Search

As more land is altered by human activity and more species become at risk of extinction, it is essential that we understand the requirements for conserving threatened species across human-modified landscapes. Owing to their rarity and often sparse distributions, threatened species can be difficult to study and efficient methods to sample them across wide temporal and spatial scales have been lacking. Passive acoustic monitoring (PAM) is increasingly recognized as an efficient method for collecting data on vocal species; however, the development of automated species detectors required to analyse large amounts of acoustic data is not keeping pace. Here, we collected 35 805 h of acoustic data across 341 sites in a region over 1000 km2 to show that PAM, together with a newly developed automated detector, is able to successfully detect the endangered Geoffroy's spider monkey (Ateles geoffroyi), allowing us to show that the Geoffroy's spider monkey was absent below a threshold of 80% forest cover and within 1 km of primary paved roads and occurred equally in old growth and secondary forests. We discuss how this methodology circumvents many of the existing issues in traditional sampling methods and can be highly successful in the study of vocally rare or threatened species. Our results provide tools and knowledge for setting targets and developing conservation strategies for the protection of the Geoffroy's spider monkey.

We focus on using the predictive uncertainty signal calculated by Bayesian neural networks to guide learning in the self-same task the model is being trained on. Not opting for costly Monte Carlo sampling of weights, we propagate the approximate hidden variance in an end-to-end manner, throughout a variational Bayesian adaptation of a ResNet with attention and squeeze-and-excitation blocks, in order to identify data samples that should contribute less into the loss value calculation. We, thus, propose uncertainty-aware, data-specific label smoothing, where the smoothing probability is dependent on this epistemic uncertainty. We show that, through the explicit usage of the epistemic uncertainty in the loss calculation, the variational model is led to improved predictive and calibration performance. This core machine learning methodology is exemplified at wildlife call detection, from audio recordings made via passive acoustic monitoring equipment in the animals' natural habitats, with the future goal of automating large scale annotation in a trustworthy manner.

Habitat loss can trigger cascades of secondary extinctions, changing the organization of interacting assemblages. Until recently, most extinction models in interaction systems had limited ecological realism. Here, we estimate a realistic sequence of species extinctions resulting from habitat loss to assess its impacts on the structure of frugivory networks from the Brazilian Atlantic Forest. We show that realistic and random extinctions led to similar patterns. We also identified a threshold in the response of network structure to habitat loss. When forest cover was reduced to less than 40% of the landscape, network organization changed dramatically. Hence, the number of species being lost, rather than the order of species extinctions, is the key determinant of its impacts on the organization of frugivory networks. We highlight the need to conserve around 40% of forest cover to keep the basic organization of frugivory networks, a threshold already reached at the best‐preserved Brazilian Atlantic Forest bioregion.

