 0.001, Fig 3A) and mean future conditions (slope = -3779, p < 0.001, Fig 3B).

Download:

Fig 3.

Scatter plots of ecological representativeness against functional connectivity defined for (a) the current time period and (b) the future average conditions for the 500,000 created networks. The solid lines represent the feature values at quantile Q = 0.925 (5.68 for ecological representativeness, 4.53e⁺⁰⁵ and 4.73e⁺⁰⁵ for current and future functional connectivity, respectively). The dashed lines represent fitted linear regression models of ecological representativeness against the functional connectivity measure plotted. The white dot represents the network proposed by the Québec government.

https://doi.org/10.1371/journal.pone.0238821.g003

The Ministry’s proposed network (Fig 1) had an ecological representativeness score of 4.65 (Fig 3A), which was slightly lower than our sample mean. Its current and future functional connectivity were 4.59e⁺⁰⁵ and 4.76e⁺⁰⁵, respectively, both above our sample means (Fig 3A and 3B). However, our methodology yields many candidate networks that surpass the Ministry’s proposed design in both representativeness and functional connectivity (Fig 3A and 3B). We conclude that the Ministry’s design is suboptimal with respect to these three indicators of network effectiveness.

Priority conservation areas

The quantile Q = 0.925 (92.5%-ile) yielded a subsample of 501 candidate protected areas scoring above the corresponding sample quantiles of all three indicators simultaneously (Fig 3). The ecological representativeness quantile was 5.68. The quantiles for current and future functional connectivity were 4.53 x 10⁵ and 4.73 x 10⁵, respectively. All 501 selected designs had higher ecological representativeness scores than did the Ministry’s. Most of these 501 designs did not exceed the Ministry’s functional connectivity scores, but some did (Fig 3): 78/501 designs had higher functional connectivity under both current and future conditions. That is to say, we identified some designs for the given tradeoff that outperformed the Ministry’s design with respect to all three indicators.

All but one of the 690 candidate protected areas were included in at least one of the 501 selected networks. Selection frequencies were highly skewed (Fig 4). At the inflection point (32,35) on the graph, 32/689 candidate protected areas were included in at least 35/501 selected networks (Fig 4). These 32 priority candidate protected areas were fairly widely distributed over the study region (Fig 5), with some spatial clustering in the southwest of the Gaspésie National Park, and in the areas adjacent to the extreme west of the park (Fig 1), which includes the important high elevation caribou breeding habitats. The spatial distributions of high priority catchments under alternate conservation objectives were markedly different (S3 Material). In designs emphasizing ecological representation, priority areas were more widely distributed, but mostly in the south of the Gaspésie National Park, and none were adjacent to the Park (S3.1 Fig in S3 Material). In designs emphasizing only functional connectivity for caribou, all priority areas were adjacent to the Park (S3.2 Fig in S3 Material).

Download:

Fig 4. A subsample of 501 candidate protected areas networks selected from near one point on the tradeoff curve (Fig 3).

These candidate networks included 689 of the 690 candidate protected areas (CPAs). The selection frequencies (y-axis) of these 689 CPAs are plotted in decreasing rank order (x-axis). Only 32/689 CPAs were included in more than 35/501 selected networks. This inflection point of the curve is indicated by the light grey lines perpendicular to the axes.

https://doi.org/10.1371/journal.pone.0238821.g004

Download:

Fig 5.

The 689 Candidate Protected Areas (CPAs) included in the subsample of 501 near-Pareto-optimal networks are shown in orange scale, coloured by selection frequency, from low (light orange) to high (dark orange). The 32 CPAs included in more than 35/501 networks (Fig 4) are outlined with a thick line. The existing protected areas are shown in black.

https://doi.org/10.1371/journal.pone.0238821.g005

Discussion

We present a methodology to define an effective protected areas network based on the tradeoff between ecological representativeness and functional connectivity over time, including the potential impacts of climate change. Our methodology yields a proposed subset of protected areas that may be close to ideal from an environmental or conservation point of view, but which does not fully respect all constraints of the use of public lands. For example, the Gaspésie region is highly disturbed by human activities. Considering all the social and economic constraints would have reduced too much the area for potential new protected area implementations, giving little space for designing different network scenarios or exploring the limits of what is possible. However, this methodology could easily include more features in the choice of the best network scenarios, and the proposed top protected areas would then be represented as the best tradeoff between all selected features. It would be possible, given data availability, to add constraints in the choice of the networks with, for example, the ecological representativeness of the future landscape under climate change, the functional connectivity of several species important for the ecosystem [68], the economic cost of excluding human activities from the proposed areas, the potential benefit with tourism if protected areas act as parks [69], or any of the other factors affecting the Ministry’s design. Target features and constraints can be defined by local managers to help meet local biodiversity goals, and could easily be implemented in our method to improve the design of regional protected areas networks.

Network ecological representativeness

Gap analyses are a common tool to quantify ecological representativeness of extant networks and to identify underrepresented features [e.g. 8, 70]. Here, we adapted an alternate two-stage approach that was recently developed in Canada [31, 32] to support systematic conservation planning in the boreal region. The first stage of that method used geospatial analysis tools to construct potential or candidate protected areas, which were then assembled into networks. A multivariate distribution matching methodology identified networks that minimized gaps. Gaps were quantified as dissimilarities, with respect to the distributions of a chosen set of environmental covariates, between the network and the rest of the study region.

Network functional connectivity

Many studies have included connectivity in conservation planning using measures of distance or the cost of movement between protected areas [71] or applying graph and circuit theory techniques [72, 73]. A novel feature of this study was that we derived network functional connectivity measures from movement simulations using an individual-based model [38, 61] that reproduced known characteristics of the focal population within the study region. The model accounted for process-based complex movement behaviors (e.g., seasonal site fidelity, foray loop movement) that would be difficult to represent in static habitat-based models. Being process-based, the models reflect behavioral responses to environmental conditions and should be preferred to make predictions under future conditions compared to models based on only observed habitat where the suppositions underlying current empirical relationships may no longer hold [74].

