Acoustic bird community composition (but not richness) responds to natural coverages in a tropical agro-cultural landscape

Acoustic bird community composition (but not richness) responds to natural coverages in a tropical agro-cultural landscape | Authorea try { document.documentElement.classList.add('js'); } catch (e) { } var _gaq = _gaq || []; _gaq.push(['_setAccount', 'G-8VDV14Y67G']); _gaq.push(['_trackPageview']); (function() { var ga = document.createElement('script'); ga.type = 'text/javascript'; ga.async = true; ga.src = ('https:' == document.location.protocol ? 'https://ssl' : 'http://www') + '.google-analytics.com/ga.js'; var s = document.getElementsByTagName('script')[0]; s.parentNode.insertBefore(ga, s); })(); Skip to main content Preprints Collections Wiley Open Research IET Open Research Ecological Society of Japan All Collections About About Authorea FAQs Contact Us Quick Search anywhere Search for preprint articles, keywords, etc. Search Search ADVANCED SEARCH SCROLL This is a preprint and has not been peer reviewed. Data may be preliminary. 21 November 2025 V1 Latest version Share on Acoustic bird community composition (but not richness) responds to natural coverages in a tropical agro-cultural landscape Authors : Angela M. Mendoza-Henao 0000-0002-6752-9238 [email protected] , Juliana Tamayo-Quintero , Leonardo Fabio Rivera-Pedroza , Diego Fernando Rosero-Portilla , Hoover Pantoja-Sánchez , Daniela Martínez-Medina , Eliana Barona-Cortés , Juan Sebastián Ulloa , and Carolina Camargo-Gil Authors Info & Affiliations https://doi.org/10.22541/au.176369414.44388497/v1 261 views 128 downloads Contents Abstract Supplementary Material Information & Authors Metrics & Citations View Options References Figures Tables Media Share Abstract The rapid loss of tropical dry forests to agriculture requires effective restoration and monitoring strategies, such as ecological corridors, and robust methods to monitor their success. Birds are excellent bioindicators, yet traditional Conventional Bird Surveys (CBS) can be biased. The emergence of Passive Acoustic Monitoring (PAM) offers a complementary approach. In this study, we evaluated the efficacy of integrating CBS and PAM for assessing avian biodiversity and its relationship with habitat structure in a fragmented tropical dry forest landscape of southwestern Colombia, dominated by sugarcane. We surveyed bird communities along a riverine corridor using both methods and quantified land cover to assess the influence of closed-canopy vegetation. Our results demonstrate that bird community composition was strongly influenced by habitat, with forested sites supporting a significantly higher proportion of closed-habitat species. This relationship was more precisely revealed by PAM, which proved particularly effective at detecting forest-associated birds. While PAM detected a broader range of vocal species, CBS was crucial for identifying silent or less vocal taxa, including certain raptors and waterbirds. The integrated approach recorded 144 species, with each method contributing unique detections, confirming their complementarity. We found that community composition, specifically the proportion of forest-specialist species, was a more sensitive indicator of habitat recovery than species richness alone. We conclude that combining PAM and CBS provides a comprehensive and robust framework for biodiversity assessment in agricultural landscapes. This integrated methodology mitigates detection biases, enhances inventory accuracy, and yields reliable indicators for evaluating the ecological integrity and functional connectivity of riverine corridors. Acoustic bird community composition (but not richness) responds to natural coverages in a tropical agro-cultural landscape Abstract - The rapid loss of tropical dry forests to agriculture requires effective restoration and monitoring strategies, such as ecological corridors, and robust methods to monitor their success. Birds are excellent bioindicators, yet traditional Conventional Bird Surveys (CBS) can be biased. The emergence of Passive Acoustic Monitoring (PAM) offers a complementary approach. In this study, we evaluated the efficacy of integrating CBS and PAM for assessing avian biodiversity and its relationship with habitat structure in a fragmented tropical dry forest landscape of southwestern Colombia, dominated by sugarcane. We surveyed bird communities along a riverine corridor using both methods and quantified land cover to assess the influence of closed-canopy vegetation. Our results demonstrate that bird community composition was strongly influenced by habitat, with forested sites supporting a significantly higher proportion of closed-habitat species. This relationship was more precisely revealed by PAM, which proved particularly effective at detecting forest-associated birds. While PAM detected a broader range of vocal species, CBS was crucial for identifying silent or less vocal taxa, including certain raptors and waterbirds. The integrated approach recorded 144 species, with each method contributing unique detections, confirming their complementarity. We found that community composition, specifically the proportion of forest-specialist species, was a more sensitive indicator of habitat recovery than species richness alone. We conclude that combining PAM and CBS provides a comprehensive and robust framework for biodiversity assessment in agricultural landscapes. This integrated methodology mitigates detection biases, enhances inventory accuracy, and yields reliable indicators for evaluating the ecological integrity and functional connectivity of riverine corridors. Keywords : Aves, ecoacoustic, riverine corridors, tropical dry forest, agroecosystems Introduction Intensive land-use change has caused severe habitat fragmentation, biodiversity declines, and disruption of ecological processes (Santos and Telleria 2006, Quintero-Vallejo et al. 2017). Such impacts promote an accelerated loss of biodiversity and ecosystems, especially in tropical regions. This scenario emphasises the need of implementing restoration initiatives within agricultural landscapes, in which ecological corridors, including streamside forests, can reconnect isolated habitat patches, enhance connectivity, and support species movement and ecosystem functioning (Araujo Calçada et al. 2013). This need is particularly in tropical dry forests which are critically