Habitat fragmentation studies are producing inconsistent and complex results across which it is nearly impossible to synthesise. Consistent analytical techniques can be applied to primary datasets, if stored in a flexible database that allows simple data retrieval for subsequent analyses. Method: We developed a relational database linking data collected in the field to taxonomic nomenclature, spatial and temporal plot attributes and further environmental variables (e.g. information on biogeographic region. Typical field assessments include measures of biological variables (e.g. presence, abundance, ground cover) of one species or a set of species linked to a set of plots in fragments of a forested landscape. Conclusion: The database currently holds records of 5792 unique species sampled in 52 landscapes in six of eight biogeographic regions: mammals 173, birds 1101, herpetofauna 284, insects 2317, other arthropods: 48, plants 1804, snails 65. Most species are found in one or two landscapes, but some are found in four. Using the huge amount of primary data on biodiversity response to fragmentation becomes increasingly important as anthropogenic pressures from high population growth and land demands are increasing. This database can be queried to extract data for subsequent analyses of the biological response to forest fragmentation with new metrics that can integrate across the components of fragmented landscapes. Meta-analyses of findings based on consistent methods and metrics will be able to generalise over studies allowing inter-comparisons for unified answers. The database can thus help researchers in providing findings for analyses of trade-offs between land use benefits and impacts on biodiversity and to track performance of management for biodiversity conservation in human-modified landscapes. Fil: Pfeifer, Marion. Imperial College London; Reino Unido Fil: Lefebvre, Veronique. Imperial College London; Reino Unido Fil: Gardner, Toby A.. Stockholm Environment Institute; Suecia Fil: Arroyo Rodríguez, Víctor. Universidad Nacional Autónoma de México; México Fil: Baeten, Lander. University of Ghent; Bélgica Fil: Banks Leite, Cristina. Imperial College London; Reino Unido Fil: Barlow, Jos. Lancaster University; Reino Unido Fil: Betts, Matthew G.. State University of Oregon; Estados Unidos Fil: Brunet, Joerg. Swedish University of Agricultural Sciences; Suecia Fil: Cerezo Blandón, Alexis Mauricio. Universidad de Buenos Aires. Facultad de Agronomía. Departamento de Métodos Cuantitativos y Sistemas de Información; Argentina Fil: Cisneros, Laura M.. University of Connecticut; Estados Unidos Fil: Collard, Stuart. Nature Conservation Society of South Australia; Australia Fil: D´Cruze, Neil. The World Society for the Protection of Animals; Reino Unido Fil: Da Silva Motta, Catarina. Ministério da Ciência, Tecnologia, Inovações. Instituto Nacional de Pesquisas da Amazônia; Brasil Fil: Duguay, Stephanie. Carleton University; Canadá Fil: Eggermont, Hilde. University of Ghent; Bélgica Fil: Eigenbrod, Félix. University of Southampton; Reino Unido Fil: Hadley, Adam S.. State University of Oregon; Estados Unidos Fil: Hanson, Thor R.. No especifíca; Fil: Hawes, Joseph E.. University of East Anglia; Reino Unido Fil: Heartsill Scalley, Tamara. United State Department of Agriculture. Forestry Service; Puerto Rico Fil: Klingbeil, Brian T.. University of Connecticut; Estados Unidos Fil: Kolb, Annette. Universitat Bremen; Alemania Fil: Kormann, Urs. Universität Göttingen; Alemania Fil: Kumar, Sunil. State University of Colorado - Fort Collins; Estados Unidos Fil: Lachat, Thibault. Swiss Federal Institute for Forest; Suiza Fil: Lakeman Fraser, Poppy. Imperial College London; Reino Unido Fil: Lantschner, María Victoria. Consejo Nacional de Investigaciones Científicas y Técnicas. Centro Científico Tecnológico Conicet - Bahía Blanca; Argentina. Instituto Nacional de Tecnología Agropecuaria. Centro Regional Patagonia Norte. Estación Experimental Agropecuaria San Carlos de Bariloche; Argentina Fil: Laurance, William F.. James Cook University; Australia Fil: Leal, Inara R.. Universidade Federal de Pernambuco; Brasil Fil: Lens, Luc. University of Ghent; Bélgica Fil: Marsh, Charles J.. University of Leeds; Reino Unido Fil: Medina Rangel, Guido F.. Universidad Nacional de Colombia; Colombia Fil: Melles, Stephanie. University of Toronto; Canadá Fil: Mezger, Dirk. Field Museum of Natural History; Estados Unidos Fil: Oldekop, Johan A.. University of Sheffield; Reino Unido Fil: Overal , Williams L.. Museu Paraense Emílio Goeldi. Departamento de Entomologia; Brasil Fil: Owen, Charlotte. Imperial College London; Reino Unido Fil: Peres, Carlos A.. University of East Anglia; Reino Unido Fil: Phalan, Ben. University of Southampton; Reino Unido Fil: Pidgeon, Anna Michle. University of Wisconsin; Estados Unidos Fil: Pilia, Oriana. Imperial College London; Reino Unido Fil: Possingham, Hugh P.. Imperial College London; Reino Unido. The University Of Queensland; Australia Fil: Possingham, Max L.. No especifíca; Fil: Raheem, Dinarzarde C.. Royal Belgian Institute of Natural Sciences; Bélgica. Natural History Museum; Reino Unido Fil: Ribeiro, Danilo B.. Universidade Federal do Mato Grosso do Sul; Brasil Fil: Ribeiro Neto, Jose D.. Universidade Federal de Pernambuco; Brasil Fil: Robinson, Douglas W.. State University of Oregon; Estados Unidos Fil: Robinson, Richard. Manjimup Research Centre; Australia Fil: Rytwinski, Trina. Carleton University; Canadá Fil: Scherber, Christoph. Universität Göttingen; Alemania Fil: Slade, Eleanor M.. University of Oxford; Reino Unido Fil: Somarriba, Eduardo. Centro Agronómico Tropical de Investigación y Enseñanza; Costa Rica Fil: Stouffer, Philip C.. State University of Louisiana; Estados Unidos Fil: Struebig, Matthew J.. University of Kent; Reino Unido Fil: Tylianakis, Jason M.. University College London; Estados Unidos. Imperial College London; Reino Unido Fil: Teja, Tscharntke. Universität Göttingen; Alemania Fil: Tyre, Andrew J.. Universidad de Nebraska - Lincoln; Estados Unidos Fil: Urbina Cardona, Jose N.. Pontificia Universidad Javeriana; Colombia Fil: Vasconcelos, Heraldo L.. Universidade Federal de Uberlandia; Brasil Fil: Wearn, Oliver. Imperial College London; Reino Unido. The Zoological Society of London; Reino Unido Fil: Wells, Konstans. University of Adelaide; Australia Fil: Willig, Michael R.. University of Connecticut; Estados Unidos Fil: Wood, Eric. University of Wisconsin; Estados Unidos Fil: Young, Richard P.. Durrell Wildlife Conservation Trust; Reino Unido Fil: Bradley, Andrew V.. Imperial College London; Reino Unido Fil: Ewers, Robert M.. Imperial College London; Reino Unido

