Does hunting-mortality risk affect landscape connectivity? Insights from brown bear movement ecology Daniele Falcinelli a,*, Paolo Ciucci a, María del Mar Delgado b, Ilpo Kojola c, Samuli Heikkinen c, Alexander Kopatz d, Daniele De Angelis a, Vincenzo Penteriani e a Department of Biology and Biotechnologies “Charles Darwin” (BBCD), Sapienza University of Rome, 5 Piazzale Aldo Moro, 00185, Roma, Italy b Biodiversity Research Institute (IMIB, CSIC–University of Oviedo–Principality of Asturias), Campus Mieres, 5 Gonzalo Gutiérrez Quirós, 33600, Mieres, Spain c Natural Resources Institute Finland (LUKE), 6 Ounasjoentie, 96200, Rovaniemi, Finland d Norwegian Institute for Nature Research (NINA), 5685 Torgarden, 7485, Trondheim, Norway e Department of Evolutionary Ecology, National Museum of Natural Sciences (MNCN), Spanish National Research Council (CSIC), 2 c/José Gutiérrez Abascal, 28006, Madrid, Spain A R T I C L E I N F O Keywords: Brown bear Circuitscape Landscape connectivity Hunting mortality risk Mortality traps Step-selection functions Ursus arctos A B S T R A C T Landscape connectivity supports key ecological processes of animal populations by enabling movement and ensuring gene flow. Habitat corridors are crucial for facilitating connectivity but may expose transient in dividuals to high mortality risk in human-modified landscapes. Historically, intense hunting led to fragmentation of Scandinavian and Karelian brown bear Ursus arctos populations in Fennoscandia. Currently, the Finnish- Russian Karelian population is managed through legal hunting and shows limited gene flow toward the Scan dinavian peninsula (Sweden and Norway). Using a two-dimensional modelling framework and long-term data sets, here we assessed landscape connectivity in the Karelian population range and whether hunting-related mortality risk affects intra-population connectivity toward Scandinavia. First, we used GPS data (2002–2014) from 37 bears to perform a step-selection analysis, derive a resistance surface, and model connectivity by circuit theory across Finland and Russian Karelia. Next, we used Finnish bear harvest data (1174 locations, 2002–2014) and resource selection functions to model hunting-mortality risk across Finland, which we integrated with previously identified corridors. Connectivity was highest in forests along central-eastern Finland, southern Finnish-Russian border, and southern Russian Karelia, while it was limited in the northern Finnish reindeer husbandry region, probably due to higher cover of shrubland/open areas. Hunting-mortality risk was highest throughout central-eastern Finland. About 44% of Finland's corridor area fell within high-risk zones, which may potentially constrain connectivity to Scandinavia by acting as functional barriers. To enhance connectivity with Scandinavia, conservation actions in Finland should minimise forest fragmentation and decrease hunting pres sure, e.g., within corridors and by protecting females with offspring and solitary subadults. 1. Introduction Landscape connectivity, originally defined as the degree to which the landscape facilitates or impedes movement among resource patches (Taylor et al., 1993), is a vital component of the spatio-temporal dy namics of animal populations. Connectivity across the landscape may fulfil several ecological processes, including regular/seasonal move ments for foraging, mating and migration, colonisation of the unoccu pied habitat through dispersal from natal ranges, and gene flow among otherwise isolated populations (e.g., Benz et al., 2016; Elliot et al., 2014; Frankham, 2006; Gantchoff et al., 2021; Proctor et al., 2012; Viau et al., 2024). A stable connection among populations, through successful movements of individuals, may be essential to maintain genetic vari ability and mitigate the potentially deleterious effects of inbreeding depression and genetic drift, hence enabling species to adapt to chang ing environmental conditions (Crooks and Sanjayan, 2006; Frankham, 2006; Pearman et al., 2024; Schloss et al., 2012). Preserving landscape connectivity is especially critical in human-modified landscapes, where most species occur as partially isolated populations fragmented by habitats of limited or inadequate suitability and/or by high levels of human-caused mortality (Hanski, 1999). Despite the evident benefits for connectivity, conservationists have * Corresponding author. E-mail address: daniele.falcinelli@gmail.com (D. Falcinelli). Contents lists available at ScienceDirect Biological Conservation journal homepage: www.elsevier.com/locate/biocon https://doi.org/10.1016/j.biocon.2026.111818 Received 31 July 2025; Received in revised form 2 March 2026; Accepted 21 March 2026 Biological Conservation 317 (2026) 111818 Available online 25 March 2026 0006-3207/© 2026 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license ( http://creativecommons.org/licenses/by/4.0/ ). mailto:daniele.falcinelli@gmail.com www.sciencedirect.com/science/journal/00063207 https://www.elsevier.com/locate/biocon https://doi.org/10.1016/j.biocon.2026.111818 https://doi.org/10.1016/j.biocon.2026.111818 http://creativecommons.org/licenses/by/4.0/ also argued that habitat corridors may have undesirable consequences (Beier and Noss, 1998; Crooks and Sanjayan, 2006). In particular, edge effects might increase the exposure of transient individuals travelling within corridors to risk derived from competitors, predators, and human activities, eventually creating death traps (i.e., “mortality sinks”) (Crooks and Sanjayan, 2006; Hobbs, 1992; Simberloff et al., 1992). For instance: (a) caribou Rangifer tarandus were at higher risk of predation by wolves Canis lupus when they occupied habitat near linear corridors (James and Stuart-Smith, 2000); and (b) roadkill risk for mammalian carnivores, such as the stone marten Martes foina, the Eurasian badger Meles meles, and the golden jackal Canis aureus, was shown to be higher in high-connectivity areas (Fabrizio et al., 2019; Frangini et al., 2022; Grilo et al., 2011). Functional connectivity is an emergent property of species-landscape relations, resulting from the interaction between animal movement behaviour and the physical structure of the landscape (Taylor et al., 2006). Large carnivores are species particularly susceptible to the loss of inter-population connectivity resulting from habitat fragmentation and anthropogenic disturbance, due to their high energetic requirements, large home ranges, slow life histories, and low population densities (Crooks, 2002; Ripple et al., 2014; Woodroffe, 2000). For these reasons, conserving connectivity for large carnivores is challenging and would need large-scale planning and management interventions, often encompassing transboundary geographical areas (Chapron et al., 2014; Maiorano et al., 2019; Penteriani et al., 2018; Recio et al., 2021). However, although human-caused mortality is a key factor limiting large carnivore populations (Benson et al., 2023; Lamb et al., 2020; Woodroffe and Ginsberg, 1998), few studies have contributed an inte grated assessment of landscape connectivity and the effects of spatially heterogeneous mortality risk (e.g., Maiorano et al., 2019), and none have considered hunting as a source of mortality (but see Ghoddousi et al., 2021). This may be partially due to the difficulty of collecting mortality data over large spatial scales (Bruskotter et al., 2025), which has led to the inclusion of only-simulated mortality risk in spatially explicit models (Ash et al., 2020; Kaszta et al., 2020, 2019). In the 1800s and early 1900s, large carnivores in Europe experienced intense human persecution and substantial range contractions, surviving as small, isolated, and fragmented populations (Ripple et al., 2014; Woodroffe, 2000). The once-continuous Fennoscandian brown bear Ursus arctos population was heavily reduced by hunting in Scandinavia (i.e., Sweden and Norway) and Finland, gradually splitting into the Scandinavian and Karelian populations (Kojola et al., 2020, 2003; Swenson et al., 1995). Following the implementation of conservation- minded legislation and management practices in the second half of the 20th century, however, both bear populations have increased in number and expanded during recent decades (Chapron et al., 2014; Swenson et al., 2023). In Finland, up to the late 1960s, the brown bear range was restricted to the northern and easternmost parts of the country. Since then, however, the species initiated a gradual expansion to the west and south, and over the last 50 years, bear numbers increased by a factor of ten, due to improved protection and continuous dispersal from core areas in the Russian Karelia region (Kojola et al., 2006, 2003). Nowa days, both Karelian and Scandinavian populations are managed through legal hunting (Penteriani et al., 2018), with Finland and Sweden con trolling their bear numbers by annual harvest (Kojola et al., 2020; Milleret et al., 2024). During the recovery process, several studies have investigated the genetic diversity and population structure of brown bears in Scandi navia, Finland, and northwestern Russia (e.g., Hagen et al., 2015; Kopatz et al., 2021, 2014, 2012; Schregel et al., 2015, 2012). The latest genetic study across the entire Fennoscandia revealed the restoration of once-lost connectivity between the Scandinavian and Karelian pop ulations (Kopatz et al., 2021). However, this study showed that gene flow between Karelia and Scandinavia was asymmetrical, with a low immigration rate of individuals from the Karelian population into Scandinavia, which may not ensure the long-term maintenance of connectivity (Kopatz et al., 2021). This finding has been attributed to several factors in the reindeer husbandry region of northern Finland compared to the south, such as lower bear density, relatively higher harvest rates, and a lower proportion of females that could produce dispersing males (Kojola et al., 2020). In addition, there may be func tional habitat barriers to long-distance dispersal by Karelian bears into the Scandinavian population (Kopatz et al., 2012; Schregel et al., 2012). Nonetheless, a comprehensive analysis of the quality and heterogeneity of the habitat matrix (sensu Revilla et al., 2004) in the Karelian popu lation range has been lacking. The transboundary Finnish-Russian Karelian bear population pro vides the opportunity to explore the potential overlooked effects of hunting, which may transform connectivity areas into mortality traps. Therefore, the general aim of this study was to assess potential landscape connectivity for Karelian brown bears and examine whether hunting- mortality risk affects intra-population connectivity, potentially limiting bear settlement near the Karelian-Scandinavian population border and thereby reducing subsequent dispersal toward Scandinavia. Specifically, our objectives were to identify: (1) potential movement corridors that may facilitate gene flow from the Karelian to the Scan dinavian population, and (2) high hunting-mortality risk areas within the Finnish bear population range. To this end, we adopted a two- dimensional modelling framework (Naves et al., 2003; Nielsen et al., 2006) by making use of extensive, long-term datasets from the Karelian population. To achieve the first objective, we analysed telemetry data from bears in Finnish and Russian Karelia, conducting a step-selection analysis (Avgar et al., 2016) to estimate a resistance surface and model landscape connectivity using electrical circuit theory (McRae et al., 2008; Zeller et al., 2012). For the second objective, we used lo cations of legally harvested bears in Finland and applied resource se lection functions (Boyce and McDonald, 1999; Nielsen et al., 2004) to model hunting-mortality risk to be subsequently integrated with the previously identified corridor areas. This study provides valuable insight into the combined effects of environmental features, landscape resis tance, and a novel investigation of hunting pressure on connectivity. Ultimately, it offers spatially explicit outputs relevant to the long-term conservation and viability of the Karelian bear population. 