impacted by land transformation. In Colombia, for instance, less than 4 % of the original cover remains (García et al. 2014, Espejo and Morales 2019, Tamayo-Quintero et al. 2023) and most of the original cover has been largely converted to agricultural lands. As protected areas alone are insufficient, integrated restoration strategies that establish functional corridors and promote sustainable land use are vital to conserving biodiversity and ecosystem processes in fragmented landscapes (Xun et al. 2017). Bird communities are widely recognized as valuable indicators for assessing the success of ecological restoration, due to their sensitivity to environmental changes and their critical roles in ecosystem functioning. Their responses to variations in habitat quality, structural complexity, and resource availability make them effective proxies for biodiversity and overall ecosystem health (Kułaga and Budka 2019). Their presence and community composition often reflect the progression of vegetation recovery, the re-establishment of trophic interactions, and the functionality of restored habitats. Additionally, birds are relatively easy to monitor compared to other taxa. These traits position bird communities as practical bioindicators for evaluating restoration and guiding management (Ramírez-Soto et al. 2018, Elliot et al. 2022). Although historically, bird monitoring has been conducted through Conventional Bird Surveys (CBS), the rise of autonomous recording technology and machine learning propel forward the development of bird-based indicators of restoration success (Darras et al. 2019). Given their diverse vocalizations and widespread presence across habitats, birds serve as an excellent alternative for passive acoustic monitoring (hereafter PAM), particularly for reducing observer bias or complementing traditional survey methods (Celis-Murillo et al. 2009, Kulaga and Budka 2019). Passive acoustic monitoring is increasingly used in tropical transformed landscapes to assess biodiversity responses and soundscape homogenization linked to agricultural intensification (Mendoza-Henao et al. 2025). The use of audio recordings enhances species detection, identification, and documentation, as recordings can be replayed, analyzed by multiple interpreters, cross-validated with published references or expert opinions, and examined through sonograms (Celis-Murillo et al. 2009). Data from autonomous recording units can detect species presence and estimate vocal activity (Navine et al. 2024). Conventional bird surveys (CBS), on the other hand, remain an indispensable monitoring strategy, as visual and aural point counts provide direct confirmation of species identity, capture non-vocal or infrequently calling taxa, and supply valuable contextual information on behavior and habitat use that acoustic data alone cannot fully resolve (Suárez-García, 2017). The integration of CBS and PAM mitigates method-specific detection biases, strengthens inferences on species occupancy and habitat associations, and yields a comprehensive assessment of avian community structure in restoration landscapes. The implementation of PAM presents several challenges, particularly in tropical forests. For instance, not all bird species are vocal, and those that are may vocalize irregularly, leading to gaps in detection. Furthermore, many machine learning models for species identification are still under development, especially for the diverse and understudied avifauna of tropical regions. This limitation affects the accuracy and efficiency of automated species recognition, requiring significant effort for manual verification. Despite these challenges, PAM represents an opportunity for enhancing the development of bird-based indicators for restoration in agricultural landscapes. By continuously capturing soundscapes, PAM facilitates the detection of cryptic or elusive species, provides valuable temporal data on bird activity patterns, and offers insights into the reassembly of bird communities during ecological recovery. For instance, closed patches dominated by trees and bamboo may attract understory and forest-preferent birds, leading to differences in acoustic community composition along corridors. In this context, the agro-cultural landscape of the Cauca River valley in southwestern Colombia provides an ideal setting to test the integration of PAM and CBS for assessing the success of ecological corridor implementation. This region has experienced significant land transformation, primarily due to agricultural expansion and urban development, leading to extensive loss of tropical dry forests and consequent declines in biodiversity (Arcila-Cardona et al. 2012; Alvarado-Solano and Otero Ospina 2015). More than 70% of the valley of the Cauca River area is dedicated to extensive sugarcane monoculture (Banda-R et al. 2016). In this sense, the Caña Biodiversa initiative emerged with the aim of performing the restoration of riverine corridors in the Amaime-Nima River Basin, and to do so, a baseline measurement of biodiversity allows the monitoring of conservation efforts. Here, we address two questions 1) Does the bird species’ richness and the bird composition vary along the riverine corridor in relation to the closed habitat availability? and 2) Does passive acoustic monitoring and conventional bird surveys perform well on detecting the likely relation between bird diversity and habitat availability? Materials and Methods The Amaime-Nima river basin is located in the Cauca Valley River between Western and Central Andes, in southwestern Colombia. It spans 1042.9 ha, and the Amaime river extends 86.14 km from the western slope of the Central Cordillera to its confluence with the Cauca river (Figure 1), between the municipalities of Palmira and El Cerrito (Valle del Cauca). Pastures and sugarcane crops dominate some areas along the river while in other areas, the river is dominated by giant bamboo ( Guadua angustifolia ) and fragments with some arboreal coverage (including fruit trees) between the riverbed and the sugarcane crops. According to Holdridge (1967), the vegetation type corresponds to tropical dry forest of which only 1% is conserved in small, highly isolated forest fragments (Arcila-Cardona et al. 2012, Alvarado-Solano and Otero Ospina 2015). The basin ranges from 900 to 1200 m a.s.l., mean annual temperature > 24 °C and precipitation between 1000 and 2000 mm (Arcila-Cardona et al. 2012). Rainfall follows a bimodal pattern, with two dry (January-February and July-August) and two rainy (April-May and October-November) seasons (Armbrecht and Ulloa-Chacón 1999). Habitat estimation To delineate the study area, a 700-meter buffer was applied to the drainage networks of both rivers. Data collection in the study area involved gathering and preparing the necessary data for generating the updated 2023–2024 land cover layer. The collected data included 1) a mosaic of SkySat satellite images (2023–2024) with radiometric, geometric, and atmospheric corrections, 2) Supplementary raster data (geoservices and supporting imagery), 3) Vector data (CLC N3 2020 land cover data from CVC) and 4) field data collected by the Cenicaña technical team. In addition, three areas corresponding to 0.03 km 2 each were added to the 700 m buffer, located 1.8 km from it, specifically on the Hatico private Natural Reserve (Calle et al. 2022). The land cover layer used for the three 100-metre buffers located in El Hatico corresponds to the CVC 2020 layer, as the SkySat 2023-2024 satellite image did not reach these points. Land cover polygons (Levels 1, 2, and 3) were adjusted using the SkySat 2023–2024 orthophoto as the primary input in a Geographic Information System (GIS). Areas obscured by clouds or shadows were supplemented with secondary data, including geoservices and CLC N3 2020 vector data from CVC. Given the high resolution of the primary imagery (50 cm × 50 cm), the digitization was performed at a 1:1000 scale to capture fine details, particularly for forest and riparian gallery polygons. Topological adjustments were subsequently applied to the digitized layer. Following digitization refinement, visual interpretation was conducted to assign CLC classes (N1, N2, and N3) to each polygon. This process utilized the primary orthophoto, supporting imagery, and secondary vector data. Validation was performed by integrating field observations collected by us and drone surveys in targeted areas. The output was a Level 3 CLC land cover layer for the Amaime and Nima River ecological connectivity corridor for a final area of 64.16 km 2 . Based on this result, we selected 15 sites along the Amaime river corridor, plus one recorder in a reference forest (14 ha) in the Hatico private Natural Reserve. Finally, we generated a radius of 100 meters from the location of each sampling point and we calculated the proportion of the layers corresponding to closed habitats (dense forest, open forest, fragmented forest, bamboo forest) per point. Passive Acoustic Sampling To gather acoustic information, Audiomoth v1.2 recorders were strapped to tree trunks or shrubs, within open grasslands, bamboo groves, and areas with some fruit and native trees. The recorders were placed at least 1.5 meter above the ground, and were set to record one minute every 30 minutes (48 recordings per day) at 192 kHz and 16-bit resolution following the recommendations of Avila-García et al. (2024). The devices were programmed to record continuously for at least 12 consecutive days between May and June 2024. We conducted an exploratory analysis on randomly selected audio files to verify the proper functioning of the recorders, validate the accuracy of the configuration and assess the occurrence of the major sound sources (Ulloa et al. 2021). Random recordings were explored per sampling unit during peak acoustic activity (05−08h and 17 −20h). To obtain the bird species identity of the recordings, we employ the BirdNET algorithm (Kahl et al. 2021) and, we extract those records with a confidence value above 0.6. The initial return was filtered against the regional species pool, based on the lists of birds of eBird, the polygons available from Vélez et al. (2021), and recent local bird surveys. This initial filter was followed by a manual verification of the species correspondence in a subset of up to random 80 recordings per species. Conventional bird survey (CBS) or point count surveys Polygons (cells) of 1 km 2 were selected with reference to the sites of acoustic recorders. Within each cell, point counts were randomly established at least 200 m from each other, and ensured to cover each of the ecosystems present in the cell (Ruiz Gutiérrez et al. 2020). At each point, all birds seen or heard for a period of 10 min were recorded without a fixed observation radius. Each cell included 10 points, though numbers varied with accessibility. Surveys were carried out between October and December 2024, aiming to capture local diversity and account for bird mobility. A total of eight cells were evaluated, with two outings per cell. In each outing, two observers recorded birds, complemented by two additional visits for recognition and relocation of points. Equipment included 10x42 mm binoculars (Nikon, Vortex), a Sony ICD-UX543F recorder, a field notebook and a TideWe HR-F700 laser rangefinder to measure distances. Acoustic recordings complemented the visual records, aided by the Merlin Bird ID app (Cornell Lab), and after fieldwork identifications were confirmed by experts with BirdNet, Xeno-canto, and eBird platforms. Richness and composition analysis Statistical approaches based on replicated incidence data can estimate diversity as effectively as those that rely on abundance data (Chao and Colwell, 2017; Colwell et al., 2012). In our incidence-based model, each 1-min audio file (for PAM) and each 10-min survey (for CBS) is regarded as a ‘sampling unit’; species is ‘detected’ by the activity and vocal presence. For each study site, we calculated the observed diversity for all Hill numbers q = 0, 1, 2 (representing infrequent, frequent and highly frequent species or, respectively, species richness, Shannon diversity, Simpson diversity) with the iNEXT function from the iNEXT package (Chao et al. 2021). In addition, we calculated estimated diversity for a standardized sample coverage of 90%. For each species, we obtained the main habitat information from the AVONET data source (Tobias et al. 2022). We split the species in two classes, 1) Open, including species whose habitats are ”Wetland”, ”Human Modified”, ”Grassland”, and ”Rock”; and 2) Closed, including species whose habitats are “Forest”, “Woodland”, “Shrubland” and “Riverine”. For species composition by habitat type per site, we obtained the proportion of records (for both PAM and CBS) of open- and closed-habitat species