Estudos de fragmentação de habitat produziram resultados complexos que são desafiadores para sintetizar. Inconsistências entre os estudos podem resultar da variação a escolha de métricas de paisagem e variáveis ​​de resposta, que geralmente é composta por falta de informação estatística ou metodológica chave. Collating conjuntos de dados primários sobre as respostas da biodiversidade à fragmentação de forma consistente e banco de dados flexível permite a recuperação de dados simples para análises subseqüentes. Apresentamos um banco de dados relacional que vincula esses dados de campo à nomenclatura taxonômica, atributos de tramas espaciais e temporais e características ambientais. Campo avaliações incluem medidas da (s) resposta (s) (por exemplo, presença, abundância, cobertura do solo) de uma ou mais espécies ligadas a parcelas em fragmentos dentro de uma paisagem parcialmente florestada. O banco de dados atualmente possui 9830 exclusivos espécies registradas em parcelas de 58 paisagens únicas em seis dos oito reinos: mamíferos 315, aves 1286, herpteis 460, insetos 4521, aranhas 204, outros artrópodes 85, gastrópodes 70, anelídeos 8, platelmintos 4, Onychophora 2, vascular plantas 2112, plantas vasculares e líquenes 320, e fungos 449. Três paisagens foram amostrados como séries temporais de longo prazo (> 10 anos). Setecentos e onze espécies são encontradas em duas ou mais paisagens. Consolidando o substancial quantidade de dados primários disponíveis sobre as respostas da biodiversidade à fragmentação no contexto da mudança do uso da terra e das perturbações naturais é um parte da compreensão dos efeitos do aumento das pressões antrópicas sobre a terra. O formato consistente desse banco de dados facilita o teste de generalizações respostas biológicas à fragmentação em diversos sistemas e taxa. isto também permite o reexame de conjuntos de dados existentes com paisagem alternativa métricas e métodos estatísticos robustos, por exemplo, ajudando a resolver a pseudoreplicação problemas. A base de dados pode assim ajudar os pesquisadores a produzir amplas sínteses dos efeitos do uso da terra. O banco de dados é dinâmico e inclusivo, e contribuições de esforços de coleta de dados individuais e em grande escala são bem-vindos. Habitat fragmentation studies have produced complex results that are challenging to synthesize. Inconsistencies among studies may result from variation in the choice of landscape metrics and response variables, which is often compounded by a lack of key statistical or methodological information. Collating primary datasets on biodiversity responses to fragmentation in a consistent and flexible database permits simple data retrieval for subsequent analyses. We present a relational database that links such field data to taxonomic nomenclature, spatial and temporal plot attributes, and environmental characteristics. Field assessments include measurements of the response(s) (e.g., presence, abundance, ground cover) of one or more species linked to plots in fragments within a partially forested landscape. The database currently holds 9830 unique species recorded in plots of 58 unique landscapes in six of eight realms: mammals 315, birds 1286, herptiles 460, insects 4521, spiders 204, other arthropods 85, gastropods 70, annelids 8, platyhelminthes 4, Onychophora 2, vascular plants 2112, nonvascular plants and lichens 320, and fungi 449. Three landscapes were sampled as long-term time series (>10 years). Seven hundred and eleven species are found in two or more landscapes. Consolidating the substantial amount of primary data available on biodiversity responses to fragmentation in the context of land-use change and natural disturbances is an essential part of understanding the effects of increasing anthropogenic pressures on land. The consistent format of this database facilitates testing of generalizations concerning biologic responses to fragmentation across diverse systems and taxa. It also allows the re-examination of existing datasets with alternative landscape metrics and robust statistical methods, for example, helping to address pseudoreplication problems. The database can thus help researchers in producing broad syntheses of the effects of land use. The database is dynamic and inclusive, and contributions from individual and large-scale data-collection efforts are welcome.

