Introduction
The Andean bear (Tremarctos ornatus) is the only extant Ursid in South America (Peyton 1999), and categorized as Vulnerable (VU) to extinction by the International Union for the Conservation of Nature (IUCN) due to increasing threats such as habitat loss (deforestation and extension of agricultural activities), and illegal hunting of specimens as retaliation because human-bear conflicts (Velez-Liendo and García-Rangel 2017). The distribution of the Andean bear is strongly associated to the Andean mountain range, in discontinuous population blocks (Kattan et al. 2004) located from the Republic of Colombia and the Bolivarian Republic of Venezuela (Rodríguez et al. 2003; García-Rangel 2012) at the north, to the Plurinational State of Bolivia and the Republic of Argentina (Del Moral and Lameda 2011; Teta et al. 2018) at the south.
The northern distribution of the Andean bear in South America has been reported to extend until the Serranía de Portuguesa at Sierra de Mérida along the north-eastern Venezuelan territory (Yerena et al. 2007), and the Serranía del Perijá along the northern Colombian and Venezuelan geographic limits (Rodríguez and Lozada 1989; Peyton 1999; Rodríguez et al. 2003; Corredor-Carrillo and Muñoz-Saba 2007; Yerena et al. 2007; Solari et al. 2013). However, the current northernmost distribution of the Andean bear in South America has not been defined precisely; most authors have suggested this northernmost distribution as being located “at the Serranía del Perijá”, without accurate and confirmed distribution records along this mountain range (Rodríguez and Lozada 1989; Peyton 1999; Rodríguez et al. 2003; Corredor-Carrillo and Muñoz-Saba 2007; Yerena et al. 2007; Solari et al. 2013). The IUCN reports the current northernmost distribution of this species at the Venezuelan Serranía de Portuguesa (north-eastern portion of the Sierra de Mérida; Velez-Liendo and García-Rangel 2017). Although Mondolfi (1989) and Viloria et al. (1995) provided some confirmed distribution records of the Andean bear for the eastern slope of the Serranía del Perjá at Venezuela, these records corresponds to the 50’s, 70’s and 90’s decades.
The lack of information about the Andean bear distribution could be putting the survival of local populations at risk (García-Rangel 2012). Additionally, there is a knowledge gap about the fragmentation of Andean bear associated ecosystems (Andean forest and Paramo) along its Colombian and Venezuelan distribution. In this sense, the aim of this study is to provide precise knowledge on the current northernmost distribution of the Andean bear in South America at the Colombian and Venezuelan Serranía del Perijá, and quantify spatial trends of fragmentation processes of the Andean bear associated ecosystems at the western (Colombia) and eastern (Venezuela) slopes of this mountain range. This constitute a contribution to the knowledge of the species and its associated ecosystems, for the elaboration of future habitat monitoring initiatives and accurately planned conservation actions headed by the environmental authorities (governmental institutions) at their respective local jurisdictions.
Methods
During the years 2009-2016, 15 field surveys were carried out in the Colombian municipalities of Becerril, Chimichagua, Codazzi, Curumaní, La Jagua de Ibirico, La Paz, Manaure, San Alberto (department of Cesar), Barrancas, El Molino, La Jagua del Pilar, Urumita, and Villanueva (department of La Guajira; Figure 1). Additionally, involuntary incursions to the Bolivarian Republic of Venezuela was made during field surveys searching Andean bear traces (see Appendix 1).
In order to obtain accurate and confirmed distribution records of the Andean bear in northern Colombia along the Serranía del Perijá, the zones with recent information of possible presence of the Andean bear (based on the information from the local communities) were visited at each sampled municipality. The presence of the species was verified through the search of bear traces during field surveys in Andean forest and Paramo ecosystems (e. g., foot prints, tracks, bark of trees marked with scratches, feeding traces, feces, direct observations), and occasionally installing camera traps (just for the municipalities of Manaure and La Paz, department of Cesar). Additionally, we performed a revision of online information about presence records of the species (e. g., voucher specimens, camera trap images, direct observations and indirect traces of presence) from scientific literature, scientific collections (Alberto Cadena García Mammal Collection of the Instituto de Ciencias Naturales, ICN; Mammal Collection of the Instituto Alexander von Humboldt, IAvH) and Biodiversity databases (Global Biodiversity Information Facility, GBIF; Mammal Networked Information System, MaNIS; System of Information on Biodiversity of Colombia, SIB). The distribution records obtained through the different sampling methods were organized at Appendix 1, including information on locality, geographical coordinates and type of record (Appendix 1), also being displayed on a map (Figure 1).
