Neotropical insectivorous bats adjust echolocation calls
in response to vegetation structure

María Camila Salazar-Pérez¹* and Sergio Estrada-Villegas² .

1Laboratorio de Ecología Funcional, Unidad de Ecología y Sistemática (UNESIS), Departamento de Biología, Facultad de Ciencias, Pontificia Universidad Javeriana. Bogotá D.C., Colombia.

2Historia natural de plantas tropicales, Universidad del Rosario, Bogotá D.C., Colombia. E-mail: sergio.estradav@urosario.edu.co,

*Corresponding author: csalazarperez1@gmail.com

Echolocation is a crucial process in bat ecology, but it is influenced by different biotic and abiotic factors. The structural complexity of the vegetation can filter which species can be part of an assemblage because bats have to modify the spectral and temporal parameters of their echolocation calls. We evaluated how insectivorous bats modulate their echolocation calls among sites that vary in vegetation structural complexity in the Llanos of Colombia in three habitats. We used canonical correspondence analysis (CCA) to correlate vegetation variables per plot with the spectral and temporal variables of the echolocation signals from bats using those habitats. We found that delta frequency and delta time had the greatest variability in relation to the percentage of tree cover and diameter at breast height. Our results indicate that pulses, in terms of frequency and time, tend to fluctuate more in cluttered habitats. We then selected the two species with the highest number of recordings (Saccopteryx bilineata and Neoeptesicus cf. furinalis) and compared spectral and temporal variables among sites. For Neoeptesicus cf. furinalis showed a predictable behavior: pulses were shorter, and more frequency modulated in highly cluttered habitats. However, and contrary to our expectations, the pulses of Saccopteryx bilineata tended to be longer and with a smaller frequency modulated component in highly cluttered habitats. These results demonstrate that bats modulate their echolocation calls in different ways according to the structure of vegetation. Modulation of echolocation has important implications for sensory ecology and bat composition in tropical landscapes.

Key words: Audiomoth, functional diversity, habitat filtering, plant structural complexity, savannas.

La ecolocalización es un proceso crucial en la ecología de los murciélagos, pero está influenciada por diferentes factores bióticos y abióticos. La complejidad estructural de la vegetación puede determinar qué especies pueden formar parte de un ensamblaje, ya que los murciélagos deben modificar los parámetros espectrales y temporales de sus llamadas de ecolocalización. Evaluamos cómo los murciélagos insectívoros modulan sus llamadas de ecolocalización en lugares con diferentes niveles de complejidad estructural de la vegetación en los Llanos de Colombia en tres hábitats. Aplicamos un análisis de correspondencia canónica (CCA) para correlacionar las variables de vegetación por parcela con las variables espectrales y temporales de las señales de ecolocalización de los murciélagos que habitan esos lugares. Descubrimos que la delta de frecuencia y la delta de tiempo presentaron una mayor variabilidad en relación con el porcentaje de cobertura arbórea y el diámetro a la altura del pecho. Nuestros resultados indican que los pulsos tienden a fluctuar más en términos de frecuencia y tiempo en hábitats densamente estructurados. Posteriormente, seleccionamos las dos especies con mayor número de grabaciones (Saccopteryx bilineata y Neoeptesicus cf. furinalis) y comparamos sus variables espectrales y temporales en los distintos lugares. En Neoeptesicus cf. furinalis mostró un comportamiento predecible: los pulsos eran más cortos y presentaban una mayor modulación de frecuencia en hábitats muy densos. Sin embargo, y contrario a nuestras expectativas, los pulsos de Saccopteryx bilineata tendieron a ser más largos y con un componente de modulación de frecuencia menor en hábitats muy densos. Estos resultados demuestran que los murciélagos modulan sus llamadas de ecolocalización de diferentes maneras de acuerdo con la estructura de la vegetación. La modulación de la ecolocalización tiene implicaciones importantes para la ecología sensorial y la composición de murciélagos en paisajes tropicales.

Palabras clave: Audiomoth, complejidad estructural de la vegetación, diversidad funcional, hábitat como un filtro, sabanas.

Functional diversity plays a crucial role in regulating community structure, determining ecosystem functioning (Tilman et al. 1997; Naeem and Wright 2003), and influencing niche partitioning, guild membership and habitat selection in animals (García-Morales et al. 2016; Pigot et al. 2020). Functional diversity refers to the range of morphological or physiological traits among species, with functional traits being the key attributes that enable organisms to cope successfully with their environment (Hooper et al. 2005; Villéger et al. 2010). Therefore, functional diversity helps us understand how habitats can filter which species are found in a given area (Petchey and Gaston 2006). For insectivorous bats, the functional diversity of echolocation signals can help determine how vegetation structure at small spatial scales may filter species based on their ability to discriminate prey echoes from background clutter or emit signals that optimize detection in open areas (Schnitzler and Kalko 2001).

Echolocation is a sensory system that operates along a continuum, rather than being strictly divided into open-space and clutter-adapted signals. The spectral and temporal characteristics of echolocation calls allow bats to detect, classify and locate prey in different levels of vegetation clutter (Schnitzler and Kalko 2001). In densely vegetated environments, bats require broadband, short-duration pulses to navigate efficiently around obstacles (Broders et al. 2004; Barataud et al. 2013). In contrast, in open areas, bats emit echolocation calls with narrower bandwidths and longer durations, which enhance long-range prey detection (Kalko and Schnitzler 1993). However, some species show a high degree of plasticity, modulating their echolocation parameters to different vegetation structures and foraging conditions (Neuweiler 2003).

Bats can also modify the timing and rate of their echolocation pulses depending on environmental constraints (Fenton et al. 2016). For instance, during critical hunting moments, bats decouple their call emissions from wingbeats to enhance sensory flow, rapidly increasing their call rate when they are close to their prey (Stidsholt et al. 2021). However, this behavior comes at an energetic cost, so when not actively hunting, bats synchronize their calls with wingbeats to optimize energy efficiency (Stidsholt et al. 2021). Similarly, when navigating through highly cluttered environments, bats tend to increase their call emission rates to improve spatial resolution (Falk et al., 2014). Yet, in the presence of conspecifics, they reduce emission rates and even omit pulses to minimize interference, which can ultimately impact their navigation and collision avoidance (Adams et al. 2017).