Habitat fragmentation studies have produced complex results that are challenging to synthesize. Inconsistencies among studies may result from variation in the choice of landscape metrics and response variables, which is often compounded by a lack of key statistical or methodological information. Collating primary datasets on biodiversity responses to fragmentation in a consistent and flexible database permits simple data retrieval for subsequent analyses. We present a relational database that links such field data to taxonomic nomenclature, spatial and temporal plot attributes, and environmental characteristics. Field assessments include measurements of the response(s) (e.g., presence, abundance, ground cover) of one or more species linked to plots in fragments within a partially forested landscape. The database currently holds 9830 unique species recorded in plots of 58 unique landscapes in six of eight realms: mammals 315, birds 1286, herptiles 460, insects 4521, spiders 204, other arthropods 85, gastropods 70, annelids 8, platyhelminthes 4, Onychophora 2, vascular plants 2112, nonvascular plants and lichens 320, and fungi 449. Three landscapes were sampled as long-term time series (>10 years). Seven hundred and eleven species are found in two or more landscapes. Consolidating the substantial amount of primary data available on biodiversity responses to fragmentation in the context of land-use change and natural disturbances is an essential part of understanding the effects of increasing anthropogenic pressures on land. The consistent format of this database facilitates testing of generalizations concerning biologic responses to fragmentation across diverse systems and taxa. It also allows the re-examination of existing datasets with alternative landscape metrics and robust statistical methods, for example, helping to address pseudo-replication problems. The database can thus help researchers in producing broad syntheses of the effects of land use. The database is dynamic and inclusive, and contributions from individual and large-scale data-collection efforts are welcome. © 2014 The Authors. Ecology and Evolution published by John Wiley & Sons Ltd.

Tesis para obtener el grado de Doctor of Philosophy (PhD) en Life Sciences, del Imperial College London, en 2015 The conversion of land from natural environments to human-managed areas has been particularly pronounced in the semiarid Chaco Region. In this context, it is critical to understand the impacts that these major changes have on species diversity. The specific aims of this thesis were to understand: i) how the different human-modified environments influenced the distribution of the bird assemblages in the current agricultural matrix; ii) which local and landscape scale characteristics best explained species richness and relative abundance of mammals; and iii) what is the relationship between biodiversity and profit for both taxa, and given this relationship, to what extent can conservation and production objectives be simultaneously achieved. Bird and mammal community data was gathered in seven different habitats of an agricultural matrix and inside a National Park. Bird species richness and abundance were extremely low in highly modified environments (agriculture plots), but increased significantly in intermediate, or low modified habitats (silvopasture plots, forest outside and inside the National Park, respectively). Richness and capture frequency of mammals increased gradually across the gradient of habitat modification, from low numbers in agricultural habitats to a maximum in National Park forests. Changes in species composition with production intensification differed between birds and mammals. The bird communities were similar for low- and intermediate-intensity farming, with constant number of species, abundance and community integrity. However, further intensification led to an abrupt decline, defining a clear threshold. Mammal species richness was maintained in low-yield farming, relative abundance declined with increasing production, favoured by a wildlife-friendly farming approach. In contrast to birds, mammal community integrity decreased exponentially with increased levels of intensification, showing that only protected or well-preserved forests can support some habitat-restricted species. The results suggest that a combined strategy is the best option to achieve conservation and production targets that include both taxa. Instituto de Recursos Biológicos Fil: Decarre, Julieta. Instituto Nacional de Tecnología Agropecuaria (INTA). Instituto de Recursos Biológicos; Argentina