A characterization of the fragmentation degree of Andean bear associated ecosystems (Andean forest and Paramo) at the Colombian (western) and Venezuelan (eastern) slopes of the Serranía del Perijá was performed through the analysis of satellite images from 2008 (TERRA ASTER satellite: 3A1-70, 3A1-71, 3A1-72, 3A1-73, 3A1-74; SPOT satellite: 645-330, 646-329, 56463290503141521431A, 2A/203045101). The satellite images were interpreted using ArcMap software (v10.1; ESRI, Redlands, CA, USA), taking into account land coverages over 500 m of elevation to generate a landscape coverage map (scale 1:100,000). Three landscape coverage classes were established: Andean forest (comprising natural Andean forest coverages, considered as natural habitat of the Andean bear), Paramo (comprising natural highland shrubs and grass coverages, considered as natural habitat of the Andean bear), and transformed coverages (comprising anthropogenic or anthropogenic-natural mixed landscape coverages such as pasturelands, croplands and urban areas, not considered as natural habitat of the Andean bear). A raster image (250 m resolution) was extracted from the landscape coverage map and analyzed with FRAGSTATS software (v4.2; McGarigal et al. 2012). Twelve metrics at landscape (overall landscape configuration), class (specific landscape coverages configuration) and patch (individual patch configuration, including a core area analysis extracting a disturbance buffer of 1,000 m into the border/edge of each patch) levels were calculated (for detailed descriptions of metrics see Table 1), quantifying the fragmentation spatial trends for the Andean bear associated ecosystems at the western (Colombia) and eastern (Venezuela) slopes of the Serranía del Perijá (Table 1). Taking into account the current knowledge of home ranges for female (1,477 ha) and male (5,908 ha) Andean bears (Castellanos 2011), individual patches were also categorized per area interval as small (1 ha < patch ≤ 1,500 ha), medium (1,500 ha < patch ≤ 6,000 ha) and large (patch > 6,000 ha) patches. The Andean bear associated ecosystems fragments (patches) were displayed on a map (Figure 1). Although Andean bears do not recognize political geographic limits at its natural movements across associated ecosystems, we focused the fragmentation analysis at a bi-national (Colombia and Venezuela) scale, as a contribution to future management efforts headed by environmental authorities (governmental institutions) of both countries at their respective local jurisdictions.
Value | |||||
---|---|---|---|---|---|
Metrics level | Metric | Description | Category | Colombia | Venezuela |
Landscape level | Shannon’s Diversity Index (SHDI) | Increases as the number of different patch classes increases (patch richness). Equals “0” when the landscape contains only 1 patch of a particular class of landscape coverage (no diversity). | Landscape | 0.78 | 0.48 |
Simpson’s diversity index (SIDI) | Increases (approaching 1) as the number of different patch types (patch richness) increases. Equals “0” when the landscape contains only 1 patch of a particular class of landscape coverage (no diversity). | Landscape | 0.52 | 0.25 | |
Class level | Total class area (CA) | It is the total area of a particular class of landscape coverage. Approaches “0” when a class of landscape coverage becomes rare in the landscape. Measured as hectares (ha). | Forest | 42,6094 ha | 313,244 ha |
Paramo | 15,931 ha | 7,394 ha | |||
Percentage of landscape (PLAND) | It is the percentage of the total area of the entire landscape comprised by a particular class of landscape coverage. Approaches “0” when a class of landscape coverage becomes rare in the landscape. Measured as percentage (%). | Forest | 47.56 % | 85.37 % | |
Paramo | 1.78 % | 2.01 % | |||
Total core area (TCA) | It is the total core area of a particular class of landscape coverage. Approaches “0” when a class of landscape coverage becomes rare in the landscape. Measured as hectares (ha). | Forest | 71,462 ha | 219,350 ha | |