Accounting for climate change

It is more efficient to take future climate change impacts into account now instead of reacting to them only once changes occur [28]. To account for climate change, we defined the future functional connectivity criteria using simulated movements on hypothetical landscapes resulting from different climate change scenarios. Network effectiveness will likely depend on the realized landscape outcome which is presently unknown. Therefore, networks selected using an average functional connectivity over a wide range of climate change impacts would be suboptimal for any particular climate change scenario. However, they might perform reasonably well under a large range of possible future environmental conditions. The differences in movement predicted for the different climate change scenarios resulted from the assumptions made about the future environmental conditions. We used this relatively simple approach as the functional connectivity differences among the scenarios were not particularly large. This suggests that using coupled climate change and vegetation change dynamic models may not dramatically improve the choice of protected areas in this region. Furthermore, since no climate-sensitive model of vegetation dynamics was available for Gaspésie natural region, using the scenario approach with averaging of results seemed to be a reasonable compromise to account for uncertainty in future conditions. Our approach could readily be adapted to cases with much greater divergence among expected future conditions, at the price of increasing the dimensionality of the tradeoff surface. Management decisions in that case would ideally be based on weighting the relative costs of possible solutions and the probabilities associated with each alternate outcome.

Case study

The lack of functional connectivity during the last 20 years between the different major mountain summits of the Gaspésie National Park has led to a division of the caribou population into two genetically distinct sub-populations and is now jeopardizing the persistence of this isolated population [75]. Consequently, using an approach that could identify key elements to preserve in order to maintain functional connectivity could benefit the conservation of an endangered caribou population, especially under a changing climate.

In the Gaspésie region, there is only one caribou population occurring primarily inside the Gaspésie National Park [37, 48]. Consequently, in our analyses, networks with high functional connectivity tended to include many protected areas adjacent to or near this park. On the other hand, to achieve high ecological representativeness, protected areas needed to be more or less evenly distributed over the entire region to capture habitat diversity. There was therefore a negative relationship between representation and connectivity in our case study, as others have found [e.g. 76]. It is, in general, not possible to simultaneously optimize for multiple design criteria. Our sampling-based design methodology allows us to approximate a multidimensional tradeoff surface. Given the management decision about the choice of tradeoff, we can identify a large number of solutions that are close to this point. Our methodology showed how to design new protected areas for a regional network that satisfy multiple, divergent criteria to (nearly) the highest degree possible. Other tradeoffs between the two objectives of representation and caribou conservation could be defined. The two extremes (S3 Material), in which only one criterion is emphasised, produce very different spatial distributions of priority candidate protected areas. Our chosen tradeoff produced an intermediate result, indicating that new protected areas in the south of the study region are needed to improve ecological representation, and that an expansion of Gaspésie National Park would best improve functional connectivity for caribou. However, the designs considered in Fig 5 are not intended or represented as optimal for caribou conservation per se.

The protected areas network expansion proposed by the Québec government to achieve the 12% coverage target that we used as a comparison was different than the sample of candidate protected areas suggested from the selected best networks identified here (Figs 1 and 5). The Ministry’s network achieved a lower ecological representativeness than our selected networks. However, they had to respect design criteria and constraints, like socio-economic issues or the inclusion of rare ecosystems that we did not consider. This could explain the suboptimal ecological representativeness achieved by their network. Regarding functional connectivity, their network seems well connected from the Atlantic-Gaspésie caribou point of view. The performance of their scenario is surprisingly quite good for this feature, considering it was not explicitly part of their design. The networks we selected may not achieve such a high functional connectivity because our specific choice of intactness criteria resulted in some areas being excluded from any of our network solutions. In particular, we excluded some non-intact areas close to the Gaspésie National Park that were included in the Ministry’s design.

Supporting information

S1 Material. Construction of the potential future landscapes.

https://doi.org/10.1371/journal.pone.0238821.s001

(DOCX)

S2 Material. Caribou movements in the current and future landscapes.

https://doi.org/10.1371/journal.pone.0238821.s002

(DOCX)

S3 Material. Description of the optimal protected areas networks.

https://doi.org/10.1371/journal.pone.0238821.s003

(DOCX)

Acknowledgments

We thank the Canadian BEACONs Project (http://www.beaconsproject.ca) for developing and sharing methods and software for protected areas design, and for providing the custom hydrological catchments. We also thank BEACONs members M. Houle, K Lisgo and P. Vernier for their help with the use of the BEACONs toolkit. We thank the Ministère de l’Environnement et de la Lutte aux Changements Climatiques du Québec (D. Boisjoly, F. Brassard and S. Benoit) and the Ministère des Forêts, de la Faune et des Parcs du Québec for the data and documents provided. We finally thank K. Malcolm and three anonymous reviewers for their constructive comments on an earlier version of this manuscript.