Vegetation complexity, which serves as a proxy for clutter, can act as a strong selective force, shaping echolocation structures and filtering community composition (Yovel et al. 2009). Greater habitat heterogeneity provides more ecological niches, which promotes species diversity (Langridge et al. 2019). Conversely, a reduction in structural complexity negatively affects bat diversity (Meyer et al. 2016) and poses a threat to several taxonomic groups (Meyer and Kalko 2008; Jones et al. 2009; Estrada-Villegas et al. 2010; Cruz et al. 2016). For example, Díaz-B et al. (2023) demonstrated that environmental conditions in Colombian dry forests significantly influence functional traits of bats, including echolocation pulse structure, diet, vertical foraging behavior and trophic level. However, despite increasing research on the drivers of functional diversity in Neotropical bats, few studies have examined how vegetation complexity modulates the functional diversity of echolocation signals in insectivorous bat ensembles (Cisneros et al. 2015; Núñez et al. 2019). This knowledge gap is particularly relevant in naturally heterogeneous landscapes such as savannas. In the Colombian Llanos, savannas are often flanked by gallery forests, and sometimes there are soft forest edges with sparce tree cover in the transition between gallery forests and savannas (Romero-Ruíz et al. 2004).

To better understand how vegetation structure influences echolocation in Neotropical bats, we assessed the effect of vegetation complexity on the functional diversity of echolocation pulses in an ensemble of insectivorous bats in the savannas of Colombia. Based on Denzinger et al. (2016), we predicted that bats in cluttered habitats would echolocate at higher frequencies and shorter durations, whereas bats in open areas would emit lower frequency, longer duration calls. We also expected the acoustic parameters of the bat ensemble to vary with habitat complexity. Specifically, we predicted that species should modulate their vocalizations to the structural characteristics of their environment.

Previous studies have shown that species such as Saccopteryx bilineata and members of the genus Eptesicus modify their echolocation signals according to vegetation complexity: in cluttered habitats they emit highly modulated frequency pulses, whereas in open areas they reduce modulation (Ratcliffe et al. 2011; Jakobsen et al. 2012). In this study, we first describe the acoustic parameters of the insectivorous bat ensemble and then test our predictions by correlating vegetation clutter with the variability of spectral and temporal parameters of echolocation calls for all species in the ensemble, as well as for Saccopteryx bilineata and Neoeptesicus cf. furinalis separately.

Materials and Methods

Study area and vegetation coverage. We conducted our study in El Caduceo Natural Reserve, municipality of San Martin, department of Meta, Colombia (3.6711° N, -73.6585° W, Elevation: 377.8 MASL). The reserve is located at the piedmont of the Eastern Cordillera and the onset of the Llanos (Basto-González 2009) (Supplementary material, Figure 1). The Colombian Llanos are mainly composed of natural savannas and gallery forests (Lasso et al. 2011). Savannas are mostly covered with native grasses and are mainly used for cattle grazing, and the forest boundaries between gallery forests and savannas have not shifted due to stable land use patterns and the lack of significant encroachment or deforestation (Romero-Ruíz et al. 2004).

In El Caduceo, 70 hectares are covered with gallery forests while the remaining 103 hectares are savannas used for cattle farming and agriculture in small parcels (Casallas-Pabón et al. 2017). The area experiences a monomodal rainfall regime, with a dry season from December to March (Correa et al. 2006). We selected three habitats: open (S), cluttered (BT), and highly cluttered (BM) environments. Open areas (S) consisted mainly of pastures used for cattle grazing and agricultural crops, such as pineapples. Cluttered sites (BT) were in the transitional area between open areas and highly cluttered environments, and were composed of scattered, thin, and tall pioneer trees characteristic of early successional stages, interspersed with open savanna patches. Highly cluttered sites (BM) were located within gallery forests, which exhibited greater structural complexity, including a diverse assemblage of tall trees, palms, dense understory vegetation, as well as numerous saplings.

We measured diameter at breast height (DBH), canopy width and percent cover in 18 plots, six per habitat. We also measured the percentage of vegetation coverage with a concave spherical densiometer (Rivera-Gallego and Pinzón-Florián 2022) in each plot. Vegetation structure data were collected on the same sites where we conducted acoustic monitoring to ensure comparability between environmental variables and acoustic parameters across sites (please see below). Since vegetation structure remains stable over short time periods (months or years), data collection of vegetation variables did not require the same sampling effort as bat monitoring (Willcox et al., 2017).

Bat capture and reference calls. We captured and recorded bats from February 11th to March 9th, 2023. Four 6 × 2.5 m nets were used on each habitat and deployed from 17:30 to 23:30 for 18 nights. We relocated nets every night. Captured bats were placed in cloth bags, measured, and sexed (Supplementary material, Figure S2). We followed the Guidelines of the American Society of Mammalogists for proper handling protocols (Sikes et al. 2011).

We used taxonomic keys to identify all the captured bats (Díaz et al. 2016 ; Díaz et al. 2021). All bats from the family Phyllostomidae bats were released immediately. Although some phyllostomid bats are insectivorous, we did not include any species from this family because their echolocation calls are unidirectional (all other families are omnidirectional) and are not adequately sampled with our sound recorders (Jakobsen et al. 2013). For reference calls, we used a Pettersson M-500 USB microphone, with Pettersson-BatMicRecorder software at a sampling rate of 386 kHz and a resolution of 16 bits. We started recording when the individual was released until it was no longer in sight. The recording methodology involved releasing bats at the site where they were found on the night of capture. Our study is in compliance with Permit No. 2467 (25 October 2023) issued by the Colombian Environmental Licensing Authority for collecting and manipulating wildlife for non-commercial scientific research purposes.