Paramo | 1,131 ha | 56 ha | |||
Core percentage of landscape (CPLAND) | It is the percentage of the total area of the entire landscape comprised by a particular class of landscape coverage. Approaches “0” when a class of landscape coverage becomes rare in the landscape. Measured as percentage (%). | Forest | 7.98 % | 59.78 % | |
Paramo | 0.13 % | 0.02 % | |||
Number of patches (NP) | It is the number of patches of a particular class of landscape coverage. Equals “1” when a class of landscape coverage contains only 1 patch. | Forest | 336 | 19 | |
Paramo | 18 | 15 | |||
Mean patch size (AREA_MN) | It is the mean size of patches of a class of landscape coverage. Approaches “0” when a class of landscape coverage becomes rare in the landscape. Measured as hectares (ha). | Forest | 628 ha | 7,119 ha | |
Paramo | 1,225 ha | 528 ha | |||
Edge density (ED) | Increases as the number of patches of a class of landscape coverage increases, increasing more when patches are smaller (greater contact edge per patch). Equals “0” when a class of landscape coverage contains only 1 patch. Measured as meters per hectare (m/ha). | Forest | 13.48 m/ha | 3.13 m/ha | |
Paramo | 0.63 m/ha | 0.45 m/ha | |||
Mean shape index (SHAPE_MN) | Increases as mean patch shape of a class of landscape coverage becomes more irregular. Equals “1” (minimum value) when patch shape is circular or squared (not irregular shape). | Forest | 1.63 | 1.59 | |
Paramo | 2.24 | 1.91 | |||
Mean euclidean nearest neighbor distance (ENN_MN) | It is the mean distance between nearest patches of a class of landscape coverage. Approaches “0” when the distance between nearest neighbor patches decreases. Measured as meters (m). | Forest | 650.86 m | 520.75 m | |
Paramo | 888.37 m | 1,174.04 m | |||
Patch level | Absolute number of patches (and its area, ha) per area interval | It is the number of patches of andean bear associated ecosystems (and the area comprised by them) within a particular area interval | Small (1 < patch ≤ 1,500 ha) | 667 (57,562 ha) | 53 (7,256 ha) |
Medium (1,500 ha < patch ≤ 6,000 ha) | 13 (35,693 ha) | 3 (7,631 ha) | |||
Large (patch > 6,000 ha) | 11 (348,768 ha) | 2 (305,750 ha) | |||
Absolute number of patches (and its area, ha) per core area interval | It is the number of patches of andean bear associated ecosystems (and the core area comprised by them) within a particular core area interval | Small (1 < patch ≤ 1,500 ha) | 17 (4,512.5 ha) | 4 (200 ha) | |
Medium (1,500 ha < patch ≤ 6,000 ha) | 2 (5,493 ha) | 0 (0 ha) | |||
Large (patch > 6,000 ha) | 1 (62,587 ha) | 2 (219,206 ha) |
Results
From the field surveys, 97 confirmed presence records of the Andean bear was found along the Serranía del Perijá at the municipalities of Becerril, Chimichagua, Codazzi, Curumaní, La Jagua de Ibirico, La Paz, Manaure, San Alberto (department of Cesar), Barrancas, El Molino, La Jagua del Pilar, Urumita and Villanueva (department of La Guajira) for the western slope of the Serranía del Perijá at Colombia, and at the municipalities of Mara, Jesús Enrique Lozada, Rosario de Perijá and Machiques de Perijá (state of Zulia) for the eastern slope of the Serranía del Perijá at Venezuela (involuntary incursions to the Bolivarian Republic of Venezuela during field surveys while searching Andean bear traces). Literature review provided eight presence records of the Andean bear at the municipalities of Mara, Jesús Enrique Lozada, Machiques del Perijá and Rosario del Perijá (state of Zulia) for the eastern slope of the Serranía del Perijá at Venezuelan (Mondolfi 1989; Viloria et al. 1995; Appendix 1), and did not yield results for the western slope of the Serranía del Perijá at Colombia. The revision of online records coming from scientific collections and biodiversity databases did not yield results for presence records of the Andean bear at the north of Colombia and Venezuela along the Serranía del Perijá.