References

1. Secretariat of the Convention on Biological Diversity. Global biodiversity outlook 2. Montreal, Canada: Secretariat of the Convention on Biological Diversity; 2006.
2. Tylianakis JM, Laliberté E, Nielsen A, Bascompte J. Conservation of species interaction networks. Biol Conserv. 2010; 143(10):2270–2279.
- View Article
- Google Scholar
3. Spooner FE, Pearson RG, Freeman R. Rapid warming is associated with population decline among terrestrial birds and mammals globally. Glob Chang Biol. 2018; 24(10):4521–4531. pmid:30033551
- View Article
- PubMed/NCBI
- Google Scholar
4. Coetzee BW, Gaston KJ, Chown SL. Local scale comparisons of biodiversity as a test for global protected area ecological performance: a meta-analysis. PloS One. 2014; 9(8):e105824. pmid:25162620
- View Article
- PubMed/NCBI
- Google Scholar
5. Ervin J. Protected area assessments in perspective. Bioscience. 2003; 53(9):819–822.
- View Article
- Google Scholar
6. Geldmann J, Barnes M, Coad L, Craigie ID, Hockings M, Burgess ND. Effectiveness of terrestrial protected areas in reducing habitat loss and population declines. Biol Conserv. 2013; 161:230–238.
- View Article
- Google Scholar
7. Gaston KJ, Jackson SF, Cantú-Salazar L, Cruz-Piñón G. The ecological performance of protected areas. Annu Rev Ecol Evol Syst. 2008; 39:93–113.
- View Article
- Google Scholar
8. Rodrigues AS, Andelman SJ, Bakarr MI, Boitani L, Brooks TM, Cowling RM, et al. Effectiveness of the global protected area network in representing species diversity. Nature. 2004; 428(6983):640. pmid:15071592
- View Article
- PubMed/NCBI
- Google Scholar
9. Stewart FE, Darlington S, Volpe JP, McAdie M, Fisher JT. Corridors best facilitate functional connectivity across a protected area network. Sci Rep. 2019; 9(1):1–9. pmid:30626917
- View Article
- PubMed/NCBI
- Google Scholar
10. Nuñez TA, Lawler JJ, McRae BH, Pierce DJ, Krosby MB, Kavanagh DM, et al. Connectivity planning to address climate change. Conserv Biol. 2013; 27(2):407–416. pmid:23410037
- View Article
- PubMed/NCBI
- Google Scholar
11. Dietz MS, Belote RT, Aplet GH, Aycrigg JL. The world’s largest wilderness protection network after 50 years: An assessment of ecological system representation in the US National Wilderness Preservation System. Biol Cons. 2015; 184:431–438.
- View Article
- Google Scholar
12. Beaver D, Llewellyn G. Designing a comprehensive, adequate and representative (CAR) network of marine protected areas for Australia’s Commonwealth waters. Ultimo, Australia: WWF; 2009 Feb.
13. Kukkala AS, Moilanen A. Core concepts of spatial prioritisation in systematic conservation planning. Biol Rev Camb Philos Soc. 2013; 88(2):443–464. pmid:23279291
- View Article
- PubMed/NCBI
- Google Scholar
14. Scott JM, Davis FW, McGhie RG, Wright RG, Groves C, Estes J. Nature reserves: Do they capture the full range of America’s biological diversity? Ecol Appl. 2001; 11(4):999–1007.
- View Article
- Google Scholar
15. Joppa LN, Pfaff A. High and far: biases in the location of protected areas. PloS One. 2009; 4(12):e8273. pmid:20011603
- View Article
- PubMed/NCBI
- Google Scholar
16. Asaad I, Lundquist CJ, Erdmann MV, Costello MJ. Ecological criteria to identify areas for biodiversity conservation. Biol Conserv. 2017; 213:309–316.
- View Article
- Google Scholar
17. Venter O, Fuller RA, Segan DB, Carwardine J, Brooks T, Butchart SHM, Di Marco M et al. Targeting global protected area expansion for imperiled biodiversity. PLoS Biology 2014; 12(6):e100189.
- View Article
- Google Scholar
18. Taylor PD, Fahrig L, Henein K, Merriam G. Connectivity is a vital element of landscape structure. Oikos. 1993; 571–573.
- View Article
- Google Scholar
19. Baguette M, Van Dyck H. Landscape connectivity and animal behavior: functional grain as a key determinant for dispersal. Landsc Ecol. 2007; 22(8):1117–1129.
- View Article
- Google Scholar
20. Kadoya T. Assessing functional connectivity using empirical data. Popul Ecol. 2009; 51(1):5–15.
- View Article
- Google Scholar
21. Andrello M, Jacobi MN, Manel S, Thuiller W, Mouillot D. Extending networks of protected areas to optimize connectivity and population growth rate. Ecography. 2015; 38(3):273–282.
- View Article
- Google Scholar
22. Urban MC. Accelerating extinction risk from climate change. Science. 2015; 348(6234):571–573. pmid:25931559
- View Article
- PubMed/NCBI
- Google Scholar
23. Jantz SM, Barker B, Brooks TM, Chini LP, Huang Q, Moore RM, et al. Future habitat loss and extinctions driven by land-use change in biodiversity hotspots under four scenarios of climate-change mitigation. Conserv Biol. 2015; 29(4):1122–1131. pmid:26129841
- View Article
- PubMed/NCBI
- Google Scholar
24. Loarie SR, Duffy PB, Hamilton H, Asner GP, Field CB, Ackerly DD. The velocity of climate change. Nature. 2009; 462(7276):1052. pmid:20033047
- View Article
- PubMed/NCBI
- Google Scholar
25. Dawson TP, Jackson ST, House JI, Prentice IC, Mace GM. Beyond predictions: biodiversity conservation in a changing climate. Science. 2011; 332(6025):53–58. pmid:21454781
- View Article
- PubMed/NCBI
- Google Scholar
26. Chen I-C, Hill JK, Ohlemüller R, Roy DB, Thomas CD. Rapid range shifts of species associated with high levels of climate warming. Science. 2011; 333(6045):1024–1026. pmid:21852500
- View Article
- PubMed/NCBI
- Google Scholar
27. Lawler JJ, Ruesch AS, Olden JD, McRae BH. Projected climate-driven faunal movement routes. Ecol Lett. 2013; 16(8):1014–1022. pmid:23782906
- View Article
- PubMed/NCBI
- Google Scholar
28. Hannah L, Midgley G, Andelman S, Araújo M, Hughes G, Martinez-Meyer E, et al. Protected area needs in a changing climate. Front Ecol Environ. 2007; 5(3):131–138.
- View Article
- Google Scholar
29. Vos CC, Berry P, Opdam P, Baveco H, Nijhof B, O’Hanley J, et al. Adapting landscapes to climate change: examples of climate-proof ecosystem networks and priority adaptation zones. J Appl Ecol. 2008; 45(6):1722–1731.
- View Article
- Google Scholar
30. Klein C, Wilson K, Watts M, Stein J, Berry S, Carwardine J, et al. Incorporating ecological and evolutionary processes into continental-scale conservation planning. Ecol Appl. 2009; 19(1):206–217. pmid:19323184
- View Article
- PubMed/NCBI
- Google Scholar
31. Schmiegelow F, Lisgo K, Anderson L, Cumming S, Krawchuk M, Leroux S, et al. Canadian BEACONs Project 2009–10 Report. University of Alberta, Edmonton; 2010. Available online: www.beaconsproject.ca/documents.
32. Saucier F. Développement d’une approche de planification systématique pour l’élaboration de réseaux de conservation représentatifs faits d’aires protégées qui visent la persistance de la biodiversité à long terme: étude de cas au Moyen Nord du Québec. M.Sc. Thesis, Université Laval; 2011. Available from: https://corpus.ulaval.ca/jspui/handle/20.500.11794/22949.
33. Calkin DE, Montgomery CA, Schumaker NH, Polasky S, Arthur JL, Nalle DJ. Developing a production possibility set of wildlife species persistence and timber harvest value. Can J For Res. 2002; 32(8):1329–1342.
- View Article
- Google Scholar
34. Arthur JL, Camm JD, Haight RG, Montgomery CA, Polasky S. Weighing conservation objectives: maximum expected coverage versus endangered species protection. Ecol Appl. 2004; 14(6):1936–1945.
- View Article
- Google Scholar
35. Kennedy MC, Ford ED, Singleton P, Finney M, Agee JK. Informed multi‐objective decision‐making in environmental management using Pareto optimality. J Appl Ecol. 2008; 45(1):181–192.
- View Article
- Google Scholar
36. COSEWIC. Designatable units for caribou (Rangifer tarandus) in Canada. Ottawa, Canada: Committee on the Status of Endangered Wildlife in Canada; 2011.
37. Morin M. Inventaire aérien de la population de caribous de la Gaspésie (Rangifer tarandus caribou)–Automne 2018. Ministère des Forêts, de la Faune et des Parcs; 2018.
