Different environmental contexts, different responses: Evaluating the drivers of red fox occupancy patterns in Portugal

doi:10.21203/rs.3.rs-5200050/v1

Download PDF

Research Article

Different environmental contexts, different responses: Evaluating the drivers of red fox occupancy patterns in Portugal

https://doi.org/10.21203/rs.3.rs-5200050/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Identifying the drivers and ecological mechanisms that shape species occupancy patterns at large scales is decisive in understanding how species respond to environmental changes. In this study, we compiled camera trapping data from different scientific projects and monitoring programs to conduct the first assessment of the relevant ecological drivers of red fox (Vulpes vulpes) occupancy patterns at the national level in Portugal. Here, using single-species/single-season occupancy models, we target two regions with different environmental contexts - north and south of River Tagus. Our results show that the red fox occupancy drivers change depending on the environmental context. In the northern region (more mountainous and densely forested), occupancy patterns were mostly shaped by direct human-associated disturbance, positively associated with dogs’ relative abundance index. On the other hand, in the southern region (plain areas dominated by agroforestry systems), the variation in red fox occupancy patterns was best explained by the habitat characteristics, with foxes showing a higher probability of occupying areas with greater habitat diversity. Considering these results, we conclude that the red fox’s generalist character is highlighted by its capacity to adopt two different occupancy strategies depending on the environmental context: the use of more humanized areas, along with free-ranging dogs, in the Northern region, and the use of areas with more diversified resources in the Southern region. Therefore, our study reinforces the red fox's highly adaptive character, adding crucial information to tailor red fox management and conservation actions to the different regional environmental contexts of Portugal.

Camera-trapping

free-ranging dogs

habitat diversity

occupancy models

spatial patterns

Vulpes vulpes

The European Mediterranean region has a long-standing human presence, whose activities have shaped the landscapes for millenniums (Blondel 2006). Traditional agricultural fields have dominated the landscape for centuries, making it a highly heterogeneous and an unique seminatural ecosystem (Blondel 2006; Blondel et al. 2010). Because of this typical landscape context, Mediterranean wildlife communities have long adapted to humans and their activities, enabling their coexistence in this biodiversity hotspot (Mangas et al. 2008; Blondel et al. 2010).

However, in the last decades, the Mediterranean landscape is undergoing profound habitat transformation (Pinto-Correia 2000; Torres et al. 2014) due to the abandonment of rural lands together with the conversion of traditional subsistence agricultural fields into intensive agriculture (e.g., olive tree plantations, avocado trees) and exotic trees monocultures (e.g., Eucalyptus spp.) (Geri et al. 2010; Pereira et al. 2012; Cruz et al. 2015). Portugal, a Mediterranean southern European country, is no exception since native habitats and traditional agriculture fields are rapidly being replaced by forest plantations of Eucalyptus spp. and Pinus spp. (especially in the central and northern regions) and intensive agriculture (more evident in the southern region), transforming the land into a more homogeneous landscape (Pinto-Correia et al. 2004; Jones et al. 2011; Almeida 2020).

This habitat modification has an impact upon the entire ecosystem functioning, destroying or changing ecological niches, and largely affecting most of the native wildlife, often inducing a change in their spatial pattern (Pereira et al. 2012; Cruz et al. 2015; Curveira-Santos 2017; Alexandre et al. 2020; Castro et al. 2022; Ares-Pereira et al. 2022; Linck et al. 2023). However, animal sensitivity to disturbance varies between taxonomic groups (Burton et al. 2024). For instance, generalist mesocarnivores present some ecological features (e.g., wide ecological niche - Marneweck et al. 2022), allowing them to be more resilient to disturbance, namely habitat modification. Depending on food availability, habitat characteristics, and competitors’ presence, these species can adjust and shift their diet and habitat use patterns (Jiménez et al. 2019; Alexandre et al. 2020; Sarmento et al. 2021; Marneweck et al. 2022). All these characteristics make the mesocarnivores an adequate model group to explore the impact of global changes on species’ spatial patterns, as landscape modification tends to favor species more adaptable to change.

The red fox (Vulpes vulpes) is one of the most common carnivores in Portugal (Mathias et al. 2023). This generalist and non-threatened canid has a wide distribution in Portugal (Bencatel et al. 2018; Negrões et al. 2023), being considered a widely adaptable species, mainly due to its diverse diet and opportunistic behavior (Bencatel et al. 2019; Álvares et al. 2019). This species inhabits both natural or anthropogenic landscapes and shows multiple land use strategies depending on the landscape context (Pereira et al. 2012; Curveira-Santos et al. 2017; Alexandre et al. 2020), although more recent studies have shown that human-altered habitats may have a negative impact on this species occurrence (Alexandre et al. 2020; Castro et al. 2022). Most studies assessing the red fox occupancy patterns in Portugal have been conducted in specific local environmental contexts (Sarmento et al. 2011; Monterroso et al. 2014; Cruz et al. 2015; Curveira-Santos et al. 2019; Castro et al. 2022), but large-scale assessments are needed to add information relevant to design better management actions or tools (Toews et al. 2017). However, broad-scale assessments for non-threatened wide-ranging species are scarce (Moll et al. 2016), and few studies have assessed red fox spatial patterns and drivers at broader scales in Portugal (Teixeira et al. 2023 b). To fill this gap, this study aims to identify the factors that explain the occupancy patterns variation of red fox in Portugal. Specifically, we intended to explore how the distinct environmental contexts (i.e., the regions south and north of Tagus River) could influence the type of drivers affecting red fox occupancy patterns. We hypothesized that the red fox will adapt its occupancy pattern to the regional environmental context and, therefore, its spatial pattern will respond differently to distinct factors shaping the landscapes North and South of the Tagus River.

Study Area:

Portugal is a highly diverse and heterogeneous country with two main distinct environmental regions, north and south of the Tagus River (Fig. 1). This division is based on these regions’ distinct climatic and orographic characteristics. The climate in both regions is mainly Mediterranean, characterized by hot dry summers and cold wet winters, with the northern region presenting, in general, a strong Atlantic influence (with some exceptions near the border with Spain), with higher precipitation levels, but lower temperatures than the southern region, which presents a stronger Continental influence (Hijmans et al. 2005; Loidi 2017). Also, the north region of the Tagus River presents a more mountainous terrain and a denser and longer hydrographic network. At the same time, the south is dominated by plains irrigated by a more scattered and smaller hydrographic network, often composed of intermittent streams (SNIRH 2015). The northern region of the Tagus River is more urbanized, with a denser road network and higher population density (Alves et al. 2009). Large cultivated areas, and Eucalyptus spp. and Pinus spp. monocultures mixed with several patches of broadleaves forests and shrublands dominate this region. On the other hand, the south is dominated by shrublands and Quercus spp. agroforestry systems (Álvares et al. 2019) - “montado”, a typical Portuguese multifunctional agroforestry system (Pinto-Correia et al. 2004).

Data Collection and Tested Environmental Drivers:

Our dataset consisted of red fox records from camera traps collected from different ongoing projects and implemented at a broad scale from north to south Portugal. Hence, red fox records were collected in a wide variety of land covers, areas with distinct management schemes (e.g., forestry plantations, game reserves, and protected areas), and ecoregions within both regions (Online Resource 1). The camera trap survey designs varied for each project due to differences in objectives, implementation years (2019 to 2021), seasons (wet and dry), and sampling effort. These variations were based on project-specific constraints such as logistics and objectives (Online Resource 1).

To match the environmental division of this study, camera locations were divided into two regions (north and south) according to their location relative to Tagus River (Fig. 1): 37 sampling areas located on the north of the Tejo River (N camera traps = 765), and eight on the south (N camera traps = 190) (Fig. 1). Cameras were set at a minimum of 40cm from the soil, attached to tree or wooden poles, without the use of bait, and spaced at least 1 km apart from the nearest one. Each camera site was characterized according to several bio-environmental features (Table 1), whose data were collected remotely using the QGIS software version 3.26.3 (QGIS Development Team 2022), unless stated otherwise: ecoregions, climate, landscape composition and structure, anthropogenic disturbance, and biotic interactions (e.g., competition and predation).