to gather a single value per site ranging from 0 to 1, being a value of 0 for sites with species entirely of open areas and a value of 1 for sites with species entirely of closed areas. Considering that both variables are proportions (0 to 1), we used beta regression models to test the relationship between proportion of closed-land coverage and proportion of closed-habitat bird records, linear and logarithmic models were tested and the r-squared was compared between them. Habitat estimation Analysis of land cover in the 700-metre buffer zone shows a clear predominance of permanent herbaceous crops, which occupy an area of 4,794.6 hectares, equivalent to 74.7% of the total area. In contrast, clean pastures cover an area of 190.8 ha, corresponding to 3% of the total, followed by gallery and riparian forests, with 228.4 ha (3.6%), and shrublands, which cover 241 ha (3.8%). In terms of natural vegetation, fragmented forest covers 109 ha (1.7%) and open forest 74.5 ha (1.2%), while dense forest has a reduced presence of 14.9 ha (0.23%). Secondary or transitional vegetation covers 19.1 ha (0.3%), evidence of ecological succession processes. Likewise, water bodies and wetlands comprise significant areas: rivers cover 96.7 ha (1.5%), canals cover 17.7 ha (0.3%), and artificial water bodies cover 27.2 ha (0.4%). Swampy areas are restricted to 1.4 ha (0.02%). The spatial distribution of forest cover and natural vegetation within the 700-meter buffer zone is primarily concentrated along the Amaime and Nima rivers, forming parallel corridors on both banks. This pattern highlights the strong association between riparian zones and remaining natural habitats. In contrast, the landscape is largely dominated by permanent herbaceous crops, while natural and urban ecosystems represent only a minor fraction of the total area. Composition of coverages on the 100 m buffer of all sampling points was heterogeneous: the percentage of closed coverage varied between 0.01 in the most open herbaceous sampling points to 1.00 in the Hatico reference forest (mean = 0.46 for PAM sites and 0.64 for CBS sites). Bird composition Between 671 and 817 good-quality recordings were obtained for 17 sampling points (Supplementary Figure 2). Recording quality of the remaining sites was maintained consistently throughout the survey event, with no consequent effect on the detectability of acoustic activity. The initial output of BirdNet returned 549 species. After manual depuration, the final dataset comprises 4176 records for 112 species (Figure 2, Supplementary material 1). A total of 3015 records were obtained of Passeriformes, followed by Psittaciformes (299) and Strigiformes (188). The species with the most records was Tolmomyias sulphurescens (399 records) followed by Vireo olivaceus (226) and Troglodytes aedon (216). Two registered species Ortalis columbiana and Picumnus granadensis are endemic to Colombia. The conventional bird surveys detected 123 species along the corridor, identified by both visual and acoustic methods. 44 species were detected only by PAM while 55 species were unique for CBS. The number of species detected in each site varied between 9 and 43 species. The species with the most records were Pionus menstruus (37 records), Tyrannus melancholicus (30 records), Pitangus sulphuratus (28 records), and Daptrius chimachima (27 records). A total of 498 records were of Passeriformes followed by Columbiformes (68 records) and Psittaciformes (60 records). The preferred habitats for all species recorded were Forests (57 species) and Shrublands (21 species) while Riverine and Rock had the fewest species with only one species each. The mean proportion of closed habitat species was lower by CBS (0.65) than for PAM (0.72). The Welch t-test revealed a statistically significant difference between the two groups ( t (37.2) = -2.36, p = 0.02, n = 52). The effect size was medium-to-large (Hedges’ g = -0.67), confirming that the CBS method yielded systematically lower proportions of closed-habitat species. This suggests that conventional bird surveys might be biased towards open-habitat species, which are of easier detection (Figure 3). The beta regression model indicates a statistically significant relationship between the log-transformed proportion of closed landscape and the Proportion of closed-habitat acoustic recording birds (Supplementary Table 1, Figure 4). The model revealed a statistically significant positive coefficient for the proportion of closed-habitat species (Estimate = 0.27904, SE = 0.07455, z = 3.743, p <0.001). The precision parameter (phi) was estimated to be 31.079 (SE = 9.957, z = 3.121, p < 0.005), indicating a good model fit. The model’s log-likelihood was 21.55 on 3 degrees of freedom, the pseudo R-squared indicates that approximately 41.78% of the variability in the response was explained by the model. Discussion Our results show that bird species composition was strongly influenced by the availability of closed-canopy vegetation, with forested sites consistently supporting a greater proportion of closed-habitat species. This pattern highlights the role of closed habitats as critical reservoirs of biodiversity within fragmented agricultural landscapes. We also found that Conventional Bird Surveys (CBS) and Passive Acoustic Monitoring (PAM) provided complementary insights: while PAM detected a broad range of vocal species, CBS was effective for visually identifying silent or less vocal taxa. Integrating both methods reduced detection biases and improved the accuracy of species inventories and community metrics. Together, our findings show that combining CBS and PAM strengthens biodiversity assessments and provides relevant information to develop indicators to evaluate the use of those corridors by birds assemblages, enhancing connectivity and ecological integrity. Complementarity of Conventional Bird Surveys and Passive Acoustic Monitoring Depending on the specific conservation or management objective, PAM can serve as a standalone or complementary approach. In the cases where the goal is to determine whether restoration or conservation strategies successfully support habitat specialists such as forest-interior species, PAM provided sufficient and precise evidence (Figure 4). Conversely, broader biodiversity inventories or behavioral studies still benefit from integrating PAM with conventional survey methods. In our study area, both