Echolocation calls were recorded and analyzed using Raven Pro 1.6. (Cornell Lab of Ornithology, Cornell Univeristy, USA). Both reference and passive acoustic recordings (see next section) were processed in Raven Pro 1.6. Spectrograms were generated using a Hamming window with a 512-point FFT, 100% frame size, and 80% overlap. A sonotype can be defined as a distinct acoustic category that groups echolocation calls sharing similar structural characteristics, regardless of their signal frequency (Roemer et al. 2021). Based on this definition, we labeled each call and measured the following spectral-temporal parameters: start time (s); end time (s); low frequency (kHz); high frequency (kHz); Δ Time delta (s); 5% frequency (kHz); 95% frequency (kHz); Δ Frequency delta (kHz); peak frequency (kHz). We also labeled each recording with sonotype; taxon; sex; recording type; and quality. The audio quality was evaluated using the signal-to-noise ratio (SNR), which quantifies the quality of a signal in the presence of noise and is expressed in decibels (dB). The selection of these parameters was based on Martínez-Medina et al. (2021).

Passive acoustic sampling. We used nine Audiomoth recorders using the following parameters: 250 kHz sampling rate, medium gain, 1740 seconds of inactive time, and 60 seconds of active time. We sampled from 17:30 to 5:30 daily for three weeks. In total, we had nine recording sites, three in each habitat. Each recorder was placed four meters above ground level in all three habitats and at least 150 meters apart to avoid pseudoreplication (Pryde and Greene 2016). Recording sites were relocated weekly to avoid nights without recordings. Spectral-temporal parameters from search calls were analyzed and labeled in Raven Pro1.6. (Cornell Lab of Ornithology, Cornell University, USA), including start and end times, low and high frequencies, Δ Time delta, bats presence, sonotype, taxon, habitat classification, feeding behavior, and recording quality (Martínez-Medina et al. 2021). Sonotype identifications were informed by previous studies (Jung et al. 2007; Jung et al. 2014; Arias-Aguilar et al. 2018).

Data Analysis. We used a Principal Component Analysis (PCA) to identify patterns in vegetation structure, using the following R libraries: FactoMineR (Lê et al. 2008), factoextra (Kassambara and Mundt 2020), REdaS (Maier 2022) and ggplot2 (Wickham 2016). Then we used canonical correspondence analysis (CCA) to determine whether the spectral and temporal parameters of the insectivorous bat ensemble were modulated given the complexity of the vegetation. To do this, we first averaged the data obtained for each vegetation variable and each spectral and temporal variable per plot across the three habitats (18 plots in total). Then we created two matrices with the same number of rows, one with the vegetation data per plot, and another with the spectral and temporal variables per plot. Then, we used the function “scale” to normalize the data because all variables were in different units and then calculated the CCA using the library Vegan (Oksanen et al. 2022).

To determine differences in spectral and temporal parameters of echolocation signals among habitats, we selected species recorded at least 10 times in each habitat to ensure sample representativeness. Only two species met our criteria: Saccopteryx bilineata (Number of recordings per habitat: BT= 33, BM= 28, S= 49) and Neoeptesicus cf. furinalis (Number of recordings per habitat: BT= 15, BM= 10, S= 49).

We calculated boxplots for the following parameters: Bandwidth (Δ Frequency, kHz), Pulse Duration (Δ Time, s), and Peak Frequency, as these parameters best characterize echolocation pulses (Jung et al. 2007). Subsequently, we used a Kruskal-Wallis test to compare the means of each parameter across habitats, and used a Dunn-Bonferroni test for multiple comparisons (Supplementary material, Table S1, S2). For the aforementioned tests, we used the following R libraries: PMCMRplus (Pohlert 2021), ggplot2 (Wickham 2016), and ggpubr (Kassambara and Mundt 2020), o and the FSA library (Ogle et al. 2025). Given that vegetation variables were not used in these analyzes, our comparisons between habitats for Saccopteryx bilineata and Neoeptesicus cf. furinalis are based on our whole sampling (three weeks).

Results

We found 19 species, being Saccopteryx bilineata and Neoeptesicus cf. furinalis the most relatively abundant across all three habitats (Supplementary Material, Table S3 and Figure S4 to S6). The first PCA axis of vegetation variables explained 38.3% of the data, whereas the second axis explained 19.9%. The variables that explained most of the variance in Axis 1 and 2 were DBH, and Total Height, respectively (Supplementary material, Table S4). The PCA plot indicated that BM and BT are clustered together. Since there were no vegetation variables in habitat S, all plots from S are aggregated into a single point (Supplementary material, Figure S3).

Relationship between vegetation variables and acoustic parameters. The first axis of the CCA indicates that DBH had the strongest influence on the spectral and temporal parameters (-0.67), while percentage cover (COV) was the most influential variable on the second axis (-0.62). Among the spectral-temporal variables, Δ Frequency (kHz) was most influenced by the first axis (-0.29), while Δ Time (s) was mainly influenced by the second axis (0.14) (Figure 1).

Spectral and temporal differences in two species present in the three habitats. We found that the Δ Frequency (kHz) (P= 1.987e-05), Δ Time (ms) (P= 2.2e-16), and peak frequency (kHz) (P= 4.89e-12) of the low pulse of Saccopteryx bilineata (Figure 2) were significantly different among habitats. Multiple comparisons for peak frequency only showed a significant difference between BM-S (P= 0.002). For Δ Frequency (kHz), we observed significant differences between BM-S (P= 0.002) and BT-S (P= 0.004). For Δ Time (ms), we found significant differences between BM-BT (P= 1.32e-04), between BM-S (P= 9.34e-40), and between BT-S (P= 3.23e-13).

Moreover, we also found significant differences among habitats for the high pulse of Saccopteryx bilineata in Δ Frequency (kHz) (P= 0.0001), Δ Time (ms) (P= 1.72e-05), and peak frequency (kHz) (P= 4.89e-12). The multiple comparisons tests only showed a significant difference between BM-S (P= 0.031). For Δ Frequency (kHz), we only found significant difference between BT-S (P= 0.0001). However, for the Δ Time (s), there is a significant difference between BM-BT (P= 1.40e-03), between BM-S (P= 2.3e-34), and between BT-S (P= 1.84e-13). These results indicate that environmental factors significantly influence the echolocation calls of Saccopteryx bilineata (Supplementary material, Figure S7). The significant differences in Δ Frequency and peak frequency between habitats suggest that this species is modifying their vocalizations, probably to optimize foraging efficiency and sound detection in response to the specific characteristics of their environment.