Table 1

Explanatory variables used for modelling red fox detectability and occupancy patterns
Hypothesis	Variables Description/Code	Data Source	Prediction
Detection (p)	Season when the camera was active (dry or wet). (Season) Habitat of the exact location of the camera. (Habitat_COS) Number of distinct habitats edges in relation to the buffer* area. (Edge_Density) Shannon Diversity Index of habitats inside buffer*. (Shannon_Index) Relative Abundance Index of rabbits (proxy for red fox’s prey). (RAI_Rabbit) Relative Abundance Index of dogs. (RAI_Dogs)	Study’s dataset Land use and land cover map of continental Portugal – COS2018 (DGT, 2018) Land use and land cover map of continental Portugal – COS2018 (DGT, 2018) Study’s dataset estimation Study’s dataset estimation Study’s dataset estimation	We predicted the dry season would have a positive influence on red fox detectability due to juvenile dispersion (Sillero-Zubiri et al. 2004). We expected forested habitats to have a negative influence on red fox detection probability due to camera obstructions (Cruz et al. 2015). We hypothesized that a higher habitat fragmentation would negatively affect red fox detection probability (Cruz et al. 2015). We believed a higher habitat diversity would have a positive effect on red fox detectability because it is easier to find animals trails to install cameras and optimize detectability (Cruz et al. 2015). We hypothesized a higher abundance of prey would increase red fox detectability (Litvaitis et al. 1986; Carbone and Gittleman, 2002; Hetherington and Gorman 2007; Sarmento et al. 2011). We expected a higher abundance of dogs to negatively affect red fox detection probability (Farris et al., 2016; Murphy et al., 2019).
H1: Ecoregions	Ecoregion Classification according to Olson et al (2001). (Abbreviation_Ecorregion)	The Nature Conservancy. (2009). Terrestrial Ecoregions of the World. https://geospatial.tnc.org/datasets/b1636d640ede4d6ca8f5e369f2dc368b/about	More forested ecoregions will have a positive influence on red fox occurrence (Olson et al. 2001).
H2: Interspecific Interactions	Presence/absence of Egyptian mongoose (Herpestes ichneumon) on cameras within each five occasions. (Egyptian_M_P_A) Presence/absence of wolf in each study area. (Wolf_P_A) Relative Abundance Index of rabbits (Oryctolagus cuniculus) as a proxy for red fox’s prey. (RAI_Rabbit)	Study’s dataset Álvares et al (2019) Atlas de Mamíferos de Portugal Study’s dataset estimation	Red fox is less likely to occupy areas with greater presence of a competitor (Egyptian mongoose) (Bandeira et al. 2018; Jiménez et al. 2019; Sarmento et al. 2021; Castañeda et al. 2022) or a predator (wolf) (Petroelje et al. 2022), but more likely to occupy areas with higher density of preys (rabbits) (Litvaitis et al. 1986; Carbone and Gittleman, 2002; Hetherington and Gorman 2007; Sarmento et al. 2011).
H3: Habitat	Habitat of the exact location of the camera. (Habitat_COS) Number of distinct habitats edges in relation to the buffer* area. (Edge_Density) Shannon Diversity Index of habitats inside buffer*. (Shannon_Index)	Land use and land cover map of continental Portugal – COS2018 (DGT, 2018) Land use and land cover map of continental Portugal – COS2018 (DGT, 2018) Study’s Dataset estimation	Red fox occupancy probability would increase in landscapes with more natural, less fragmentated, and diverse habitats (Rosalino et al. 2010; Sarmento et al. 2011; Pereira et al. 2012; Alexandre et al. 2020; Linck et al. 2023).
H4: Climatic Factors	Season when the camera was active (dry or wet). (Season) Annual Temperature at camera location. (Annual_Mean_Temp) Minimum Temperature of Coldest Month at camera location. (MinTempColM) Maximum Temperature of Warmest Month at camera location. (MaxTempWarmM) Annual Precipitation at camera location. (Annual_Pre) Precipitation of Driest Month at camera location. (PreDryM) Precipitation of Wettest Month at camera location. (PreRainM)	Study’s dataset Fick and Hijmans (2017)	Seasonality, precipitation, and temperature may affect red fox’s occupancy due to changes in resource availability throughout the year (Vilella et al. 2020; Linck et al. 2021; Castañeda et al. 2022). Therefore, red fox occupancy would decrease during the dry season because the restriction of food resources available to specific patches; and challenging weather conditions (extreme temperatures) may restrain foxes’ movements (Santos et al. 2011; Shamoon et al. 2017; Teixeira et al. 2023a).
H5: Human Disturbance	Relative Abundance Index of dogs. (RAI_Dogs) Human Footprint Index at camera location. (HFI) Distance from camera location to the nearest urban area. (Dist_Urb) Distance from camera location to the nearest road. (Dist_Roads)	Study’s dataset estimation Socioeconomic Data and Applications Centre SEDAC, NASA (https://sedac.ciesin.columbia.edu) Land use and landcover map of continental Portugal – COS2018 (DGT, 2018) Land use and landcover map of continental Portugal – COS2018 (DGT, 2018)	Red foxes are less likely to occupy areas where domestic dogs are more abundant (Farris et al. 2016; Murphy et al. 2019). A high Human Footprint Index and a shorter distance to urban areas would decrease occupancy probability of red fox (Hewison et al. 2001; Sillero-Zubiri et al. 2004). Areas closer to roads may be avoided due to a higher chance of vehicle collision, which may constrain red fox occupancy (Sillero-Zubiri et al. 2004; Van Der Ree et al. 2011; Torres et al. 2014).
*200 meters radius buffer

We tested the influence of ecoregions in red fox occupancy based on the ecoregion classification proposed by Olson et al. (2001). Ecoregions data for each sampling point were gathered from The Nature Conservancy Terrestrial Ecoregions of the World database (2009) (Table 1), by overlaying the location of each camera-trap with the ecoregion shapefile layer. All ecoregions whose range overlapped the territory of continental Portugal (Cantabrian Mixed Forests, Iberian Sclerophyllous and Semi-Deciduous Forests, Northwest Iberian Montane Forests, and Southwest Iberian Mediterranean Sclerophyllous and Mixed Forests) were included in the dataset.

Mammals’ occupancy and activity patterns may be shaped by climatic factors such as seasonality and temperature variations (Vilella et al. 2020; Linck et al. 2021; Ares-Pereira et al. 2022). To assess the influence of weather conditions on red fox occurrence, we collected data from Fick and Hijimas (2017). Each camera trap site was characterized according to annual mean temperature, maximum temperature of warmest month, minimum temperature of coldest month, annual precipitation, precipitation of wettest month, and precipitation of driest month. We also considered the season when the data was collected – wet (November to April) or dry (from May to October) (Table 1). When cameras were active between seasons (e.g., between April – May and October-November), the month in which the camera was active during more days was considered as the predominant season.

Landscape composition and structure (i.e. habitat type, heterogeneity, and diversity) may directly influence the species' ecological dynamics, as it drives the variety and abundance of available resources, which can shape the spatial ecology of individuals (Rosalino et al. 2010; Verdade et al. 2011). Furthermore, different landscapes can also provide a variety of protection conditions against predation risks, disturbance factors, and adverse weather conditions (Sarmento et al. 2011; Pereira et al. 2012; Cruz et al. 2015; Alexandre et al. 2020). The dominant land cover types in each sample point were collected from the Land Use Map of Continental Portugal for 2018 - COS2018 (DGT 2018) (Table 1). We grouped the available land cover types into the following categories, according to their structure and broad land use characteristics: agriculture lands, Pinus spp., and Eucalyptus spp. stands, deciduous forests, pastures, shrublands, areas with sparse vegetation, agroforests, and other less represented habitats. To estimate habitat heterogeneity and diversity, we defined a 200-meter radius buffer around each camera trap site (Castro et al. 2022) (Table 1). The buffer’s radius size corresponds approximately to the radius of a theoretical circular red fox core area, estimated for Mediterranean areas (i.e., 0.11 km²; Cavallini 1996; Pandolfi et al. 1997).

Anthropogenic disturbance, and especially human activities, have been increasingly altering natural habitats, shaping how species use the landscape (Pereira et al. 2012; Cruz et al. 2015; Ares-Pereira et al. 2022; Linck et al. 2023). A denser human population, road network, and the development/expansion of more urban centers represent great disturbance factors for wildlife, as they reduce habitat quantity and quality and increase noise pollution, artificial lighting, edge effect, and collision risks (Hewison et al. 2001; Sillero-Zubiri et al. 2004; Van Der Ree et al. 2011; McClure et al. 2013; Torres et al. 2014). To incorporate disturbance risk into the analysis, we measured the distance from each camera site to the nearest road and the nearest urban area based on the Land Use Map of Continental Portugal for 2018 - COS2018 (DGT 2018). We also used the Human Footprint Index (HFI- Venter et al. 2016) as a surrogate of humans’ impact on the environment, by overlapping the map from the Socioeconomic Data and Applications Centre (SEDAC) from NASA (2005) with the sampling points (Table 1).

The establishment of interspecific interactions influences species occurrence patterns within the landscape. Therefore, we evaluated the influence of Egyptian mongoose (Herpestes ichneumon) presence on red fox occupancy because intraguild competition with other carnivores for food resources may affect the red fox's occupancy pattern (Bandeira et al. 2018; Jiménez et al. 2019; Sarmento et al. 2021; Castañeda et al. 2022). Furthermore, both species may act as direct competitors for one another (Soto and Palomares 2015; Sanglas and Palomares 2022). Egyptian mongoose presence in each camera was assessed from our dataset (Table 1).

However, the presence of a mesocarnivore is not only constrained by competitor species. Top predators can also shape foxes’ spatial behavior. To our knowledge, in Portugal, there are no studies that investigate the effects of the Iberian wolf (Canis lupus signatus) presence, the larger predator in the country, on red fox occurrence. Nevertheless, it is already known that the wolf might kill or indirectly exclude red foxes, where the two species are sympatric (Petroelje et al. 2022), according to the concept of landscape of fear (Bommel and Johnson 2016) (for more information about wolf and red fox interaction see Ferretti et al. 2021, 2023; Rossa et al. 2021). To test if the presence of the Iberian wolf affects the red fox occupancy patterns, we overlapped the wolf distribution range available in Álvares et al (2019) with the red fox detection data (Table 1). Thus, for each camera-trap site, we were able to register if they were located in wolf presence areas. We could not use our dataset to confirm wolf presence because the wolf detection rate was too low. Additionally, this variable was not tested for the south region dataset, because there have been no registers of wolves in this region in the last decades (Álvares et al. 2019; Pimenta et al. 2023).