Effective monitoring and protection of tropical ecosystems has the potential to conserve vast amounts of the earth’s biodiversity. Yet logistical and technical challenges associated with species rarity, large distributions and home ranges, low detectability and inaccessibility of study sites can hinder monitoring efforts. If we are to mitigate against the threats to biodiversity and restore natural ecosystems, then we urgently need rapid and cost-effective methods to evaluate the current state of species, communities, ecosystems and the effectiveness of interventions. Here, we use passive acoustic monitoring (PAM) combined with computational approaches, as a method to effectively measure trends in biodiversity at a wide spatial scale across a biodiverse region in the tropics. In Chapter Two we used soundscape analysis to evaluate changes in acoustic diversity across the diel cycle over a gradient of land use change. We aimed to assess if soundscape indices can reveal changes in the biological community across the diel cycle and whether loss of native forests affect acoustic diversity in tropical ecosystems. In disturbed habitats, we found a loss of the characteristic dawn and dusk peaks in the diel cycle; known as the dawn and dusk chorus. This was especially prominent in palm oil plantations and grasslands, which showed a complete loss of these peaks. This suggests that in disturbed ecosystems there is likely a loss of species diversity, a shift in species composition, where forest specialists are being replaced by disturbance tolerant species, or that there are modifications in species behaviour, reinforcing the value of native old growth forests in maintaining ecosystem functionality. This loss in dawn and dusk peaks was not apparent when analysing acoustic diversity at specific times during the diel cycle, showing that evaluating acoustic diversity at this temporal scale can be misleading, but in assessing trends across the diel cycle, we can gain a much better representation of changes to biotic communities. In Chapter Three we determined if PAM and a newly developed automated detection and classification system was effective at retrieving information on the Geoffroy’s spider monkey at a wide spatial scale. We assessed how this endangered primate responds to habitat loss and human influence across a gradient of disturbance. We found that the Geoffroy’s spider monkey was absent below 80% forest cover and within 1 km of paved roads, yet was found to some extent in areas of secondary forest and near unpaved roads and buildings. The success of this methodology in the study of a vocal rare species suggests that similar rare species could be studied in the same way. Threshold values for percent forest cover and paved roads will be valuable in developing conservation strategies for the protection of this species. In Chapter Four, we investigated the effectiveness of a sustainable use forest reserve in facilitating connectivity for the Geoffroy’s Spider monkey between two National Parks. We specifically evaluated occurrence across the reserve, habitat suitability, barriers to connectivity and potential mitigation strategies to improve connectivity in the region. We found that the Golfo Dulce Forest Reserve is acting as a buffer to Corcovado National Park and is able to support populations of the Geoffroy’s spider monkey, however, as occurrence was limited to the area surrounding Corcovado, it is possible that it is not facilitating connectivity as intended. Primary road and low forest cover were the most important predictors of poor habitat suitability, both acting as barriers to connectivity and potentially impeding the conservation of an endangered species. This is problematic since Piedras Blancas National Park serves as a connection between the Osa Peninsula and populations of the Geoffroy’s spider monkey in other areas of Costa Rica. In summary, we have shown how PAM, combined with computational approaches, can be used to effectively monitor trends at both fine temporal scales and wide spatial scales across a tropical ecosystem. PAM has provided an effective and rapid approach to monitor trends in biological communities across disturbance gradients, to study rare species across a challenging environment and to evaluate the effectiveness of current management interventions, overcoming many of the key logistical and technical challenges associated with biodiversity monitoring. These methods have revealed important information regarding how anthropogenic disturbance, related to land use change and human development, are threatening both species and communities, which can contribute to setting targets and developing conservation strategies for the protection of biodiversity in the Osa region and beyond. Open Access

