The Spectral Bat (Vampyrum spectrum) is classified as Near Threatened by the International Union for Conservation of Nature. In Colombia, it has not been included in any risk assessments, because it is a poorly documented species in the country. Using new occurrence data and ecological niche modelling we evaluate the distribution and analyze the species’ ecological and conservation needs in Colombia. With the new records, V. spectrum shows a distribution over six biogeographic provinces in the country, along an elevational range between 96 and 1750 m. The Spectral Bat has been reported in a wide variety of ecosystems from dry forest and fragmented landscapes, to forested areas. Predictive models suggest a wider presence of this species in the west of the country, center, and southern Andes. The limited number of specimens limits our attempt to evaluate morphometric variation among Colombian populations; however, we detect a variation in the mean value of the forearm and ear length regarding northeast South American populations. Areas with the greatest potential distribution detected in this study could be prioritized in conservation strategies, and make plans accordingly for this and other species that coexist in its range. Finally, we highlight the low representativeness of Protected Areas for the Spectral Bat and add some ecological notes about this rare species.

Carnivorous bat, ecological niche modelling, Phyllostominae, Protected Areas, rare species, Spectral Bat

The Spectral Bat Vampyrum spectrum (Linnaeus, 1758) is the largest bat species in the Neotropical region. It has a wide distribution ranging from southern Mexico to northern and central Bolivia (Da Silva and Rossi 2011; Acosta et al. 2019), east to the Amazon and central and northern Brazil (Nunes et al. 2005; Nogueira et al. 2007; Sousa et al. 2016). Despite its wide distribution, it is considered a rare species due to its low abundance and scarce representation in biological collections (Greenhall 1968; Solari 2018). Little is known about its natural history, ecology, and population dynamics (Acosta and Azurduy 2006; Acosta et al. 2019). It inhabits well-preserved environments (Solari 2018), where it acts as a top predator of small birds and mammals (rodents and bats), and possibly its diet includes some insects and fruits too (Peterson and Kirmse 1969; Navarro and Wilson 1982; Bonato et al. 2004; Gardner 2008). In addition, some studies have provided information on its echolocation, home range (Vehrencamp et al. 1977), and local distribution (Sousa et al. 2016; Esquivel and Rodríguez-Bolaños 2018).

Globally, the Spectral Bat is listed as Near Threatened (NT) according to the IUCN Red List (Solari 2018) due to population decline associated with loss and fragmentation of natural habitats (Aguirre et al. 2008). In Colombia, as in other countries, there are relatively few records, limiting our knowledge about its actual distribution and conservation status. The first record for Colombia was made by Hall and Kelson (1959) based on one specimen collected at “El Boquerón (de Napipi), Chocó (6°43'N, 77°30'W)”, however, this locality is actually located in Panama, causing confusion about the species presence in Colombia (Aellen 1970; Gardner 2008). Later, Cuervo et al. (1986) included V. spectrum as present in humid forest up to 2800 m a. s. l. in Colombia, but they provided no additional information. Alberico et al. (2000) followed Cuervo et al. (1986) by including the species in an elevational range between 0 to 2800 m, and mentioned voucher specimens in the collections of the Instituto de Investigaciones Biológicas Alexander von Humboldt (IAvH), Villa de Leyva and the Instituto de Ciencias Naturales (ICN), Bogotá.

Since then, the presence of the species has been confirmed in different geographic regions of Colombia including: i) The Andean Region: departments of Antioquia (Cuartas-Calle and Muñoz-Arango 2003), Caldas (Castaño et al. 2003; Escobar-Lasso et al. 2013), Cauca (Ramírez-Chaves and Pérez 2010), Huila (Rivas-Pava et al. 2007), Tolima (Maria 2004; Gutiérrez-Díaz et al. 2010; García-Herrera et al. 2019; Esquivel et al. 2020), Nariño (Ramírez-Chaves and Noguera-Urbano 2010), Cundinamarca, and Valle del Cauca (Mantilla-Meluk et al. 2009). ii) The Amazon Region: departments of Amazonas (Montenegro and Romero-Ruiz 1999; Rojas et al. 2004; Mantilla-Meluk et al. 2009), and Caquetá (Montenegro and Romero-Ruiz 1999; Marín-Vásquez and Aguilar-González 2005). iii) The Pacific Region: Department of Chocó (Asprilla et al. 2016). iv) The Caribbean Region: departments of Sucre (Cabrera 2011) and La Guajira (Esquivel and Rodríguez-Bolaños 2018), and v) the Orinoquian Region: departments of Meta and Vichada (Mantilla-Meluk et al. 2009; Esquivel and Rodríguez-Bolaños 2018). However, the elevational range distribution is not clearly defined, and has been mentioned as occurring from 0 to 2800 m (Cuervo et al. 1986; Alberico et al. 2000), from 10 to 1065 m (Mantilla-Meluk et al. 2009), and from 10 a 2150 m (Solari et al. 2013).