Finally, we tested the effect of prey availability (rabbits) and potential competitors/predators (free-ranging dogs) (Palomares and Delibes 1993; Palomares et al. 1995; Ritchie et al. 2013; Jiménez et al. 2019). We used the relative abundance of rabbits (Oryctolagus cuniculus) as a proxy of red foxes’ prey availability (Table 1), a factor that can influence the occupancy of carnivores (Litvaitis et al. 1986; Carbone and Gittleman 2002; Hetherington and Gorman 2007; Sarmento et al. 2011). Furthermore, rabbits were the most often detected red fox prey (Castañeda et al. 2022) in our dataset, most likely because they are easier to identify using camera traps, comparatively to smaller mammal species.

On the other hand, it has been reported that free-ranging dogs change the activity and spatial patterns of native species (Lacerda et al. 2009; Farris et al. 2016; Murphy et al. 2019). The presence of dogs in wild habitats is likely harmful to native carnivore species, namely due to increased competition and predation (Butler and du Toit 2002; Ritchie et al. 2013; Vilela and Lamim-Guedes 2014) or because of direct disturbance and a higher risk of transmission of contagious diseases and parasites (Sillero-Zubiri et al. 2004). We extracted all the independent registers (i.e., photos of the same species, in the same camera, taken at least 30 minutes apart) of rabbits and dogs from our dataset to calculate their relative abundance using the Relative Abundance Index (RAI; O’Brien et al 2003) (Table 1). This index was calculated for 30-day periods because the minimum period a camera-trap was active in each sampling session was 30 days.

The RAI was estimated per camera trap and species as:

$$\:RAI=\:\frac{\text{n}\text{u}\text{m}\text{b}\text{e}\text{r}\:\text{o}\text{f}\:\text{i}\text{n}\text{d}\text{e}\text{p}\text{e}\text{n}\text{d}\text{e}\text{n}\text{t}\:\text{r}\text{e}\text{c}\text{o}\text{r}\text{d}\text{s}\:\text{o}\text{f}\:\text{t}\text{h}\text{e}\:\text{s}\text{p}\text{e}\text{c}\text{i}\text{e}\text{s}\:\text{d}\text{u}\text{r}\text{i}\text{n}\text{g}\:\text{t}\text{h}\text{e}\:\text{f}\text{i}\text{r}\text{s}\text{t}\:30\:\text{d}\text{a}\text{y}\text{s}\:\text{t}\text{h}\text{e}\:\text{c}\text{a}\text{m}\text{e}\text{r}\text{a}\:\text{w}\text{a}\text{s}\:\text{a}\text{c}\text{t}\text{i}\text{v}\text{e}}{30\:\left(\text{n}\text{u}\text{m}\text{b}\text{e}\text{r}\:\text{o}\text{f}\:\text{d}\text{a}\text{y}\text{s}\:\text{t}\text{h}\text{e}\:\text{c}\text{a}\text{m}\text{e}\text{r}\text{a}\:\text{w}\text{a}\text{s}\:\text{a}\text{c}\text{t}\text{i}\text{v}\text{e}\right)}\:\text{x}\:100$$

The use of RAI as a surrogate of abundance has been criticized (e.g., Sollmann et al. 2013) based on the possible detection variability between sites that may bias RAI values. However, other studies found a correlation between RAI estimates and absolute densities if cameras are placed randomly (Tanwar et al. 2021), and not only in trails. Our sampling schemes were not targeting trails, and we believe this could minimize the bias detections and validate the use of RAI.

Statistical analysis

From all camera trap images of red foxes obtained, we only used independent detection events within the same camera-site, i.e. two red fox records registered with a minimum interval of 30 minutes (Davis et al. 2011). Independent events were then grouped into ten periods of three days each, totalizing 30 days of sampling per camera. We considered periods with detections as “1”, periods without detections as “0”, and the absence of data as “NA” when the camera, for some unknown reason, was not active during the whole sampling period.

Considering that red fox detection is imperfect, we started our modelling procedure by testing what variables could influence red fox detection probability, prior to assessing the drivers of occupancy, based on a single-species, single-season occupancy model framework (MacKenzie et al. 2002). This model type considers two separate parameters: detection (hereafter p), defined as the probability a species is detected at a site, given its presence (Rovero and Zimmermann 2016), and occupancy (hereafter psi or ψ), defined as the probability a site is occupied by the target species given a detection probability that is smaller than 1.0 (Mackenzie 2002).

Additionally, for this study, we considered that, at each site, populations were closed for changes in occupancy during the survey period, i.e. we assumed that there was no relevant emigration/immigration during the study. Even though the surveys were conducted in different years, the study period (30 days) in each study site is probably too short for the emigration/immigration rate to be an important process that may affect detectability estimates and for changes in occupancy to occur.

The following steps were the same for both regions, north and south. First, we tested which variables could be more relevant to influence p while keeping psi constant (i.e., psi = 1). We assessed the effect of season, habitat at camera location, number of distinct habitats inside a 200 m radius buffer, Shannon Diversity Index for habitats inside a 200 m radius buffer, RAI of rabbits and RAI of dogs (Table 1) on p (Litvaitis et al. 1986; Carbone and Gittleman, 2002; Sillero-Zubiri et al. 2004; Hetherington and Gorman 2007; Sarmento et al. 2011; Cruz et al. 2015; Farris et al., 2016; Murphy et al., 2019). Those variables were tested for collinearity, using the “corvif” function in the “AED” R package (Zuur et al. 2009) to calculate the Variance Inflation Factors (VIF). Variables with VIF > 3 in each run were considered collinear and excluded from the analysis. This process was repeated until no more variables reached values of VIF > 3 (Zuur et al. 2009). To produce models corresponding to all possible combinations of the candidate variables influencing p, we used the “dredge” function in “MuMIn” R package (Barton 2015). All the models produced were ranked by their Akaike Information Criterion, corrected for small samples (AICc) (Burnham and Anderson 2002), and the one with the lowest AICc value was considered the best model to explain detectability variations. The variables included in the best-supported models for p were then selected and incorporated into all the subsequent occupancy models (Mackenzie 2002).

The next step was to build the occupancy models per se using the “occu” function in the “unmarked” R package (Fiske and Chandler 2011). For this analysis, we grouped the candidate variables into five initial ecological hypotheses, related to the drivers that may influence red fox occupancy, as follows: (H1) Ecoregion; (H2) Interspecific Interactions; (H3) Habitat; (H4) Climatic Factors; (H5) Human Disturbance (see Table 1). Within each hypothesis, all variables were tested for collinearity using the same process done for p (Zuur et al. 2009).

To perform the occupancy models accounting for all hypotheses, we included the variables identified previously as influential for p in the detection section of all models and the group of non-collinear variables from each hypothesis in the psi section of the model. For each hypothesis, we produced models corresponding to all possible combinations of candidate variables, and all models with a ΔAICc ≤ 2 were considered as best models for each one of the tested hypotheses (Burnham and Anderson 2002). When there was more than one model with ΔAICc ≤ 2, an average modeling procedure was implemented to obtain average variables coefficients (MacKenzie and Bailey 2004).

Variables included in the best model, in each one of the hypotheses, whose confident interval of its coefficient did not include 0 (i.e., we could infer with higher degree of certainty their positive or negative influence on occupancy) were considered as variables with well-supported effects (Arnold 2010). Those variables were used as candidates to test an additional “combined” hypothesis (H6) to understand if the combination of variables, used as proxies of different hypotheses, would better explain red fox occupancy probability variation (Castro et al. 2022).

After running the models for these six initial hypotheses and identifying the best overall variables in each hypothesis, we decided to test if the interaction between those variables would improve the models and better explain the red fox occupancy pattern. Thus we created two additional models, representing the interaction between these best overall variables from the previous six hypotheses (Table 1): H7 (Habitat_COS : RAI_Dogs) and H8 (HFI : RAI_Dogs) (see results).

The models for H6, H7, and H8 were built and selected using the same process described above for the other models. So, of all the models created, the one with the lowest overall AICc value was considered the best overall model (Burnham and Anderson 2002).

The goodness-of-fit of the best overall model for the red fox occupancy patterns at both regions was tested for overdispersion by calculating the c-hat value (MacKenzie and Bailey 2004). If there was overdispersion in the models (i.e., the models presented c-hat values distant from 1), then we restarted the process of model building and selection, but using the Quasi Akaike Information Criterion, corrected for small samples sizes (QAICc), to select the best models, considering the threshold of ΔQAICc ≤ 2 to identify the best models (MacKenzie and Bailey 2004). When there was more than one model with ΔQAICc ≤ 2, we implemented a model averaging procedure for models fulfilling this criterion (MacKenzie and Bailey 2004). All statistical analyses were performed using RStudio version 4.3.1 (R Development Core Team 2022).

Red-fox detection

We compiled data from 955 camera- traps, in which we registered 6668 independent records from red-fox: 765 camera-traps (488 of which had at least one period with detection of red fox) and 5333 independent detections in north of the Tagus River; and 190 camera-traps (133 of which detected red foxes) and 1335 independent detections in the southern region.

Occupancy models

The results of the collinearity assessment among variables were similar in both datasets (north and south of River Tagus). Edge_Density was excluded as a candidate variable for explaining detectability (p) because it was associated with a very high VIF (> 3). Only the variables Shannon_Index, RAI_Rabbit, RAI_Dogs, Season, and Habitat_COS were considered for the subsequent p modelling.

