Biology of Tropical Birds

INTRODUCTION

Tropical rainforests are home to the majority of the world’s bird species, making them the most valuable terrestrial ecosystems for avian conservation. However, deforestation and forest fragmentation threaten to destroy large areas of formerly intact rainforests, endangering many bird species (Morris 2010, Laurance 2015, Turubanova et al. 2018). Despite this threat, the scientific literature dealing with bird biodiversity in human-modified rainforests is subject to significant regional bias. Most studies have been conducted in the Neotropics, while other regions, such as Africa and Oceania, are neglected in comparison (Sodhi et al. 2008, Matuoka et al. 2020). Findings from one biogeographical region are not always globally applicable due to the high regional specificity of bird communities and the ecosystems they inhabit (Corlett and Primack 2006, Burivalova et al. 2019, Silva et al. 2020), which often leads to an insufficient understanding of anthropogenic impacts (Sodhi et al. 2008).

The island of New Guinea is home to the third largest contiguous primary rainforest in the world (Brooks et al. 2006), and over five percent of global biodiversity is contained within its boundaries (Sekhran et al. 1995). More than 800 bird species live on the island, their diversity peaks at low and medium elevations (Sam et al. 2019, 2024). Until the beginning of the 20th century, the rainforests in the lowlands of New Guinea were virtually untouched by commercial logging (Fox 2011). In Papua New Guinea (hereafter PNG), which occupies the eastern half of the island, over 70 percent of primary forests have retained their continuity until today (Gamoga et al. 2021). In recent decades, however, forest degradation and deforestation have increased significantly (Bryan and Shearman 2015, Turia et al. 2022). This effect is most pronounced in lowland regions, which have the greatest potential for biodiversity loss (Shearman and Bryan 2011). Despite this alarming trend, the impact of deforestation, forest fragmentation or degradation on birds in New Guinea remains poorly understood.

Several studies have discussed the interaction between agricultural land use and bird diversity in PNG. Elevation-dependent changes in bird community structure were found with increasing agricultural exploitation (Marsden et al. 2006, Marsden and Symes 2008). In addition, a preliminary survey identified decreases in species richness and abundances in secondary forest patches regenerating after subsistence agriculture (Tvardíková 2010). To date, only one study has examined the effects of large-scale forest disturbance on bird communities in New Guinea (Sam et al. 2014). The study found that forest fragmentation reduces the species richness of insectivorous and frugivorous birds in particular. These results show comparable effects of large-scale fragmentation as reported in other tropical regions (Thiollay 1997, Renjifo 1999, Şekercioḡlu et al. 2002). However, this study did not consider the full range of altered forest ecosystems in lowland PNG.

In particular, secondary forests form the most widespread human-modified landscape of PNG (Gamoga et al. 2021), yet their bird communities have never been extensively surveyed. Studies from neighboring tropical landscapes have suggested that secondary forests can maintain high bird diversity. Response of large canopy frugivores to selective logging and subsistence agriculture was examined on the island of New Britain, finding mainly that some species sought out gardens and logged forests due to increased proportion of flowering and fruiting trees (Marsden and Pilgrim 2003). In Sulawesi, understory birds were similarly abundant in secondary and primary forests (Waltert et al. 2005), including endemic and red-list species (Martin and Blackburn 2014). Studies from Sumatra and Borneo found similar bird species richness in primary and secondary forests, although this is partially determined by high forest connectivity (Imron et al. 2022, Prabowo et al. 2016).

On the other hand, New Guinean birds are reportedly sensitive to hunting pressure, (Pangau-Adam et al. 2015), as well as changes to habitat structure (Tvardíková 2010), and edge effect (Sam et al. 2014), all of which can be associated with secondary forests (Watson et al. 2004). Many species, such as canopy frugivores or terrestrial insectivores, cannot survive in secondary forests due to limitations associated with variation in habitat structure (Şekercioḡlu et al. 2002). Furthermore, secondary forests are often subject to constant disturbance from adjacent human settlements, such as timber gathering, fire, or hunting (Whitworth et al. 2018, Chmel et al. 2018). Researchers should also consider the regional specificity of forest disturbances in PNG.