Correlative environmental niche modelling (ENM) attempts to model a species’ relationship with environmental variables and predict habitat suitability for that species based on known observations of its presence. Although ENM techniques are inherently limited by imperfect knowledge, they have nonetheless proven extremely useful in assessing the effectiveness of Protected Areas distribution (Catullo et al. 2008), assessing species’ vulnerability to proposed local land use change (Santos et al. 2013), predicting distributions of rare species (Marino et al. 2011; Rheingantz et al. 2014) and possible responses to climate change by species and ecosystems (Moor et al. 2015; Sobral-Souza et al. 2018). Due to the limited information on the distribution of V. spectrum in Colombia, environmental niche models could be extremely useful for predicting the distribution of this species in several areas of the country. In addition, a predictive model of its distribution could be calculated based not only on records gathered from this country, but also from other nations in its range.

In this study we (1) report new localities of V. spectrum from the Andean, Caribbean and Orinoquian regions, (2) update the distribution of the species in Colombia, (3) use the distribution data together with Ecological Niche Modelling (ENM) to generate a map of the potential distribution of the species, and (4) include information on its morphometric variation, conservation status and some ecological notes that contribute to the knowledge of this rare species.

In Colombia, Vampyrum spectrum is known from 37 records in 15 departments along six biogeographic provinces: a) Peri-Caribbean Arid Belt, b) Orinoco, c) Guyana, d) Amazonia e) Chocó-Magdalena and f) Nor-Andean. Likewise, the species has been recorded in different ecoregions as Montane forests (19 records), Moist forests (8), Dry forests (4), Llanos (4) and Xeric scrub (2). Based on vouchers, the Spectral Bat is distributed in an elevational range between 96 and 1750 m, with most records (67.5%; N = 25) below 600 m. Of the total, four are new localities, recently obtained in fieldwork activities (Fig. 1; Table 1) and presented below.

Figure 1. 

Localities of records of Vampyrum spectrum in Colombia. Red circles show previous records and blue squares show the new records. The numbers correspond to those described in Table 1.

The first record from Mutatá, Department of Antioquia – north Urabá sub-region, “Vereda” Chontadural (7°13'50.4"N, 76°26'58.8"W; 163 m a. s. l) – is a specimen found as a result of a biological characterization in a requirement for obtaining environmental permits for the development of a mining project (stone), in June 2019. The specimen is an adult male (CZUC-M 0409), found entangled in a wire fencing at ground level on a gutter created for runoff. Because the skin was already totally destroyed, with only the skull preserved, a fracture in the braincase was presumed to be the cause of death. The area containing this specimen was a wooded cover of tropical humid forest (bh-T) crossed by the Caño Bijará, where secondary vegetation prevails with a wide canopy providing spaces for the species’ flight capacity. The second record comes from Paz de Ariporo, Department of Casanare (5°29'24.36"N, 71°7'4.79"W; 330 m a. s. l), and corresponds to an adult male captured as result of a biological characterization during 2017. The specimen was photographed and then released in the same area, which has gallery forests and extensive plains. The third record comes from San Alberto, Department of Cesar (7°45'0"N, 73°22'58.7"W; 125 m a. s. l); it is an individual found in the development of a biological characterization during 2015. The individual corresponds to an adult female recorded in a gallery forest fragment, in a very narrow strip of secondary vegetation on the banks of a stream, near to oil palm crop. The area where this specimen was found is strongly pressured by the oil crop, but there are still some remnants of gallery forest that allow the survival of the species. The fourth record comes from the municipality of Gualanday, Department of Tolima (4°17'24"N, 75°2'9.6"W; 541 m a. s. l), and represents an adult female found as a result of the development of an environmental impact study during 2015. The individual was recorded by Tatiana Toro-Sánchez in a dry tropical forest remnant near water bodies and agricultural crops (Fig. 2).

Figure 2. 

Individual of Vampyrum spectrum captured in Gualanday, Department of Tolima, Colombian Central Andes. Photo by Tatiana Toro-Sánchez.

Ecological niche modelling

Our models had a mean TSS (0.49 for BRT, 0.48 for Maxent, 0.47 for RF and 0.46 for SVM) and mean AUC (0.78 for BRT, 0.76 for Maxent, Random Forest and SVM) indicating the models fit well to the data. All individual model runs had values above our minimum accuracy values also. The relative habitat suitability for V. spectrum in Colombia based on our models, and the areas converted into a binary threshold and subsequently filtered to remove areas above 2000 m in elevation, are shown in Fig. 3A, B, respectively. The outputs indicate that habitat suitability across Colombia based on ENM ensemble was generally predicted to be much higher in the west of the country, in the Central Andes and the departments of the southern Andes.