For Neoeptesicus cf. furinalis (Figure 3), there were significant differences among habitats in the Δ Frequency (kHz) (P= 0.004), Δ Time (s) (P= 2.20e-16), and peak frequency (kHz) (P= 1.39e-05). Regarding Δ Frequency (kHz), there was a significant difference among all three habitats: BT-S (P= 2.44e-06), BM-S (P= 5.99e-02), and BM-BT (P= 4.70e-02). For the Δ Time (s), there is a significant difference between BM-S (P= 1.13e-30), between BT-S (P= 4.15e-17), and finally between BT-BM (P= 6.81e-01) (Supplementary material, Figure S8).

Discussion

Our results confirm that bats can modify their echolocation calls according to the vegetation structure of their surroundings. These results suggest that vegetation structure, particularly tree size (i.e., DBH), plays a key role in the variation of bat call frequency, whereas canopy cover affects the temporal aspects of calls, such as pulse duration. Our results indicate that different aspects of habitat complexity influence different acoustic parameters, reflecting how bats modify their echolocation to navigate and forage in environments with varying levels of clutter.

The most variable parameters were Δ Frequency (kHz) and Δ Time (s) for the whole ensemble. For Saccopteryx bilineata, we found rather unusual results that contradicted our predictions (Fenton et al. 2016). Initially, we assumed that in a more cluttered environment, pulses would be shorter and with a wider Δ Frequency delta to better avoid environmental obstacles (Moss et al. 2011). However, our results suggest that S. bilineata does the opposite: we recorded a shift to larger frequencies and a shorter in open spaces compared to cluttered environments. In comparison, and as we expected, Neoeptesicus cf. furinalis had longer duration calls with narrower Δ Frequency delta in open spaces, and short pulses with a higher Δ Frequency delta in highly cluttered areas (Moss et al. 2011).

Bats adjust their spectral and temporal parameters according to vegetation structure. To understand how the temporal and spectral parameters of echolocation pulses vary in bats, it is key to first determine how vegetation structure varies. The cluttered area (BT) is a transitional forest that has been restored for about 15 years. In this area, tall trees predominate, but with a lower DBH and with a smaller canopy width. In comparison, the highly cluttered space (BM) is a mature forest with very little disturbance in the last 50 years, according to landowners. Trees present in the highly cluttered area have different heights and diameters, with an average height of 25 m, and an average DBH higher than BT. Therefore, the highly cluttered space has a thick intertwined canopy whereas the cluttered space did not.

Bats adjusted the spectral and temporal parameters of their echolocation calls according to the complexity of vegetation structure. The type of acoustic environment that bats need to detect, classify and avoid is directly related to structural variables such as tree height, canopy density, and DBH because these parameters determine the number and type of obstacles and need to be sorted out (Schnitzler and Kalko 2001). For instance, Δ Frequency increased in sites with higher DBH, while Δ Time increased in areas with lower percentage cover. These results indicate that echolocation pulses become more modulated in more cluttered environments, supporting our predictions.

Larger trees generally contribute to greater clutter because of their sheer size and because they promote a denser and taller understory (Chazdon et al. 2010). Moreover, pulse interval became longer in less cluttered spaces, also supporting our prediction. Sites with lower percentage of canopy coverage, by definition, are much more open and less obstructed. In these types of conditions, bats are not forced to increase their rate of sound emission to update their perception of their environment (Fenton et al. 2016). We do not exclude other factors that may influence how bats modulate their echolocation signals. For instance, bat size (López‐Cuamatzi et al. 2020) and feeding behavior (Fenton et al. 2016) also play a role in structuring echolocation calls. Yet, our results support our prediction that bats will echolocate at higher frequencies for shorter durations in cluttered areas (Denzinger et al. 2016).

Differences in spectral-temporal parameters in two species across three habitats. Saccopteryx bilineata is known for its vocal learning abilities (Knörnschild et al. 2006) and exhibits complex acoustic social interactions (Davidson and Wilkinson 2004). As shown by Ratcliffe et al. (2011), we observed that S. bilineata can alter its echolocation signals as needed, especially in relation to the type of habitat where it is foraging. For instance, Jakobsen et al. (2012) showed that S. bilineata significantly reduced the intensity and duration of its calls when flying in a flight cage, a proxy for a closed space environment. This aligns with our findings; Δ Time was higher in open environments (S) compared to sites with some obstacles (BT) or to highly cluttered environments (BM). In other words, pulses were longer in open spaces than in cluttered spaces. Our results support the idea that bats reduce the duration of their echolocation calls in cluttered environments. In cluttered habitats, the speed of receiving and processing environmental information is crucial, and this modification in call duration could facilitate such processing (Denzinger et al. 2016).

On the other hand, Δ Frequency and peak frequency showed a surprising behavior. We expected that Δ Frequency should be highest in cluttered spaces as bats increase their frequency modulation so they can effectively differentiate prey from vegetation (Denzinger et al. 2016). However, and contrary to our predictions, we observed that Δ Frequency and peak frequency were significantly higher in open spaces compared to cluttered spaces. Both results contradict existing literature. We think there are three different, not mutually exclusive hypotheses that can help us explain these unforeseen results.

First, call structure could change due to prey size (Aldridge and Rautenbach 1987). It is possible that smaller insects are found in greater numbers in the savanna (open spaces), so increasing both peak frequency and the range of frequency modulation of the echolocation signal would be advantageous and facilitate the detection of smaller prey at closed distances. Further studies assessing prey size across habitats would help us evaluate this hypothesis.