Subsistence agriculture is a key driver of deforestation and secondary succession (Corlett 1994). PNG is no exception, but as the lowland areas of the country have historically been sparsely populated due to the prevalence of malaria, there are significant spatial constraints to agricultural activities (Gamoga et al. 2021). Subsistence agriculture is known to have less severe impacts on bird populations, as it is defined by creation of smaller forest clearings and regrowths, which in PNG is accompanied by selective logging of adjacent forest (Blankespoor 1991, Borges 2007). Another widespread form of disturbance leading to secondary forest regrowth in lowland PNG is large-scale deforestation, which has been introduced in the last century (Shearman and Bryan 2011). This type of disturbance is characterized by clear-cut logging of large swathes of primary forest (Gamoga et al. 2021). Bird colonization is subsequently hampered not only by restricted dispersal but also by the absence of important habitat characteristics for many decades after the onset of succession (Watson et al. 2004). The resulting secondary forest regrowth may, therefore, host severely impoverished bird assemblages (Sodhi 2008)

Conservation efforts in PNG are facing many obstacles associated with the country’s natural resources, fractured land ownership, and low level of societal and economic development (Laurance et al. 2012). Furthermore, financial alternatives to the selling of forest land to foreign agents are rare, and mostly unsupported by the national government (Novotny 2010). Education and employment of native para-ecologists is often dependent on non-governmental organizations, as well as researchers bringing expatriate funding (Basset et al. 2000). Moreover, PNG is among the most expensive areas in the world for field research, a problem exacerbated by low forest accessibility and infrastructure development (Bartlett 2018). These issues compound the poorly explored ecosystem functioning with increasing intensity of land use change and deforestation. As especially secondary forests after clear-cutting become a prevalent land cover in the coming decades, it is crucial that their conservation value for birds is quantified.

Our study is the first in-detail examination of bird communities in human-modified lowland rainforests in PNG. We compare key metrics such as bird species richness and density, as well as community structure represented by guild composition, in four forest types: continuous primary forests (P1), isolated primary forest fragments (P2), secondary forests regrown after subsistence agriculture (S1), and secondary forest after large scale clear-cutting deforestation (S2). We aim to test the following predictions:

1) Bird diversity will decrease from P1 forest to P2 forests due to dispersal limitation and loss of microhabitats in forest fragments (Sam et al. 2014). We predict that particularly insectivores will be affected, as they are sensitive to fragmentation (Thiollay 1997).

2) Similarly, S1 and S2 forests will feature declines of bird diversity due to changes in habitat structure. We predict that this effect will be associated with habitat loss for many insectivore species, as well as disappearance of frugivores due to hunting, reduced food supply and increased predation (Waltert et al. 2005, Chmel et al. 2018). We predict that this decline will be more severe in S2 forests regrown after clear-cutting.

METHODS

Study sites

The fieldwork was conducted from January 2010 to October 2012. Our study took place in the lowlands (<200 m asl) of Madang province in Papua New Guinea (Fig. 1a). Study sites included: P1) a continuous primary forest area located in the middle of >10,000 ha continuous lowland primary forest in Wanang Conservation Area, which itself is embedded within ∼100,000 ha of selectively logged, but largely continuous rainforest, P2) isolated primary rainforest fragments of varying sizes (300–1200 ha) located within a formerly clear-cut landscape, S1) small-scale (ca. 400 ha) secondary forest area created by subsistence-level slash-and-burn agriculture, surrounded by continuous primary forest, S2) secondary forests regrown after large-scale clear-cutting which occurred circa 50 years ago (Shearman and Bryan 2011). Some secondary sites were located close (min. 4 km) to surveyed primary forest fragments (Fig. 1a). For a list of surveyed study sites, their centroid coordinates, general locations, mean elevations, and survey periods refer to Appendix 1, Table S1.