methodologies found that a predominance of the order Passeriformes in the surveys aligns with expectations for Neotropical assemblages, where this group is typically the most diverse and abundant. Additionally, the families Tyrannidae, Vireonidae and Troglodytidae accounted for the highest number of vocal records, as species within these families are diverse and abundant across environments with varying degrees of disturbance (Tamayo-Quintero and Cruz-Bernate 2015). In total, 144 species were recorded using both methodologies. However, visual monitoring allowed the detection of certain taxa that are not typically highly vocal, unlike species such as Contopus spp ., Tolmomyias sulphurescens , Empidonax virescens , Legatus leucophaius , Myiopagis viridicata and some furnarids such as Xiphorhynchus susurrans that can be identified more easily only when they vocalize and, in addition, they tend to vocalize frequently. Comparing our findings with other surveys in the region, the number of species detected was similar to former conventional samplings in the Hatico (109 species by Hurtado-Giraldo et al. 2016; 135 species by Cardenas, 1998), but slightly higher to the 98 species of a recent census on natural vegetation strips along sugarcane crops (Rivera-Pedroza et al. 2019). These findings suggest that, for a more comprehensive assessment into avian biodiversity, CBS and PAM provide complementary insights and should be used in combination for a complete bird survey. The differences observed between both methodologies coincide with results of Celis-Murillo et al. (2012) and Budka et al. (2022). In the northern Yucatan Peninsula were found similar species richness but different composition between methods, while no significant differences in the number of meadow and farmland bird species were detected by recorders and observers within a 100 m radius in Poland. Differences between humans and automatic species detection depends on habitat types: observers recorded more species on farms but not in forests (Kułaga and Budka 2019). The main limitation of using PAM alone for a comprehensive species inventory is that not all species can be effectively detected. Some species, such as storks, herons, and birds of prey, are often silent or produce vocalizations too low to be recorded, making them more reliably detected through visual surveys. Additionally, species that vocalize primarily with call notes are difficult to identify and are often underrepresented due to misidentification or classification as unknowns. Fortunately, the ability to replay recordings and verify sonograms can improve species identification accuracy. Conversely, point counts provide direct confirmation of species identity, capture non-vocal or infrequently calling taxa, and offer valuable habitat and behavioral context that acoustic data may overlook (Suárez-García 2017, Klingbeil and Willig 2015). However, a meta-analysis found that recorders record a 11% higher species richness than point counts Darras et al. (2019). The performance of CBS is better in open-habitat where birds are more conspicuous (Fontúrbel et al. 2020) and are influenced by observer bias, often favoring large-bodied and conspicuous species, while dense vegetation can hinder the detection of smaller or more cryptic species (Anderson et al. 2015). Here, our results show that the proportion of forest-associated species detected was higher in PAM, indicating that conventional methods may underestimate forest specialists due to detectability biases favoring open-habitat species. Although both methods can yield similar estimates of richness and composition overall, their relative performance depends on survey strategy, habitat context, and the balance between vocally active and cryptic taxa. This implies that a single method may underestimate or mischaracterize certain components of the avian community (Klingbeil and Willig 2015), while combining auditory and visual cues in the field can enhance identification accuracy especially for rare species (Celis-Murillo et al. 2009). Together, a complementary use integrates the strengths of both conventional methods and mitigates their intrinsic biases, enhancing inventory completeness, and improving the reliability of community metrics. When the goal is a comprehensive assessment of bird assemblages the combined use of PAM and CBS provides a more robust framework for monitoring, restoration evaluation, and biodiversity management. This integration mitigates method-specific detection biases and enhances the completeness of species inventories. Composition as an Indicator of Habitat Recovery Several studies emphasize total species richness as an indicator of habitat recovery. For instance, forest remnants in the Yucatán Peninsula harbored higher species richness than modified environments (Ramírez-Albores et al. 2018), and Costa Rican coffee plantations supported more bird species with increased tree cover (Şekercioğlu et al. 2019). However, our results indicate that community composition, specifically the proportion of forest-dependent species, provides a more sensitive measure of ecological integrity and reflects a site’s ecological condition better than species richness. This pattern is consistent across ecosystems: in Colombian sugarcane landscapes, species lists were similar between monocultures and vegetation strips, yet the richness and composition of functional guilds differed markedly (Rivera-Pedroza et al. 2019), while studies in the Indonesian rainforest margin (Maas et al. 2009), Western Kenyan farmlands (Mulwa et al. 2012), and the Ecuadorian Amazon (Bare and Danner 2017) also reported stable species richness but strong compositional shifts as forest specialists declined and generalists increased. Attributes such as community composition and functional structure are therefore more sensitive indicators of ecosystem integrity than richness (Socolar et al. 2016). These results emphasize that biodiversity monitoring in human-transformed landscapes should prioritize representativeness and functional composition over raw species counts (Anderle et al. 2024). Bird composition based on standard habitat preferences, and revealed by PAM surveys, provides a robust metric for early conservation initiatives on highly transformed landscapes and this information can be easily obtained and interpreted for decision-makers. Moreover, the positive association between the