We report the current northernmost distribution records of the Andean bear in South America for the locality Monte Fresco at the western slope of northern Serranía del Perijá in Colombia (municipality of Barrancas, department of La Guajira), and the adjacent eastern slope of northern Serranía del Perijá in Venezuela (municipality of Mara, state of Zulia; Figure 1; Appendix 1). These northernmost records corresponded to recent scratches on trees of Nigüito (Miconia sp.), Aguacatillo (Persea sp.), Pringamoza (Urera baccifera), Guamo (Inga sp.) and feces on natural bear trails into the Andean forest, as well as direct observations of an individual feeding on corn and plantain crops (Appendix 1).
The results of the characterization of the spatial trends of fragmentation degree of the Andean bear associated ecosystems (Andean forest and Paramo) along the Serranía del Perijá are summarized in Table 1. The Colombian slope possess a much greater fragmentation degree regarding the Venezuelan one, represented by greater patches diversity index values (SHDI: 0.78 Colombia vs 0.48 Venezuela; SIDI: 0.52 vs 0.25) at the landscape level analyses (Table 1), a greater number of patches (NP: 336 vs 19; class level analysis), smaller size of patches (AREA_MN: 628.46 ha vs 7,119.18 ha) and bigger isolation between patches (ENN_MN: 650.86 m vs 520.75 m) for the Andean forest (the dominant landscape coverage for both slopes in terms of surface and patch occurrence, Table 1), as well as greater edge amount for the Andean forest (ED: 13.48 m/ha vs 3.13 m/ha) and Paramo (ED: 0.63 m/ha vs 0.45 m/ha), and a greater irregular shape of patches for the Andean forest (SHAPE_MN: 1.63 vs 1.59) and Paramo (SHAPE_MN: 2.24 vs 1.91), at the class level analyses (Table 1).
In terms of patch level analyses, when core area analyses were performed (extracting a disturbance buffer of 1,000 m into the border/edge of each patch, Table 1), for the Colombian slope of the Serranía del Perijá, the total area comprised by small (1 ha < patch ≤ 1,500 ha), medium (1,500 ha < patch ≤ 6,000 ha) and large (patch > 6,000 ha) patches of Andean bear associated ecosystems is reduced from 57,562.5 ha to 4,512.5 ha (small patches), 35,693.75 ha to 5,493.75 ha (medium), and 348,768.75 ha to 62,587.5 ha (large patches; Table 1). On the other hand, for the Venezuelan slope of the Serranía del Perijá, those reductions were from 7,256.25 ha to 200 ha (small patches), 7,631.25 ha to “0” (medium), and 305,750.01 ha to 219,206.25 ha (large patches; Table 1).
Discussion and conclusions
The presence records of the Andean bear obtained during the field surveys made for this study and from the revised literature, reveals at least three distinguishable sectors with bears presence along the eastern (Venezuela) and western (Colombia) slopes of the Serranía del Perijá (Figure 1; Appendix 1). The first sector corresponds to the southernmost portion of the Serranía del Perijá at the municipality of San Alberto (department of Cesar) at Colombia, coinciding with the only Andean bear distribution currently recognized by the IUCN for the Serranía del Perijá (Velez-Liendo and García-Rangel 2017; Figure 1). The second sector corresponds to the center of the Serranía del Perijá, and includes the municipalities of Chimichagua and Curumaní (department of Cesar) at Colombia, which is most probably connected with the Andean bear distribution block recognized by the IUCN for the Colombian National Natural Park Catatumbo Barí in the department of Norte de Santander (adjacent, but not belonging to the Serranía del Perijá; Velez-Liendo and García-Rangel 2017; Figure 1). The third sector correspond to the north of the Serranía del Perijá and includes the municipalities of La Jagua de Ibirico, Becerril, Codazzi, La Paz, Manaure (department of Cesar), La Jagua del Pilar, Urumita, Villanueva, El Molino and Barrancas (department of La Guajira) at Colombia, and the municipalities of Mara, Jesús Enrique Lozada, Rosario de Perijá and Machiques de Perijá (state of Zulia) at Venezuela (Figure 1). Although there is abundant information from local communities on the presence of the species at other municipalities from the western slope of the Serranía del Perijá at Colombia, that could complement the distribution information for the species at the south (Municipalities of Abrego, San Martín and Cachira at the department of Norte de Santander), middle (El Carmen Municipality at the department of Norte de Santander, and the municipalities of Chiriguaná, La Gloria, Pelaya and Pailitas at the department of Cesar) and north (municipalities of San Juan del Cesar and Fonseca at the department of La Guajira) of the Colombian Serranía del Perijá (Torres 2008; Rodríguez and Reyes 2009; Rodríguez et al. 2010; Rodríguez et al. 2011), those areas were not sampled for this study.