Second, S. bilineata might be increasing the frequency of their calls in open spaces to increase the directionality of their sonar beam. Jakobsen et al. (2012) found that S. billineata emits at higher frequencies to achieve more directional sonar beams, which increases the resolution and level of detail in their acoustic field of view. Seems contradictory to increase frequency and frequency modulation in open spaces because a reduction in call intensity also reduces the chances that a bat will detect an echo from a distant prey (Jakobsen et al. 2012). However, a narrower sound beam produces more intense sounds at the center of the beam, increasing the range at which bats can receive echoes, thus increasing the chances of detecting prey (Jakobsen et al. 2013). It is commonly assumed that bats use more energy when emitting echolocation calls at higher frequencies.

Initial studies by Speakman et al. (1989) showed that bats expend more energy at rest as vocal frequency increases. However, later studies showed the energy demand of echolocation was negligible compared to the high energy demand of flight (Speakman and Racey 1991; Voigt and Lewanzik 2012). Currie et al. (2020) found that increased signal intensity, but not frequency, increased energy expenditure even during flight. Therefore, more detailed studies in echolocation physiology will help us determine whether S. billineata, and maybe other species, increase their frequency range of emission to produce narrower beams in open spaces.

A third option as to why S. bilineata increased its frequency and decreased its pulse duration in open areas could be to avoid acoustic interference. Saccopteryx leptura, Peropteryx macrotis, and Molossops temminckii forage in similar areas as S. bilineata, and there could be an overlap with the echolocation calls of these species and S. bilineata. These three other species produce calls in similar frequency ranges (Supplementary material, Table 5), which could lead to acoustic interference.

Acoustic interference is related to acoustic niche partitioning, where coexisting species adjust their acoustic signals to minimize masking and enhance foraging efficiency (Siemers and Schnitzler 2004; Ey and Fischer 2009; Roemer et al. 2019). However, we believe this hypothesis seems unlikely because the abundance of these three species in the sampled habitats was low, reducing the likelihood of significant acoustic interference with S. bilineata.

Alternatively, the shift in S. bilineata’s call structure might not be solely related to acoustic overlap but could also be a strategy to reduce direct competition with other insectivorous bats in cluttered environments (Beilke et al. 2021). In cluttered habitats, where prey availability and species overlap may be higher, S. bilineata might exploit open spaces to reduce competition with species that hunt similar prey sizes. In these cluttered environments, where dense vegetation makes prey capture more difficult (Rainho et al. 2010), S. bilineata often shows a proportional aggregative response to increased prey abundance, whereas clutter-adapted species may remain unaffected by such variations (Müller et al. 2012). These findings support the hypothesis that S. bilineata may shift to open spaces as a strategy to access prey more efficiently and reduce competitive interactions.

Identifying the specific Molossidae species present in each habitat would help clarify whether this behavioral shift is driven by competition avoidance or by other adaptive factors. In sum, our results suggest that S. bilineata may have greater plasticity in its echolocation structure than previously thought. Further studies on prey size, sonar beam adjustments, and call overlap avoidance would be needed to better understand the potential costs of this plasticity.

Contrary to our results for S. bilineata, the spectral and temporal parameters of N. cf. furinalis, were consistent with theory (Barataud et al. 2013). Δ Frequency was higher in cluttered spaces compared to open spaces, suggesting that this species indeed benefits from having a broader frequency range in cluttered environments to efficiently navigate obstacles. Moreover, N. cf. furinalis showed shorter pulses in cluttered environments and longer pulses with lower frequency modulation in open spaces. The peak frequency for this species did not show significant differences between habitats.

Limitations. This study had two major limitations. First, our sampling period is short. We only sampled for one month and during the peak of the dry season. We do not know whether our results would be different in other times of the year where insect population abundance would be different across habitats, or if tree deciduousness would alter the acoustic space between habitats that bats must sort. It would be advisable that other studies integrate different climatic seasons to get a more complete picture of how vegetation complexity affects the echolocation structure of the insectivorous bat ensemble. The second limitation was the lack of data on insect size present in the different habitats, and the prey size that S. bilineata prefers among habitats. More information on the foraging ecology of this species would help us understand the unusual behavior we found.

Conclusions

Vegetation complexity significantly influences temporal and spectral parameters of bat echolocation calls in an insectivorous bat ensemble in the Llanos of Colombia. The structure of echolocation calls were significantly affected by diameter at breast height (DBH) and percentage of vegetation coverage (COV). We found unexpected call behavior in Saccopteryx bilineata, potentially influenced by various factors such a s prey sizes, sonar beam adjustments, and overlap avoidance, including direct competition with other species of similar size, even in the absence of acoustic overlap. Sorting among these hypotheses would help elucidate the potential costs of echolocation plasticity. Contrary to S. bilineata, the echolocation behavior of N. cf. furinalis followed the expected pattern according to theory. Understanding how bats change their spectral and temporal parameters according to different levels of vegetation complexity can provide insights into how species manage, or fail to manage, naturally open areas, such as the Colombian Llanos, or open areas resulting from anthropogenic activities.

Acknowledgments

We would like to extend our gratitude to everyone who contributed to this process: Jairo Peres-Torres for his support and guidance, as well as Inga Geipel, Adrià Lopez Baucells and Carlos Restrepo for their valuable reviews and feedback.

Literature cited

Adams AM, Davis K, and Smotherman M. 2017. Suppression of emission rates improves sonar performance by flying bats. Scientific Reports 7:41641. https://doi.org/10.1038/srep41641

Aldridge HD, and Rautenbach IL. 1987. Morphology, echolocation and resource partitioning in insectivorous bats. Journal of Animal Ecology 56:763–778. https://doi.org/10.2307/4947

Arias-Aguilar A, Hintze F, Aguiar LMS, Rufray V, Bernard E, and Pereira MJR. 2018. Who’s calling? Acoustic identification of Brazilian bats. Mammal Research 63:231–253. https://doi.org/10.1007/s13364-018-0367-z

Barataud M, Giosa S, Leblanc F, Rufray V, Disca T, Tillon L, et al. 2013. Identification et écologie acoustique des chiroptères de Guyane française. Le Rhinolophe 19:103–145.

Basto-González MA. 2009. Interacciones sociales en un grupo de Callicebus ornatus, ubicado en un fragmento de bosque de galería en San Martín, Meta, Colombia. [Bachelor’s Thesis]. [Bogotá (COL)]: Pontificia Universidad Javeriana.