2. Materials and methods 2.1. Study area and brown bear data The boundaries of Finland and the Russian Karelia region defined our study area, totalling more than 500,000 km2 (Fig. 1). However, we focused on Finland alone for the hunting-mortality risk analysis (see below). Approximately 81% of the entire study area consists of low-lying land below 200 m altitude. About 77% of the territory is covered by forests: located in the boreal vegetation zone (Ahti et al., 1968), they are composed mainly of coniferous Scots pine Pinus sylvestris and Norway spruce Picea abies, mixed with broad-leaved trees, such as birches Betula spp. and Eurasian aspen Populus tremula. Lakes and wetlands, such as swamps and peat bogs, also characterise the landscape extensively, covering around 13% of the land area. Russian Karelia has an overall lower anthropogenic impact than Finland, with a percentage of anthropogenic land and road density roughly ten and five times lower, respectively (e.g., average road density within Finland: 1.77 km/km2; average density within Russian Karelia: 0.35 km/km2). From 2002 to 2013, 71 brown bears were captured throughout the study area (115 total captures, including recaptures), and each bear was fitted with a GPS collar (Televilt, Lindesberg, Sweden; Vectronic Aerospace GmbH, Berlin, Germany). Bear capture, handling, and collaring (for details see e.g., Penteriani et al., 2021) met the guidelines issued by the Animal Care and Use Committee at the University of Oulu, with all permits provided by the provincial government of Oulu and the Regional State Administrative Agency (OYEKT-6-99, OLH-01951/Ym-23, ESAVI/ 3229/04.10.07/2013). The GPS collars were programmed to collect one D. Falcinelli et al. Biological Conservation 317 (2026) 111818 2 location at a rate of 1, 2, or 4 h (i.e., 6–24 locations/day; Falcinelli et al., 2024), with these schedules allowed to vary within the same collar's deployment. Signals from satellite transmitters were recorded by the ARGOS satellite system (https://www.cls.fr/en/cls-group/). Following D'Eon et al. (2002), we excluded all 2-D fixes (~8%) to screen data from high location errors (Bjørneraas et al., 2010). Additionally, we excluded fixes collected during the denning period (i.e., most typically, mid- November to April). Although estimates of landscape connectivity are most reliable when based on species dispersal data (Elliot et al., 2014; Vasudev et al., 2015), such data are notoriously difficult to obtain (Zeller et al., 2018); in our sample, for instance, only three collared individuals were confirmed to disperse (see subsection 2.4.3). Therefore, in line with several landscape-connectivity studies (e.g., Abrahms et al., 2017; Maiorano et al., 2019; Ziółkowska et al., 2016), we used intra-range direct movements to model resistance to movement (see below), as they are considered a useful proxy for dispersal movement (Zeller et al., 2018, 2012). This approach was particularly suitable for our study population, as during the study period, it was expanding into the Finnish range (Kojola et al., 2006, 2003), with many bears exhibiting extensive movements typical of non-resident individuals (Appendix Table S1). Additionally, movement patterns in Karelian bears have previously been shown to be largely unaffected by sex, age, or reproductive status (Lamamy et al., 2022; Penteriani et al., 2025, 2022, 2021), although these studies did not include many subadult dispersing individuals. For subsequent Step Selection Analysis (iSSA), we thus prepared a consistent dataset by filtering and resampling all locations at a 2-h fix rate with a flex of 15 min to include those locations on the margin. The final dataset included 38,907 GPS locations collected from 2002 to 2014 from 37 brown bears: adult females (9562), adult males (5524), females with cubs (11,815), subadult females (7076), and subadult males (4930; mean number of locations per individual ± SD = 1052 ± 884) (Ap pendix Table S1). In Finland, the bear annual hunting season starts on 20 August and stops on 31 October, with harvest limited by regional quotas (Kojola et al., 2020). Hunters generally fill a form about every harvested bear, including information about the geographic coordinates of the kill site and the bear's sex, and it is sent to the Natural Resources Institute Finland (Luke) (Kojola et al., 2021, 2020). Although data on legally harvested bears within Russian Karelia are available (see Kopatz et al., 2014), these are few, and kill site coordinates are missing or represent very rough, approximate locations. Additionally, information for esti mating hunting-related covariates (see subsection 2.5.1) is not available for the Russian territory. For these reasons, we investigated hunting- mortality risk for bears in Finland alone. We filtered bear harvest loca tions for the same years of our GPS data (i.e., 2002–2014) and excluded cubs-of-the-year from the dataset because (a) they are not present in the GPS data, and (b) their space use is not independent of their mother (i.e., autocorrelated data; Steyaert et al., 2016a). After filtering, a total of Fig. 1. The study area is defined by the boundaries of Finland (~338,000 km2 land area) and the Russian Karelia (~170,000 km2). GPS locations of 71 brown bears captured and monitored in central-western and eastern Finland and Finnish-Russian Karelia between 2002 and 2014 are shown, overlaid on the five-category land-use layer derived from the 2015 Copernicus Global Land Cover (see Appendix A.2). Although GPS data occur across three clusters, they belong to the same Karelian population (Kopatz et al., 2014, 2012; Schregel et al., 2012), which is distributed virtually continuously throughout Russian Karelia, north-eastern Finland, and central-western Finland (Heikkinen et al., 2024; Kojola et al., 2020). Basemap credits: © 2009 Esri – World Terrain Base. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Fig. 2. Average bear density (number of bears/km2) across regions of the Finnish Wildlife Agency (see Appendix A.8 for all details and the other hunting- related covariates). Red dots represent all retained bear hunting mortality lo cations (n = 1174; 2002–2014, and cubs-of-the-year excluded). Note that, despite the low bear density in northern Finland (i.e., Lapland; Fig. 1) — similar to that in the two more southern regions of central Finland — this area still shows a relatively high number of mortality locations compared to those central regions. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) D. Falcinelli et al. Biological Conservation 317 (2026) 111818 3 https://www.cls.fr/en/cls-group/ 1174 hunting mortality locations were included in the final dataset (Fig. 2). 2.2. Digital environmental covariates For landscape connectivity and hunting-mortality risk analyses, we selected an extensive set of environmental covariates according to their potential ecological relevance for our species in the study area. All covariates were derived from free-downloadable spatial datasets (Ap pendix Table S2) and converted to raster layers with a spatial resolution of 100 m. We aggregated these covariates into four categories: (1) topographic, which included Terrain Ruggedness Index and Topo graphic Position Index; (2) landcover (Copernicus Global Land Cover; Buchhorn et al., 2020), which comprised the four continuous land-use covariates anthropogenic areas, forest, natural open areas, and shrub land, all expressed as proportional coverages within a 700-m radius buffer. The radius value was chosen based on the average 2-h step length pooled over all individuals (Osipova et al., 2019); (3) forest-related, which contained Tree Canopy Cover and forest height (also a proxy for forest age) (GLAD laboratory; https://glad.umd.edu/), as well as distance from forest edge; and (4) human-related, which included four distance variables, i.e., distance from main roads, secondary roads, all (combined) roads, and human habitations (road network from Open StreetMap; OpenStreetMap contributors, 2023) (see Appendix A.2 for complete information on environmental covariates, including their preparation and source). We processed all data and performed all spatial/statistical analyses in R version 4.4 (R Core Team, 2024) and QGIS version 3.3 (QGIS Development Team, 2023). 2.3. Habitat selection modelling 2.3.1. Generation of available steps For each observed step, i.e., the line connecting two successive GPS locations (Turchin, 1998), we first calculated its length and turning angle using the R package adehabitatLT (Calenge, 2006). As we were interested in characterising habitat selection by bears associated with active movements rather than limited movements or stationary states (e. g., feeding and resting), we excluded consecutive locations below a threshold distance from each other from the analysis. Specifically, we used the step length-based approach of Ziółkowska et al. (2016) to select the “active” steps indicating directional movement, i.e., all steps ≥300 m (n = 15,480) (Appendix A.3). Next, we paired each retained step with 10 random steps generated using a parametric sampling method and an iSSA approach (Avgar et al., 2016). In detail, step lengths were drawn through a uniform distribution within a minimum distance of 300 m and a maximum corresponding to the 99th percentile of the observed step length distribution (i.e., 5410 m). Turning angles of random steps were sampled from a von Mises distribution (μ = − 0.019, κ = 0.713) fitted by maximum likelihood estimation to observed turning angles using the R package circular (Agostinelli and Lund, 2024). We finally extracted the above-mentioned environmental covariates at the endpoints of each observed and random step, as averaging predictors along steps may overestimate landscape resistance, especially at reduced fix rates (Thurfjell et al., 2014; Ziółkowska et al., 2016). To verify that using 10 random steps per used step adequately characterised availability (Benson, 2013; Northrup et al., 2013), we refitted models (see below) using increasing numbers of available locations (1− 20) and found that coefficient estimates stabilised at this value (Appendix A.3). 2.3.2. Step-selection analysis We checked all environmental covariates for collinearity using Pearson's correlation coefficient threshold of r ≥ |0.65| (Ziółkowska et al., 2016). The pairs of correlated variables were tree canopy cover with forest height (r = 0.76), distance to main roads with distance to human habitations (r = 0.65), and distance to secondary roads with distance to all roads (r = 0.98). We took correlation into account in building the candidate model set, such that pairs of correlated variables were not included in the same model (Corradini et al., 2021; Peck et al., 2017) (Appendix Table S4). We fitted a set of 12 ecologically plausible step-selection function models using mixed conditional logistic regression (Duchesne et al., 2010) (Table 1). Specifically, we implemented the method proposed by Muff et al. (2020) via the R package glmmTMB (Brooks et al., 2017), making use of the Poisson trick and treating intercepts as a random ef fect with a large, fixed variance. The model structure was as follows: (1) used (1) and available (0) steps as the response variable; (2) each used step with its associated 10 random steps as the stratum; (3) five “core” environmental covariates as predictor variables, i.e., the four land-use covariates and forest height, representing the effects of land use and forest structure; (4) one or more additional environmental covariates characterising the different candidate models; (5) step length as a pre dictor, to reduce bias in estimated model coefficients (Forester et al., 2009). All continuous predictors were standardised using Min-Max normalisation (i.e., subtraction of the minimum value and division by the range) to facilitate the comparison of parameter estimates and convergence of models (Grueber et al., 2011); and (6) bear ID as a random slope with respect to all included covariates, to accommodate inter-individual heterogeneity (Muff et al., 2020) (further details in Appendix A.4). We evaluated all candidate models through corrected Akaike's Information Criterion (AICC) scores (Burnham and Anderson, 2002), selecting only the most parsimonious model for inference. To assess the predictive power of this model, we finally used k-fold cross- validation as introduced by Boyce et al. (2002), modified by splitting the data by individual, i.e., cross-validation blocked by individual (Roberts et al., 2017). 