When the same collinearity test was performed in the context of occupancy modelling for H1 (Ecoregion), H2 (Interspecific Interactions), H5 (Human Disturbance), H6 (Mixed Model), H7 (Habitat_COS : RAI_Dogs) and H8 (HFI : RAI_Dogs), none of the variables were excluded (i.e., VIF < 3, in all cases). For H3 (Habitat), the variable Edge_Density was excluded from the analysis due to VIF > 3. For H4 (Climatic Factors), we excluded the variables Annual_Pre, PreDryM, Annual_Mean_Temp (only for cameras north of the Tagus River), and MaxTempWarmM (only for cameras south of the Tagus River) due to VIF > 3.

North Region

The back-transformed linear combination of occupancy and detection estimates, considering the lack of effect of all tested variables, was 0.643 ± 0.017 and 0.377 ± 0.007, respectively.

Initially, the hypothesis with the best-supported model (i.e., lowest AIC value) was H6 (Mixed Model) (AICc: 7038.9; Online Resource 2). However, this best model presented overdispersion (c-hat = 2.47). Therefore, we repeated the modelling procedure using the QAICc (Quasi Akaike Information Criterion) as a model selection criterion.

For p, the best model included the variables Habitat_COS, RAI_Dogs, RAI_Rabbit and Season (QAICc: 3080.0; Table 2). Eucalyptus spp. plantations and the wet season presented a negative correlation with detectability (Table 3). On the other hand, the variables forests, RAI_Dogs, and RAI_Rabbit had the opposite effect, showing a positive correlation/association (Table 3).

Table 2

Red fox occupancy models north of the Tagus River ranked by QAICc, ∆ QAICc and model cumulative weight. Only models with ΔQAIC ≤ 2 for the most supported models of detection and occupancy are represented. The best models (lowest QAIC) are highlighted in bold. (df – degree of freedom; logLik – log-Likelihood; QAICc – Quasi Akaike’s Information Criterion corrected for small samples; delta – delta of QAICc; weight – model cumulative weight; Overall delta – the difference in QAICc between the best model within each hypothesis and the best overall model). See Table 1 for acronyms descriptions.
Hypothesis	Model	df	logLik	QAICc	delta	weight	Overall delta
Detectability c-hat = 2.33	p (Habitat_COS + RAI_Dogs + RAI_Rabbit + Season) ~ psi (1) p (Habitat_COS + RAI_Dogs + RAI_Rabbit + Season + Shannon_Index) ~ psi (1)	12 13	-3557.857 -3556.461	3080.0 3080.8	0.00 0.80	0.516 0.346	0.00 0.80
H5 (Human Disturbance) c-hat = 3.24	p (Habitat_COS + RAI_Dogs + RAI_Rabbit + Season) ~ psi (RAI_Dogs) p (Habitat_COS + RAI_Dogs + RAI_Rabbit + Season) ~ psi (HFI + RAI_Dogs)	13 14	-3549.628 -3549.628	2219.1 2221.1	0.00 2.00	0.602 0.221	0.00 2.00

Table 3

Coefficients estimates of each variable included in the best detection and occupancy models’ averages (ΔQAIC ≤ 2) from north of the Tagus River. (SE- Standard Error; z- z-score; P(>|z|) – p-value of the test).
Hypothesis	Variables	Estimate	SE	z	P(>\|z\|)
Detectability	(Intercept)	-0.455	0.091	4.966	7.0e-07
	Habitat_COS (Eucalyptus spp. Plantations)	-1.276	0.183	6.955	< 2e-16
	Habitat_COS (Native Broadleaf and Mixed Forests)	0.260	0.098	2.643	0.008
	Habitat_COS (Other Habitats)	0.505	0.306	1.648	0.099
	Habitat_COS (Pastures)	-0.365	0.354	1.033	0.301
	Habitat_COS (Pinus spp. Plantations)	-0.011	0.116	0.102	0.918
	Habitat_COS (Shrublands)	0.073	0.102	0.715	0.474
	Habitat_COS (Sparse Vegetation)	-0.573	0.312	1.835	0.066
	Relative Abundance Index of dogs	0.140	0.027	5.070	4.0e-07
	Relative Abundance Index of rabbit	0.120	0.026	4.581	4.6e-06
	Wet Season	-0.256	0.066	3.849	0.000
	Shannon Index	0.056	0.033	1.671	0.094
H5 (Human Disturbance)	(Intercept)	0.822	0.165	4.963	7.00e-07
	Relative Abundance Index of Dogs	0.457	0.132	3.450	0.000
	HFI (Human Footprint Index)	0.000	0.008	0.016	0.987

Regarding occupancy, of all tested hypotheses for the north of the Tagus River dataset, red fox occupancy pattern was best described by hypothesis H5 (Human Disturbance), in which two models presented Δ QAICc ≤ 2 (Table 2). The best model (QAICc: 2219.1) included only the variable RAI_Dogs, while the second-best model included the variables RAI_Dogs and HFI (QAICc: 2221.1; Table 2). The average coefficients of variables included in red fox top-ranked models showed a positive effect of RAI_Dogs on red fox site occupancy probability, while HFI did not present a significant effect (p > 0.05; CI95% of the variable's coefficient included the 0) (Table 3).

Additional information on the detectability and occupancy models built using the north region dataset is provided in Online Resource 2, 3, and 4.

South Region

The back-transformed linear combination of occupancy and detection estimates, considering no effect of all tested variables, was 0.705 ± 0.033 and 0.398 ± 0.013 respectively.

For this region, the c-hat value of the best model did not present overdispersion (c-hat = 1.01), i.e. it was not necessary to use QAICc as a model selection criterion. Regarding the best detection model (AICc: 1970.6), it included the variables Habitat_COS and Season (Table 4). Agriculture land habitats presented a positive influence on red fox detectability, while the variable Season did not show a significant effect (p > 0.05; CI95% of the variable's coefficient included the 0) (Table 5).

Table 4

Red fox occupancy models south of the Tagus River ranked by AICc, ∆ AICc and model cumulative weight. Only models with ΔAIC ≤ 2 for the most supported models of detection and occupancy are represented. The best models (lowest AICc) are highlighted in bold. (df – degree of freedom; logLik – log-Likelihood; AICc – Akaike’s Information Criterion corrected for small samples; delta – delta of AICc; weight – model cumulative weight; Overall delta – the difference in AICc between the best model within each hypothesis and the best overall model). See Table 1 for acronyms descriptions.
Hypothesis	Model	df	logLik	AICc	delta	weight	Overall delta
Detectability	p (Habitat_COS + Season) ~ psi (1)	9	-976.320	1970.6	0.00	0.160	0.00
	p (Habitat_COS + Season + Shannon_Index) ~ psi (1)	10	-975.491	1971.0	0.34	0.135	0.34
	p (Habitat_COS + Shannon_Index) ~ psi (1)	9	-976.755	1971.5	0.87	0.104	0.87
	p (Habitat_COS + RAI_Rabbit + Season) ~ psi (1)	10	-976.206	1972.4	1.77	0.066	1.77
	p (Habitat_COS + RAI_Dogs + Season) ~ psi (1)	10	-976.292	1972.6	1.94	0.061	1.94
	p (Habitat_COS + RAI_Rabbit + Season + Shannon_Index) ~ psi (1)	11	-975.320	1972.6	2.00	0.059	2.00
H3 (Habitat)*	p (Habitat_COS + Season) ~ psi (Shannon_Index)	10	-973.241	1966.5	0.00	0.602	0.00
	p (Habitat_COS + Season) ~ psi (Habitat_COS + Shannon_Index)	16	-968.002	1968.0	1.52	0.281	1.5
* c-hat = 1.02

Table 5

Coefficients estimates of each variable included in the best detection and occupancy models’ averages (ΔAIC ≤ 2) south of the Tagus River. (SE- Standard Error; z- z-score; P(>|z|) – p-value of the test).
Hypothesis	Variables	Estimate	SE	z	P(>\|z\|)
	(Intercept)	-0.360	0.187	-1.926	0.054
Detectability	Habitat_COS (Agriculture)	0.812	0.262	3.104	0.001
	Habitat_COS (Eucalyptus spp. Plantations)	-0.553	0.298	-1.855	0.063
	Habitat_COS (Native Broadleaf and Mixed Forests)	0.061	0.196	0.314	0.753
	Habitat_COS (Other Habitats)	-0.205	0.358	-0.575	0.564
	Habitat_COS (Pinus spp. Plantations)	0.001	0.239	0.005	0.995
	Habitat_COS (Shrublands)	0.245	0.220	1.111	0.266
	Wet Season	-0.218	0.129	-1.693	0.090
	(Intercept)	1.219	0.641	1.901	0.057
H3 (Habitat)	Shannon Index	0.435	0.184	2.356	0.018
	Habitat_COS (Agriculture)	0.403	1.322	0.305	0.760
	Habitat_COS (Eucalyptus spp. Plantations)	-0.727	1.064	0.683	0.494
	Habitat_COS (Native Broadleaf and Mixed Forests)	-0.862	0.859	1.003	0.315
	Habitat_COS (Other Habitats)	-1.645	1.110	1.482	0.138
	Habitat_COS (Pinus spp. Plantations)	-1.796	0.888	2.023	0.043
	Habitat_COS (Shrublands)	-0.619	0.926	0.669	0.503

Of all tested hypotheses for the south of the Tagus River dataset, the red fox occupancy pattern drivers were best described by hypothesis H3 (Habitat), in which two models presented Δ AICc ≤ 2. The best occupancy model included only the variable Shannon_Index (AICc: 1966.5; Table 4). The second-best occupancy model included the variables Habitat_COS and Shannon_Index (AICc: 1968.0; Table 4). The average coefficients of variables included on red fox top-ranked models for the south of the Tagus River dataset showed that the variable Shannon_Index had a positive influence on red fox occupancy probability, while pine plantation habitats presented the opposite effect (Table 5). Additional information on the detectability and occupancy models for the south region dataset is provided in Online Resources 5 and 6.