38. Bauduin S, McIntire E, St-Laurent M-H, Cumming S. Overcoming challenges of sparse telemetry data to estimate caribou movement. Ecol Modell. 2016; 335:24–34.
- View Article
- Google Scholar
39. Bouchard M-A. Portrait du réseau d’aires protégées au Québec—Analyse de carence écorégionale, Gaspésie (Terrestre). Ministère du Développement Durable de l’Environnement et de la Lutte aux Changements Climatiques; 2014.
- View Article
- Google Scholar
40. Saucier J-P, Grondin P, Robitaille A, Bergeron J-F. Zones de végétation et les domaines bioclimatiques du Québec. Ministère des ressources naturelles,de la Faune et des Parcs; 2003.
41. Brassard F, Bouchard AR, Boisjoly D, Poisson F, Bazoge A, Bouchard M-A, et al. Portrait du réseau d’aires protégées au Québec: période 2002–2009. Ministère du Développement durable, de l’Environnement et des Parcs du Québec; 2010.
42. LeSage L, Paquin P. Historique, géographie physique et biogéographie du Parc de conservation de la Gaspésie, Québec. Proc Ent Soc Ont. 2000; 131:17–66.
- View Article
- Google Scholar
43. Species at Risk Public Registry [Internet]. Environment Canada; c2019 [cited 2019 06 18]. Available from: http://www.sararegistry.gc.ca/.
44. Lesmerises F, Déry F, Johnson CJ, St-Laurent M-H. Spatiotemporal response of mountain caribou to the intensity of backcountry skiing. Biol Conserv. 2018; 217:149–156.
- View Article
- Google Scholar
45. Boudreau M. Impacts de 25 ans d’aménagement forestier intensif sur l’habitat du caribou de la Gaspésie et de ses prédateurs. M.Sc. Thesis, Université du Québec à Rimouski; 2017. Available from: http://semaphore.uqar.ca/1289/.
46. Frenette J, Pelletier F, St-Laurent M-H. Linking habitat, predators and alternative prey to explain recruitment variations of an endangered caribou population. Glob Ecol Conserv. 2020; 22:e00920.
- View Article
- Google Scholar
47. Ouellet J-P, Ferron J, Sirois L. Space and habitat use by the threatened Gaspé caribou in southeastern Quebec. Can J Zool. 1996; 74(10):1922–1933.
- View Article
- Google Scholar
48. Mosnier A, Ouellet J-P, Sirois L, Fournier N. Habitat selection and home-range dynamics of the Gaspé caribou: a hierarchical analysis. Can J Zool. 2003; 81(7):1174–1184.
- View Article
- Google Scholar
49. Mosnier A, Boisjoly D, Courtois R, Ouellet J-P. Extensive predator space use can limit the efficacy of a control program. J Wildl Manage. 2008; 72(2):483–491.
- View Article
- Google Scholar
50. Dumais C, Ropars P, Denis M-P, Dufour-Tremblay G, Boudreau S. Are low altitude alpine tundra ecosystems under threat? A case study from the Parc National de la Gaspésie, Québec. Environ Res Lett. 2014; 9(9):094001.
- View Article
- Google Scholar
51. Renard S, Isabel C, McIntire E. Le vent, un facteur déterminant pour le maintien de la toundra alpine. In: Bulletin de conservation 2014–2015. Québec, Canada : Société des Établissements de Plein Air du Québec; 2015. p. 27–30
52. Périé C, de Blois S, Lambert M-C, Casajus N. Effets anticipés des changements climatiques sur l’habitat des espèces arborescentes au Québec. Ministère des Ressources naturelles du Québec; 2014.
53. Linke S, Norris RH, Pressey RL. Irreplaceability of river networks: towards catchment-based conservation planning. J Appl Ecol. 2008; 45(5):1486–1495.
- View Article
- Google Scholar
54. Hermoso V, Filipe AF, Segurado P, Beja P. Effectiveness of a large reserve network in protecting freshwater biodiversity: a test for the Iberian Peninsula. Freshw Biol. 2015; 60(4):698–710.
- View Article
- Google Scholar
55. Adams VM, Setterfield SA, Douglas MM, Kennard MJ, Ferdinands K. Measuring benefits of protected area management: trends across realms and research gaps for freshwater systems. Philos Trans R Soc Lond B. 2015; 370(1681):20140274.
- View Article
- Google Scholar
56. BEACONs. Quebec Catchments 1:50,000 Version 1. BEACONs Project, University of Alberta, Edmonton, AB. http://beaconsproject.ca/
57. Toews M, Juanes F, Burton AC. Mammal responses to the human footprint vary across species and stressors. J Environ Manage. 2018; 217:690–699. pmid:29654972
- View Article
- PubMed/NCBI
- Google Scholar
58. Gaudry W. Impacts des structures anthropiques linéaires sur la sélection d’habitat du caribou, de l’ours noir et du coyote en Gaspésie. MSc Thesis, Université du Québec à Rimouski; 2013.
59. Lowe WH, Likens GE. Moving headwater streams to the head of the class. BioScience. 2005; 55(3):196–197
- View Article
- Google Scholar
60. Polak T, Watson JE, Bennett JR, Possingham HP, Fuller RA, Carwardine J. Balancing ecosystem and threatened species representation in protected areas and implications for nations achieving global conservation goals. Conserv Lett. 2016; 9(6):438–445.
- View Article
- Google Scholar
61. Bauduin S, McIntire E, St-Laurent M-H, Cumming SG. Compensatory conservation measures for an endangered caribou population under climate change. Sci Rep. 2018; 8(1):16438. pmid:30401921
- View Article
- PubMed/NCBI
- Google Scholar
62. Manly BFJ, MacDonald LL, Thomas DL, MacDonald TL, Erickson WP. Resource selection by animals: Statistical analysis and design for field studies. Dordrecht, The Netherlands; 2002.
63. Boyce MS, Vernier PR, Nielsen SE, Schmiegelow FK. Evaluating resource selection functions. Ecol Modell. 2002; 157(2–3):281–300.
- View Article
- Google Scholar
64. Lesmerises F, Johnson CJ, St-Laurent M-H. Effect of behavioral marginality on survival of an alpine ungulate. Ecosphere. 2019; 10(7):e02806.
- View Article
- Google Scholar
65. Logan T. Scénarios climatiques pour les régions naturelles de la péninsule de la Gaspésie et la dépression de La Tuque. Consortium Ouranos; 2012.
66. Gray DR. The relationship between climate and outbreak characteristics of the spruce budworm in eastern Canada. Clim Change. 2008; 87(3–4):361–383.
- View Article
- Google Scholar
67. Régnière J, St-Amant R, Duval P. Predicting insect distributions under climate change from physiological responses: spruce budworm as an example. Biol Invasions. 2012; 14(8):1571–1586.
- View Article
- Google Scholar
68. Minor ES, Urban DL. A graph-theory framework for evaluating landscape connectivity and conservation planning. Conserv Biol. 2008; 22(2):297–307. pmid:18241238
- View Article
- PubMed/NCBI
- Google Scholar
69. Mayer M. Can nature-based tourism benefits compensate for the costs of national parks? A study of the Bavarian Forest National Park, Germany. J Sustain Tourism. 2014; 22(4):561–583.
- View Article
- Google Scholar
70. Wiersma YF, Nudds TD. Efficiency and effectiveness in representative reserve design in Canada: the contribution of existing protected areas. Biol Conserv. 2009; 142(8):1639–1646.
- View Article
- Google Scholar
71. Sundblad G, Bergström U, Sandström A. Ecological coherence of marine protected area networks: a spatial assessment using species distribution models. J Appl Ecol. 2011; 48(1):112–120.
- View Article
- Google Scholar
72. Baldwin RF, Perkl RM, Trombulak SC, Burwell WB. Modeling ecoregional connectivity. In: Trombulak SC, Baldwin RF, editors. Landscape-scale conservation planning. Dordrecht, Netherlands: Springer; 2010. p. 349–367.
73. Rayfield B, Fortin M-J, Fall A. Connectivity for conservation: a framework to classify network measures. Ecology. 2011; 92(4):847–858. pmid:21661548
- View Article
- PubMed/NCBI
- Google Scholar
74. Bélisle M. Measuring landscape connectivity: the challenge of behavioral landscape ecology. Ecology. 2005; 86(8):1988–1995.
- View Article
- Google Scholar
75. Pelletier F, Turgeon G, Bourret A, Garant D, St-Laurent M-H. Genetic structure and effective size of an endangered population of woodland caribou. Conserv Genet. 2019; 20(2):203–213.
- View Article
- Google Scholar
76. Almany GR, Connolly SR, Heath DD, Hogan JD, Jones GP, McCook LJ, et al. Connectivity, biodiversity conservation and the design of marine reserve networks for coral reefs. Coral Reefs. 2009; 28(2):339–351.
- View Article
- Google Scholar