2.4. Landscape connectivity modelling 2.4.1. Calculation of the resistance surface Using the coefficients (β) of the top iSSF model and spatial Table 1 Model selection results based on the corrected Akaike's Information Criterion (AICC) for step-selection function models. For each candidate model, the included environmental covariates, degrees of freedom (df), difference in AICC values between models (ΔAICC), Akaike weight (wi), and log-likelihood (logL) are reported. Model M10 outperformed all other models (ΔAICC > 2) and received a weight of one. Model numbers correspond to Appendix Table S4. Model ID Covariates df AICC ΔAICC wi logL M10 Core + DFE + DMR + DSR 18 299,203.3 0.0 1.00 − 149,583.6 M7 Core + DMR + DSR 16 299,304.5 101.3 0.00 − 149,636.3 M12 Core + TRI + TPI + DFE + DR + DHH 22 299,393.2 189.9 0.00 − 149,674.6 M11 Core + DFE + DR + DHH 18 299,433.2 229.9 0.00 − 149,698.6 M9 Core + DR + DHH 16 299,516.4 313.1 0.00 − 149,742.2 M5 Core + DMR 14 299,618.5 415.2 0.00 − 149,795.2 M6 Core + DSR 14 299,777.7 574.4 0.00 − 149,874.8 M8 Core + DHH 14 299,856.9 653.6 0.00 − 149,914.5 M4 Core + DFE 14 300,053.3 850.0 0.00 − 150,012.6 M3 Core + TRI + TPI 16 300,103.9 900.6 0.00 − 150,035.9 M1 Core + TRI 14 300,112.3 909.0 0.00 − 150,042.2 M2 Core + TPI 14 300,131.3 928.0 0.00 − 150,051.6 Note: Core covariates comprised the four land-use covariates (anthropogenic areas, forest, natural open areas, and shrubland) and forest height. TRI = Terrain Ruggedness Index; TPI = Topographic Position Index; DFE = Distance to Forest Edge; DMR = Distance to Main Roads; DSR = Distance to Secondary Roads; DR = Distance to all Roads; DHH = Distance to Human Habitations. All models also included step length as a predictor to reduce bias in estimates of model co efficients (see text). D. Falcinelli et al. Biological Conservation 317 (2026) 111818 4 https://glad.umd.edu/ environmental covariates (x), we first predicted a movement perme ability surface spanning the entire study area, calculating the selection score for each raster cell by the equation w(x) = exp.(β1x1 + β2x2 + … + βnxn) (Hofmann et al., 2021; Thurfjell et al., 2014). Following Squires et al. (2013), we then included only the range of predicted scores within the 5th–95th percentiles of their original values. For ease of interpre tation, we normalised the final permeability surface to a range between 0 and 1 (Hofmann et al., 2021). Our permeability model was also based on within-home-range movements, whereas dispersing bears may be more tolerant of lower-quality habitat compared to resident/non- dispersing bears (Abrahms et al., 2017; Elliot et al., 2014; Mateo- Sánchez et al., 2015). To account for this difference, we thus trans formed the permeability surface using the equation R = 100–99*((1 − exp.(− c*H))/(1 − exp.(− c))) where R is the resistance, H is the habitat permeability, and c is a constant (i.e., c8 negative exponential trans formation; Keeley et al., 2016). This transformation is well suited to calculate resistance derived from telemetry data as it assigns high resistance values to only the lowest habitat permeability values (Carroll et al., 2020; Zeller et al., 2018). 2.4.2. Potential connectivity Landscape connectivity was modelled by applying circuit theory in Circuitscape.jl (Anantharaman et al., 2020; McRae et al., 2008) (see Appendix A.5 for details on circuit theory). To focus on general con nectivity (i.e., regional movement corridors; García-Sánchez et al., 2022), we adopted an omnidirectional approach based on the method in Boudreau et al. (2022). Therefore, we placed a set of source points randomly across the study area to represent locations of individual bears travelling randomly through the landscape. Since connectivity is always higher near focal nodes (McRae et al., 2008), we performed random sampling for source points weighted on their habitat permeability score. We selected 250 source points (i.e., density of 0.05/100 km2), corre sponding roughly to 15% of the number of bears estimated in Finland during our study period. As Circuitscape models movement between each pair of nodes separately, we used the all-to-one mode to acquire cumulative connectivity between all point pairs: the resulting map represents the overall potential connectivity for a bear moving randomly through the landscape, regardless of the movement direction. The cu mulative current density map was log-transformed (e.g., Buchholtz et al., 2020; Merrick and Koprowski, 2017; Zeller et al., 2016) and then linearly rescaled from 0 to 100 to ease its interpretation. To assess the sensitivity of the final map as a function of the number/location of source points, we repeated the analysis with 200 and 300 points (further details and corresponding results in Appendix A.5). We used zonal statistics to analyse landscape features associated with the highest current densities (García-Sánchez et al., 2022). For this zonal analysis, we reclassified the final current map into five categories using the Jenks natural breaks classification method, where 1 indicates the category most constraining bear movement (i.e., current flow) and 5 the most favourable category. For management purposes, corridors are often identified from a connectivity model by taking an upper percent age of connectivity values (Dutta et al., 2022; Maiorano et al., 2019; Zeller et al., 2018). Following this approach, we finally delineated cor ridors by taking the top 15% of the final current surface (Appendix Table S5). 2.4.3. Model evaluation We assessed the fit of our connectivity model using dispersal data from three subadult males and two complementary evaluation methods, i.e., a null model comparison and a random path comparison (McClure et al., 2016; Zeller et al., 2018). Specifically, model performance was evaluated by: (1) comparing empirical connectivity values at observed dispersal locations against those expected under a null (isolation-by- distance) model, and (2) testing whether observed dispersal paths fol lowed areas of higher predicted connectivity than randomly-generated alternative paths (for all details, see Appendix A.7). All connectivity analyses were performed in Circuitscape version 5.13 through Julia version 1.10 (Anantharaman et al., 2020). 2.5. Modelling hunting mortality risk 2.5.1. Hunting-related covariates Using available data on Finnish wildlife hunting, we derived three hunting-related covariates indicating, for each Finnish Wildlife Agency region: (1) average bear density (Fig. 2), estimated annually from 2007 to 2014; (2) average density of registered hunters, estimated annually from 2007 to 2014; and (3) the bear hunting quota for 2014, i.e., the only year available (see Appendix A.8 for details). We considered as predictors the above hunting-related covariates along with land-use covariates, tree canopy cover, distance from forest edge, and distance from all roads, i.e., the environmental covariates expected to exert the greatest effect on bear occurrence (Grueber et al., 2011; Zuur et al., 2010). 2.5.2. Resource selection functions To model hunting-mortality risk, we used a Resource Selection Function (RSF) approach following the guidelines in Fieberg et al. (2021). We sampled 100 times as many random locations as the number of mortality locations within the 100% minimum convex hull of all mortality locations after masking water bodies. We then checked for collinearity among the selected covariates using the same criterion described above (see subsection 2.3.2). Average bear density correlated with bear hunting quota (r = 0.81), forest correlated with both natural open areas (r = − 0.72) and distance to forest edge (r = − 0.68); there fore, we retained bear density, open areas, and distance to forest edge for modelling due to their easier interpretation in a management context. We fitted a global RSF model using weighted logistic regression with the R package glmmTMB (Brooks et al., 2017). The model structure was as follows: (1) used (1) and available (0) locations as the response var iable; (2) the above covariates as predictor variables, standardised with Min-Max normalisation; and (3) a weight of 1 and 5000 assigned to used and available locations, respectively, as this ensures that RSF co efficients are estimated correctly according to an inhomogeneous Pois son point process model (Fieberg et al., 2021; Fithian and Hastie, 2013; Warton and Shepherd, 2010). For model selection, we employed a multi- model inference approach through full averaging of models with ΔAICC < 2 (Burnham and Anderson, 2002; Grueber et al., 2011) using the R package MuMIn (Bartoń, 2024). The final model predictive performance was evaluated using k-fold cross-validation by dividing the data randomly, i.e., random cross-validation (Boyce et al., 2002). Based on the model's coefficients and the above equation (subsection 2.4.1), we predicted hunting-mortality risk for bears across Finland, curtailing predicted scores between the 5th and 95th percentile of their original values and then linearly rescaling from 0 to 100 to ease inter pretation. Accordingly, we identified areas with the highest hunting- mortality risk as those comprising the top 15% of predicted risk values (Dutta et al., 2022). We then overlapped these areas with previously modelled corridors (Ghoddousi et al., 2021) and calculated the per centage of corridor area within Finland that included high-mortality-risk areas. All manuscript maps were created with the R package tmap (Tennekes, 2018). 3. Results 3.1. Brown bear habitat selection Model M10 was the most parsimonious habitat selection model; in addition to land-use covariates and forest height (i.e., “core” covariates), this model included distance to forest edge, main roads, and secondary roads (Table 1). According to this model, Karelian brown bears selected forests with high canopy height when travelling (Table 2). They also selected anthropogenic and natural open areas, but did not seem to D. Falcinelli et al. Biological Conservation 317 (2026) 111818 5 avoid proximity to any roads, as well as to the forest edge (Table 2). Such a model had high predictive power (mean Spearman correlation coef ficient across folds ± SD = 0.973 ± 0.03). 