In this study, we aimed to evaluate the drivers of red fox occupancy patterns in Portugal, exploring how different environmental contexts (i.e., north and south of the Tagus River) may determine which factors can shape red fox occupancy patterns. Our research is, to our knowledge, the first to consider data obtained from camera traps for assessing red fox occupancy patterns in Portuguese, and at a national level, within these two broader and environmentally distinct regions.

At a local scale, other studies have concluded that red fox occupancy is influenced by drivers such as Eucalyptus spp. plantations (Cruz et al. 2015; Castro et al. 2022) and topography (Monterroso et al. 2014; Curveira-Santos et al. 2019). However, at a broader scale, our results showed that the red fox is capable of adopting different ecological strategies to take advantage of a wider variety of environmental contexts (and the resources each provide), corroborating the often-assumed view of this species as a generalist an extremely common mesocarnivore in Portugal (Negrões et al. 2023), and within Europe (Sillero-Zubiri et al. 2004). These detected patterns aligned with our hypothesis and confirmed what has already been described for the north of Portugal, but on a smaller geographic scale (Alexandre et al. 2020). Moreover, we have identified important occupancy drivers that go beyond those already tested and highlighted in previous studies, confirming the idea that the set of drivers influencing a species at small spatial range analysis can change at large-scale assessments due to their capacity of increasing its niche width (Moll et al. 2016; Toews et al. 2017).

Foxes’ detection probability in north of the Tagus River was lower in Eucalyptus spp. plantations and during the wet season, but higher in forest areas where rabbits and free-ranging dogs were more common. The lower detectability within Eucalyptus spp. plantations may be associated with its tree line structure. Within this simpler and more open and homogeneous habitat, red foxes not only have additional areas where movement is facilitated (between the innumerous tree lines where vegetation can be sparse), but it is also more difficult to find animals’ trails where cameras can be installed nearby to optimize red fox detections (Cruz et al. 2015), which makes less likely for a fox to activate the camera trigger. The negative effect of the wet season in detection can be linked to the fact that this season is usually characterized by an increase in rainfall (which may directly affect camera’s activation as well - Newey et al. 2015) and is the breeding season period of the red fox (Sillero-Zubiri et al. 2004; Loureiro et al. 2012). Both factors concur to a change in individual behavior, which become more territorial (Monterroso et al. 2014), thus moving less (i.e. less wider movements) and therefore reducing the likelihood of encounters with camera traps. A higher abundance of prey at a camera site implies a favorable area for hunting, which can attract foxes, thus increasing its detection probability (Curveira-Santos et al. 2019). This prey abundance can also attract other potential carnivores, such as free-ranging dogs. Even though studies have shown that dogs represent a disturbance factor for red foxes (Richie et al. 2013; Bommel and Johnson 2016; Farris et al. 2016; Murphy et al. 2019; Gálvez et al. 2021; Teixeira et al. 2023 a) and contrary to our initial hypothesis (see Table 1), our data revealed that the RAI of dogs had a positive correlation with occupancy and detection of the red fox at north of the Tagus River. One of the reasons for this result might be related to the fact that free-ranging dogs consume lagomorphs (Lino et al. 2023) and therefore may be using areas with higher rabbit abundance alongside foxes. Because of this, and the fact that dogs can present a surplus killing behavior (Richie et al. 2013), it is also possible that dogs could be promoting a facilitation relationship with the red fox. Thus, as this extra food resource linked to dogs hunting behaviors may overrule the disturbance effect of their presence, foxes could be taking advantage of carcasses left by dogs (Lino et al. 2023). A similar pattern was also detected in Italy where foxes benefit from the presence of wolves (Rossa et al. 2021; Ferretti et al. 2021; Ferretti et al. 2023). However, other explanations could justify those results from north of the Tagus River. For example, higher occupancy of foxes could be promoting a greater dog RAI (i.e., dogs respond to foxes’ occupancy), which could be tested by performing a temporal analysis of the camera trap records. Another possibility is that both species could be responding to the same variable that was not included in our analysis. Therefore, to better understand the ecological relationship between free-ranging dogs and the red fox, more information on their spatial-temporal co-occurrence patterns is needed and an effort should be made to generate this type of data.

In the region south of the Tagus River, our results showed a different pattern, with foxes being more detected in open habitats, such as agricultural lands. The lack of vegetation or the less dense and mostly herbaceous crop plant species (commonly found in southern areas) may facilitate red fox detection (Pereira et al. 2012; Cruz et al. 2015). As for the best southern occupancy model, a similar pattern was revealed. Red fox showed a lower probability of occupying Pine spp. plantations when compared to agroforestry habitats. We believe this suggests the species' preference for more heterogeneous areas (i.e. sites with different habitat types in balanced proportions) rather than single habitat types, since Mediterranean agroforestry areas are typically multiuse and thus multifunctional and heterogeneous environments (Pinto-Correia and Vos, 2004). Red foxes may also avoid pine plantations because, in general, those simplified habitats may represent an area lower food availability (Paviolo et al. 2018; Iezzi et al. 2018; da Silva et al. 2019). More heterogeneous and diverse landscapes (i.e. with higher diversity indexes) provide more shelter opportunities and food resources for the red fox, which can create the conditions for those sites to be more often occupied by foxes (Rosalino et al. 2010; Pereira et al. 2012; Linck et al. 2023).

Even though we intended to include in our models as many relevant factors as possible, the results presented here have some limitations and should be interpreted with caution. In order to standardize information from different research projects and considering the use of camera-trapping methodology, it was not possible to analyze some environmental factors that could be important as drivers of red fox occupancy and should be tested in future modelling. For example, we could not test the effect of other food items abundance (e.g., small mammals, insects, and fruits; Curveira-Santos et al. 2019) on red fox occupancy due to the impossibility of sampling them in the field. Therefore, there was a need to use as a proxy other fox’s prey item that we could evaluate via camera trapping (i.e. the rabbit). Nevertheless, the distinct patterns of red fox occupancy that we detected for the regions north and south of Tagus River can be linked to the unique characteristics of each region that create different environmental contexts. For example, the effect of dogs’ relative abundance on red fox occupancy reflects the intensity of human presence in the Northern region (Alves et al. 2009; Lino et al. 2023), while in the Southern region, the dominant “montado” system, which is a heterogenous habitat type (Álvares et al. 2019), shapes this canid’s occupancy patterns.

Our results revealed that foxes adopt distinct ecological strategies to face different environmental contexts, leading to distinct underlying ecological mechanisms shaping the occupancy pattern of the red fox within the northern and southern regions of Continental Portugal. These results corroborate the view that a generalist species, like the red fox, must be highly adaptable to respond to the variation of environmental conditions. Such adaptability enables the red fox to optimize the use of resources and cope with a myriad of disturbance factors. Thus, these results add information to the body of knowledge supporting red foxes’ adaptative character, which can support management and conservation plans targeting this species. However, it also stimulates the implementation of more studies at these broader scales to gather information on foxes’ responses to different environmental conditions at a local level, as they may be context-specific. The identification of this species´ preferences in resource usage or conditions, and of what is shaping this generalist canid spatial patterns is a crucial knowledge to manage biodiversity in a global change scenario, which in Europe is forecasted to affect drastically Mediterranean regions (Moatti and Thiébault 2016).

Conflict of interest: The authors have no conflicts of interest to declare.

Funding:

We acknowledge financial support of FCT – Fundação para a Ciência e a Tecnologia to CESAM (UIDP/50017/2020+UIDB/50017/2020+ LA/P/0094/2020), to cE3c (DOI: 10.54499/UIDB/00329/2020), through national funds, and the co-funding by the FEDER, within the PT2020 Partnership Agreement and Compete 2020, and to CHANGE (LA/P/0121/2020). Eduardo Ferreira is funded by national funds (OE) through FCT in the scope of the framework contract foreseen in article 23, Decree-Law 57/2016. This research was funded by the project JAVALI.PT “Plano Estratégico e de Ação do Javali em Portugal”, from Fundo Florestal Permanente/ICNF. R. T. Torres and J. Carvalho were supported by a research contract (2021.00690.CEECIND; CEECIND/01428/2018, respectively) through FCT – Fundação para a Ciência e a Tecnologia, I.P. M. Rossa was supported by PhD grant from FCT - Fundação para a Ciência e Tecnologia (2021.05387.BD). This research was also supported by LIFE WolFlux (LIFE17 NAT/PT/000554), funded by the LIFE Programme of the European Union. We also thank FCT for its financial support to the projects WildForests (Ref. POCI-01-0145-FEDER-028204) and ForCe (https://doi.org/10.54499/2022.03253.PTDC). Data collected by the project “Revisão do Livro Vermelho de Mamíferos de Portugal” received support from POSEUR (Operational Sustainability and Efficient Pro- gram in the Use of Resources) and Fundo Ambiental (Ministry of Environment and Climatic Action), project Ref: POSEUR-03-2215-FC-000097. We also acknowledge the Instituto da Conservação da Natureza e das Florestas (ICFN, I.P.) for providing the data from Programa de Monitorização de Mamíferos.