[ref1] 1. Secretariat of the Convention on Biological Diversity. Global biodiversity outlook 2. Montreal, Canada: Secretariat of the Convention on Biological Diversity; 2006.

[ref2] 2. Tylianakis JM, Laliberté E, Nielsen A, Bascompte J. Conservation of species interaction networks. Biol Conserv. 2010; 143(10):2270–2279.
View Article
Google Scholar

[3] View Article

[4] Google Scholar

[ref3] 3. Spooner FE, Pearson RG, Freeman R. Rapid warming is associated with population decline among terrestrial birds and mammals globally. Glob Chang Biol. 2018; 24(10):4521–4531. pmid:30033551
View Article
PubMed/NCBI
Google Scholar

[6] View Article

[7] PubMed/NCBI

[8] Google Scholar

[ref4] 4. Coetzee BW, Gaston KJ, Chown SL. Local scale comparisons of biodiversity as a test for global protected area ecological performance: a meta-analysis. PloS One. 2014; 9(8):e105824. pmid:25162620
View Article
PubMed/NCBI
Google Scholar

[10] View Article

[11] PubMed/NCBI

[12] Google Scholar

[ref5] 5. Ervin J. Protected area assessments in perspective. Bioscience. 2003; 53(9):819–822.
View Article
Google Scholar

[14] View Article

[15] Google Scholar

[ref6] 6. Geldmann J, Barnes M, Coad L, Craigie ID, Hockings M, Burgess ND. Effectiveness of terrestrial protected areas in reducing habitat loss and population declines. Biol Conserv. 2013; 161:230–238.
View Article
Google Scholar

[17] View Article

[18] Google Scholar

[ref7] 7. Gaston KJ, Jackson SF, Cantú-Salazar L, Cruz-Piñón G. The ecological performance of protected areas. Annu Rev Ecol Evol Syst. 2008; 39:93–113.
View Article
Google Scholar

[20] View Article

[21] Google Scholar

[ref8] 8. Rodrigues AS, Andelman SJ, Bakarr MI, Boitani L, Brooks TM, Cowling RM, et al. Effectiveness of the global protected area network in representing species diversity. Nature. 2004; 428(6983):640. pmid:15071592
View Article
PubMed/NCBI
Google Scholar

[23] View Article

[24] PubMed/NCBI

[25] Google Scholar

[ref9] 9. Stewart FE, Darlington S, Volpe JP, McAdie M, Fisher JT. Corridors best facilitate functional connectivity across a protected area network. Sci Rep. 2019; 9(1):1–9. pmid:30626917
View Article
PubMed/NCBI
Google Scholar

[27] View Article

[28] PubMed/NCBI

[29] Google Scholar

[ref10] 10. Nuñez TA, Lawler JJ, McRae BH, Pierce DJ, Krosby MB, Kavanagh DM, et al. Connectivity planning to address climate change. Conserv Biol. 2013; 27(2):407–416. pmid:23410037
View Article
PubMed/NCBI
Google Scholar

[31] View Article

[32] PubMed/NCBI

[33] Google Scholar

[ref11] 11. Dietz MS, Belote RT, Aplet GH, Aycrigg JL. The world’s largest wilderness protection network after 50 years: An assessment of ecological system representation in the US National Wilderness Preservation System. Biol Cons. 2015; 184:431–438.
View Article
Google Scholar

[35] View Article

[36] Google Scholar

[ref12] 12. Beaver D, Llewellyn G. Designing a comprehensive, adequate and representative (CAR) network of marine protected areas for Australia’s Commonwealth waters. Ultimo, Australia: WWF; 2009 Feb.