Figure 3. 

A Potential distribution of Spectral Bat Vampyrum spectrum in Colombia using an ensemble modelling approach with five different algorithms. Blue to red color scales indicate less to greater suitability area for the species. B Ensemble suitability map converted to a binary of above (suitable) and below (not suitable) threshold, manually reclassified to exclude high elevation areas (>2000 m a. s. l). Areas considered suitable were focused in the west and south of the country, although eastern Colombia also had significant areas above the threshold.

Regarding variable importance, although the degree of importance differed between the two importance assessment metrics (TSS/AUC), the order was constant in both. Under both importance assessment methods, forest cover was by far the most important variable, followed by Bio14 (precipitation in the driest month), then slope, and finally Bio05 (maximum temperature of the warmest month) (Table 2). Suitability for V. spectrum declined sharply in very high temperatures (> 35 °C approximately), whilst higher precipitation increased suitability. Areas with high slope were less suitable for the species and the forest cover generally had a positive relationship with suitability, although this one decreased after 65% approximate forest cover (Fig. 4).

Table 2.

Download as

CSV

XLSX

Mean variable importance scores across all models and replicates.

Variable	AUC metric	Pearson correlation
Bio05	12.8	7.1
Bio14	40.4	19.7
EDForest	53.7	29
Slope	17.6	7.5

Figure 4. 

Variable response curves for Vampyrum spectrum. High values for temperature (bio05) and slope are related to less available habitats for the species.

Vampyrum spectrum presents a wide but discontinuous distribution in Colombia, probably related to its low detection. Information about this species is limited and mainly restricted to vouchers’ records in natural history collections. From the first record in 1952 (Museo de Historia Natural de la Universidad del Cauca – MHNUC M042E) until 2020 only 37 records have been obtained. Although this number is low, when we analyze their temporality, we find that 70.2% were made before the year 2010, and in areas that are currently deforested, as in Tolima [reported in Guamo by Niceforo Maria in 1968, (Maria 2004)], Antioquia (in Zaragosa by N. Peterson in 1972), Caldas [via La Victoria-Mariquita by E. Osorno and A. Morales in 1967, (Castaño et al. 2003)], among others.

The number of records occurring after 2010 (N = 11), indicate the presence of the species in well-preserved areas in Amazonas, Caquetá, Meta, and Nariño, and its subsistence in patches of forest with strong anthropogenic pressure in the Central Andes. Although efforts are required to determine the state of the populations and confirm their presence in other areas, it is possible that the populations of the Andes and, in particular, northern Colombia, are most threatened given the high rates of deforestation that affect these regions (Etter et al. 2006). The habitat loss and fragmentation, and transformation of natural environments may disconnect populations, favoring possible local extinction events (Lino et al. 2019). Furthermore, intentional killings at the hands of local people who perceive bats as a threat, as occurs for other species of bats, could be causing the loss of a considerable number of individuals, mainly associated with the lack of knowledge and disinformation of the ecosystem services they provide (O’Shea et al. 2016; Bhattacharjee et al. 2018).

In Colombia, there is little information about the ecosystems occupied by V. spectrum. Herein, we recognize it in diverse ecosystems, but mainly associated with primary and secondary forests (Cabrera 2011; Esquivel and Rodríguez-Bolaños 2018), tropical dry forests (Maria 2004), and sub-Andean forests. Information about habitat preferences in the western region of the country is scarce, but, curiously, this area is predicted to have the most habitat suitability for the species. Finding the species in diverse environments with different degrees of anthropogenic disturbance, seems to indicate a high tolerance of the Spectral Bat to habitat transformation. Additional information about diet and prey consumption, breeding patterns and home range in Colombia is non-existent; in the same way, it was not possible to recover many field notes and data about the collection events.

Some recent records confirm the low detectability and local rarity of the Spectral Bat. The new record for the Department of Antioquia confirms the presence of the species that has not been registered since 1959 for the Urabá subregion (Cuartas-Calle and Muñoz-Arango 2003). Similarly, records after 35 years have been reported for the Department of Meta (Esquivel and Rodríguez-Bolaños 2018), despite several sampling efforts in both departments.

Forest cover is the most important factor behind V. spectrum’s distribution, and the response curves for this variable support this view – as has been suggested in other countries (Solari 2018). However, extremely high levels of forest cover reduced suitability, perhaps indicating a possible preference of the Spectral Bat for mosaic habitats. Precipitation in the driest month (Bio 14) was also very important for the species, with the response curves indicating higher levels of suitability for areas with high precipitation. These results suggest a potential vulnerability of V. spectrum to climate change effects in areas where temperature increases or precipitation decreases.