Bird surveys

We surveyed bird communities using point counts and mist-netting. Each study site was surveyed three times per month of study (respectively January, June, October; Appendix 1, Table S1). At each site, counts were conducted at 16 points semi-regularly (ca. 150 m apart) spaced along a 2250 m transect (Fig. 1b). Transects at each site were directed through representative microhabitats (e.g., ridges, valleys, rivulets, emergent trees, small natural forest gaps, and flat land forest; Sam et al. 2014). For forest fragments, the transects were located within the forest interior (<250 m from edge), to minimize immediate edge effect. Coordinates of each point were marked down. All birds seen or heard within 50 m radius of the point were recorded. We measured distances to bird individuals from the center of the point with a laser rangefinder, with each recorded bird being classified within one of five distance bands (0–10, 11–20, 31–40, 41–50, as per Sam et al. 2014). We started censuses 15 minutes before sunrise (5:45 AM) at a randomly selected starting point, and we then continued counts in a randomly selected direction along a circular route, sampling points in a site. We counted birds for 15 minutes at each point so all 16 points were surveyed before 11:00 AM. To minimize double-counting, we attempted to accurately track movements of birds, and recorded more individuals of the same species only when they called at the same time or from distinctively different directions. The total sampling effort amounted to 160 counting points over 90 days of fieldwork, as each counting point has been surveyed nine times. All surveys were conducted by B. Koane, S. Jeppy, and K. Sam, all of whom had previous experience with bird surveys in Papua New Guinea. Surveys were conducted by a team of two of rotating membership, ensuring equal contribution by each author present in the field.

We mist-netted birds in two blocks of 3 days (Appendix 1, Table S1). At each study site, we mist-netted birds along a 200m-long line of nets (2.5 m high × 1–18 m long each, 16-mm mesh) for 6 days from 05:30 to 17:30, with checks every 20 minutes. During the first 3 days, nets were placed between the first three points of the point-count transect, and then transferred to the last three points of the point-count transect for the next 3 days. We identified all mist-netted birds to species, marked them with color bands, and released them within 10 minutes.

Bird community datasets

We used the bird data collected during the point counts as the basis for all our analyses of the bird community. We used bird records from mist-netting to determine whether we omitted any bird species during point counts, but we did not include mist-netting data in our analyses because they provide an incomplete record of bird community composition, as they primarily catch birds that live in the understory, but fail to detect birds of other forest strata (Mulvaney and Cherry 2020). Prior to data analysis, we excluded all raptors and swift species from our datasets, due to their wide home ranges or aerial lifestyles (Beehler and Pratt 2016).

All observed bird species were categorized into four broad feeding guilds, including insectivores (invertebrates as main food), frugivores (fruit and seed eaters), omnivores (feed on plant material and invertebrates in similar ratios), and nectarivores, based on empirical data by (Sam et al. 2017), supplemented by information obtained from literature (Beehler and Pratt 2016, Billerman et al. 2022). We used the 14.1 version of the IOC world bird list as taxonomical authority (Gill et al. 2024).

Point-count data were used to create two datasets: first, we created a table noting down the presence/absence of each species over all counting points. If a species was recorded at a counting point over any of nine replications in time, we considered it present at this point for further analyses. We used data on bird species presence/absence to calculate the following variables: bird species richness, measured as the number of species present per counting point over all nine visits (for example, if five species of birds were recorded at one visit, and one additional species was recorded at the next visit, the total species richness would be six, et cetera). Using the same approach, we calculated the species richness of insectivore, frugivore, nectarivore, and omnivore birds.

A second dataset was based on the abundance of each species at each point, averaged over all visits. This number was further adjusted using a distance sampling method for each species recorded, in order to provide an unbiased abundance estimate (Buckland et al. 2005). We used the package RDistance (Mcdonald et al. 2019) in R 4.3.2 (R Core Team 2024) to perform these adjustments. Furthermore, we converted the number of individuals per counting point (i.e., 0.78 ha) to the number of individuals per hectare, so that our data represented actual density of bird populations that can be observed during a bird survey. We then used the population density of each bird species averaged over the nine visits to calculate total assemblage density of birds per counting point. The same way we calculated the assemblage density of insectivores, frugivores, nectarivores, and omnivores.

Habitat survey

To characterize the habitats at our study sites, we recorded at each point: canopy height (using the heights of three randomly selected trees within the point-count circle and measured with a laser meter); shrub foliage density (five measurements at randomly selected locations within each point-count circle using scatterplot estimates; Creagh et al. 2004); and percent cover of shrubs, grass, bare ground, and litter (within 1 m² at 15 randomly selected locations within each point-count circle). We also measured canopy density (based on three photos taken from a 1.5 m tall tripod at the places where canopy height was also measured and analyzed).