The combined effect of habitat degradation and shifts in climate are pushing many species towards local and global extinctions. Often, the most sensitive species to disturbance are the most important for the maintenance of healthy ecosystems. With conservation efforts focussing on biodiversity, can we assume that ecosystem functions will also be preserved? We aimed to test these assumptions by disrupting trophic interactions along natural and anthropogenic gradients of diversity with vertebrate exclusions. In chapter two, we focus on shifts in biotic distribution that occur naturally with elevation. Vertebrate exclusion caused strong trophic cascades on herbivory in pristine tropical forests, but only at high altitudes. In contrast, at lower elevations, forests showed little response to predator exclusion, with no effect on herbivore density or subsequent plant damage. We hypothesis that habitat in the warmer and wetter lowlands harbours higher levels of diversity within mid and low trophic levels and this may confer some resilience to disturbance. In chapters three and four, we investigate how distance to edge and habitat loss effects the above processes. Edge effects significantly mediated trophic cascades, which occurred only within the forest interior. Changes in vertebrate community composition along interior to edge gradients are likely to be responsible for these findings. The effect of exclusion on both arthropod densities was mediated by habitat loss, although the proportion of plant damage was not. The generality of plant-herbivore interactions may make them robust to perturbation, hence the neutral response of herbivory to habitat loss. Our findings provide insight into how the provision of ecosystem function is mediated by biodiversity loss. This information can be vital in identifying which areas of habitat might lose functionality in the wake of disturbance. We 6 suggest that conservation efforts are refocussed to target ecosystems in which functions are reliant on vulnerable specialist species. Open Access

The microbes living inside vertebrates’ guts have highly important consequences for host health and fitness. In turn, microbial communities can rapidly respond to changes in host diet, for example, or environmental conditions. Hence, this relationship appears to be an essential ecological feature. However, our knowledge on how wild birds interact with microbial communities is poor, and little research has been done on non-model organisms. In this PhD thesis, I investigate factors that shape microbiota and utilize high-throughput sequencing techniques to paint a broader picture of the relationship between birds and their microbes by using three populations from two British passerines. My PhD comprised two related projects: (1) a methodological framework to best store faecal samples to analyse gut microbiomes of avian hosts, and (2) an analysis of factors influencing gut microbial communities. First, we evaluated the influence of five widespread storing methods on microbial composition of house sparrow (Passer domesticus) faecal samples (Chapter II). We were then able to analyse the influences of some intrinsic and extrinsic factors that shape the gut microbiota on blue tits (Cyanestes caeruleus), focusing on their development during nestling life. We evaluated how age impacts diversity and composition of bacterial communities and assessed the age-related changes on diet and compared them with simultaneous changes in the microbiota (Chapter III). Finally, we used experimental cross-fostering design to examine the relative contributions of environmental and genetic factors (i.e., nest of rearing and kinship) to the variation of gut bacterial assemblages (Chapter IV). We found that there is no panacea for storing bird faeces, and until then we show that regardless of the protocol chosen, consistency throughout is a prime practice to achieve replicable and reliable results for microbial ecology. We demonstrate that the gut microbial community of blue tits colonises the gut of nestlings very quickly, with multiple colonisation events that lead to high interindividual variation, which is then reduced in adulthood. We further demonstrated that genetics have a minor role in determining microbial composition; in contrast, there are significant similarities in the compositions of the microbiota of genetically unrelated individuals who share a nest, highlighting the importance of local environment shaping gut communities. In summary, this work offers new insights into host-microbiome interactions by enabling a greater understanding of the multifaceted relationship between birds and their microbes. Open Access

The high pressure that anthropogenic activities places on tropical forests are pushing species and communities to the brink of local, regional or global extinctions. However, the consequences of the loss of species and their ecosystem functions are not well known. My thesis addresses the causes and consequences of human disturbance on vascular epiphytes in the Brazilian Atlantic forest (BAF). I sampled understory and canopy, across a gradient of habitat loss, i.e. pastureland, human-modified forests (HMFs) and old-growth forest. I found that habitat loss has driven a net loss of 91% of species exclusive to old-growth, and 90% of individuals. I also found the edge effect to be a ubiquitous landscape process extending at least 500 m within forests habitats, leaving just 19.4% of the whole of the BAF suitable for forest-dependant epiphytes. However, whilst endemic species (habitat specialised species) are more prone to disappear as consequence of habitat loss, species with larger dispersal ranges and the ability to colonize different forest types flourish or endure the harsh conditions after disturbance. Consequently, HMFs have low functional richness and low functional redundancy, meaning low resilience, whereas old-growth forests exhibit high ecological resilience. Habitat transformation also leads to the loss of a large set of ecosystem functions related to pollination and water cycling across strata. Hence preserving large continuous forests are probably the only pragmatic conservation strategy for vascular epiphytes in highly human-modified landscapes. However, human-modified forests still provide limited ecosystem functions that may increase because of initiatives to improve habitat amount. Open Access