extent of closed landscapes and the proportion of forest species detected acoustically highlights the sensitivity of PAM to structural habitat complexity. Representativeness might capture shifts in functional or habitat specialist groups that may remain hidden under other approaches. For instance, a possible increase of forest-preferring species within closed-land sites in the future can indicate habitat recovery even when total species counts remain stable (Darras et al. 2019; Anderle et al. 2024). An essential factor to consider in this scenario is that many bird species of the regional pool are generalist and exhibit high dispersal capacity, meaning that along the riverine corridor, the species detected at each sampling point represent not an assemblage, but a continuum of populations using the entire system. Although these narrow systems sometimes cover a small area, they still provide habitat for diverse plant and bird communities and function as local corridors (Pulido-Santacruz and Rengifo-Ruiz, 2011). Forest birds, both habitat specialists and generalists, may find suitable conditions within restored corridors facilitating movement and even providing habitat to carry out critical life cycle activities (Montealegre-Talero et al. 2017). Our results provide evidence for this pattern. Through PAM, we detected a greater number of forest species in closed-land points, which highlights the relevance of maintaining and restoring connectivity within agricultural landscapes to support avian biodiversity. Our findings support the robustness of acoustic monitoring as a reliable, observer bias-free tool for assessing bird assemblages and evaluating restoration outcomes, particularly in small habitat patches where it effectively reflects ecosystem condition. While some species avoid farmland, many birds provide essential ecosystemic services in these farmlands like pest control, pollination, and seed dispersal (Şekercioğlu et al. 2019, Rivera-Pedroza et al. 2019); in some cases, even used as a method of biological control of field rats (Díaz-Siefer et al. 2021). Therefore, the most effective strategy for conserving avian biodiversity and its associated benefits in agricultural landscapes is to preserve and restore connectivity (Bhakti et al. 2018). For maintaining long-term connectivity at both large forest fragments and smaller, strategically located patches and corridors, a robust monitoring xof the entire bird community, from forest interior specialists to wide-ranging generalists is crucial. References Anderle, M., Brambilla, M., Angelini, L., Guariento, E., Paniccia, C., Plunger, J., … & Hilpold, A. (2024). Efficiency of birds as bioindicators for other taxa in mountain farmlands. Ecological Indicators, 158, 111569. Anderson, A. S., Marques, T. A., Shoo, L. P., & Williams, S. E. (2015). Detectability in audio-visual surveys of tropical rainforest birds: the influence of species, weather and habitat characteristics. PloS one, 10(6), e0128464. Alvarado-Solano, D. P., & Otero Ospina, J. T. (2015). Distribución espacial del bosque seco tropical en el Valle del Cauca, Colombia. Acta biológica colombiana, 20(3), 141-153. Araujo Calçada, E., Closset‐Kopp, D., Gallet‐Moron, E., Lenoir, J., Rêve, M., Hermy, M., & Decocq, G. (2013). Streams are efficient corridors for plant species in forest metacommunities. Journal of Applied Ecology, 50(5), 1152-1160. Arcila Cardona, A. M., Valderrama Ardila, C., & Chacón de Ulloa, P. (2012). Estado de fragmentación del bosque seco de la cuenca alta del río Cauca, Colombia. Biota Colombiana, 13(2). Armbrecht I, & Ulloa-Chacón P. 1999. Rareza y diversidad de hormigas en fragmentos de bosque seco colombianos y sus matrices. Biotropica.;31(4):646-653. Doi: 10.1111/j.1744-7429.1999.tb00413.x Ávila-García, A. M., Martínez-Medina, D., Buitrago-Cardona, A., Ulloa, J. S., Mejía-Cepeda, N., Barona-Cortés, E., Mendoza-Henao, A. y Pantoja Sánchez, H. (2024). Protocolo de monitoreo, procesamiento y etiquetado manual para grabaciones pasivas. Instituto Humboldt. Banda-R, K., A. Delgado-Salinas, K.G. Dexter, R. Linares-Palomino, A. Oliveira-Filho, D. Prado, M. Pullan, C. Quintana, R. Riina, G.M. Rodríguez M, J. Weintritt, P. Acevedo-Rodríguez, J. Adarve, E. Álvarez, A. Aranguren B, J. Camilo, G. Aymard, A. Castaño, N. Ceballos-Mago, Á. Cogollo, H. Cuadros, F. Delgado, W. Devia, H. Dueñas, L. Fajardo, Á. Fernández, M.A. Fernández, J. Franklin, E.H. Freid, L.A. Galetti, R. Gonto, R. González-M, R. Graveson, E.H. Helmer, Á. Idárraga, R. López, H. Marcano-Vega, O.G. Martínez, H.M. Maturo, M. McDonald, K. McLaren, O. Melo, F. Mijares, V. Mogni, D. Molina, N.P. Moreno, J.M. Nassar, D.M. Neves, L.J. Oakley, M. Oatham, A.R. Olivera-Luna, F.F. Pezzini, O.J. Reyes, M.E. Ríos, O. Rivera, N. Rodríguez, A. Rojas, T. Särkinen, R. Sánchez, M. Smith, C. Vargas, B. Villanueva and R.T. Pennington. 2016. Plant diversity patterns in neotropical dry forests and their conservation implications. Science, 353: 1383-1387 Bare, M. C., & Danner, R. M. (2017). Bird diversity in actively and naturally restored tropical forests in an urban-agricultural landscape. Ecological Restoration, 35(2), 102-111. Budka, M., Jobda, M., Szałański, P., & Piórkowski, H. (2022). Acoustic approach as an alternative to human-based survey in bird biodiversity monitoring in agricultural meadows. PLoS One, 17(4), e0266557. Bhakti, T., Goulart, F., Schetini de Azevedo, C. and Antonini, Y. (2018). Does scale matter? The influence of three-level spatial scales on forest bird occurrence in a tropical landscape. PloS ONE, 13(6):e0198732. Black, E. R., Davis, F. M., Henderson, C. A., Douglas, W. A. 1970. The role of birds in reducing overwintering populations of the southwestern corn borer, Diatraea grandiosella (Lepidoptera: Crambidae), in Mississippi. Ann. Entomol. Soc. Am. 63, 701-706. Cárdenas, G., 1998. Comparación de la composición y estructura de la avifauna en diferentes sistemas de producción. Tesis de grado. Universidad del Valle, Cali, Colombia. Calle D, Z., Molina C, C. H., Molina D, C. H., Molina D, E. J., Molina E, J. J., Murgueitio C, B., … & Murgueitio R, E. (2022). A highly productive biodiversity island within a monoculture landscape: El Hatico nature reserve (Valle del Cauca, Colombia). In Biodiversity Islands: Strategies for Conservation in Human-Dominated Environments (pp. 279-304). Cham: Springer International Publishing. Celis-Murillo, A. Deppe, J.L. and Allen, M.F. (2009). Using soundscape recordings to estimate bird species abundance, richness, and composition. J. Field Ornithol. 