[ref13] 13. Kukkala AS, Moilanen A. Core concepts of spatial prioritisation in systematic conservation planning. Biol Rev Camb Philos Soc. 2013; 88(2):443–464. pmid:23279291
View Article
PubMed/NCBI
Google Scholar

[39] View Article

[40] PubMed/NCBI

[41] Google Scholar

[ref14] 14. Scott JM, Davis FW, McGhie RG, Wright RG, Groves C, Estes J. Nature reserves: Do they capture the full range of America’s biological diversity? Ecol Appl. 2001; 11(4):999–1007.
View Article
Google Scholar

[43] View Article

[44] Google Scholar

[ref15] 15. Joppa LN, Pfaff A. High and far: biases in the location of protected areas. PloS One. 2009; 4(12):e8273. pmid:20011603
View Article
PubMed/NCBI
Google Scholar

[46] View Article

[47] PubMed/NCBI

[48] Google Scholar

[ref16] 16. Asaad I, Lundquist CJ, Erdmann MV, Costello MJ. Ecological criteria to identify areas for biodiversity conservation. Biol Conserv. 2017; 213:309–316.
View Article
Google Scholar

[50] View Article

[51] Google Scholar

[ref17] 17. Venter O, Fuller RA, Segan DB, Carwardine J, Brooks T, Butchart SHM, Di Marco M et al. Targeting global protected area expansion for imperiled biodiversity. PLoS Biology 2014; 12(6):e100189.
View Article
Google Scholar

[53] View Article

[54] Google Scholar

[ref18] 18. Taylor PD, Fahrig L, Henein K, Merriam G. Connectivity is a vital element of landscape structure. Oikos. 1993; 571–573.
View Article
Google Scholar

[56] View Article

[57] Google Scholar

[ref19] 19. Baguette M, Van Dyck H. Landscape connectivity and animal behavior: functional grain as a key determinant for dispersal. Landsc Ecol. 2007; 22(8):1117–1129.
View Article
Google Scholar

[59] View Article

[60] Google Scholar

[ref20] 20. Kadoya T. Assessing functional connectivity using empirical data. Popul Ecol. 2009; 51(1):5–15.
View Article
Google Scholar

[62] View Article

[63] Google Scholar

[ref21] 21. Andrello M, Jacobi MN, Manel S, Thuiller W, Mouillot D. Extending networks of protected areas to optimize connectivity and population growth rate. Ecography. 2015; 38(3):273–282.
View Article
Google Scholar

[65] View Article

[66] Google Scholar

[ref22] 22. Urban MC. Accelerating extinction risk from climate change. Science. 2015; 348(6234):571–573. pmid:25931559
View Article
PubMed/NCBI
Google Scholar

[68] View Article

[69] PubMed/NCBI

[70] Google Scholar

[ref23] 23. Jantz SM, Barker B, Brooks TM, Chini LP, Huang Q, Moore RM, et al. Future habitat loss and extinctions driven by land-use change in biodiversity hotspots under four scenarios of climate-change mitigation. Conserv Biol. 2015; 29(4):1122–1131. pmid:26129841
View Article
PubMed/NCBI
Google Scholar

[72] View Article

[73] PubMed/NCBI

[74] Google Scholar

[ref24] 24. Loarie SR, Duffy PB, Hamilton H, Asner GP, Field CB, Ackerly DD. The velocity of climate change. Nature. 2009; 462(7276):1052. pmid:20033047
View Article
PubMed/NCBI
Google Scholar

[76] View Article

[77] PubMed/NCBI

[78] Google Scholar

[ref25] 25. Dawson TP, Jackson ST, House JI, Prentice IC, Mace GM. Beyond predictions: biodiversity conservation in a changing climate. Science. 2011; 332(6025):53–58. pmid:21454781
View Article
PubMed/NCBI
Google Scholar

[80] View Article

[81] PubMed/NCBI

[82] Google Scholar

[ref26] 26. Chen I-C, Hill JK, Ohlemüller R, Roy DB, Thomas CD. Rapid range shifts of species associated with high levels of climate warming. Science. 2011; 333(6045):1024–1026. pmid:21852500
View Article
PubMed/NCBI
Google Scholar

[84] View Article

[85] PubMed/NCBI

[86] Google Scholar

[ref27] 27. Lawler JJ, Ruesch AS, Olden JD, McRae BH. Projected climate-driven faunal movement routes. Ecol Lett. 2013; 16(8):1014–1022. pmid:23782906
View Article
PubMed/NCBI
Google Scholar

[88] View Article

[89] PubMed/NCBI

[90] Google Scholar

[ref28] 28. Hannah L, Midgley G, Andelman S, Araújo M, Hughes G, Martinez-Meyer E, et al. Protected area needs in a changing climate. Front Ecol Environ. 2007; 5(3):131–138.
View Article
Google Scholar

[92] View Article

[93] Google Scholar

[ref29] 29. Vos CC, Berry P, Opdam P, Baveco H, Nijhof B, O’Hanley J, et al. Adapting landscapes to climate change: examples of climate-proof ecosystem networks and priority adaptation zones. J Appl Ecol. 2008; 45(6):1722–1731.
View Article
Google Scholar

[95] View Article

[96] Google Scholar

[ref30] 30. Klein C, Wilson K, Watts M, Stein J, Berry S, Carwardine J, et al. Incorporating ecological and evolutionary processes into continental-scale conservation planning. Ecol Appl. 2009; 19(1):206–217. pmid:19323184
View Article
PubMed/NCBI
Google Scholar

[98] View Article

[99] PubMed/NCBI

[100] Google Scholar

[ref31] 31. Schmiegelow F, Lisgo K, Anderson L, Cumming S, Krawchuk M, Leroux S, et al. Canadian BEACONs Project 2009–10 Report. University of Alberta, Edmonton; 2010. Available online: www.beaconsproject.ca/documents.