These results exhibit a current Andean bear population block located along the eastern (Venezuela) and western (Colombia) slopes of the Serranía del Perijá (Appendix 1; Figure 1), previously suggested as the “Perijá population nucleus” by Rodríguez and Lozada (1989). The current geographical distribution recognized by the IUCN for the Andean bear in South America states that its northernmost limit is reached at the Venezuelan Serranía de Portuguesa (Sierra de Mérida, north-eastern Venezuela; Velez-Liendo and García-Rangel 2017; Figure 1), including just the suthernmost portion of the Serranía del Perijá in the distribution range of this species, and dismissing previous distributions mentioned by Mondolfi (1989), Rodríguez and Lozada (1989), Viloria et al. (1995), Peyton (1999), Rodríguez et al. (2003), Corredor-Carrillo and Muñoz-Saba (2007), Yerena et al. (2007), and Solari et al. (2013), that suggested the northern presence of the Andean bear as reaching at least the middle portion of the Serranía del Perijá. Thus, the presence records provided at this study extend the distribution of the species 180 km at the north of Colombia and Venezuela until the north of the Serranía del Perijá regarding the current distribution recognized by the IUCN (Velez-Liendo and García-Rangel 2017; Figure 1), provide confirmed records that supports previously suggested distributions of the Andean bear along the Serranía del Perijá (Rodríguez and Lozada 1989; Peyton 1999; Rodríguez et al. 2003; Corredor-Carrillo and Muñoz-Saba 2007; Yerena et al. 2007; Solari et al. 2013), update previous records reported for the Venezuelan (eastern) slope of the Serranía del Perijá (Mondolfi 1989; Viloria et al. 1995), and elucidate a more accurate northernmost distribution limit for this species in South America at Colombia and Venezuela. These findings locate the northernmost distribution range limit of the Andean bear at the municipality of Barrancas (locality Monte Fresco, department of La Guajira) in Colombia, and the municipality of Mara (state of Zulia) in Venezuela (Figure 1; Appendix 1). These northernmost records also confirm that feeding traces on plantain bushes include not just fruit consumption, but also buds consumption, in agreement with popular information previously collected (Rodríguez and Reyes 2009, Rodríguez et al. 2010).
Although there is some general information about Andean bear distribution range fragmentation and habitat availability for the northern Andes (Venezuela, Colombia and Ecuador; Kattan et al. 2004), this information focuses at a large scale analysis of potential distribution discontinuity for the species, with landscape coverage information from 1994 for some sectors and from 2000 for others. Although the characterization of the spatial trends of fragmentation degree of the Andean bear associated ecosystems (Andean forest and Paramo) along the Serranía del Perijá made for this study used satellite images from 2008 (and fragmentation trends could have slightly changed), this study provides the first fine-scale information on Andean bear associated ecosystems fragmentation for Colombia and Venezuela, with traceable landscape measurements permitting long-term comparative analyses in the future (Table 1). These analyses exhibit that the Colombian (western) and Venezuelan (eastern) slopes differ in patch number, size, shape, edge and isolation between parches, so that the Colombian slope possess a much greater fragmentation degree regarding the Venezuelan one (Table 1).
Although the Colombian slope of the Serranía del Perijá comprise a bigger raw area of Andean bear associated ecosystems (442,025 ha vs 320,638 ha), the Venezuelan slope comprise much less fragmented ones (Table 1), so that the Andean bear associated ecosystems resulting from core area analyses exhibit much more extension (area) for the Venezuelan slope (219,406.25 ha) than the Colombian one (72,593.75 ha; Table 1). This phenomenon is related to the high quantity of small and irregularly shaped patches present at the Colombian slope of the Serranía del Perijá (that disappear at the core area analyses) (Figure 1); smaller patches with greater shape complexity have less internal area, which constitutes less habitat availability and greater edge effect (McGarigal et al. 2012). In the case of the Colombian slope of the Serranía del Perijá, the loss of natural landscape coverage is a result of extensive agricultural activities for legal (food production for self-maintenance or trade) and illegal (crops to produce illegal drugs) purposes, as well as wood extraction to generate grazing pastures (livestock) or build houses (Torres 2008; Rodríguez and Reyes 2009; Rodríguez et al. 2010; Rodríguez et al. 2011). In this sense, Kattan et al. (2004) previously mentioned that large habitat blocks encompassing a continuous elevational gradient remain in the Venezuelan (eastern) slope of the Serranía de Perijá.