Acknowledgments:

We thank all the researchers involved in the projects we used data from to carry out this study, namely to the coordinators and team of the projects: WildForests (Ref. POCI-01-0145-FEDER-028204), Life WolFlux (Ref. LIFE17-NAT/PT/000554), Plano Estratégico e de Ação do Javali em Portugal (Javali.Pt), Livro Vermelho dos Mamíferos de Portugal (Ref. POSEUR-03-2215-FC-000097), and Programa de Monitorização de Mamíferos do ICNF.

Alexandre M, Hipólito D, Ferreira E, Fonseca C, Rosalino LM (2020) Humans do matter: determinants of red fox (Vulpes vulpes) presence in a western Mediterranean landscape. Mamm Res 65: 203–214. https://doi.org/10.1007/s13364-019-00449-y
Almeida, MA (2020) The use of rural areas in Portugal: Historical perspective and the new trends. Revista Galega de Economia 29 (2): 1–17. https://doi.org/10.15304/rge.29.2.6750.
Álvares F, Ferreira CC, Barbosa AM, Rosalino LM, Pedroso NM, Bencatel J (2019) Carnívoros. In: Bencatel J, Álvares F, Moura AE, Barbosa AM (eds.) Atlas de Mamíferos de Portugal 2nd edn. Universidade de Évora, pp 271. https://ambiogeo.shinyapps.io/atlasmampor/. Accessed 22 Dec 2023.
Alves JMS, Espírito-Santo MD, Costa JC, Capelo JH, Lousã MF (2009) Habitats Naturais e Seminaturais de Portugal Continental: Tipos de Habitats Mais Significativos e Agrupamentos Vegetais Característicos. 2nd edn. Assírio & Alvim. https://www.researchgate.net/publication/255721679_Habitats_Naturais_e_Se minaturais_de_Portugal_Continental. Accessed in 26 Dec 2023.
Ares-Pereira G, Rosalino LM, Teixeira D, Castro G, Magalhães A, Lima C, Fonseca C, Torres RT (2022) Eucalyptus plantations alter spatiotemporal relationships of wild ungulates. Agriculture, Ecosystems and Environment 340. https://doi.org/10.1016/j.agee.2022.108174
Bandeira V, Virgós E, Carvalho J, Barros T, Cunha MV, Fonseca C (2018) Diet footprint of Egyptian mongoose along ecological gradients: effects of primary productivity and life history traits. Mamm Biol 88: 16–25. https://doi.org/10.1016/j.mambio.2017.11.004
Barton K (2015). MuMIn: Multi-model inference – R package version 1.43.17. http://cran.r-project.org/package=MuMIn. Accessed in 26 Dec 2023.
Bencatel J, Ferreira CC, Barbosa AM, Rosalino LM, Álvares F (2018) Research trends and geographical distribution of mammalian carnivores in Portugal (SW Europe). PLoS ONE 13 (11). https://doi.org/10.1371/journal.pone.0207866.
Bencatel J, Sabino-Marques H, Álvares F, Moura AE, Barbosa AM (2019) Atlas de Mamiferos de Portugal (2nd edition). Universidade de Évora, Portugal https://ambiogeo.shinyapps.io/atlasmampor/ Accessed in 26 Dec 2023.
Blondel J (2006) The "design" of Mediterranean landscapes: a millennial story of Humans and ecological systems during the historic period. Human Ecology 34: 713–729. doi:10.1007/s10745-006-9030-4.
Blondel J, Aronson J, Bodiou JY, Boeuf G (2010) The Mediterranean region: biological diversity in space and time. Oxford University Press. https://global.oup.com/academic/product/the-mediterranean-region9780199557998?cc=us&lang=en&. Accessed in 26 Dec 2023.
Bommel L., Johnson CN (2016) Livestock guardian dogs as surrogate top predators? How Maremma sheepdogs affect a wildlife community. Ecol. Evol., 6 (18): 6702-6711. https://doi.org/10.1002/ece3.2412.
Burnham K, Anderson D (2002) A practical information-theoretic approach. Model Selection and Multimodel Inference, 2nd edn. Springer, New York. http://sutlib2.sut.ac.th/sut_contents/H79182.pdf. Accessed in 26 Dec 2023.
Burton, A.C., Beirne, C., Gaynor, K.M. et al. Mammal responses to global changes in human activity vary by trophic group and landscape. Nat Ecol Evol (2024). https://doi.org/10.1038/s41559-024-02363-2.
Butler J, du Toit J (2002) Diet of free-ranging domestic dogs (Canis familiaris) in rural Zimbabwe: implications for wild scavengers on the periphery of wildlife reserves. Anim Conserv 5: 29–37. https://doi.org/10.1017/S136794300200104X
Carbone C, Gittleman JL (2002) A common rule for the scaling of carnivore density. Science 295 (5563): 2273-2276. https://doi.org/10.1126/science.1067994
Castañeda I, Doherty TS, Fleming PA, Stobo-Wilson AM, Woinarski JCZ, Newsome TM (2022) Variation in red fox Vulpes vulpes diet in five continents. Mammal Rev. 52: 328–342. https://doi.org/10.1111/mam.12292
Castro G, Teixeira D, Ares-Pereira G, Lima C, Magalhães A, Camarinha C, Guillera‐Arroita G, Fonseca C, Rosalino LM (2022) Drivers of occupancy patterns for the red fox, Vulpes vulpes, in Mediterranean eucalyptus plantations. Forest Ecology and Management 519: 120293. https://doi.org/10.1016/j.foreco.2022.120293
Cavallini P 1996. Ranging behaviour of red foxes during the mating and breeding seasons. Ethology, Ecology & Evolution 8: 57–65. https://doi.org/10.1080/08927014.1996.9522935
Cruz J, Sarmento P, White PCL (2015) Influence of exotic forest plantations on occupancy and co-occurrence patterns in a Mediterranean carnivore guild. Journal of Mammalogy 96 (4): 854-865. https://doi.org/10.1093/jmammal/gyv109
Curveira-Santos G, Marques TA, Björklund M, Santos-Reis M (2017) Mediterranean mesocarnivores in spatially structured managed landscapes: community organization in time and space. Ecosyst. Environ. 237: 280- 289, https://doi.org/10.1093/jmammal/gyv109
Curveira-Santos G, Pedroso NM, Barros AL, Santos-Reis M (2019) Mesocarnivore community structure under predator control: Unintended patterns in a conservation context. PLoS ONE 14 (1). https://doi.org/10.1371/journal.pone.0210661
Davis ML, Kelly MJ, Stauffer DF (2011) Carnivore co-existence and habitat use in the Mountain Pine Ridge Forest Reserve, Belize. Animal Conservation 14 (1): 56-65. https://doi.org/10.1111/j.1469-1795.2010.00389.x
DGT (2018) Carta de Uso e Ocupação do Solo de Portugal continental para 2018. Direção de Serviços de Geodesia, Cartografia e Informação Geográfica. Direção-Geral Do Território, Lisboa. https://snig.dgterritorio.gov.pt/rndg/srv/por/catalog.search#/metadata/b498e89c1093-4793-ad22-63516062891b. Accessed in 26 Dec 2023.
Doherty TS, Dickman CR, Glen AS, Newsome TM, Nimmo DG, Ritchie EG, Vanak AT, Wirsing AJ (2017) The global impacts of domestic dogs on threatened vertebrates. Biol Conserv 210: 56–59. https://doi.org/10.1016/j.biocon.2017.04.007
Farris ZJ, Kelly MJ, Karpanty S, Ratelolahy F, Zach Farris CJ, Hall C (2016) Patterns of spatial co-occurrence among native and exotic carnivores in north-eastern Madagascar. Animal Conservation 2: 189–198. https://doi.org/10.1111/acv.12233
Ferretti F, Pacini G, Belardi I, Cate B, Sensi M, Oliveira R, Rossa M, Burrini L, Lovari S (2021) Recolonizing wolves and opportunistic foxes: interference or facilitation?. Biological Journal of the Linnean Society 132(1): 196–210, https://doi.org/10.1093/biolinnean/blaa139
Ferretti F, Oliveira R, Rossa M, Belardi I, Pacini G, Mugnai S, Fattorini N, Lazzeri L (2023) Interactions between carnivore species: limited spatiotemporal partitioning between apex predator and smaller carnivores in a Mediterranean protected area. Front Zool 20 (20). https://doi.org/10.1186/s12983-023-00489-w
Fick SE, Hijmans RJ (2017) WorldClim 2: new 1km spatial resolution climate surfaces for global land areas. International Journal of Climatology 37 (12): 4302-4315. https://www.worldclim.org/data/worldclim21.html. Accessed in 26 Dec 2023
Fiske IJ, Chandler RB (2011) Unmarked: An R Package for Fitting Hierarchical Models of Wildlife Occurrence and Abundance. Journal of Statistical Software 43 (10): 1-23. https://www.jstatsoft.org/article/view/v043i10. Accessed in 26 Dec 2023
Gálvez N, Meniconi P, Infante J, Bonacic C (2021) Response of mesocarnivores to anthropogenic landscape intensification: activity patterns and guild temporal interactions. J Mammal 102: 1149–1164. https://doi.org/10.1093/jmammal/gyab074
Geri F, Amici V, Rocchini D (2010) Human activity impact on the heterogeneity of a Mediterranean landscape. Applied geography 30 (3): 370- 379. https://doi.org/10.1016/j.apgeog.2009.10.006