[ref32] 32. Saucier F. Développement d’une approche de planification systématique pour l’élaboration de réseaux de conservation représentatifs faits d’aires protégées qui visent la persistance de la biodiversité à long terme: étude de cas au Moyen Nord du Québec. M.Sc. Thesis, Université Laval; 2011. Available from: https://corpus.ulaval.ca/jspui/handle/20.500.11794/22949.

[ref33] 33. Calkin DE, Montgomery CA, Schumaker NH, Polasky S, Arthur JL, Nalle DJ. Developing a production possibility set of wildlife species persistence and timber harvest value. Can J For Res. 2002; 32(8):1329–1342.
View Article
Google Scholar

[104] View Article

[105] Google Scholar

[ref34] 34. Arthur JL, Camm JD, Haight RG, Montgomery CA, Polasky S. Weighing conservation objectives: maximum expected coverage versus endangered species protection. Ecol Appl. 2004; 14(6):1936–1945.
View Article
Google Scholar

[107] View Article

[108] Google Scholar

[ref35] 35. Kennedy MC, Ford ED, Singleton P, Finney M, Agee JK. Informed multi‐objective decision‐making in environmental management using Pareto optimality. J Appl Ecol. 2008; 45(1):181–192.
View Article
Google Scholar

[110] View Article

[111] Google Scholar

[ref36] 36. COSEWIC. Designatable units for caribou (Rangifer tarandus) in Canada. Ottawa, Canada: Committee on the Status of Endangered Wildlife in Canada; 2011.

[ref37] 37. Morin M. Inventaire aérien de la population de caribous de la Gaspésie (Rangifer tarandus caribou)–Automne 2018. Ministère des Forêts, de la Faune et des Parcs; 2018.

[ref38] 38. Bauduin S, McIntire E, St-Laurent M-H, Cumming S. Overcoming challenges of sparse telemetry data to estimate caribou movement. Ecol Modell. 2016; 335:24–34.
View Article
Google Scholar

[115] View Article

[116] Google Scholar

[ref39] 39. Bouchard M-A. Portrait du réseau d’aires protégées au Québec—Analyse de carence écorégionale, Gaspésie (Terrestre). Ministère du Développement Durable de l’Environnement et de la Lutte aux Changements Climatiques; 2014.
View Article
Google Scholar

[118] View Article

[119] Google Scholar

[ref40] 40. Saucier J-P, Grondin P, Robitaille A, Bergeron J-F. Zones de végétation et les domaines bioclimatiques du Québec. Ministère des ressources naturelles,de la Faune et des Parcs; 2003.

[ref41] 41. Brassard F, Bouchard AR, Boisjoly D, Poisson F, Bazoge A, Bouchard M-A, et al. Portrait du réseau d’aires protégées au Québec: période 2002–2009. Ministère du Développement durable, de l’Environnement et des Parcs du Québec; 2010.

[ref42] 42. LeSage L, Paquin P. Historique, géographie physique et biogéographie du Parc de conservation de la Gaspésie, Québec. Proc Ent Soc Ont. 2000; 131:17–66.
View Article
Google Scholar

[123] View Article

[124] Google Scholar

[ref43] 43. Species at Risk Public Registry [Internet]. Environment Canada; c2019 [cited 2019 06 18]. Available from: http://www.sararegistry.gc.ca/.

[ref44] 44. Lesmerises F, Déry F, Johnson CJ, St-Laurent M-H. Spatiotemporal response of mountain caribou to the intensity of backcountry skiing. Biol Conserv. 2018; 217:149–156.
View Article
Google Scholar

[127] View Article

[128] Google Scholar

[ref45] 45. Boudreau M. Impacts de 25 ans d’aménagement forestier intensif sur l’habitat du caribou de la Gaspésie et de ses prédateurs. M.Sc. Thesis, Université du Québec à Rimouski; 2017. Available from: http://semaphore.uqar.ca/1289/.

[ref46] 46. Frenette J, Pelletier F, St-Laurent M-H. Linking habitat, predators and alternative prey to explain recruitment variations of an endangered caribou population. Glob Ecol Conserv. 2020; 22:e00920.
View Article
Google Scholar

[131] View Article

[132] Google Scholar

[ref47] 47. Ouellet J-P, Ferron J, Sirois L. Space and habitat use by the threatened Gaspé caribou in southeastern Quebec. Can J Zool. 1996; 74(10):1922–1933.
View Article
Google Scholar

[134] View Article

[135] Google Scholar

[ref48] 48. Mosnier A, Ouellet J-P, Sirois L, Fournier N. Habitat selection and home-range dynamics of the Gaspé caribou: a hierarchical analysis. Can J Zool. 2003; 81(7):1174–1184.
View Article
Google Scholar

[137] View Article

[138] Google Scholar

[ref49] 49. Mosnier A, Boisjoly D, Courtois R, Ouellet J-P. Extensive predator space use can limit the efficacy of a control program. J Wildl Manage. 2008; 72(2):483–491.
View Article
Google Scholar

[140] View Article

[141] Google Scholar

[ref50] 50. Dumais C, Ropars P, Denis M-P, Dufour-Tremblay G, Boudreau S. Are low altitude alpine tundra ecosystems under threat? A case study from the Parc National de la Gaspésie, Québec. Environ Res Lett. 2014; 9(9):094001.
View Article
Google Scholar

[143] View Article

[144] Google Scholar

[ref51] 51. Renard S, Isabel C, McIntire E. Le vent, un facteur déterminant pour le maintien de la toundra alpine. In: Bulletin de conservation 2014–2015. Québec, Canada : Société des Établissements de Plein Air du Québec; 2015. p. 27–30

[ref52] 52. Périé C, de Blois S, Lambert M-C, Casajus N. Effets anticipés des changements climatiques sur l’habitat des espèces arborescentes au Québec. Ministère des Ressources naturelles du Québec; 2014.