The Andean bear associated ecosystem Paramo is scarcely represented along the Serranía del Perijá, constituting 1.78 % of the landscape in the Colombian slope, and 2.01 % in the Venezuelan slope of this mountain range (PLAN; Table 1). Due to its natural distribution at Andean mountain peaks separated between them by large distances (Cuatrecasas 1958), this ecosystem is intrinsically fragmented (Jiménez-Rivillas et al. 2018). When Paramo is present, Andean bear uses it, but it does not constitute its main habitat (Rodríguez 1991; Ríos-Uzeda et al. 2006), a fact supported by the small amount of Paramo ecosystem comprised by the distribution of the persistent Andean bear population block of Perijá (Mondolfi 1989; Viloria et al. 1995; Appendix 1). In this sense, many authors have previously indicated the preference of the Andean bears to use the Andean forest rather than the Paramo ecosystem (Walker 1964; Borrero 1967; Knight 1967; Mondolfi 1971; Peyton 1980; Yerena 1987; Rodríguez 1991; Ríos-Uzeda et al. 2006).
The elongated shape of Andean bear distribution range and patterns of anthropic land use in the northern Andes make the distribution range of this species particularly prone to fragmentation (Kattan et al. 2004). Taking into account the high degree of fragmentation for the Andean bear associated ecosystems at the Colombian slope of the Serranía del Perijá (Table 1), presumably when bears belonging to the Perijá population block transit between the two slopes of the Serranía del Perijá, they pass from the Venezuelan (eastern) slope with mature and relative continuous forests, to the Colombian (western) slope where the continuity of the forest is broken due to intensive and extended human activities and the ecological integrity is very compromised (Figure 1; Table 1). Those intensive and extended human activities at the Colombian slope increases the risk of human-bear conflicts and hunting retaliations from settler communities, as well as subsistence hunting activities by indigenous communities.
The Andean bear is categorized as vulnerable to extinction (VU) at a general scale by the IUCN (Velez-Liendo and García-Rangel 2017), and at Colombia by the Ministerio de Ambiente, Vivienda y Desarrollo Territorial (MADS 2017), also categorized as endangered (EN) by Rodríguez et al. (2015) at Venezuela. There is no information on the population dynamics and conservation threats of the northernmost portion of the Andean bear population block of Perijá at the municipalities of Barrancas (locality Monte Fresco, department of La Guajira) in Colombia and Mara (state of Zulia) in Venezuela (Figure 1; Appendix 1), a concerning matter since this sector of the Serranía del Perijá is subjected to intense wood extraction and hunting activities (Rodríguez et al. 2010), and located just 12 kilometers southeast from the largest open-pit coal mine in Colombia (El Cerrejón; Figure 1).
There are generalized and not confirmed assumptions about the northern distribution of the Andean bear in South America, as the case of the lacking-evidence distribution reported for the species at the Colombian Sierra Nevada de Santa Marta at the departments of Cesar, Magdalena and La Guajira (north-western Colombia; Solari et al. 2013), refuted by Reyes-Amaya (2015), and the wrong assumption of not presence of the species at the Colombian and Venezuelan Serranía del Perijá (Velez-Liendo and García-Rangel 2017; Figure 1) refuted in this study (Figure 1; Appendix 1). Likewise, the absence of fine-scale information on Andean bear associated ecosystems fragmentation for Colombia and Venezuela could represent a major challenge for the conservation of this Ursid. The lack of information about the Andean bear distribution (García-Rangel 2012) and its associated ecosystems fragmentation could be putting the survival of local populations at risk, while a clear knowledge on these issues could generate planned conservation actions for this species by the governmental environmental authorities at their respective local jurisdictions. The absence of information on the conservation state of the northernmost population block of this species at South America (at the Serranía del Perijá) makes it necessary to carry out biological research on its conservation and threats.