Hetherington DA, Gorman ML (2007) Using prey densities to estimate the potential size of reintroduced populations of Eurasian lynx. Biol Conserv. 137 (1): 37- 44. https://doi.org/10.1016/j.biocon.2007.01.009
Hewison AJM, Vincent JP, Joachim J, Angibault JM, Cargnelutti B, Cibien C (2001) The effects of woodland fragmentation and human activity on roe deer distribution in agricultural landscapes. Canadian Journal of Zoology 79 (4): 679–689. https://doi.org/10.1139/z01-032
Hijmans RJ, Cameron SE, Parra JL, Jones PG, Jarvis A (2005) Very high resolution interpolated climate surfaces for global land areas. Int. J. Climatol. 25: 1965–1978. https://doi.org/10.1002/joc.1276
Iezzi ME, Cruz P, Varela D, De Angelo C, Di Bitetti MS (2018) Tree monocultures in a biodiversity hotspot: Impact of pine plantations on mammal and bird assemblages in the Atlantic Forest. Forest Ecology and Management 424: 216-227. https://doi.org/10.1016/j.foreco.2018.04.049
Jiménez J, Nuñez-Arjona JC, Mougeot F, Ferreras P, González LM, García-Domínguez F, Muñoz-Igualada J, Palacios MJ, Pla S, Rueda C, Villaespesa F, Nájera F, Palomares F, López-Bao JV (2019) Restoring apex predators can reduce mesopredator abundances. Biological Conservation 238: 1–10. https://doi.org/10.1016/j.biocon.2019.108234
Jones N, Graaff J de, Rodrigo I, Duarte F (2011) Historical review of land use changes in Portugal (before and after EU integration in 1986) and their implications for land degradation and conservation, with a focus on Centro and Alentejo regions. Appl. Geography. 31 (3): 1036- 1048, https://doi.org/10.1016/j.apgeog.2011.01.024
Lacerda A, Tomas W, Marinho-Filho J (2009) Domestic dogs as an edge effect in the Brasília National Park, Brazil: interactions with native mammals. Anim. Conserv. 12: 477–487. https://doi.org/10.1111/j.1469- 1795.2009.00277.x
Linck P, Tirelli FP, Bastos MC, Fonseca, AN, Cardoso LF, Trigo TC (2021) Daily activity patterns and occurrence of Leopardus guttulus (Carnivora, felidae) in Lami Biological Reserve, southern Brazil. Iheringia - Serie Zoologia 111: 1–9. https://doi.org/10.1590/1678-4766e2021006
Linck P, Palomares F, Negrões N, Rossa M, Fonseca C, Couto A, Carvalho J (2023) Increasing homogeneity of Mediterranean landscapes limits the co‐occurrence of mesocarnivores in space and time. Landsc Ecol. 38: 3657-3673 https://doi.org/10.1007/s10980-023-01749-0
Lino S, Rossa M, Fernandes JM, Barros T, Lino A, Hipólito D, Ferreira E, Aliácar SC, Cadete D, Fonseca C, Torres RT, Rosalino LM, Carvalho J (2023) Dog in sheep’s clothing: livestock depredation by freeranging dogs may pose new challenges to wolf conservation. Eur J Wildl Res 69 (107). https://doi.org/10.1007/s10344-023-01740-9.
Litvaitis JA, Sherburne JA, Bissonette JA (1986) Bobcat habitat use and home range size in relation to prey density. J. Wildl Manage 50 (1): 110-117. https://doi.org/10.2307/3801498
Loidi J (2017) The Vegetation of the Iberian Peninsula. Springer 1. https://doi.org/10.1007/978-3-319-54784-8
MacKenzie DI, Bailey LL (2004) Assessing the fit of site-occupancy models. Journal of Agricultural, Biological, and Environmental Statistics 9 (3): 300–318. https://doi.org/10.1198/108571104X3361
MacKenzie DI, Nichols JD, Lachman GB, Droege S, Royle AA, Langtimm CA (2002) Estimating site occupancy rates when detection probabilities are less than one. Ecology 83 (8): 2248–2255. https://doi.org/10.1890/0012-9658(2002)083[2248:ESORWD]2.0.CO;2
Mangas J, Lozano J, Cabezas-Díaz S, Virgós E (2008) The priority value of scrubland habitats for carnivore conservation in Mediterranean ecosystems, Biodiversity and Conservation 17: 43–51. https://doi.org/10.1007/s10531-007- 9229-8
Mathias ML, Fonseca C, Rodrigues L, Grilo C, Lopes-Fernandes M, Palmeirim JM, Santos-Reis M, Alves PC, Cabral JA, Ferreira M, Mira A, Eira C, Negroes N, Pauperio J, Pita R, Rainho A, Rosalino LM, Tapisso JT, Vingada J (2023) Livro Vermelho dos Mamíferos de Portugal Continental. FCiencias.ID, ICNF, Lisboa. https://dspace.uevora.pt/rdpc/handle/10174/35224. Accessed in 26 Dec 2023. McClure CJ, Ware HE, Carlisle J, Kaltenecker G, Barber JR (2013) An experimental investigation into the effects of traffic noise on distributions of birds: avoiding the phantom road. Proc. R. Soc. Lond. B Biol. Sci. 280: 2013– 2290. https://doi.org/10.1098/rspb.2013.2290
Moatti JP, Thiébault S (2016) The Mediterranean Region under Climate Change. IRD Éditions. https://doi.org/10.4000/books.irdeditions.22908.
Moll RJ, Kilshaw K, Montgomery RA, Abade L, Campbell RD, Harrington LA, Millspaugh JJ, Birks JDS, Macdonald DW (2016) Clarifying habitat niche width using broad-scale, hierarchical occupancy models: A case study with a recovering mesocarnivore. Journal of Zoology 300 (3): 177–185. https://doi.org/10.1111/jzo.12369.
Monterroso P, Rich LN, Serronha A, Ferreras P, Alves PC (2014) Efficiency of hair snares and camera traps to survey mesocarnivore populations. European Journal of Wildlife Research 60 (2): 279–289. https://doi.org/10.1007/s10344-013-0780-1
Murphy A, Kelly MJ, Karpanty SM, Andrianjakarivelo V, Farris ZJ (2019) Using camera traps to investigate spatial co-occurrence between exotic predators and native prey species: a case study from northeastern Madagascar. Journal of Zoology 307 (4): 264–273. https://doi.org/10.1111/jzo.12645
Negrões N, Álvares F, Petrucci-Fonseca F (2023) Vulpes vulpes raposa. In: Mathias ML, Fonseca C, Rodrigues L, Grilo C, Lopes-Fernandes M, Palmeirim JM, Santos-Reis M, Alves PC, Cabral JA, Ferreira M, Mira A, Eira C, Negrões N, Pauperio J, Pita R, Rainho A, Rosalino LM, Tapisso JT, Vingada J (eds) Livro Vermelho dos Mamíferos de Portugal Continental. FCiencias.ID, ICNF, Lisboa. https://dspace.uevora.pt/rdpc/handle/10174/35224. Accessed in 26 Dec 2023
Newey S, Davidson P, Nazir S, Fairhurst G, Verdicchio F, Irvine RJ, van der Wal R (2015) Limitations of recreational camera traps for wildlife management and conservation research: A practitioner’s perspective. Ambio 44 (4): 624-635. https://doi.org/10.1007/s13280-015-0713-1
O’Brien TG, Kinnaird MF, Wibisono HT (2003) Crouching tigers, hidden prey: Sumatran tiger and prey populations in a tropical forest landscape. Animal Conservation 6: 131–139.
Olson DM, Dinerstein E, Wikramanayake ED, Burgess ND, Powell GVN, Underwood EC, D’Amico JA, Itoua I, Strand HE, Morrison JC, Loucks CJ, Allnutt TF, Ricketts TH, Kura Y, Lamoreux JF, Wettengel WW, Hedao P, Kassem KR (2001) Terrestrial ecoregions of the world: A new map of life on Earth. BioScience 51 (11): 933–938. https://doi.org/10.1641/0006-3568(2001)051[0933:TEOTWA]2.0.CO;2
Palomares F, Delibes M (1993) Key habitat for Egyptian mongoose in Doñana National Park, south - western Spain. Journal of Applied Ecology 30: 752-758. https://doi.org/10.2307/2404253
Palomares F, Gaona P, Ferreras P, Delibes M (1995) Positive effects on game species of top predators by controlling smaller predator populations: an example with Lynx, mongooses, and rabbits. Conserv. Biol. 9: 295–305. https://doi.org/10.1046/j.1523-1739.1995.9020295.x
Pandolfi M, Forconi P, Montecchiari L (1997) Spatial behaviour of the red fox (Vulpes vulpes) in a rural area of central Italy. Italian Journal of Zoology: 64: 351–358. https://doi.org/10.1080/11250009709356222
Paviolo A, Cruz P, Iezzi ME, Pardo JM, Varela D, De Angelo C et al (2018) Barriers, corridors or suitable habitat? Effect of monoculture tree plantations on the habitat use and prey availability for jaguars and pumas in the Atlantic Forest. Forest Ecology and Management 430: 576-586. https://doi.org/10.1016/j.foreco.2018.08.029
Pereira P, Silva AA, Alves J, Matos M, Fonseca C (2012) Coexistence of carnivores in a heterogeneous landscape: habitat selection and ecological niches. Ecol Res 27: 745–753. https://doi.org/10.1007/s11284-012- 0949-1
Petroelje TR, Fowler NL, Orning EK, Patterson BR, Romanski MC, Belant JL (2022) Interspecific Killing of Vulpes vulpes (Red Fox) Kits at a Den Site by Canis lupus (Gray Wolf) in Isle Royale National Park, Michigan. Northeastern Naturalist 29 (1): 18-26. https://doi.org/10.1656/045.029.0113