Beilke EA, Blakey RV, and O’Keefe JM. 2021. Bats partition activity in space and time in a large, heterogeneous landscape. Ecology and Evolution 11:6513–6526. https://doi.org/10.1002/ece3.7504

Broders HG, Findlay CS, and Zheng L. 2004. Effects of clutter on echolocation call structure of Myotis septentrionalis and M. lucifugus. Journal of Mammalogy 85:273–281. https://doi.org/10.1644/BWG-102

Casallas-Pabón D, Calvo-Roa N, and Rojas-Robles R. 2017. Murciélagos dispersores de semillas en gradientes sucesionales de la Orinoquia (San Martín, Meta, Colombia). Acta Biológica Colombiana 22:348–358. http://dx.doi.org/10.15446/abc.v22n3.63561

Chazdon RL, Finegan B, Capers RS, Salgado-Negret B, Casanoves F, Boukili V, and Norden N. 2010. Composition and dynamics of functional groups of trees during tropical forest succession in Northeastern Costa Rica. Biotropica 42:31–40. https://doi.org/10.1111/j.1744-7429.2009.00566.x

Cisneros LM, Fagan ME, and Willig MR. 2015. Effects of human-modified landscapes on taxonomic, functional and phylogenetic dimensions of bat biodiversity. Diversity and Distributions 21:523–533. https://doi.org/10.1111/ddi.12277

Correa HD, Ruíz SL, and Arévalo LM. 2006. Plan de acción en biodiversidad de la cuenca del Orinoco 2005–2015: propuesta técnica. Bogotá (COL): Instituto de Investigación de Recursos Biológicos Alexander von Humboldt.

Cruz J, Sarmento P, Santos H, and Rebelo H. 2016. Bats like vintage: managing exotic eucalypt plantations for bat conservation in a Mediterranean landscape. Animal Conservation 19:53–64. https://doi.org/10.1111/acv.12216

Currie SE, Boonman A, Troxell S, Yovel Y, and Voigt CC. 2020. Echolocation at high intensity imposes metabolic costs on flying bats. Nature Ecology & Evolution 4:1174–1177. https://doi.org/10.1038/s41559-020-1249-8

Davidson SM, and Wilkinson GS. 2004. Function of male song in the greater white-lined bat, Saccopteryx bilineata. Animal Behaviour 67:883–891. https://doi.org/10.1016/j.anbehav.2003.06.016

Denzinger A, Schnitzler HU, Grinnell AD, Popper AN, and Fay RR. 2016. Guild structure and niche differentiation in echolocating bats. In: Fenton MB, Grinnell AD, Popper AN, and Fay RR, editors. Bat Bioacoustics. New York (USA): Springer; p. 141–166.

Díaz MM, Solari S, Aguirre LF, Aguiar LMS, and Barquez RM. 2016. Clave de identificación de los murciélagos de Sudamérica. Publicación Especial No. 2. Tucumán (ARG): PCMA (Programa de Conservación de los Murciélagos de Argentina).

Díaz MM, Solari S, Aguirre LF, Aguiar LMS, and Barquez RM. 2021. Clave de identificación de los murciélagos neotropicales. Tucumán (ARG): PCMA (Programa de Conservación de los Murciélagos de Argentina).

Díaz-B CA, Otálora-Ardila A, Valdés-Cardona MA, López-Arévalo HF, and Montengro OL. 2023. Bat functional traits associated with environmental, landscape, and conservation variables in Neotropical dry forests. Frontiers in Forests and Global Change 6:1082427. https://doi.org/10.3389/ffgc.2023.1082427

Estrada-Villegas S, Meyer CFJ, and Kalko EKV. 2010. Effects of tropical forest fragmentation on aerial insectivorous bats in a land-bridge island system. Biological Conservation 143:597–608. https://doi.org/10.1016/j.biocon.2009.11.009

Ey E, and Fischer J. 2009. The “acoustic adaptation hypothesis”– a review of the evidence from birds, anurans, and mammals. Bioacoustics 19:21–48. https://doi.org/10.1080/09524622.2009.9753613

Falk B, Jakobsen L, Surlykke A, and Moss CF. 2014. Bats coordinate sonar and flight behavior as they forage in open and cluttered environments. Journal of Experimental Biology 217:4356–4364. https://doi.org/10.1242/jeb.114132

Fenton MB, Grinnell AD, Popper AN, and Fay RR, editors. 2016. Bat Bioacoustics. New York (USA): Springer.

García-Morales C, Moreno CE, Badano EI, Zuria I, Galindo-González J, Rojas-Martínez AE, et al. 2016. Deforestation impacts on bat functional diversity in tropical landscapes. PLoS One 11:e0166765. https://doi.org/10.1371/journal.pone.0166765

Hooper DU, Chapin FS, Ewel JJ, Hector A, Inchausti P, Lavorel S, et al. 2005. Effects of biodiversity on ecosystem functioning: a consensus of current knowledge. Ecological Monographs 75:3–35. https://doi.org/10.1890/04-0922

Jakobsen L, Brinkløv S, and Surlykke A. 2013. Intensity and directionality of bat echolocation signals. Frontiers in Physiology 4:89. https://doi.org/10.3389/fphys.2013.00089

Jakobsen L, Kalko EKV, and Surlykke A. 2012. Echolocation beam shape in emballonurid bats, Saccopteryx bilineata and Cormura brevirostris. Behavioral Ecology and Sociobiology 66:1493–1502. https://doi.org/10.1007/s00265-012-1404-6

Jones G, Jacobs DS, Kunz TH, Willig MR, and Racey PA. 2009. Carpe noctem: the importance of bats as bioindicators. Endangered Species Research 8:93–115. https://doi.org/10.3354/esr00182

Jung K, Kalko EKV, and von Helversen O. 2007. Echolocation calls in Central American emballonurid bats: signal design and call frequency alternation. Journal of Zoology 272:125–137. https://doi.org/10.1111/j.1469-7998.2006.00250.x