3.2. Landscape connectivity Our prediction of the resistance surface showed marked differences across the study area (Fig. 3). Specifically, landscape resistance was lowest throughout central and southern Finland, as well as in the southernmost and westernmost parts of Russian Karelia. Here, low- resistance areas were fragmented mainly due to the widespread occur rence of lakes. Large, isolated areas of high resistance were present in the northernmost half of Russian Karelia (i.e., White Karelia) and northern Lapland in Finland. Accordingly, based on the final cumulative current density map, low-current-density areas were found throughout the northernmost regions of the study area (Fig. 4a). In contrast, areas of high connectivity occurred in central-eastern Finland, along the south ernmost Finnish-Russian border, and southern Russian Karelia (Fig. 4a). Similarly, corridors ran along the central-west coast of Finland and throughout central-eastern Finland, although more continuous and broader corridors stretched from the bordering area toward southern Russian Karelia beside Lakes Ladoga and Onega (Fig. 4b). The identified corridors featured, on average, areas with >90% forest cover and a canopy cover of about 80%, and were relatively close (i.e., < 300 m) to roads (Appendix Table S5). On the contrary, low connectivity was mainly due to a combination of higher shrubs and open areas cover and lower occurrence of anthropogenic infrastructure. Model evaluation through dispersal data yielded partially contrasting results: the null isolation-by-distance model slightly outperformed the empirical con nectivity surface at dispersal points, whereas observed dispersal paths showed higher mean connectivity values than the majority of rando mised paths for all three individuals (see Appendix A.7 for details). 3.3. Hunting mortality risk Four RSF models best described hunting-mortality risk for bears in the study area (Table 3). The probability of bears being hunted increased with bear density, at lower hunter density, and in proximity to roads (Table 3). Hunting-mortality risk was also lower in open areas at closer distances to the forest edge and increased at higher canopy cover (Table 3). Cross-validation suggested a high predictive performance for this model (mean Spearman correlation coefficient across folds ± SD = 0.936 ± 0.03). Our map of the predicted hunting-mortality risk (Fig. 5a) depicts this is highest mainly throughout central-eastern Finland and, to a lesser extent, in south-western Finland and central Lapland. By over lapping connectivity areas and the hunting mortality map, we found that about 44% of the Finland corridor area fell within zones of high mor tality risk (Fig. 5b). 4. Discussion By analysing movement data from Karelian brown bears combined with remotely-sensed environmental covariates, we found that land scape connectivity was greatest throughout central-eastern Finland and the southernmost Finnish-Russian border area, with major corridors running into southern Russian Karelia. Moreover, by analysing the spatial distribution of bear harvest locations within Finland, we esti mated that hunting-mortality risk was highest in central-eastern Finland. These findings indicate that potential movement corridors in the northernmost regions of Finland are indeed limited, limiting the expansion of the Karelian population northwards and, in turn, reducing potential connectivity to Scandinavia. Remarkably, our results also suggest that hunting-mortality risk may further constrain intra- population connectivity by acting as a functional barrier (Ghoddousi et al., 2021; Taylor et al., 2006) to movements by Karelian bears. 4.1. Landscape connectivity Karelian bears strongly selected large and high stands of mature forest when travelling, possibly because these habitats offer refuge during movement (Swenson et al., 2023). This preference resulted in a good matching of high-connectivity areas with “closed” forests (i.e., with a canopy cover of more than 70%; Buchhorn et al., 2020), such as those occurring along the Finnish-Russian border in the so-called Green Belt area (Karivalo and Butorin, 2006; Linden et al., 2000). Although there are border fences all along the Russian border (Kopatz et al., 2012), the extensive transboundary movements observed for our collared bears (Fig. 1; Appendix Figs. S4, S6) indicate that these did not act as physical barriers and seem not to restrict movement and dispersal, contrary to what was reported for wolves (Aspi et al., 2009). Additionally, brown bears preferred to move along areas near anthropogenic infrastructure, which may be associated with lower Table 2 Values of coefficients (β) and 95% confidence intervals (CI) for the most parsi monious mixed conditional logit model (i.e., M10, Table 1) comparing used steps to randomly generated steps. Numbers in bold represent effects with a P- value <0.01. Covariate β CI Anthropogenic areas 4.87 2.88; 6.85 Forest 5.57 4.41; 6.73 Natural open areas 1.76 0.42; 3.11 Shrubland − 1.13 − 3.64; 1.37 Forest height 0.60 0.43; 0.77 Distance to forest edge 0.05 − 0.78; 0.89 Distance to main roads − 0.45 − 2.51; 1.62 Distance to secondary roads − 0.99 − 3.38; 1.39 Step length − 9.59 − 10.86; − 8.31 Fig. 3. Resistance surface, scaled from 1 to 100, as estimated by applying a c8 negative exponential transformation (sensu Keeley et al., 2016) to the final permeability surface (see text for details). Landscape resistance was lowest in the whole of central-eastern Finland and the most southern and western parts of Russian Karelia. D. Falcinelli et al. Biological Conservation 317 (2026) 111818 6 movement costs (e.g., avoiding crossing numerous lakes and wetlands). This pattern agrees with our recent research showing that adult males in Finland used habitats close to anthropogenic areas to travel during the mating season (Falcinelli et al., 2024). More generally, this behaviour aligns with previous findings indicating that linear infrastructure and nearby areas may function as efficient travel routes for large carnivores (e.g., Dickie et al., 2020; Dickson et al., 2005; James and Stuart-Smith, 2000), even in our study area (Barry et al., 2020; Gurarie et al., 2011). In addition, developed and agricultural areas may represent relatively low- risk environments during the hunting season, as hunting-mortality risk is presumably lower near these areas (i.e., a human-shield effect; Steyaert et al., 2016b). This interpretation is consistent with the rela tively high habitat suitability observed in areas closer to settlements during the hunting season (authors' unpublished data). Therefore, a reduced anthropogenic-area presence and a more extensive network of shrublands, open areas, and wetlands resulted in lower connectivity observed mostly within the northern Lapland and White Karelia (Russia) regions. As high bear density has been reported within Russian Karelia (i.e., 17 bears/1000 km2; Danilov, 2005), our results suggest this region may be less favourable for bears' long-range movement but still suitable for their presence, highlighting the importance of basing connectivity assessments on movement rather than habitat models (Ziółkowska et al., 2016). The observed landscape connectivity patterns are supported by various studies showing high gene flow between Finland and Russian Karelia (Kopatz et al., 2014, 2012; Schregel et al., 2015; Tammeleht et al., 2010), as well as by historical bear monitoring data from both territories. In fact, bear immigration from Russian Karelia was primarily recorded in the eastern Finnish regions of Northern Karelia, Kainuu, and Koillismaa, whereas the immigration rate from the Kola Peninsula into eastern Lapland was much lower (Pulliainen, 1986). Also, corridors we detected within Russian Karelia were along the three isthmuses between the White Sea, Lakes Onega and Ladoga, and the Baltic Sea, assumed to connect Fennoscandia forests to Russian forests (Linden et al., 2000; Fig. 4. Cumulative current density map (a), obtained by log10 transforming and rescaling from 0 to 100 the Circuitscape output, and corridors map (b), as derived by taking the top 15% of current density values. The purple line in (b) delineates the border between Finland and Russian Karelia. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Table 3 Model selection results based on the corrected Akaike's Information Criterion (AICC) for resource selection function models. For each of the four competing models, the degrees of freedom (df), difference in AICC values between models (ΔAICC), Akaike weight (wi), and log-likelihood (logL) are reported. For each covariate, the coefficient (β), 95% confidence intervals (CI), and relative importance value (RIV) obtained by averaging this top model set (ΔAICC < 2) are reported. Numbers in bold represent effects with a P-value <0.01. Competing models df AICC ΔAICC wi logL Bear density + DFE + DR + Hunters density + Open areas + TCC 7 31,212.5 0.00 0.36 − 15,599.3 Bear density + DFE + DR + Hunters density + Open areas + Shrubland + TCC 8 31,213.2 0.66 0.26 − 15,598.6 Anthropogenic areas + Bear density + DFE + DR + Hunters density + Open areas + TCC 8 31,213.4 0.90 0.23 − 15,598.7 Anthropogenic areas + Bear density + DFE + DR + Hunters density + Open areas + Shrubland + TCC 9 31,214.3 1.79 0.15 − 15,598.2 Covariate β CI RIV Intercept − 12.60 − 13.13; − 12.06 Bear density 3.06 2.90; 3.23 1.00 Distance to forest edge − 3.08 − 4.75; − 1.41 1.00 Distance to roads − 4.21 − 6.27; − 2.15 1.00 Hunters density − 2.16 − 2.54; − 1.78 1.00 Natural open areas − 1.60 − 2.65; − 0.54 1.00 Tree canopy cover 0.61 0.37; 0.85 1.00 Shrubland 0.38 − 0.64; 2.48 0.41 Anthropogenic areas − 0.15 − 1.22; 0.41 0.38 Note: TCC = Tree Canopy Cover; DFE = Distance to Forest Edge; DR = Distance to all Roads. D. Falcinelli et al. Biological Conservation 317 (2026) 111818 7 Mayer et al., 2005; Muukkonen et al., 2009), and potentially used also by other large carnivores in the area (see Aspi et al., 2009; Danilov et al., 2018; Kojola et al., 2009). 4.2. Hunting mortality as a functional barrier Bear harvest locations were mainly concentrated in prime forest habitats close to roads, with a high bear density and low hunter density. Similar patterns have been documented for several other harvested species, such as the cougar Puma concolor, leopard Panthera pardus, moose Alces alces, and caribou (Brown et al., 2023; Dougherty et al., 2025; Pitman et al., 2015; Plante et al., 2017; Stoner et al., 2013), although these studies did not include direct measures of hunting effort. Almost half (i.e., ~44%) of Finland's corridor area overlapped with high mortality-risk areas, thus representing potential mortality traps. This finding suggests higher exposure to hunting-mortality risk for bears within corridors, supporting the conceptual framework of Panzacchi et al. (2016). According to these authors, corridors and barriers stand at the opposite ends of a continuum of permeability in which barriers may arise, for example, by reducing the width of passageways or increasing their resistance beyond critical thresholds (Panzacchi et al., 2016). Nevertheless, a large percentage of harvest locations (~65%) occurred outside corridors, indicating that high-mortality-risk areas as a whole may act as barriers limiting connectivity in our study system. This effect may result from both direct bear killing and spatio-temporal avoidance by bears of these high-risk areas (i.e., indirect behavioural effect), given their capability to perceive human-derived risk (e.g., Brown et al., 2023; Corradini et al., 2024; Ordiz et al., 2011). Our connectivity analysis pooled GPS locations collected both during and outside the hunting season, whereas the mortality-risk model directly reflects hunting pressure. This temporal mismatch should be acknowledged when interpreting the spatial overlap between corridors and high-risk areas. Moreover, in brown bears, the mating season (approximately May–July) is a key period for functional connectivity, as adult males in particular increase their movements and subadults initiate natal dispersal (Swenson et al., 2023); yet this period does not overlap with the hunting season. Nevertheless, subadults may experi ence delayed dispersal and initiate long-distance movements after the mating season (Støen et al., 2006), potentially travelling during the hunting period; notably, one of our dispersing males began its move ment toward Russian Karelia during the hunting season (Appendix A.7). Additionally, adult males in Finland have been previously shown to perform large daily displacements also during the post-mating period, which fully overlaps with the hunting season, and to select areas farther from roads only during the hunting season (Falcinelli et al., 2024). This behaviour, which suggests active avoidance of high-risk areas during this period, may partly influence the extent of overlap between corridors and high-risk areas. Therefore, the asymmetrical gene flow detected between the Karelian and Scandinavian bear populations could pre sumably be explained by a combination of factors, primarily the lower bear density and reduced landscape connectivity in Finnish Lapland (Kopatz et al., 2021), as well as higher hunting risk in central-eastern Finland. Interpopulation connectivity and gene flow in large carnivores are primarily mediated by dispersing subadult individuals, whose move ment behaviour and habitat selection may differ markedly from those of resident animals, including in brown bears (Elliot et al., 2014; Thorsen et al., 2022; Vasudev et al., 2015). Consequently, the limited inclusion of dispersing individuals in our dataset represents a limitation when interpreting connectivity in a strict demographic sense. Nevertheless, many collared bears performed wide-ranging, non-resident movements, and predicted connectivity showed encouraging agreement with Fig. 5. Map of hunting-mortality risk (rescaled from 0 to 100) predicted for the whole of Finland (a), and high mortality-risk areas (delineated by taking the 15% maximum risk values) overlaid on current-based corridors in Finland (b). D. Falcinelli et al. Biological Conservation 317 (2026) 111818 8 movements from the few dispersers available for model evaluation (Appendix A.7), supporting the relevance of the inferred permeability patterns. Future research should build upon our results by obtaining GPS-based movement data from Karelian bears during natal dispersal for a more complete understanding of inter-population connectivity and gene flow across Fennoscandia. 