[ref53] 53. Linke S, Norris RH, Pressey RL. Irreplaceability of river networks: towards catchment-based conservation planning. J Appl Ecol. 2008; 45(5):1486–1495.
View Article
Google Scholar

[148] View Article

[149] Google Scholar

[ref54] 54. Hermoso V, Filipe AF, Segurado P, Beja P. Effectiveness of a large reserve network in protecting freshwater biodiversity: a test for the Iberian Peninsula. Freshw Biol. 2015; 60(4):698–710.
View Article
Google Scholar

[151] View Article

[152] Google Scholar

[ref55] 55. Adams VM, Setterfield SA, Douglas MM, Kennard MJ, Ferdinands K. Measuring benefits of protected area management: trends across realms and research gaps for freshwater systems. Philos Trans R Soc Lond B. 2015; 370(1681):20140274.
View Article
Google Scholar

[154] View Article

[155] Google Scholar

[ref56] 56. BEACONs. Quebec Catchments 1:50,000 Version 1. BEACONs Project, University of Alberta, Edmonton, AB. http://beaconsproject.ca/

[ref57] 57. Toews M, Juanes F, Burton AC. Mammal responses to the human footprint vary across species and stressors. J Environ Manage. 2018; 217:690–699. pmid:29654972
View Article
PubMed/NCBI
Google Scholar

[158] View Article

[159] PubMed/NCBI

[160] Google Scholar

[ref58] 58. Gaudry W. Impacts des structures anthropiques linéaires sur la sélection d’habitat du caribou, de l’ours noir et du coyote en Gaspésie. MSc Thesis, Université du Québec à Rimouski; 2013.

[ref59] 59. Lowe WH, Likens GE. Moving headwater streams to the head of the class. BioScience. 2005; 55(3):196–197
View Article
Google Scholar

[163] View Article

[164] Google Scholar

[ref60] 60. Polak T, Watson JE, Bennett JR, Possingham HP, Fuller RA, Carwardine J. Balancing ecosystem and threatened species representation in protected areas and implications for nations achieving global conservation goals. Conserv Lett. 2016; 9(6):438–445.
View Article
Google Scholar

[166] View Article

[167] Google Scholar

[ref61] 61. Bauduin S, McIntire E, St-Laurent M-H, Cumming SG. Compensatory conservation measures for an endangered caribou population under climate change. Sci Rep. 2018; 8(1):16438. pmid:30401921
View Article
PubMed/NCBI
Google Scholar

[169] View Article

[170] PubMed/NCBI

[171] Google Scholar

[ref62] 62. Manly BFJ, MacDonald LL, Thomas DL, MacDonald TL, Erickson WP. Resource selection by animals: Statistical analysis and design for field studies. Dordrecht, The Netherlands; 2002.

[ref63] 63. Boyce MS, Vernier PR, Nielsen SE, Schmiegelow FK. Evaluating resource selection functions. Ecol Modell. 2002; 157(2–3):281–300.
View Article
Google Scholar

[174] View Article

[175] Google Scholar

[ref64] 64. Lesmerises F, Johnson CJ, St-Laurent M-H. Effect of behavioral marginality on survival of an alpine ungulate. Ecosphere. 2019; 10(7):e02806.
View Article
Google Scholar

[177] View Article

[178] Google Scholar

[ref65] 65. Logan T. Scénarios climatiques pour les régions naturelles de la péninsule de la Gaspésie et la dépression de La Tuque. Consortium Ouranos; 2012.

[ref66] 66. Gray DR. The relationship between climate and outbreak characteristics of the spruce budworm in eastern Canada. Clim Change. 2008; 87(3–4):361–383.
View Article
Google Scholar

[181] View Article

[182] Google Scholar

[ref67] 67. Régnière J, St-Amant R, Duval P. Predicting insect distributions under climate change from physiological responses: spruce budworm as an example. Biol Invasions. 2012; 14(8):1571–1586.
View Article
Google Scholar

[184] View Article

[185] Google Scholar

[ref68] 68. Minor ES, Urban DL. A graph-theory framework for evaluating landscape connectivity and conservation planning. Conserv Biol. 2008; 22(2):297–307. pmid:18241238
View Article
PubMed/NCBI
Google Scholar

[187] View Article

[188] PubMed/NCBI

[189] Google Scholar

[ref69] 69. Mayer M. Can nature-based tourism benefits compensate for the costs of national parks? A study of the Bavarian Forest National Park, Germany. J Sustain Tourism. 2014; 22(4):561–583.
View Article
Google Scholar

[191] View Article

[192] Google Scholar

[ref70] 70. Wiersma YF, Nudds TD. Efficiency and effectiveness in representative reserve design in Canada: the contribution of existing protected areas. Biol Conserv. 2009; 142(8):1639–1646.
View Article
Google Scholar

[194] View Article

[195] Google Scholar

[ref71] 71. Sundblad G, Bergström U, Sandström A. Ecological coherence of marine protected area networks: a spatial assessment using species distribution models. J Appl Ecol. 2011; 48(1):112–120.
View Article
Google Scholar

[197] View Article

[198] Google Scholar

[ref72] 72. Baldwin RF, Perkl RM, Trombulak SC, Burwell WB. Modeling ecoregional connectivity. In: Trombulak SC, Baldwin RF, editors. Landscape-scale conservation planning. Dordrecht, Netherlands: Springer; 2010. p. 349–367.

[ref73] 73. Rayfield B, Fortin M-J, Fall A. Connectivity for conservation: a framework to classify network measures. Ecology. 2011; 92(4):847–858. pmid:21661548
View Article
PubMed/NCBI
Google Scholar

[201] View Article

[202] PubMed/NCBI

[203] Google Scholar

[ref74] 74. Bélisle M. Measuring landscape connectivity: the challenge of behavioral landscape ecology. Ecology. 2005; 86(8):1988–1995.
View Article
Google Scholar

[205] View Article

[206] Google Scholar

[ref75] 75. Pelletier F, Turgeon G, Bourret A, Garant D, St-Laurent M-H. Genetic structure and effective size of an endangered population of woodland caribou. Conserv Genet. 2019; 20(2):203–213.
View Article
Google Scholar

[208] View Article

[209] Google Scholar

[ref76] 76. Almany GR, Connolly SR, Heath DD, Hogan JD, Jones GP, McCook LJ, et al. Connectivity, biodiversity conservation and the design of marine reserve networks for coral reefs. Coral Reefs. 2009; 28(2):339–351.
View Article
Google Scholar

[211] View Article

[212] Google Scholar

Figures

Abstract

Introduction

Methods

Overview

Study area

Building protected areas networks

Network ecological representativeness

Network functional connectivity

Identifying priority conservation areas

Results

Protected areas networks

Priority conservation areas

Discussion

Network ecological representativeness

Network functional connectivity

Accounting for climate change

Case study

Supporting information

S1 Material. Construction of the potential future landscapes.

S2 Material. Caribou movements in the current and future landscapes.

S3 Material. Description of the optimal protected areas networks.

Acknowledgments

References