Jung K, Molinari J, and Kalko EKV. 2014. Driving factors for the evolution of species-specific echolocation call design in New World free-tailed bats (Molossidae). PLoS One 9:e85279. https://doi.org/10.1371/journal.pone.0085279

Kalko EKV, and Schnitzler HU. 1993. Plasticity in echolocation signals of European pipistrelle bats in search flight: implications for habitat use and prey detection. Behavioral Ecology and Sociobiology 33:415–428. https://doi.org/10.1007/BF00170257

Kassambara A, and Mundt F. 2020. factoextra: extract and visualize the results of multivariate data analyses. R package version 1.0.7. Montpellier (FRA): STHDA. Available from: https://CRAN.R-project.org/package=factoextra

Knörnschild M, Behr O, and von Helversen O. 2006. Babbling behavior in the sac-winged bat (Saccopteryx bilineata). Naturwissenschaften 93:451–454. https://doi.org/10.1007/s00114-006-0127-9

Langridge J, Pisanu B, Laguet S, Archaux F, and Tillon L. 2019. The role of complex vegetation structures in determining hawking bat activity in temperate forests. Forest Ecology and Management 448:559–571. https://doi.org/10.1016/j.foreco.2019.04.053

Lasso CA, Rial BA, Matallana CL, Ramírez W, Señaris JC, Díaz-Pulido A, et al., editors. 2011. Biodiversidad de la cuenca del Orinoco II: áreas prioritarias para la conservación y uso sostenible. Bogotá (COL): Instituto de Investigación de Recursos Biológicos Alexander von Humboldt, Ministerio de Ambiente, Vivienda y Desarrollo Territorial, WWF Colombia, Fundación Omacha, Fundación La Salle de Ciencias Naturales e Instituto de Estudios de la Orinoquia (Universidad Nacional de Colombia).

Lê S, Josse J, and Husson F. 2008. FactoMineR: an R package for multivariate analysis. Journal of Statistical Software 25:1–18. https://doi.org/10.18637/jss.v025.i01

López-Cuamatzi IL, Vega-Gutiérrez VH, Cabrera-Campos I, Ruiz-Sanchez E, Ayala-Berdon J, and Saldaña-Vázquez RA. 2020. Does body mass restrict call peak frequency in echolocating bats? Mammal Review 50:304–313. https://doi.org/10.1111/mam.12196

Maier MJ. 2022. REdaS: companion package to the book “R: Einführung durch angewandte Statistik.” R package version 0.9.4. Vienna (AUT): Program distributed by the author. Available from: https://CRAN.R-project.org/package=REdaS

Martínez-Medina D, Sánchez J, Zurc D, Sánchez F, Otálora-Ardila A, Restrepo-Giraldo C, et al. 2021. Estándares para registrar señales de ecolocalización y construir bibliotecas de referencia de murciélagos en Colombia. Biota Colombiana 22:36–56. https://doi.org/10.21068/c2021.v22n01a03

Meyer CFJ, and Kalko EKV. 2008. Assemblage-level responses of phyllostomid bats to tropical forest fragmentation: land-bridge islands as a model system. Journal of Biogeography 35:1711–1726. https://doi.org/10.1111/j.1365-2699.2008.01916.x

Meyer CFJ, Struebig MJ, and Willig MR. 2016. Responses of tropical bats to habitat fragmentation, logging, and deforestation. In Voigt CC, and Kingston T, editors. Bats in the Anthropocene: Conservation of Bats in a Changing World. New York (NY): Springer; p. 63–103. https://doi.org/10.1007/978-3-319-25220-9_4

Moss CF, Chiu C, and Surlykke A. 2011. Adaptive vocal behavior drives perception by echolocation in bats. Current Opinion in Neurobiology 21:645–652. https://doi.org/10.1016/j.conb.2011.05.028

Müller J, Mehr M, Bässler C, Fenton MB, Hothorn T, Pretzsch H, et al. 2012. Aggregative response in bats: prey abundance versus habitat. Oecologia 169:673–684. https://doi.org/10.1007/s00442-011-2247-y

Naeem S, and Wright JP. 2003. Disentangling biodiversity effects on ecosystem functioning: deriving solutions to a seemingly insurmountable problem. Ecology Letters 6:567–579. https://doi.org/10.1046/j.1461-0248.2003.00471.x

Neuweiler G. 2003. Evolutionary aspects of bat echolocation. Journal of Comparative Physiology A 189:245–256. https://doi.org/10.1007/s00359-003-0406-2

Núñez SF, López-Baucells A, Rocha R, Farneda FZ, Bobrowiec PED, Palmerim JM, et al. 2019. Echolocation and stratum preference: key trait correlates of vulnerability of insectivorous bats to tropical forest fragmentation. Frontiers in Ecology and Evolution 7:373. https://doi.org/10.3389/fevo.2019.00373

Ogle DH, Wheeler P, and Dinno A. 2025. FSA v0.10.0: Simple fisheries stock assessment methods. R package version 0.10.0. Available from: https://CRAN.R-project.org/package=FSA.

Oksanen J, Simpson GL, Blanchet FG, Kindt R, Legendre P, Minchin PR, et al. 2022. vegan: community ecology package. R package version 2.6-2. Available from: https://CRAN.R-project.org/package=vegan

Petchey OL, and Gaston KJ. 2006. Functional diversity: back to basics and looking forward. Ecology Letters 9:741–758. https://doi.org/10.1111/j.1461-0248.2006.00924.x

Pigot AL, Sheard C, Miller ET, Bregman TP, Freeman BG, Roll U, et al. 2020. Macroevolutionary convergence connects morphological form to ecological function in birds. Nature Ecology & Evolution 4:230–239. https://doi.org/10.1038/s41559-019-1070-4

Pohlert T. 2021. PMCMR: calculate pairwise multiple comparisons of mean rank sums. R package version 4.0. Available from: https://CRAN.R-project.org/package=PMCMR

Pryde MA, and Greene TC. 2017. Determining the spacing of acoustic call count stations for monitoring a widespread forest owl. New Zealand Journal of Ecology 40:100–107. https://doi.org/10.20417/nzjecol.40.11