4.3. Management and conservation implications Our integrated analyses provide information on the effect of land scape structure and hunting-mortality risk on bear movements to help inform conservation measures and management actions for Karelian brown bears and, more generally, for the Fennoscandian bear popula tion. Our spatially explicit outputs can be directly used to guide policy- making, including determining where to focus on maintaining/restoring connectivity or refining harvest management (e.g., quotas and territory restrictions) across jurisdictions with different levels of mortality risk. In this context, a key conservation area seems to be central-western Finland and the Oulu sub-region, characterised by extensive corridor areas and a relatively low mortality risk level. Since these regions are located just south of the western Finland-Sweden border, they could actually boost connectivity between the two bear populations in Fen noscandia. By contrast, connectivity toward Russia is mediated by fewer, more spatially defined corridors than in central/southern Finland, suggesting that these pathways warrant strong protection, particularly given Russia's high bear density and their potential role in facilitating movements into Fennoscandia. Given bears' preference to move along large and mature forest patches, forest management should minimise extensive clear-cutting and fragmentation. Boreal forests in Finland underwent extensive clear-cutting operations mostly from 1950 to 1980, while the human influence on Russian Karelia forests has grown over recent decades, mainly due to increasing demand for wood material from Finland itself (Mayer et al., 2005; Muukkonen et al., 2009). To mitigate fragmentation effects, Linden et al. (2000) emphasised the strategic role of establishing a continuous “forest bridge” across Finland toward Sweden. Considering the brown bear umbrella effect (Steenweg et al., 2023), conservation and restoration of large-scale forest connections would provide protec tion and benefits even to other forest-dwelling species of the taiga (Linden et al., 2000; Määttänen et al., 2022), e.g., by ensuring contin uous gene flow between Fennoscandian and Russian populations. Based on the overall results of this work, it may be desirable to reduce hunting pressure throughout Finland by (a) limiting harvest in areas identified as critical for connectivity, and (b) specifically giving legal protection to all bear family groups (Kojola et al., 2021, 2020; Kopatz et al., 2021), as well as to solitary subadults. These demographic groups play a crucial role as potential sources of immigrant individuals from Finland to Scandinavia, given that most juvenile male bears disperse from their natal ranges as 2-year-olds (Støen et al., 2006; Zedrosser et al., 2007). However, Finnish legislation currently grants legal protection only to family groups with cubs-of-the-year, a measure that remains ineffective because cubs are frequently mistaken for year lings and consequently shot by hunters (Kojola et al., 2021). Moreover, subadult males dominate harvest statistics, and male-biased mortality is particularly pronounced in the reindeer husbandry region (Kojola et al., 2020). Although hunting-mortality risk in northern Finland was mainly concentrated in central Lapland, overall mortality risk for bears may still be higher in these northernmost areas. Within reindeer husbandry re gions, brown bears and other large carnivores can kill substantial numbers of reindeer calves, and the consequent illegal hunting can have a significant effect on their populations (Andrén et al., 2006; Liberg et al., 2012; Persson et al., 2009; Sivertsen et al., 2016). Future research should thus quantify the extent of brown bear illegal hunting within the Finnish reindeer area while increasing predator tolerance by local communities through the development of new damage prevention and compensation strategies (Rasmus et al., 2020). Our two-dimensional modelling approach could be extended to the whole Scandinavian peninsula, using bear movement and mortality data from that region too (see Steyaert et al., 2016a). Finally, this approach should possibly be combined with spatially explicit human dimensions models (e.g., Behr et al., 2017; Mayer et al., 2023) in order to foster coexistence between humans and brown bears within the Fennoscandian reindeer husbandry area. CRediT authorship contribution statement Daniele Falcinelli: Writing – review & editing, Writing – original draft, Visualization, Software, Methodology, Formal analysis, Data curation, Conceptualization. Paolo Ciucci: Writing – review & editing, Methodology, Funding acquisition, Data curation, Conceptualization. María del Mar Delgado: Writing – review & editing, Methodology, Funding acquisition, Data curation, Conceptualization. Ilpo Kojola: Writing – review & editing, Data curation. Samuli Heikkinen: Writing – review & editing, Data curation. Alexander Kopatz: Writing – review & editing, Data curation. Daniele De Angelis: Writing – review & editing, Investigation. Vincenzo Penteriani: Writing – review & editing, Methodology, Funding acquisition, Data curation, Conceptualization. Funding sources During this research, D.F. was supported by a doctorate scholarship from Sapienza University of Rome. V.P. was financially supported by the I + D + i Project PID2020-114181GB-I00 funded by MCIN/AEI/10.130 39/501100011033 and by the European Union. P.C. was supported by the European Union - NextGenerationEU National Biodiversity Future Center. Declaration of competing interest All authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgements We wish to thank Miikka Husa and the Finnish Wildlife Agency for support/metadata on wildlife hunting statistics. We thank the Sapienza Statistics Department for rendering the TeraStat2 supercomputer avail able for Circuitscape.jl analyses, and Luigi Maiorano for his helpful suggestions. We are also grateful to Antero Hakala, Leo Korhonen, Reima Ovaskainen, Seppo Ronkainen, and Markus Suominen for assis tance in capturing and collaring brown bears. We thank Jon Swenson, Andrés Ordiz, John Benson, and two anonymous Reviewers for their useful suggestions, which improved our manuscript. Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi. org/10.1016/j.biocon.2026.111818. Data availability The datasets used in this study are available from the corresponding author upon reasonable request due to conservation reasons. References Abrahms, B., Sawyer, S.C., Jordan, N.R., McNutt, J.W., Wilson, A.M., Brashares, J.S., 2017. Does wildlife resource selection accurately inform corridor conservation? J. Appl. Ecol. 54, 412–422. https://doi.org/10.1111/1365-2664.12714. Agostinelli, C., Lund, U., 2024. R package ‘circular’: circular statistics. Version 0.5–1. https://CRAN.R-project.org/package=circular. D. Falcinelli et al. Biological Conservation 317 (2026) 111818 9 https://doi.org/10.13039/501100011033 https://doi.org/10.13039/501100011033 https://doi.org/10.1016/j.biocon.2026.111818 https://doi.org/10.1016/j.biocon.2026.111818 https://doi.org/10.1111/1365-2664.12714 https://CRAN.R-project.org/package=circular Ahti, T., Hämet-Ahti, L., Jalas, J., 1968. Vegetation zones and their actions in North Western Europe. Ann. Bot. Fenn. 5, 169–211. Anantharaman, R., Hall, K., Shah, V.B., Edelman, A., 2020. Circuitscape in Julia: high performance connectivity modelling to support conservation decisions. JuliaCon Proc. 1, 58. https://doi.org/10.21105/jcon.00058. Andrén, H., Linnell, J.D.C., Liberg, O., Andersen, R., Danell, A., Karlsson, J., Odden, J., Moa, P.F., Ahlqvist, P., Kvam, T., Franzén, R., Segerström, P., 2006. Survival rates and causes of mortality in Eurasian lynx (Lynx lynx) in multi-use landscapes. Biol. Conserv. 131, 23–32. https://doi.org/10.1016/j.biocon.2006.01.025. Ash, E., Cushman, S.A., Macdonald, D.W., Redford, T., Kaszta, Ż., 2020. How important are resistance, dispersal ability, population density and mortality in temporally dynamic simulations of population connectivity? A case study of tigers in Southeast Asia. Land. https://doi.org/10.3390/land9110415. Aspi, J., Roininen, E., Kiiskilä, J., Ruokonen, M., Kojola, I., Bljudnik, L., Danilov, P., Heikkinen, S., Pulliainen, E., 2009. Genetic structure of the northwestern Russian wolf populations and gene flow between Russia and Finland. Conserv. Genet. 10, 815–826. https://doi.org/10.1007/s10592-008-9642-x. Avgar, T., Potts, J.R., Lewis, M.A., Boyce, M.S., 2016. Integrated step selection analysis: bridging the gap between resource selection and animal movement. Methods Ecol. Evol. 7, 619–630. https://doi.org/10.1111/2041-210X.12528. Barry, T., Gurarie, E., Cheraghi, F., Kojola, I., Fagan, W.F., 2020. Does dispersal make the heart grow bolder? Avoidance of anthropogenic habitat elements across wolf life history. Anim. Behav. 166, 219–231. https://doi.org/10.1016/j. anbehav.2020.06.015. Bartoń, K., 2024. MuMIn: Multi-model inference. R package version 1 (48), 4. https:// CRAN.R-project.org/package=MuMIn. Behr, D.M., Ozgul, A., Cozzi, G., 2017. Combining human acceptance and habitat suitability in a unified socio-ecological suitability model: a case study of the wolf in Switzerland. J. Appl. Ecol. 54, 1919–1929. https://doi.org/10.1111/1365- 2664.12880. Beier, P., Noss, R.F., 1998. Do habitat corridors provide connectivity? Conserv. Biol. 12, 1241–1252. https://doi.org/10.1111/j.1523-1739.1998.98036.x. Benson, J.F., 2013. Improving rigour and efficiency of use-availability habitat selection analyses with systematic estimation of availability. Methods Ecol. Evol. 4, 244–251. https://doi.org/10.1111/2041-210x.12006. Benson, J.F., Dougherty, K.D., Beier, P., Boyce, W.M., Cristescu, B., Gammons, D.J., Garcelon, D.K., Higley, J.M., Martins, Q.E., Nisi, A.C., Riley, S.P.D., Sikich, J.A., Stephenson, T.R., Vickers, T.W., Wengert, G.M., Wilmers, C.C., Wittmer, H.U., Dellinger, J.A., 2023. The ecology of human-caused mortality for a protected large carnivore. Proc. Natl. Acad. Sci. 120, e2220030120. https://doi.org/10.1073/ pnas.2220030120. Benz, R.A., Boyce, M.S., Thurfjell, H., Paton, D.G., Musiani, M., Dormann, C.F., Ciuti, S., 2016. Dispersal ecology informs design of large-scale wildlife corridors. PloS One 11, 1–20. https://doi.org/10.1371/journal.pone.0162989. Bjørneraas, K., Moorter, B., Van, Rolandsen, C.M., Herfindal, I., 2010. Screening global positioning system location data for errors using animal movement characteristics. J. Wildl. Manag. 