Although the ENM models produced high TSS and AUC scores, the ensemble maps should not be considered infallible. Factors such as human-wildlife conflict, competing species, dispersal limitations or stochastic processes can render climatically suitable areas unviable for V. spectrum. ENM methods can also be affected by unknown prevalence (the proportion of suitable sites occupied by the species), which can lead to misleading results (Guillera-Arroita et al. 2015). Our spatial rarefaction of the presence data should help to reduce the issue of sampling bias; however, the issue should not be considered completely resolved. Whereas they should not be considered infallible (especially for the thresholded map), they could nonetheless prove useful for planning future research and conservation actions.

Reviewed specimens show similar morphology and morphometrics, indicating little variability among the Colombian populations. The differences observed in the ear length and mean value of the forearm could be explained by the way the collectors historically took this measurement in the former and, in the latter, with clinal variation also observed in other bat species (Kelly et al. 2018). Specimens also showed variation in external characteristics such as fur coloration, ranging from grayish to reddish brown as have been previously recorded (Navarro and Wilson 1982). We are not able to define whether the differences between Colombian populations and northeast South America are associated with cryptic diversity within V. spectrum. However, further evidence and broader data analyses of the ear and forearm lengths should be explored to evaluate whether the differences are associated with clinal variation.

The Spectral Bat is listed as Near Threatened by IUCN (Solari 2018), Vulnerable in Bolivia and Ecuador (Aguirre et al. 2009; Burneo et al. 2015) and Least Concern in Brazil (Instituto Chico Mendes de Conservação da Biodiversidade 2018). In Colombia, the species is not included in any conservation plan, or a framework or protection policy. It is not included in the Red Book of Threatened Species (Rodríguez-Mahecha et al. 2006) or in the Threatened Fauna List of the Colombian government (Ministerio de Medio Ambiente y Desarrollo Sostenible 2017). Lack of information about natural history may be one of the main reasons why it is difficult to assess its conservation status. Likewise, the effectiveness of Protected Areas for the conservation of this species is not well known. Only three records (among 37) have been reported in these areas and our analyses show that only 13.06% of its suitable habitat area in Colombia is protected.

The key localities with capacity to contribute to the long-term survival of this species are mostly outside the boundaries of Protected Areas. In particular, we highlight the lack of PAs in the Andean region, because it is the region with the largest area of distribution in Colombia and an area quite threatened by the increasing human land use and habitat transformation. Protected areas are considered a central axis of national and international conservation strategies, becoming a fundamental tool to ensure the conservation of species and their ecosystems (Balmford et al. 2005). However, this study suggests that PAs are not providing efficient repositories for the species target conservation. This result is intended to help decision makers by identifying priority areas in which conservation efforts should focus.

Future studies should consider contributing to the knowledge of the natural history of this rare species. Widely unknown aspects such as habitat use and preference, population density, occupancy area and response to climate change should be a priority. The conservation of the Spectral Bat will depend on actions dedicated to the preservation and recovery of its habitat, to the formulation of Protected Areas and the development of environmental education activities. Likewise, the species should be included in plans of public or private areas where their presence has been recorded, as well as the protection of its identified shelters.

We thank the curators and assistants of the collections for providing data, and particularly the Kurupira Foundation for logistical and financial support. We extend our appreciation to Camilo Ramos Madera for his support in the revision of the document and Cristian Moreno for the distributional records map. Finally, we thank the reviewers for their suggestions and revision to the manuscript. Tatiana Toro-Sánchez by allowing the use of the photographs from Department of Tolima.

Natural History Museums examined. Reviewed personally * and online ** collections.

Colección de Mamíferos del Instituto Alexander von Humboldt (IAvH), Villa de Leyva*

Colección de Mamíferos de la Universidad de Nariño (PSO-CZ), Pasto*

Colección Mastozoológica del Chocó (CMCH), Quibdó**

Colección de Mastozoología Universidad del Quindío (CMUQ), Armenia**

Colección Teriológica de la Universidad de Antioquia (CTUA), Medellín*

Colección Mastozoológica Universidad Industrial de Santander (UIS-MHN), Bucaramanga**

Colección Zoológica Universidad de Córdoba (CZUC), Montería*

Colección Zoológica Universidad del Tolima (CZUT), Ibagué*

Instituto de Ciencias Naturales (ICN), Universidad Nacional de Colombia, Bogotá*

Museo de Historia Natural Universidad Distrital Francisco José de Caldas (MUD), Bogotá*

Museo Javeriano de Historia Natural “Lorenzo Uribe, S.J.” (MPUJ), Bogotá*

Museo de La Salle (MLS), Bogotá*

Museo de Historia Natural Universidad de Los Andes (Andes-M), Bogotá*

Museo de Historia Natural Universidad de Caldas (MHN-UCa), Manizales*

Museo de Historia Natural Universidad del Cauca (MHNUC), Popayán*

Universidad del Valle-Vertebrados Mamíferos (UV), Cali**.