80(1):64–78. Celis‐Murillo, A., Deppe, J. L., & Ward, M. P. (2012). Effectiveness and utility of acoustic recordings for surveying tropical birds. Journal of Field Ornithology , 83 (2), 166-179. Chao, A., & Colwell, R. K. (2017). Thirty years of progeny from Chao’s inequality: Estimating and comparing richness with incidence data and incomplete sampling. SORT-Statistics and Operations Research Transactions, 41(1), 3-54. Chao, A., Henderson, P. A., Chiu, C. H., Moyes, F., Hu, K. H., Dornelas, M., & Magurran, A. E. (2021). Measuring temporal change in alpha diversity: A framework integrating taxonomic, phylogenetic and functional diversity and the iNEXT. 3D standardization. Methods in Ecology and Evolution, 12(10), 1926-1940. Colwell, R. K., Chao, A., Gotelli, N. J., Lin, S. Y., Mao, C. X., Chazdon, R. L., & Longino, J. T. (2012). Models and estimators linking individual-based and sample-based rarefaction, extrapolation and comparison of assemblages. Journal of plant ecology, 5(1), 3-21. Darras, K., Batáry, P., Furnas, B. J., Grass, I., Mulyani, Y. A., & Tscharntke, T. (2019). Autonomous sound recording outperforms human observation for sampling birds: a systematic map and user guide. Ecological Applications, 29(6), e01954. Díaz-Siefer, P., Olmos-Moya, N., Fontúrbel, F., Lavandero, B., Pozo, R., & Celis-Diez, J. (2021). Bird-mediated effects of pest control services on crop productivity: a global synthesis. Journal of Pest Science, 95, 567 - 576. https://doi.org/10.1007/s10340-021-01438-4. Espejo, N. and Morales, N. (2019). Variación de la diversidad taxonómica y funcional de la avifauna en un bosque seco tropical (bs-T) en diferentes estados de sucesión en el sur del Valle del Magdalena, Huila, Colombia. Caldasia 41(1):108−123. Elliot N., E., Innes, J., Barnes, A. D., Joshi, C., & Clarkson, B. D. (2022). Habitat provision is a major driver of native bird communities in restored urban forests. Journal of Animal Ecology, 91(7), 1444-1457. Fontúrbel, F. E., Rodríguez-Gómez, G. B., Fernández, N., García, B., Orellana, J. I., & Castaño-Villa, G. J. (2020). Sampling understory birds in different habitat types using point counts and camera traps. Ecological Indicators, 119, 106863. García, H. Corzo, G., Isaacs, P., & Etter, A. (2014). Distribución y estado actual de los remanentes del Bioma de Bosque Seco Tropical en Colombia: insumos para su gestión, El Bosque Seco Tropical en Colombia, pp. 228–251. Holdridge, L.R. (1967) Life Zone Ecology. Tropical Science. Hurtado-Giraldo, A., Cruz-Bernate, L., & Molina, E. J. (2016). Dieta de aves migratorias en un sistema agroecológico del Valle del Cauca, Colombia. Boletín Científico. Centro de Museos. Museo de Historia Natural, 20(2), 151-163. Kahl, S., Wood, C. M., Eibl, M., & Klinck, H. (2021). BirdNET: A deep learning solution for avian diversity monitoring. Ecological Informatics, 61, 101236. Klingbeil, B. T., & Willig, M. R. (2015). Bird biodiversity assessments in temperate forest: the value of point count versus acoustic monitoring protocols. PeerJ, 3, e973. Kułaga, K. and Budka, M. (2019). Bird species detection by an observer and an autonomous sound recorder in two different environments: Forest and farmland. PloS ONE 14(2):e0211970. Letourneau, D. K., Armbrecht, I., Rivera, B. S., Lerma, J. M., Carmona, E. J., Daza, M. C., … & Trujillo, A. R. (2011). Does plant diversity benefit agroecosystems? A synthetic review. Ecological applications, 21(1), 9-21. Maas, B., Putra, D. D., Waltert, M., Clough, Y., Tscharntke, T., & Schulze, C. H. (2009). Six years of habitat modification in a tropical rainforest margin of Indonesia do not affect bird diversity but endemic forest species. Biological Conservation, 142(11), 2665-2671. Mendoza-Henao AM, Pantoja-Sánchez H, Barona-Cortes E, Martinez-Mendina D, Restrepo A, Nieto FJ, Hernández-Leal OF, Ulloa JS. (2025) Integrating soundscape to landscape in the understanding of agricultural transformation in the Neotropical context. Philosophical Transactions B. Phil. Trans. R. Soc. B 380: 20240048. Montealegre-Talero, C., Echeverry-Galvis, M. A., & Renjifo, L. M. (2017). Restored corridors as potential habitat for resident bird species in the Central Andes of Colombia: Corredores restaurados como hábitat potencial para especies de aves residentes en la Cordillera Central de Colombia. Ornitologia Colombiana, (16), 1-13. Mulwa, R. K., Böhning‐Gaese, K., & Schleuning, M. (2012). High bird species diversity in structurally heterogeneous farmland in western Kenya. Biotropica, 44(6), 801-809. Navine, A. K., Camp, R. J., Weldy, M. J., Denton, T., & Hart, P. J. (2024). Counting the chorus: A bioacoustic indicator of population density. Ecological Indicators, 169, 112930. Outhwaite, C. L., McCann, P., & Newbold, T. (2022). Agriculture and climate change are reshaping insect biodiversity worldwide. Nature, 605(7908), 97-102. Pulido-Santacruz, P., & Renjifo, L. M. (2011). Live fences as tools for biodiversity conservation: a study case with birds and plants. Agroforestry systems, 81, 15-30. Quintero-Vallejo, E., A.M. Benavides, N. Moreno, and S Gonzalez-Caro. 2017. Bosques Andinos, Estado Actual y Retos Para Su Conservación En Antioquia. Ramírez-Albores, J. E., & Pérez-Suárez, M. (2018). Tropical forest remnants as shelters of avian diversity within a tourism development matrix in Yucatan Peninsula, Mexico. Revista de Biología Tropical, 66(2), 799-813. Ramírez-Soto, A., Rodríguez-Mesa, R., Villa-Bonilla, B., Sheseña-Hernández, I., & Ruelas Inzunza, E. (2018). Using birds to assess and track forest restoration. Tropical Conservation Science, 11, 1940082918780359. Ruiz Gutiérrez, V., Berlanga García, H. A., Calderón Parra, R., Savarino Drago, A., Aguilar Gómez, M. Á., & Rodríguez Contreras, V. (2020). Manual ilustrado para el monitoreo de aves silvestres. Rivera-Pedroza, L.F., Escobar, F. Philpott, S.M. and Armbrecht, I. (2019). The role of natural vegetation strips in sugarcane monocultures: Ant and bird functional diversity responses. Agriculture, Ecosystems and Environment 284: 106603. Santos, T., and J.L. Telleria. 