74, 1361–1366. https://doi.org/10.2193/2009-405. Boudreau, M.R., Gantchoff, M.G., Ramirez-Reyes, C., Conlee, L., Belant, J.L., Iglay, R.B., 2022. Using habitat suitability and landscape connectivity in the spatial prioritization of public outreach and management during carnivore recolonization. J. Appl. Ecol. 59, 757–767. https://doi.org/10.1111/1365-2664.14090. Boyce, M.S., McDonald, L.L., 1999. Relating populations to habitats using resource selection functions. Trends Ecol. Evol. 14, 268–272. https://doi.org/10.1016/ S0169-5347(99)01593-1. Boyce, M.S., Vernier, P.R., Nielsen, S.E., Schmiegelow, F.K.A., 2002. Evaluating resource selection functions. Ecol. Model. 157, 281–300. Brooks, M.E., Kristensen, K., van Benthem, K.J., Magnusson, A., Berg, C.W., Nielsen, A., Skaug, H.J., Mächler, M., Bolker, B.M., 2017. glmmTMB balances speed and flexibility among packages for zero-inflated generalized linear mixed modeling. R J. 9, 378–400. https://doi.org/10.32614/rj-2017-066. Brown, L., Zedrosser, A., Arnemo, J.M., Fuchs, B., Kindberg, J., Pelletier, F., 2023. Landscape of fear or landscape of food? Moose hunting triggers an antipredator response in brown bears. Ecol. Appl. 33. https://doi.org/10.1002/eap.2840. Bruskotter, J.T., Carter, N.H., Berl, R., Hinton, J., Murphy, J., Mark Elbroch, L., Vucetich, J.A., 2025. Bridging social and ecological science to create spatially explicit models of human-caused mortality of carnivores. Ambio. https://doi.org/ 10.1007/s13280-025-02165-1. Buchholtz, E.K., Stronza, A., Songhurst, A., McCulloch, G., Fitzgerald, L.A., 2020. Using landscape connectivity to predict human-wildlife conflict. Biol. Conserv. 248, 108677. https://doi.org/10.1016/j.biocon.2020.108677. Buchhorn, M., Smets, B., Bertels, L., Lesiv, M., Tsendbazar, N.E., Masiliunas, D., Linlin, L., Herold, M., Fritz, S., 2020. Copernicus Global Land Service: Land Cover 100m: Collection 3: epoch 2015: Globe (Version V3.0.1) [Data set]. Zenodo. https:// doi.org/10.5281/zenodo.3939038. Burnham, K., Anderson, D., 2002. Model Selection and Multimodel Inference: A Practical Information-Theoretic Approach, 2nd ed. Springer Verlag, Berlin, Germany. Calenge, C., 2006. The package “adehabitat” for the R software: a tool for the analysis of space and habitat use by animals. Ecol. Model. 197, 516–519. https://doi.org/ 10.1016/j.ecolmodel.2006.03.017. Carroll, K.A., Hansen, A.J., Inman, R.M., Lawrence, R.L., Hoegh, A.B., 2020. Testing landscape resistance layers and modeling connectivity for wolverines in the western United States. Glob. Ecol. Conserv. 23, e01125. https://doi.org/10.1016/j. gecco.2020.e01125. Chapron, G., Kaczensky, P., Linnell, J.D.C., von Arx, M., Huber, D., Andrén, H., López- Bao, J.V., Adamec, M., Álvares, F., Anders, O., Balčiauskas, L., Balys, V., Bedő, P., Bego, F., Blanco, J.C., Breitenmoser, U., Brøseth, H., Bufka, L., Bunikyte, R., Ciucci, P., Dutsov, A., Engleder, T., Fuxjäger, C., Groff, C., Holmala, K., Hoxha, B., Iliopoulos, Y., Ionescu, O., Jeremić, J., Jerina, K., Kluth, G., Knauer, F., Kojola, I., Kos, I., Krofel, M., Kubala, J., Kunovac, S., Kusak, J., Kutal, M., Liberg, O., Majić, A., Männil, P., Manz, R., Marboutin, E., Marucco, F., Melovski, D., Mersini, K., Mertzanis, Y., Mysłajek, R.W., Nowak, S., Odden, J., Ozolins, J., Palomero, G., Paunović, M., Persson, J., Potočnik, H., Quenette, P.-Y., Rauer, G., Reinhardt, I., Rigg, R., Ryser, A., Salvatori, V., Skrbinšek, T., Stojanov, A., Swenson, J.E., Szemethy, L., Trajçe, A., Tsingarska-Sedefcheva, E., Váňa, M., Veeroja, R., Wabakken, P., Wölfl, M., Wölfl, S., Zimmermann, F., Zlatanova, D., Boitani, L., 2014. Recovery of large carnivores in Europe’s modern human-dominated landscapes. Science 80 (346), 1517–1519. https://doi.org/10.1126/science.1257553. Corradini, A., Randles, M., Pedrotti, L., van Loon, E., Passoni, G., Oberosler, V., Rovero, F., Tattoni, C., Ciolli, M., Cagnacci, F., 2021. Effects of cumulated outdoor activity on wildlife habitat use. Biol. Conserv. 253, 108818. https://doi.org/ 10.1016/j.biocon.2020.108818. Corradini, A., Falcinelli, D., Pedrotti, L., Tattoni, C., Ranc, N., Bragalanti, N., Groff, C., Ciolli, M., Cagnacci, F., 2024. Human-induced risk drives behavioural decisions in a recovering brown bear population. Anim. Conserv. https://doi.org/10.1111/ acv.12965. Crooks, K.R., 2002. Relative sensitivities of mammalian carnivores to habitat fragmentation. Conserv. Biol. 16, 488–502. Crooks, K.R., Sanjayan, M., 2006. Connectivity conservation: Maintaining connections for nature. In: Connectivity Conservation. Cambridge University Press, pp. 1–20. https://doi.org/10.1017/CBO9780511754821.001. Danilov, P.I., 2005. Game Animals of Karelia – Ecology, Resources, Management, Protection. Nauka, Moscow. Danilov, P.I., Panchenko, D.V., Tirronen, K.F., Fyodorov, F.V., Belkin, V.V., 2018. Changes in mammal Fauna of the northern Palearctic and living range dynamics of the component species. Biol. Bull. 45, 265–277. https://doi.org/10.1134/ S1062359018030032. D’Eon, R.G., Serrouya, R., Smith, G., Kochanny, C.O., 2002. GPS Radiotelemetry error and Bias in mountainous terrain. Wildl. Soc. Bull. 30, 430–439. Dickie, M., McNay, S.R., Sutherland, G.D., Cody, M., Avgar, T., 2020. Corridors or risk? Movement along, and use of, linear features varies predictably among large mammal predator and prey species. J. Anim. Ecol. 89, 623–634. https://doi.org/10.1111/ 1365-2656.13130. Dickson, B.G., Jenness, J.S., Beier, P., 2005. Influence of vegetation, topography, and roads on cougar movement in Southern California. J. Wildl. Manag. 69, 264–276. https://doi.org/10.2193/0022-541x(2005)069<0264:iovtar>2.0.co;2. Dougherty, K.D., Dellinger, J.A., Cristescu, B., Gammons, D.J., Garcelon, D.K., Higley, J. M., Martins, Q.E., Riley, S.P.D., Sikich, J.A., Stephenson, T.R., Vickers, T.W., Wengert, G.M., Wilmers, C.C., Wittmer, H.U., Benson, J.F., 2025. A functional response in resource selection links multiscale responses of a large carnivore to human mortality risk. Ecol. Lett. 28, e70035. https://doi.org/10.1111/ele.70035. Duchesne, T., Fortin, D., Courbin, N., 2010. Mixed conditional logistic regression for habitat selection studies. J. Anim. Ecol. 79, 548–555. https://doi.org/10.1111/ j.1365-2656.2010.01670.x. Dutta, T., Sharma, S., Meyer, N.F.V., Larroque, J., Balkenhol, N., 2022. An overview of computational tools for preparing, constructing and using resistance surfaces in connectivity research. Landsc. Ecol. https://doi.org/10.1007/s10980-022-01469-x. Elliot, N.B., Cushman, S.A., Macdonald, D.W., Loveridge, A.J., 2014. The devil is in the dispersers: predictions of landscape connectivity change with demography. J. Appl. Ecol. 51, 1169–1178. https://doi.org/10.1111/1365-2664.12282. Fabrizio, M., Di Febbraro, M., D’Amico, M., Frate, L., Roscioni, F., Loy, A., 2019. Habitat suitability vs landscape connectivity determining roadkill risk at a regional scale: a case study on European badger (Meles meles). Eur. J. Wildl. Res. 65. https://doi.org/ 10.1007/s10344-018-1241-7. Falcinelli, D., del Mar Delgado, M., Kojola, I., Heikkinen, S., Lamamy, C., Penteriani, V., 2024. The use of anthropogenic areas helps explain male brown bear movement rates and distance travelled during the mating season. J. Zool. https://doi.org/ 10.1111/jzo.13199. Fieberg, J., Signer, J., Smith, B., Avgar, T., 2021. A ‘how to’ guide for interpreting parameters in habitat-selection analyses. J. Anim. Ecol. 90, 1027–1043. https://doi. org/10.1111/1365-2656.13441. Fithian, W., Hastie, T., 2013. Finite-sample equivalence in statistical models for presence-only data. Ann. Appl. Stat. 7, 1917–1939. https://doi.org/10.1214/13- AOAS667. Forester, J.D., Im, H.K., Rathouz, P.J., 2009. Accounting for animal movement in estimation of resource selection functions: sampling and data analysis. Ecology 90, 3554–3565. https://doi.org/10.1890/08-0874.1. Frangini, L., Sterrer, U., Franchini, M., Pesaro, S., Rüdisser, J., Filacorda, S., 2022. Stay home, stay safe? High habitat suitability and environmental connectivity increases road mortality in a colonizing mesocarnivore. Landsc. Ecol. 37, 2343–2361. https:// doi.org/10.1007/s10980-022-01491-z. Frankham, R., 2006. Genetics and landscape connectivity. In: Crooks, K.R., Sanjayan, M. (Eds.), Connectivity Conservation, Conservation Biology. Cambridge University Press, Cambridge, pp. 72–96. https://doi.org/10.1017/CBO9780511754821.005. Gantchoff, M.G., Erb, J.D., MacFarland, D.M., Norton, D.C., Price Tack, J.L., Roell, B.J., Belant, J.L., 2021. Potential distribution and connectivity for recolonizing cougars in the Great Lakes region, USA. Biol. Conserv. 257. https://doi.org/10.1016/j. biocon.2021.109144. García-Sánchez, M.P., González-Ávila, S., Solana-Gutiérrez, J., Popa, M., Jurj, R., Ionescu, G., Ionescu, O., Fedorca, M., Fedorca, A., 2022. Sex-specific connectivity modelling for brown bear conservation in the Carpathian Mountains. Landsc. Ecol. 37, 1311–1329. https://doi.org/10.1007/s10980-021-01367-8. D. Falcinelli et al. Biological Conservation 317 (2026) 111818 10 http://refhub.elsevier.com/S0006-3207(26)00129-1/rf0015 http://refhub.elsevier.com/S0006-3207(26)00129-1/rf0015 https://doi.org/10.21105/jcon.00058 https://doi.org/10.1016/j.biocon.2006.01.025 https://doi.org/10.3390/land9110415 https://doi.org/10.1007/s10592-008-9642-x https://doi.org/10.1111/2041-210X.12528 https://doi.org/10.1016/j.anbehav.2020.06.015 https://doi.org/10.1016/j.anbehav.2020.06.015 https://CRAN.R-project.org/package=MuMIn https://CRAN.R-project.org/package=MuMIn https://doi.org/10.1111/1365-2664.12880 https://doi.org/10.1111/1365-2664.12880 https://doi.org/10.1111/j.1523-1739.1998.98036.x https://doi.org/10.1111/2041-210x.12006 https://doi.org/10.1073/pnas.2220030120 https://doi.org/10.1073/pnas.2220030120 https://doi.org/10.1371/journal.pone.0162989 https://doi.org/10.2193/2009-405 https://doi.org/10.1111/1365-2664.14090 https://doi.org/10.1016/S0169-5347(99)01593-1 https://doi.org/10.1016/S0169-5347(99)01593-1 http://refhub.elsevier.com/S0006-3207(26)00129-1/rf0095 http://refhub.elsevier.com/S0006-3207(26)00129-1/rf0095 https://doi.org/10.32614/rj-2017-066 https://doi.org/10.1002/eap.2840 https://doi.org/10.1007/s13280-025-02165-1 https://doi.org/10.1007/s13280-025-02165-1 https://doi.org/10.1016/j.biocon.2020.108677 https://doi.org/10.5281/zenodo.3939038 https://doi.org/10.5281/zenodo.3939038 http://refhub.elsevier.com/S0006-3207(26)00129-1/rf0125 http://refhub.elsevier.com/S0006-3207(26)00129-1/rf0125 https://doi.org/10.1016/j.ecolmodel.2006.03.017 https://doi.org/10.1016/j.ecolmodel.2006.03.017 https://doi.org/10.1016/j.gecco.2020.e01125 https://doi.org/10.1016/j.gecco.2020.e01125 https://doi.org/10.1126/science.1257553 https://doi.org/10.1016/j.biocon.2020.108818 https://doi.org/10.1016/j.biocon.2020.108818 https://doi.org/10.1111/acv.12965 https://doi.org/10.1111/acv.12965 http://refhub.elsevier.com/S0006-3207(26)00129-1/rf0155 http://refhub.elsevier.com/S0006-3207(26)00129-1/rf0155 https://doi.org/10.1017/CBO9780511754821.001 http://refhub.elsevier.com/S0006-3207(26)00129-1/rf0165 http://refhub.elsevier.com/S0006-3207(26)00129-1/rf0165 https://doi.org/10.1134/S1062359018030032 https://doi.org/10.1134/S1062359018030032 http://refhub.elsevier.com/S0006-3207(26)00129-1/rf0175 http://refhub.elsevier.com/S0006-3207(26)00129-1/rf0175 https://doi.org/10.1111/1365-2656.13130 https://doi.org/10.1111/1365-2656.13130 https://doi.org/10.2193/0022-541x(2005)069<0264:iovtar>2.0.co;2 https://doi.org/10.1111/ele.70035 https://doi.org/10.1111/j.1365-2656.2010.01670.x https://doi.org/10.1111/j.1365-2656.2010.01670.x