Rainho A, Augusto AM, and Palmeirim JM. 2010. Influence of vegetation clutter on the capacity of ground foraging bats to capture prey. Journal of Applied Ecology 47:850–858. https://doi.org/10.1111/j.1365-2664.2010.01820.x

Ratcliffe JM, Jakobsen L, Kalko EKV, and Surlykke A. 2011. Frequency alternation and an offbeat rhythm indicate foraging behavior in the echolocating bat, Saccopteryx bilineata. Journal of Comparative Physiology A 197:413–423. https://doi.org/10.1007/s00359-011-0630-0

Rivera-Gallego ID, and Pinzón-Florián OP. 2022. Fruit-feeding butterflies (Lepidoptera: Nymphalidae) in eucalyptus plantations and gallery forests in the Colombian Orinoquia. Revista de Biología Tropical 70:768–786. https://doi.org/10.15517/rev.biol.trop..v70i1.50860

Roemer C, Julien JF, Ahoudji PP, Chassot JM, Genta M, Colombo R, et al. 2021. An automatic classifier of bat sonotypes around the world. Methods in Ecology and Evolution 12:2432–2444. https://doi.org/10.1111/2041-210X.13721

Roemer C, Coulon A, Disca T, and Bas Y. 2019. Bat sonar and wing morphology predict species vertical niche. Journal of the Acoustical Society of America 145:3242–3251. https://doi.org/10.1121/1.5102166

Romero-Ruiz MH, Armenteros PD, Galondo GG, and Otero GJ, editors. 2004. Ecosistemas de la cuenca del Orinoco colombiano. Bogotá (COL): Instituto de Investigación de Recursos Biológicos Alexander von Humboldt.

Schnitzler HU, and Kalko EKV. 2001. Echolocation by insect-eating bats. BioScience 51:557–569. https://doi.org/10.1641/0006-3568(2001)051[0557:EBIEB]2.0.CO;2

Siemers BM, and Schnitzler HU. 2004. Echolocation signals reflect niche differentiation in five sympatric congeneric bat species. Nature 429:657–661. https://doi.org/10.1038/nature02547

Sikes RS, Gannon WL, and the Animal Care and Use Committee of the American Society of Mammalogists. 2011. Guidelines of the American Society of Mammalogists for the use of wild mammals in research. Journal of Mammalogy 92:235–253. https://doi.org/10.1644/10-MAMM-F-355.1

Speakman JR, and Racey PA. 1991. No cost of echolocation for bats in flight. Nature 350:421–423. https://doi.org/10.1038/350421a0

Speakman JR, Anderson ME, and Racey PA. 1989. The energy cost of echolocation in pipistrelle bats (Pipistrellus pipistrellus). Journal of Comparative Physiology A 165:679–685. https://doi.org/10.1007/BF00610999

Stidsholt L, Johnson M, Goerlitz HG, and Madsen PT. 2021. Wild bats briefly decouple sound production from wingbeats to increase sensory flow during prey captures. iScience 24:102896. https://doi.org/10.1016/j.isci.2021.102896

Tilman D, Knops J, Wedin D, Reich P, Ritchie M, and Siemann E. 1997. The influence of functional diversity and composition on ecosystem processes. Science 277:1300–1302. https://doi.org/10.1126/science.277.5330.1300

Villéger S, Miranda JR, Hernández DF, and Mouillot D. 2010. Contrasting changes in taxonomic vs. functional diversity of tropical fish communities after habitat degradation. Ecological Applications 20:1512–1522. https://doi.org/10.1890/09-1310.1

Voigt CC, and Lewanzik D. 2012. “No cost of echolocation for flying bats” revisited. Journal of Comparative Physiology B 182:831–840. https://doi.org/10.1007/s00360-012-0663-x

Wickham H. 2016. ggplot2: elegant graphics for data analysis. New York (USA): Springer-Verlag. https://doi.org/10.1007/978-3-319-24277-4

Willcox EV, Giuliano WM, Watine LN, Mills DJ, and Andreu MG. 2017. Forest structure and composition affect bats in a tropical evergreen broadleaf forest. Forests 8:317. https://doi.org/10.3390/f8090317

Yovel Y, Falk B, Moss CF, and Ulanovsky N. 2009. What a plant sounds like: the statistics of vegetation echoes as received by echolocating bats. PLoS Computational Biology 5:e1000429. https://doi.org/10.1371/journal.pcbi.1000429

Associated editor: Alonso Panti May

Submitted: June 6, 2025; Reviewed: September 4, 2025

Accepted: September 8, 2025; Published on line: September 30, 2025

THERYA, 2025, Vol. 16(3):397-408

DOI:10.12933/therya-25-6208 ISSN 2007-3364

Figure 1. Canonical correspondence analysis (CCA) showing that spectral and temporal parameters vary according to the vegetation variables. The parameters showing the most significant changes were Δ Frequency (DF) and Δ Time (DT). The eigenvalues for the first and second axes were 0.3790 and 0.0474, respectively. Spectral and temporal parameters in red. LF: Low frequency, High frequency: HF, Δ Time: DT, Frequency 5%: F5, Frequency 95%: F95, Δ Frequency: DF, Peak frequency: FP. Vegetation variables in blue. Total height: TH, Stem height: SH, Canopy width: CW, DBH, % Coverage: COV.

Figure 2. Spectral and temporal parameters of the two Saccopteryx bilineata pulses vary between the different sites, Open (S), Clutter (BT) and High Clutter (BD), as indicated by the lack of overlap in the notches of the box plots. a. Δ Frequency of the low pulse b. Δ Time of the low pulse c. Peak frequency of the low pulse d. Δ Frequency of the high pulse e. Δ Time of the high pulse f. Peak frequency of the high pulse.

Figure 3. Spectral and temporal parameters of the Neoeptesicus cf. furinalis pulses vary between the different sites, Open (S), Clutter (BT) and High Clutter (BD), as indicated by the lack of overlap in the notches of the box plots. a. Δ Frequency (kHz) b. Δ Time (s) c. Peak frequency (kHz).

Supplementary material

This article contains Supplementary Material or Data Files, which can be downloaded from the journal website.