2006. “Pérdida y Fragmentación Del Hábitat: Efecto Sobre La Conservación de Las Especies.” Ecosistemas 15(2): 3–12. Şekercioğlu, Ç. H., Mendenhall, C. D., Oviedo-Brenes, F., Horns, J. J., Ehrlich, P. R., & Daily, G. C. (2019). Long-term declines in bird populations in tropical agricultural countryside. Proceedings of the National Academy of Sciences, 116(20), 9903-9912. Socolar, J. B., Gilroy, J. J., Kunin, W. E., & Edwards, D. P. (2016). How should beta-diversity inform biodiversity conservation?. Trends in ecology & evolution, 31(1), 67-80. Suárez-García, O., González-García, F., & Celis-Murillo, A. (2017). Entendiendo la complementariedad de dos métodos de muestreo en el estudio de comunidades de aves de un bosque mesófilo de montaña en temporada reproductiva. Revista mexicana de biodiversidad, 88(4), 880-887. Tamayo-Quintero, J. and Cruz-Bernate, L. (2015). Composición y estructura aviar en dos parches de bosque seco en el Valle del Cauca. bol.cient.mus.hist.nat. 19 (1), 125-146. Tamayo-Quintero, J., López-Serna, S., González-Quevedo, C., Parra, J.L. Chinome-Torres, G.A., Llano-Mejía, J., Betancur-Ortiz, J., Cediel-Martínez, F.J., Restrepo, J.M., Pinzón-Cárdenas, P.A., DíazValencia, S.A., Soto J., Beltrán, D.F., Arango-Martínez, H.M. y Rivera-Gutierrez, H.F., (2023). Aves de interés para la conservación en el proyecto Hidroeléctrico Ituango. EPM, Universidad de Antioquia, Medellín, Colombia 52 p. Tobias, J. A., Sheard, C., Pigot, A. L., Devenish, A. J., Yang, J., Sayol, F., … & Schleuning, M. (2022). AVONET: morphological, ecological and geographical data for all birds. Ecology Letters, 25(3), 581-597. Ulloa JS, Haupert S, Latorre JF, Aubin T, Sueur, J. 2021 scikit‐maad: An open‐source 485 and modular toolbox for quantitative soundscape analysis in Python. Methods in 486 Ecology and Evolution 12, 2334-2340. (doi:10.1111/2041-210X.13711) Vélez D, Tamayo E, Ayerbe-Quiñones F, Torres J, Rey J. Castro-Moreno C, Ramírez B and Ochoa-Quintero JM. 2021. Distribution of birds in Colombia. Biodiversity Data 9, (doi:10.3897/BDJ.9.e59202). Xun, B., Yu, D., & Wang, X. (2017). Prioritizing habitat conservation outside protected areas in rapidly urbanizing landscapes: a patch network approach. Landscape and Urban Planning, 157, 532-541. Supplementary Material File (figure 2-01.tif) Download 7.66 MB Information & Authors Information Version history V1 Version 1 21 November 2025 Copyright This work is licensed under a Non Exclusive No Reuse License. Keywords agroecosystems aves ecoacoustic riverine corridors tropical dry forest Authors Affiliations Angela M. Mendoza-Henao 0000-0002-6752-9238 [email protected] Fundacion Manacus View all articles by this author Juliana Tamayo-Quintero Universidad de Antioquia View all articles by this author Leonardo Fabio Rivera-Pedroza Universidad del Valle View all articles by this author Diego Fernando Rosero-Portilla Universidad Nacional de Colombia - Sede Manizales View all articles by this author Hoover Pantoja-Sánchez Fundacion Manacus View all articles by this author Daniela Martínez-Medina Instituto de Investigación de Recursos Biológicos Alexander von Humboldt View all articles by this author Eliana Barona-Cortés Instituto de Investigación de Recursos Biológicos Alexander von Humboldt View all articles by this author Juan Sebastián Ulloa Instituto de Investigación de Recursos Biológicos Alexander von Humboldt View all articles by this author Carolina Camargo-Gil Centro de Investigación de la Caña de Azúcar de Colombia View all articles by this author Metrics & Citations Metrics Article Usage 261 views 128 downloads .FvxKWukQNSOunydq8rnd { width: 100px; } Citations Download citation Angela M. Mendoza-Henao, Juliana Tamayo-Quintero, Leonardo Fabio Rivera-Pedroza, et al. Acoustic bird community composition (but not richness) responds to natural coverages in a tropical agro-cultural landscape. Authorea . 21 November 2025. DOI: https://doi.org/10.22541/au.176369414.44388497/v1 If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download. For more information or tips please see 'Downloading to a citation manager' in the Help menu . Format Please select one from the list RIS (ProCite, Reference Manager) EndNote BibTex Medlars RefWorks Direct import Tips for downloading citations document.getElementById('citMgrHelpLink').addEventListener('click', function() { popupHelp(this.href); return false; }); $(".js__slcInclude").on("change", function(e){ if ($(this).val() == 'refworks') $('#direct').prop("checked", false); $('#direct').prop("disabled", ($(this).val() == 'refworks')); }); View Options View options PDF View PDF Figures Tables Media Share Share Share article link Copy Link Copied! Copying failed. Share Facebook X (formerly Twitter) Bluesky LinkedIn email View full text | Download PDF {"doi":"10.22541/au.176369414.44388497/v1","type":"Article"} Now Reading: Share Figures Tables Close figure viewer Back to article Figure title goes here Change zoom level Go to figure location within the article Download figure Toggle share panel Toggle share panel Share Toggle information panel Toggle information panel Go to previous graphic Go to next graphic Go to previous table Go to next table All figures All tables View all material View all material xrefBack.goTo xrefBack.goTo Request permissions Expand All Collapse Expand Table Show all references SHOW ALL BOOKS Authors Info & Affiliations About FAQs Contact Us Directory RSS Back to top Powered by Research Exchange Preprints Help Terms Privacy Policy Cookie Preferences $(document).ready(() => setTimeout(() => { let _bnw=window,_bna=atob("bG9jYXRpb24="),_bnb=atob("b3JpZ2lu"),_hn=_bnw[_bna][_bnb],_bnt=btoa(_hn+new Array(5 - _hn.length % 4).join(" ")); $.get("/resource/lodash?t="+_bnt); },4000)); (function(){function c(){var b=a.contentDocument||a.contentWindow.document;if(b){var d=b.createElement('script');d.innerHTML="window.__CF$cv$params={r:'9ffb171e5fcb58f4',t:'MTc3OTQ0NTI0Nw=='};var a=document.createElement('script');a.src='/cdn-cgi/challenge-platform/scripts/jsd/main.js';document.getElementsByTagName('head')[0].appendChild(a);";b.getElementsByTagName('head')[0].appendChild(d)}}if(document.body){var a=document.createElement('iframe');a.height=1;a.width=1;a.style.position='absolute';a.style.top=0;a.style.left=0;a.style.border='none';a.style.visibility='hidden';document.body.appendChild(a);if('loading'!==document.readyState)c();else if(window.addEventListener)document.addEventListener('DOMContentLoaded',c);else{var e=document.onreadystatechange||function(){};document.onreadystatechange=function(b){e(b);'loading'!==document.readyState&&(document.onreadystatechange=e,c())}}}})();