https://doi.org/10.1007/s10980-022-01469-x https://doi.org/10.1111/1365-2664.12282 https://doi.org/10.1007/s10344-018-1241-7 https://doi.org/10.1007/s10344-018-1241-7 https://doi.org/10.1111/jzo.13199 https://doi.org/10.1111/jzo.13199 https://doi.org/10.1111/1365-2656.13441 https://doi.org/10.1111/1365-2656.13441 https://doi.org/10.1214/13-AOAS667 https://doi.org/10.1214/13-AOAS667 https://doi.org/10.1890/08-0874.1 https://doi.org/10.1007/s10980-022-01491-z https://doi.org/10.1007/s10980-022-01491-z https://doi.org/10.1017/CBO9780511754821.005 https://doi.org/10.1016/j.biocon.2021.109144 https://doi.org/10.1016/j.biocon.2021.109144 https://doi.org/10.1007/s10980-021-01367-8 Ghoddousi, A., Buchholtz, E.K., Dietsch, A.M., Williamson, M.A., Sharma, S., Balkenhol, N., Kuemmerle, T., Dutta, T., 2021. Anthropogenic resistance: accounting for human behavior in wildlife connectivity planning. One Earth 4, 39–48. https:// doi.org/10.1016/j.oneear.2020.12.003. Grilo, C., Ascensão, F., Santos-Reis, M., Bissonette, J.A., 2011. Do well-connected landscapes promote road-related mortality? Eur. J. Wildl. Res. 57, 707–716. https:// doi.org/10.1007/s10344-010-0478-6. Grueber, C.E., Nakagawa, S., Laws, R.J., Jamieson, I.G., 2011. Multimodel inference in ecology and evolution: challenges and solutions. J. Evol. Biol. 24, 699–711. https:// doi.org/10.1111/j.1420-9101.2010.02210.x. Gurarie, E., Suutarinen, J., Kojola, I., Ovaskainen, O., 2011. Summer movements, predation and habitat use of wolves in human modified boreal forests. Oecologia 165, 891–903. https://doi.org/10.1007/s00442-010-1883-y. Hagen, S.B., Kopatz, A., Aspi, J., Kojola, I., Geir Eiken, H., 2015. Evidence of rapid change in genetic structure and diversity during range expansion in a recovering large terrestrial carnivore. Proc. R. Soc. B Biol. Sci. 282, 13–16. https://doi.org/ 10.1098/rspb.2015.0092. Hanski, I., 1999. Metapopulation ecology. Oxford University Press. https://doi.org/ 10.1093/oso/9780198540663.001.0001. Heikkinen, S., Kojola, I., Mäntyniemi, S., 2024. Karhukanta Suomessa 2023. In: Luonnonvara- ja biotalouden tutkimus 19/2024. Luonnonvarakeskus, Helsinki, p. 17. Hobbs, R.J., 1992. The role of corridors in conservation: solution or bandwagon? Trends Ecol. Evol. 7, 389–392. Hofmann, D.D., Behr, D.M., McNutt, J.W., Ozgul, A., Cozzi, G., 2021. Bound within boundaries: do protected areas cover movement corridors of their most mobile, protected species? J. Appl. Ecol. 58, 1133–1144. https://doi.org/10.1111/1365- 2664.13868. James, A.C.R., Stuart-Smith, A.K., 2000. Distribution of Caribou and wolves in relation to linear corridors. J. Wildl. Manag. 64, 154–159. Karivalo, L., Butorin, A., 2006. The fennoscandian green belt. In: Terry, A., Ullrich, K., Riecken, U. (Eds.), The Green Belt of Europe–From Vision to Reality. IUCN, Gland, pp. 37–45. Kaszta, Ż., Cushman, S.A., Hearn, A.J., Burnham, D., Macdonald, E.A., Goossens, B., Nathan, S.K.S.S., Macdonald, D.W., 2019. Integrating Sunda clouded leopard (Neofelis diardi) conservation into development and restoration planning in Sabah (Borneo). Biol. Conserv. 235, 63–76. https://doi.org/10.1016/j. biocon.2019.04.001. Kaszta, Ż., Cushman, S.A., Htun, S., Naing, H., Burnham, D., Macdonald, D.W., 2020. Simulating the impact of belt and road initiative and other major developments in Myanmar on an ambassador felid, the clouded leopard, Neofelis nebulosa. Landsc. Ecol. 35, 727–746. https://doi.org/10.1007/s10980-020-00976-z. Keeley, A.T.H., Beier, P., Gagnon, J.W., 2016. Estimating landscape resistance from habitat suitability: effects of data source and nonlinearities. Landsc. Ecol. 31, 2151–2162. https://doi.org/10.1007/s10980-016-0387-5. Kojola, I., Danilov, P.I., Laitala, H.-M., Belkin, V., Yakimov, A., 2003. Brown bear population structure in core and periphery: analysis of hunting statistics from Russian Karelia and Finland. Ursus 14, 17–20. Kojola, I., Hallikainen, V., Nygren, T., Pesonen, M., Ruusila, V., 2006. Recent trends and harvest in Finland’s brown bear population. Ursus 17, 159–164. https://doi.org/ 10.2192/1537-6176(2006)17[159:RTAHIF]2.0.CO;2. Kojola, I., Kaartinen, S., Hakala, A., Heikkinen, S., Voipio, H.-M., 2009. Dispersal behavior and the connectivity between wolf populations in northern Europe. J. Wildl. Manag. 73, 309–313. https://doi.org/10.2193/2007-539. Kojola, I., Hallikainen, V., Heikkinen, S., Nivala, V., 2020. Has the sex-specific structure of Finland’s brown bear population changed during 21 years? Wildl. Biol. 2020. https://doi.org/10.2981/wlb.00575. Kojola, I., Hallikainen, V., Heikkinen, S., Nivala, V., 2021. Inadvertent shooting of brown bear cubs in Finland: what can managers do to reduce it? Wildl. Biol. 2021. https:// doi.org/10.2981/wlb.00773. Kopatz, A., Eiken, H.G., Hagen, S.B., Ruokonen, M., Esparza-Salas, R., Schregel, J., Kojola, I., Smith, M.E., Wartiainen, I., Aspholm, P.E., Wikan, S., Rykov, A.M., Makarova, O., Polikarpova, N., Tirronen, K.F., Danilov, P.I., Aspi, J., 2012. Connectivity and population subdivision at the fringe of a large brown bear (Ursus arctos) population in North Western Europe. Conserv. Genet. 13, 681–692. https:// doi.org/10.1007/s10592-012-0317-2. Kopatz, A., Eiken, H.G., Aspi, J., Kojola, I., Tobiassen, C., Tirronen, K.F., Danilov, P.I., Hagen, S.B., 2014. Admixture and gene flow from Russia in the recovering northern European brown bear (Ursus arctos). PloS One 9, 1–10. https://doi.org/10.1371/ journal.pone.0097558. Kopatz, A., Kleven, O., Kojola, I., Aspi, J., Norman, A.J., Spong, G., Gyllenstrand, N., Dalén, L., Fløystad, I., Hagen, S.B., Kindberg, J., Flagstad, Ø., 2021. Restoration of transborder connectivity for Fennoscandian brown bears (Ursus arctos). Biol. Conserv. 253, 108936. https://doi.org/10.1016/j.biocon.2020.108936. Lamamy, C., Delgado, M.M., Kojola, I., Heikkinen, S., Penteriani, V., 2022. Does moonlight affect movement patterns of a non-obligate carnivore? Brown bears do not mind that the moon exists. J. Zool. 316, 128–138. https://doi.org/10.1111/ jzo.12938. Lamb, C.T., Ford, A.T., Mclellan, B.N., Proctor, M.F., Mowat, G., Ciarniello, L., Nielsen, S. E., Boutin, S., Estes, J.A., 2020. The ecology of human-carnivore coexistence. Proc. Natl. Acad. Sci. 117, 17876–17883. https://doi.org/10.1073/pnas.1922097117/-/ DCSupplemental. Liberg, O., Chapron, G., Wabakken, P., Pedersen, H.C., Thompson Hobbs, N., Sand, H., 2012. Shoot, shovel and shut up: cryptic poaching slows restoration of a large carnivore in Europe. Proc. R. Soc. B Biol. Sci. 279, 910–915. https://doi.org/ 10.1098/rspb.2011.1275. Linden, H., Danilov, P.I., Gromtsev, A.N., Helle, P., Ivanter, E.V., Kurhinen, J., 2000. Large-scale forest corridors to connect the taiga fauna to Fennoscandia. Wildl. Biol. 6, 179–188. https://doi.org/10.2981/wlb.2000.007. Määttänen, A.M., Virkkala, R., Leikola, N., Heikkinen, R.K., 2022. Increasing loss of mature boreal forests around protected areas with red-listed forest species. Ecol. Process. 11. https://doi.org/10.1186/s13717-022-00361-5. Maiorano, L., Chiaverini, L., Falco, M., Ciucci, P., 2019. Combining multi-state species distribution models, mortality estimates, and landscape connectivity to model potential species distribution for endangered species in human dominated landscapes. Biol. Conserv. 237, 19–27. https://doi.org/10.1016/j. biocon.2019.06.014. Mateo-Sánchez, M.C., Balkenhol, N., Cushman, S., Pérez, T., Domínguez, A., Saura, S., 2015. Estimating effective landscape distances and movement corridors: comparison of habitat and genetic data. Ecosphere 6, 1–16. https://doi.org/10.1890/ES14- 00387.1. Mayer, A.L., Kauppi, P.E., Angelstam, P.K., Zhang, Y., Tikka, P.M., 2005. Importing timber, exporting ecological impact. Science 80 (308), 359–360. https://doi.org/ 10.1126/science.1109476. Mayer, P., Grêt-Regamey, A., Ciucci, P., Salliou, N., Stritih, A., 2023. Mapping human- and bear-centered perspectives on coexistence using a participatory Bayesian framework. J. Nat. Conserv. 73. https://doi.org/10.1016/j.jnc.2023.126387. McClure, M.L., Hansen, A.J., Inman, R.M., 2016. Connecting models to movements: testing connectivity model predictions against empirical migration and dispersal data. Landsc. Ecol. 31, 1419–1432. https://doi.org/10.1007/s10980-016-0347-0. McRae, B.H., Dickson, B.G., Keitt, T.H., Shah, V.B., 2008. Using circuit theory to model connectivity in ecology, evolution, and conservation. Ecology 89, 2712–2724. https://doi.org/10.1890/07-1861.1. Merrick, M.J., Koprowski, J.L., 2017. Circuit theory to estimate natal dispersal routes and functional landscape connectivity for an endangered small mammal. Landsc. Ecol. 32, 1163–1179. https://doi.org/10.1007/s10980-017-0521-z. Milleret, C., Dupont, P., Chapron, G., Swenson, J.E., Bischof, R., 2024. Sweden is shooting brown bears in the dark. Science 80 (385), 1052. https://doi.org/10.1126/ science.adq7487. Muff, S., Signer, J., Fieberg, J., 2020. Accounting for individual-specific variation in habitat-selection studies: efficient estimation of mixed-effects models using Bayesian or frequentist computation. J. Anim. Ecol. 89, 80–92. https://doi.org/10.1111/ 1365-2656.13087. Muukkonen, P., Takala, T., Virtanen, T., 2009. Differences in the forest landscape structure along the Finnish-Russian border in southern Karelia. Scand. J. For. Res. 24, 140–148. https://doi.org/10.1080/02827580902758851. Naves, J., Wiegand, T., Revilla, E., Delibes, M., 2003. Endangered species constrained by natural and human factors: the case of Brown bears in northern Spain. Conserv. Biol. 17, 1276–1289. https://doi.org/10.1046/j.1523-1739.2003.02144.x. Nielsen, S.E., Herrero, S., Boyce, M.S., MacE, R.D., Benn, B., Gibeau, M.L., Jevons, S., 2004. Modelling the spatial distribution of human-caused grizzly bear mortalities in the central Rockies ecosystem of Canada. Biol. Conserv. 120, 101–113. https://doi. org/10.1016/j.biocon.2004.02.020. Nielsen, S.E., Stenhouse, G.B., Boyce, M.S., 2006. A habitat-based framework for grizzly bear conservation in Alberta. Biol. Conserv. 130, 217–229. https://doi.org/10.1016/ j.biocon.2005.12.016. Northrup, J.M., Hooten, M.B., Anderson Jr., C.R., Wittemyer, G., 2013. Practical guidance on characterizing availability in resource selection functions under a use–availability design. Ecology 94, 1456–1463. https://doi.org/10.1890/12- 1688.1. OpenStreetMap contributors, 2023. Planet Dump. retrieved from. https://planet.osm. org. Accessed 2023-05-12 from. https://www.openstreetmap.org. Ordiz, A., Støen, O.-G., Delibes, M., Swenson, J.E., 2011. Predators or prey? Spatio- temporal discrimination of human-derived risk by brown bears. Oecologia 166, 59–67. https://doi.org/10.1007/s00442-011-1920-5. Osipova, L., Okello, M.M., Njumbi, S.J., Ngene, S., Western, D., Hayward, M.W., Balkenhol, N., 2019. Using step-selection functions to model landscape connectivity for African elephants: accounting for variability across individuals and seasons. Anim. Conserv. 