Pimenta V, Barroso I, Alvares F, Petrucci-Fonseca F (2023) Canis lupus lobo. In: Mathias ML, Fonseca C, Rodrigues L, Grilo C, Lopes-Fernandes M, Palmeirim JM, Santos-Reis M, Alves PC, Cabral JA, Ferreira M, Mira A, Eira C, Negrões N, Pauperio J, Pita R, Rainho A, Rosalino LM, Tapisso JT, Vingada J (eds) Livro Vermelho dos Mamíferos de Portugal Continental. FCiencias.ID, ICNF, Lisboa. https://dspace.uevora.pt/rdpc/handle/10174/35224. Accessed in 26 Dec 2023
Pinto-Correia T (2000) Future development in Portuguese rural areas: How to manage agricultural support for landscape conservation?. Landscape and Urban Planning 50: 95–106. https://doi.org/10.1016/S0169-2046(00)00082-7
Pinto-Correia T, Vos W (2004) Multifunctionality in Mediterranean landscapes – past and future. In: Jongman R (ed.) The New Dimensions of the European Landscape (pp. 135-164). Dordrecht: Springer-Verlag.
QGIS Development Team (2022) QGIS geographic information system. Open Source Geospatial Foundation Project. https://qgis.org/en/site/ Accessed in 26 Dec 2023.
R Development Core Team (2022) R: a language and environment for statistical computing computer program, version 4.3.1. R Development Core Team, Vienna, Austria. http://www.r-project.org. Accessed in 26 Dec 2023
Ritchie EG, Dickman CR, Letnic M, Vanak AT (2013) Dogs as predators and trophic regulators. In: Gompper ME (ed) Free-Ranging Dogs and Wildlife Conservation Oxford Academic, pp 55-68). https://doi.org/10.1093/acprof:osobl/9780199663217.003.0002
Rosalino LM, Rosa S, Santos-Reis M (2010) The Role of Carnivores as Mediterranean Seed Dispersers. Ann. Zool. Fennici 47: 195–205. https://doi.org/10.5735/086.047.0304
Rossa M, Lovari S, Ferretti F (2021) Spatiotemporal patterns of wolf, mesocarnivores and prey in a Mediterranean area. Behav Ecol Sociobiol 75 (32). https://doi.org/10.1007/s00265-020-02956-4
Rovero F, Zimmermann F (2016) Camera trapping for Wildlife Research. In: Pelagic Publishing Ltd. Sala OE et al (2000) Global biodiversity scenarios for the year 2100. Science. 10;287(5459): 1770-4. https://doi.org/10.1126/science.287.5459.1770
Sanglas A, Palomares F (2022) Response of a mesocarnivore community to a new food resource: recognition, exploitation, and interspecific competition. Eur J Wildl Res 68 (51). https://doi.org/10.1007/s10344-022-01597-4
Santos MJ, Matos HM, Palomares F, Santos-Reis M (2011) Factors affecting mammalian carnivore use of riparian ecosystems in Mediterranean climates. Journal of Mammalogy 92 (5): 1060–1069. https://doi.org/10.1644/10-MAMM-A-009.1
Sarmento PB, Cruz J, Eira C, Fonseca C (2011) Modeling the occupancy of sympatric carnivorans in a Mediterranean ecosystem. European Journal of Wildlife Research 57 (1): 119–131. https://doi.org/10.1007/s10344- 010-0405-x
Sarmento P, Bandeira V, Gomes P, Carrapato C, Eira C, Fonseca C (2021) Adapt or perish: how the Iberian lynx reintroduction affects fox abundance and behaviour. Hystrix, the Italian Journal of Mammalogy: 1–7. https://doi.org/doi:10.4404/hystrix–00372-2020
Shamoon H, Saltz D, Dayan T (2017) Fine-scale temporal and spatial population fluctuations of medium sized carnivores in a Mediterranean agricultural matrix. Landscape Ecology 32 (6): 1243–1256. https://doi.org/10.1007/s10980-017-0517-8
Sillero-Zubiri C, Hoffmann M, Macdonald DW (2004) Canids: Foxes, Wolves, Jackals and Dogs. IUCN, Gland, Switzerland and Cambridge, UK. https://doi.org/10.1515/9781400844913.56
Silva LP, Heleno RH, Costa JM, Valente M, Mata VA, Gonçalves SC, da Silva AA, Alves J, Ramos JA (2019) Natural woodlands hold more diverse, abundant, and unique biota than novel anthropogenic forests: a multi-group assessment. Eur J Forest Res 138: 461– 472. https://doi.org/10.1007/s10342-019-01183-5
SNIRH Sistema Nacional de Informação de Recursos Hídricos (2015) https://snirh.apambiente.pt/. Accessed in 27 Dec 2023.
Soto C, Palomares F (2015) Coexistence of sympatric carnivores in relatively homogeneous Mediterranean landscapes: functional importance of habitat segregation at the fine-scale level. Oecologia. 179 (1):223-35. https://doi.org/10.1007/s00442-015-3311-9
Teixeira DF, Ares-Pereira G, Camarinha C, Lima C, Magalhães A, Castro G, Fonseca C, Rosalino LM (2023 a) Effect of anthropic disturbances on the activity pattern of two generalist mesocarnivores inhabiting Mediterranean forestry plantations. Biodivers Conserv 32: 1251–1270. https://doi.org/10.1007/s10531-023-02548-4
Teixeira DF, Carpio AJ, Rosalino LM, Carniato D, Fonseca C, Acevedo P (2023 b) Can Eucalyptus plantations influence the distribution range of mesocarnivores?. Landscape Ecology. https://doi.org/10.1007/s10980- 023-01787-8
The Nature Conservancy (2009) Terrestrial Ecoregions of the World. https://geospatial.tnc.org/datasets/b1636d640ede4d6ca8f5e369f2dc368b/about. Accessed in 27 Dec 2023
Toews M, Juanes F, Burton AC (2017) Mammal responses to human footprint vary with spatial extent but not with spatial grain. Ecosphere 8 (3): 1- 28. https://doi.org/10.1002/ecs2.1735
Torres RT, Santos J, Fonseca C (2014) Factors influencing red deer occurrence at the southern edge of their range: A Mediterranean ecosystem. Mammalian Biology 79 (1): 52–57. https://doi.org/10.1016/j.mambio.2013.09.002. 75
Van Der Ree R, Jaeger JA, Van Der Grift EA, Clevenger AP (2011) Effects of roads and traffic on wildlife populations and landscape function: road ecology is moving toward larger scales. Ecology and Society 16 (1): 48–48. http://dx.doi.org/10.5751/ES-03982-160148
Venter O, Sanderson EW, Magrach A, Allan JR, Beher J, Jones KR, Possingham HP, Laurance WF, Wood P, Fekete BM, Levy MA, Watson JEM (2016) Sixteen years of change in the global terrestrial human footprint and implications for biodiversity conservation. Nature Communications 7: 1–11. https://doi.org/10.1038/ncomms12558
Verdade LM, Rosalino LM, Gheler-costa C, Pedroso NM, Lyra-jorge M C (2011). Adaptation of mesocarnivores (Mammalia: Carnivora) to agricultural landscapes in Mediterranean Europe and Southeastern Brazil: a trophic perspective. In: Rosalino LM, Gheler-Costa C (eds) Middle-Sized Carnivores in Agricultural Landscapes. Nova Science Publishers pp 1–38. https://www.researchgate.net/publication/232417256_Adaptation_of_mesocarni vores_Mammalia_Carnivora_to_agricultural_landscapes_in_Mediterranean_Eur ope_and_Southeastern_Brazil_a_trophic_perspective. Accessed in 27 Dec 2023.
Vilela ALO, Lamim-Guedes V (2014) Cães domésticos em unidades de conservação: impactos e controle. Holos Environment 14 (2): 198-210. https://doi.org/10.14295/holos.v14i2.8192
Vilella M, Ferrandiz-Rovira M, Sayol F (2020) Coexistence of predators in time: Effects of season and prey availability on species activity within a Mediterranean carnivore guild. Marc. Ecology and Evolution: 1–15. https://doi.org/10.1002/ece3.6778
Wildlife Conservation Society and Center for International Earth Science Information Network Columbia University. Last of the Wild Project, Version2,2 005 (LWP2): Global Human Footprint Dataset (Geographic). Palisades: NASA Socioeconomic Data and Applications Center. http://sedac.ciesin.columbia.edu/data/set/wildareas-v2-human-footprintgeographic. 2005. Accessed in 27 Dec 2023
Zuur AF, Ieno EN, Walker NJ, Saveliev AA, Smith GM (2009) Mixed Effects Models and Extensions in Ecology with R. Journal of the Royal Statistical Society. https://link.springer.com/book/10.1007/978-0-387-87458-6. Accessed in 27 Dec 2023.

QuezadoetalSupplementaryMaterial.docx

Download PDF

Reviewers agreed at journal
16 Oct, 2024
Reviewers invited by journal
11 Oct, 2024
Editor assigned by journal
04 Oct, 2024
First submitted to journal
03 Oct, 2024

You are reading this latest preprint version

Different environmental contexts, different responses: Evaluating the drivers of red fox occupancy patterns in Portugal

Status:

Version 1

Abstract

Figures

Introduction

Materials and Methods

Study Area:

Data Collection and Tested Environmental Drivers:

Statistical analysis

Results

Red-fox detection

Occupancy models

North Region

South Region

Discussion

Conclusions

Declarations

References

Supplementary Files

Status:

Version 1