22, 35–48. https://doi.org/10.1111/acv.12432. Panzacchi, M., Van Moorter, B., Strand, O., Saerens, M., Kivimäki, I., St. Clair, C.C., Herfindal, I., Boitani, L., 2016. Predicting the continuum between corridors and barriers to animal movements using step selection functions and randomized shortest paths. J. Anim. Ecol. 85, 32–42. https://doi.org/10.1111/1365- 2656.12386. Pearman, P.B., Broennimann, O., Aavik, T., Albayrak, T., Alves, P.C., Aravanopoulos, F. A., Bertola, L.D., Biedrzycka, A., Buzan, E., Cubric-Curik, V., Djan, M., Fedorca, A., Fuentes-Pardo, A.P., Fussi, B., Godoy, J.A., Gugerli, F., Hoban, S., Holderegger, R., Hvilsom, C., Iacolina, L., Kalamujic Stroil, B., Klinga, P., Konopiński, M.K., Kopatz, A., Laikre, L., Lopes-Fernandes, M., McMahon, B.J., Mergeay, J., Neophytou, C., Pálsson, S., Paz-Vinas, I., Posledovich, D., Primmer, C.R., Raeymaekers, J.A.M., Rinkevich, B., Rolečková, B., Ruņģis, D., Schuerz, L., Segelbacher, G., Kavčič Sonnenschein, K., Stefanovic, M., Thurfjell, H., Träger, S., Tsvetkov, I.N., Velickovic, N., Vergeer, P., Vernesi, C., Vilà, C., Westergren, M., Zachos, F.E., Guisan, A., Bruford, M., 2024. Monitoring of species’ genetic diversity in Europe varies greatly and overlooks potential climate change impacts. Nat. Ecol. Evol. 8, 267–281. https://doi.org/10.1038/s41559-023-02260-0. Peck, C.P., VanManen, F.T., Costello, C.M., Haroldson, M.A., Landenburger, L.A., Roberts, L.L., Bjornlie, D.D., Mace, R.D., 2017. Potential paths for male-mediated gene flow to and from an isolated grizzly bear population. Ecosphere 8. https://doi. org/10.1002/ecs2.1969. Penteriani, V., Huber, D., Jerina, K., Krofel, M., Lopez-Bao, J.V., Ordiz, A., Zarzo- Arias, A., Dalerum, F., 2018. Trans-boundary and trans-regional management of a D. Falcinelli et al. Biological Conservation 317 (2026) 111818 11 https://doi.org/10.1016/j.oneear.2020.12.003 https://doi.org/10.1016/j.oneear.2020.12.003 https://doi.org/10.1007/s10344-010-0478-6 https://doi.org/10.1007/s10344-010-0478-6 https://doi.org/10.1111/j.1420-9101.2010.02210.x https://doi.org/10.1111/j.1420-9101.2010.02210.x https://doi.org/10.1007/s00442-010-1883-y https://doi.org/10.1098/rspb.2015.0092 https://doi.org/10.1098/rspb.2015.0092 https://doi.org/10.1093/oso/9780198540663.001.0001 https://doi.org/10.1093/oso/9780198540663.001.0001 http://refhub.elsevier.com/S0006-3207(26)00129-1/rf0285 http://refhub.elsevier.com/S0006-3207(26)00129-1/rf0285 http://refhub.elsevier.com/S0006-3207(26)00129-1/rf0285 http://refhub.elsevier.com/S0006-3207(26)00129-1/rf0290 http://refhub.elsevier.com/S0006-3207(26)00129-1/rf0290 https://doi.org/10.1111/1365-2664.13868 https://doi.org/10.1111/1365-2664.13868 http://refhub.elsevier.com/S0006-3207(26)00129-1/rf0300 http://refhub.elsevier.com/S0006-3207(26)00129-1/rf0300 http://refhub.elsevier.com/S0006-3207(26)00129-1/rf0305 http://refhub.elsevier.com/S0006-3207(26)00129-1/rf0305 http://refhub.elsevier.com/S0006-3207(26)00129-1/rf0305 https://doi.org/10.1016/j.biocon.2019.04.001 https://doi.org/10.1016/j.biocon.2019.04.001 https://doi.org/10.1007/s10980-020-00976-z https://doi.org/10.1007/s10980-016-0387-5 http://refhub.elsevier.com/S0006-3207(26)00129-1/rf0325 http://refhub.elsevier.com/S0006-3207(26)00129-1/rf0325 http://refhub.elsevier.com/S0006-3207(26)00129-1/rf0325 https://doi.org/10.2192/1537-6176(2006)17[159:RTAHIF]2.0.CO;2 https://doi.org/10.2192/1537-6176(2006)17[159:RTAHIF]2.0.CO;2 https://doi.org/10.2193/2007-539 https://doi.org/10.2981/wlb.00575 https://doi.org/10.2981/wlb.00773 https://doi.org/10.2981/wlb.00773 https://doi.org/10.1007/s10592-012-0317-2 https://doi.org/10.1007/s10592-012-0317-2 https://doi.org/10.1371/journal.pone.0097558 https://doi.org/10.1371/journal.pone.0097558 https://doi.org/10.1016/j.biocon.2020.108936 https://doi.org/10.1111/jzo.12938 https://doi.org/10.1111/jzo.12938 https://doi.org/10.1073/pnas.1922097117/-/DCSupplemental https://doi.org/10.1073/pnas.1922097117/-/DCSupplemental https://doi.org/10.1098/rspb.2011.1275 https://doi.org/10.1098/rspb.2011.1275 https://doi.org/10.2981/wlb.2000.007 https://doi.org/10.1186/s13717-022-00361-5 https://doi.org/10.1016/j.biocon.2019.06.014 https://doi.org/10.1016/j.biocon.2019.06.014 https://doi.org/10.1890/ES14-00387.1 https://doi.org/10.1890/ES14-00387.1 https://doi.org/10.1126/science.1109476 https://doi.org/10.1126/science.1109476 https://doi.org/10.1016/j.jnc.2023.126387 https://doi.org/10.1007/s10980-016-0347-0 https://doi.org/10.1890/07-1861.1 https://doi.org/10.1007/s10980-017-0521-z https://doi.org/10.1126/science.adq7487 https://doi.org/10.1126/science.adq7487 https://doi.org/10.1111/1365-2656.13087 https://doi.org/10.1111/1365-2656.13087 https://doi.org/10.1080/02827580902758851 https://doi.org/10.1046/j.1523-1739.2003.02144.x https://doi.org/10.1016/j.biocon.2004.02.020 https://doi.org/10.1016/j.biocon.2004.02.020 https://doi.org/10.1016/j.biocon.2005.12.016 https://doi.org/10.1016/j.biocon.2005.12.016 https://doi.org/10.1890/12-1688.1 https://doi.org/10.1890/12-1688.1 https://planet.osm.org https://planet.osm.org https://www.openstreetmap.org https://doi.org/10.1007/s00442-011-1920-5 https://doi.org/10.1111/acv.12432 https://doi.org/10.1111/1365-2656.12386 https://doi.org/10.1111/1365-2656.12386 https://doi.org/10.1038/s41559-023-02260-0 https://doi.org/10.1002/ecs2.1969 https://doi.org/10.1002/ecs2.1969 http://refhub.elsevier.com/S0006-3207(26)00129-1/rf0490 http://refhub.elsevier.com/S0006-3207(26)00129-1/rf0490 large carnivore: Managing brown bears across national and regional borders in Europe. In: Hovardas, T. (Ed.), Large Carnivore Conservation and Management: Human Dimensions and Governance. Taylor & Francis Group, London and New York, pp. 291–302. Penteriani, V., Lamamy, C., Kojola, I., Heikkinen, S., Bombieri, G., del Mar Delgado, M., Delgado, M., 2021. Does artificial feeding affect large carnivore behaviours? The case study of brown bears in a hunted and tourist exploited subpopulation. Biol. Conserv. 254, 108949. https://doi.org/10.1016/j.biocon.2021.108949. Penteriani, V., Lamamy, C., Kojola, I., Heikkinen, S., Vermeulen, C., Delgado, M. del M., 2022. Age Ain’t nothing but a number: factors other than age shape brown bear movement patterns. Anim. Behav. 183, 61–67. https://doi.org/10.1016/j. anbehav.2021.10.020. Penteriani, V., Delgado, M. del M., Kojola, I., Heikkinen, S., Fedorca, A., García- Sánchez, P., Fedorca, M., Find’o, S., Skuban, M., Balbontín, J., Zarzo-Arias, A., Falcinelli, D., Ordiz, A., Swenson, J.E., 2025. Mating from a female perspective: do brown bear females play an active role in mate searching? Mov. Ecol. 13, 24. https://doi.org/10.1186/s40462-025-00553-6. Persson, J., Ericsson, G., Segerström, P., 2009. Human caused mortality in the endangered Scandinavian wolverine population. Biol. Conserv. 142, 325–331. https://doi.org/10.1016/j.biocon.2008.10.028. Pitman, R.T., Swanepoel, L.H., Hunter, L., Slotow, R., Balme, G.A., 2015. The importance of refugia, ecological traps and scale for large carnivore management. Biodivers. Conserv. 24, 1975–1987. https://doi.org/10.1007/s10531-015-0921-9. Plante, S., Dussault, C., Côté, S.D., 2017. Landscape attributes explain migratory caribou vulnerability to sport hunting. J. Wildl. Manag. 81, 238–247. https://doi.org/ 10.1002/jwmg.21203. Proctor, M.F., Paetkau, D., McLellan, B.N., Stenhouse, G.B., Kendall, K.C., MacE, R.D., Kasworm, W.F., Servheen, C., Lausen, C.L., Gibeau, M.L., Wakkinen, W.L., Haroldson, M.A., Mowat, G., Apps, C.D., Ciarniello, L.M., Barclay, R.M.R., Boyce, M. S., Schwartz, C.C., Strobeck, C., 2012. Population fragmentation and inter-ecosystem movements of grizzly bears in Western Canada and the northern United States. Wildl. Monogr. 1–46. https://doi.org/10.1002/wmon.6. Pulliainen, E., 1986. Brown bear immigration into Finland from the east. Int. Conf. Bear Res. Manag. 6, 15–20. QGIS Development Team, 2023. QGIS geographic information system. Open Source Geospatial Foundation Project. http://qgis.osgeo.org. R Core Team, 2024. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing. https://www.R-project.org/. Rasmus, S., Kojola, I., Turunen, M., Norberg, H., Kumpula, J., Ollila, T., 2020. Mission impossible? Pursuing the co-existence of viable predator populations and sustainable reindeer husbandry in Finland. J. Rural. Stud. 80, 135–148. https://doi.org/ 10.1016/j.jrurstud.2020.08.017. Recio, M.R., Knauer, F., Molinari-Jobin, A., Huber, Filacorda, S., Jerina, K., 2021. Context-dependent behaviour and connectivity of recolonizing brown bear populations identify transboundary conservation challenges in Central Europe. Anim. Conserv. 24, 73–83. https://doi.org/10.1111/acv.12624. Revilla, E., Wiegand, T., Palomares, F., Ferreras, P., Delibes, M., 2004. Effects of matrix heterogeneity on animal dispersal: from individual behavior to metapopulation-level parameters. Am. Nat. 164. https://doi.org/10.1086/424767. Ripple, W.J., Estes, J.A., Beschta, R.L., Wilmers, C.C., Ritchie, E.G., Hebblewhite, M., Berger, J., Elmhagen, B., Letnic, M., Nelson, M.P., Schmitz, O.J., Smith, D.W., Wallach, A.D., Wirsing, A.J., 2014. Status and ecological effects of the world’s largest carnivores. Science (80-. ). https://doi.org/10.1126/science.1241484. Roberts, D.R., Bahn, V., Ciuti, S., Boyce, M.S., Elith, J., Guillera-arroita, G., Hauenstein, S., Lahoz-monfort, J.J., Schröder, B., Thuiller, W., Warton, D.I., Wintle, B.A., Hartig, F., Dormann, C.F., 2017. Cross-validation strategies for data with temporal, spatial, hierarchical, or phylogenetic structure. Ecography 913–929. https://doi.org/10.1111/ecog.02881. Schloss, C.A., Nuñez, T.A., Lawler, J.J., 2012. Dispersal will limit ability of mammals to track climate change in the Western hemisphere. Proc. Natl. Acad. Sci. 109, 8606–8611. https://doi.org/10.1073/pnas.1116791109. Schregel, J., Kopatz, A., Hagen, S.B., Broseth, H., Smith, M.E., Wikan, S., Wartiainen, I., Aspholm, P.E., Aspi, J., Swenson, J.E., Makarova, O., Polikarpova, N., Schneider, M., Knappskog, P.M., Ruokonen, M., Kojola, I., Tirronen, K.F., Danilov, P.I., Eiken, H.G., 2012. Limited gene flow among brown bear populations in far northern Europe? Genetic analysis of the east-west border population in the Pasvik Valley. Mol. Ecol. 21, 3474–3488. https://doi.org/10.1111/j.1365-294X.2012.05631.x. Schregel, J., Eiken, H.G., Grøndahl, F.A., Hailer, F., Aspi, J., Kojola, I., Tirronen, K., Danilov, P., Rykov, A., Poroshin, E., Janke, A., Swenson, J.E., Hagen, S.B., 2015. Y chromosome haplotype distribution of brown bears (Ursus arctos) in northern Europe provides insight into population history and recovery. Mol. Ecol. 24, 6041–6060. https://doi.org/10.1111/mec.13448. Simberloff, D., Farr, J.A., Cox, J., Mehlman, D.W., 1992. Movement corridors: conservation bargains or poor investments? Conserv. Biol. 6, 493–504. Sivertsen, T.R., Åhman, B., Steyaert, S.M.J.G., Rönnegård, L., Frank, J., Segerström, P., Støen, O.G., Skarin, A., 2016. Reindeer habitat selection under the risk of brown bear predation during calving season. Ecosphere 7, 1–17. https://doi.org/10.1002/ ecs2.1583. Squires, J.R., DeCesare, N.J., Olson, L.E., Kolbe, J.A., Hebblewhite, M., Parks, S.A., 2013. Combining resource selection and movement behavior to predict corridors for Canada lynx at their southern range periphery. Biol. Conserv. 157, 187–195. https:// doi.org/10.1016/j.biocon.2012.07.018. Steenweg, R., Hebblewhite, M., Burton, C., Whittington, J., Heim, N., Fisher, J.T., Ladle, A., Lowe, W., Muhly, T., Paczkowski, J., Musiani, M., 2023. Testing umbrella species and food-web properties of large carnivores in the Rocky Mountains. Biol. Conserv. 278, 109888. https://doi.org/10.1016/j.biocon.2022.109888. Steyaert, S.M.J.G., Leclerc, M., Pelletier, F., Kindberg, J., Brunberg, S., Swenson, J.E., Zedrosser, A., 2016b. Human shields mediate sexual conflict in a top predator. Proc. R. Soc. B Biol. Sci. 283. https://doi.org/10.1098/rspb.2016.0906. Steyaert, Sam M.J.G., Zedrosser, A., Elfström, M., Ordiz, A., Leclerc, M., Frank, S.C., Kindberg, J., Støen, O.G., Brunberg, S., Swenson, J.E., 2016a. Ecological implications from spatial patterns in human-caused brown bear mortality. Wildl. Biol. 22, 144–152. https://doi.org/10.2981/wlb.00165. Støen, O.G., Zedrosser, A., Sæbø, S., Swenson, J.E., 2006. Inversely density-dependent natal dispersal in brown bears Ursus arctos. Oecologia 148, 356–364. https://doi. org/10.1007/s00442-006-0384-5. Stoner, D.C., Wolfe, M.L., Rieth, W.R., Bunnell, K.D., Durham, S.L., Stoner, L.L., 2013. De facto refugia, ecological traps and the biogeography of anthropogenic cougar mortality in Utah. Divers. Distrib. 19, 1114–1124. https://doi.org/10.1111/ ddi.12035. Swenson, J.E., Wabakken, P., Sandegren, F., Bjärvall, A., Franzén, R., Söderberg, A., 1995. The near extinction and recovery of brown bears in Scandinavia in relation to the bear management policies of Norway and Sweden. Wildl. Biol. 1,