Preliminary analysis of habitat structure

Because most vegetation characteristics are likely to be correlated with each other (Sam et al. 2014), we did not include them as predictors in bird community analyses to avoid the confounding effect of multicollinearity (Alin 2010). Instead, we summarized the variation in the habitat characteristics listed above by performing a principal components analysis (PCA) in the Canoco 5 program (ter Braak and Šmilauer 2012). This analysis allowed us to reduce the dimensionality of the variables while minimizing information loss (Abdi and Williams 2010). For each sample (point where birds were surveyed), we included the values of each habitat characteristic as dependent variables. Each counting point was categorized based on the forest type where it was located: P1 (continuous primary forest); P2 (fragmented primary forest); S1 (secondary forest after subsistence agriculture); and S2 (secondary forest after clear-cutting). Following this, we performed an unconstrained PCA analysis on the vegetation characteristics, and projected forest type categories as supplementary variables in ordination space (Šmilauer and Lepš 2014). This process allowed us to identify which vegetation characteristics were more associated with particular types of forest, and determine whether forest type is an accurate proxy for effects of vegetation structure on birds.

Bird community analyses

As our core method of data analysis, we built mixed-effect linear and generalized linear models (glmms; Dean and Nielsen 2007), using the R packages lme4 (Bates et al. 2015) and glmmTMB (Brooks et al. 2017). Within each of our models, one sample was represented by the sum of bird species richness/assemblage density at one counting point over all visits. After this, we established two categorical variables. First, we created a factor describing the forest type in which a study site was located. This factor had four levels: P1 (continuous primary forest); P2 (fragmented primary forest); S1 (secondary forest after subsistence agriculture); and S2 (secondary forest after clear-cutting). This factor was used as a fixed effect in all models. Following this, we created a factor describing the general area where study sites were located (Fig. 1a). This grouping variable had six levels, each describing the same general area where respective study sites were located, and was used as a random effect predictor to account for various confounding effects, including pseudo-replication, as points from the same/adjacent study sites are not independent observations (Heffner et al. 1996), as well as spatial autocorrelation between observations from the same or closely adjacent study sites (Dormann et al. 2012). All of our models thus used a continuous dependent variable (such as bird species richness), the fixed effect categorical predictor describing forest type for each sample (point), and the random effect predictor describing the study site where the sample was located.

We used the mean-parametrized Conway-Maxwell distribution with the “log” link function (Huang 2017) with correction for overdispersion (Brooks et al. 2017) in glmmTMB for modeling five count variables (total bird species richness, as well as species richness of insectivores, frugivores, nectarivores, and omnivores). We used package lme4 to build linear mixed-effects models for continuous dependent variables (log-transformed total assemblage density, as well as that of insectivores, frugivores, nectarivores, and omnivores). We performed model residual diagnostics in the Dharma package (Hartig 2018), testing for between-site spatial autocorrelation by calculating Moran’s I of clustered residuals (Chen 2016). We used the emmeans package (Lenth 2022) to calculate pairwise post hoc tests. We found no indication of spatial autocorrelation in model residuals (Appendix 1, Table S2).

Species-level community composition

To summarize variation in abundances of individual bird species, we performed unconstrained ordination analyses. Out of the community abundance dataset, we extracted the abundances of 1) insectivores, 2) frugivores, 3) nectarivores, and 4) omnivores. We used these abundances to perform four additional Principal Components Analyses (PCA; Abdi and Williams 2010), with projected supplementary variables describing forest type (P1, P2, S1, S2). The resulting ordination diagram showed which species were associated with each type of forest based on their co-occurrence (Šmilauer and Lepš 2014). These unconstrained ordinations aim to highlight particular species which showed the most pronounced response to habitat modification and to what extent was species composition shared among forest types.

RESULTS

Habitat structure

In our PCA ordination, we found that P1 (continuous primary) and P2 (continuous fragmented) forests were associated with greater cover of bare ground, plant litter, as well as increased canopy density, while S1 (secondary after subsistence agriculture) and S2 forests (secondary after clear-cutting) were associated with increased herb cover, shrub cover, and shrub leaf density. This is visualized by the ordination diagram output of our PCA with projected supplementary variables (Fig. 2, variation explained by supplementary variables: 30.36%, variation explained by first two PCA axes: 99.7%). These results show that primary forests (P1, P2) were in general characterized by greater vertical density and stratification of vegetation, and more consistent canopy coverage, while secondary forests (S1, S2) had a less complex vertical structure of vegetation, and thus denser vegetation grew in the understory. We found that the average daily temperatures ranged from 24.1°C in the early morning to 25.6°C in the late afternoon in the P2 forest and from 24.2°C to 27.8°C in the S2 forest. These results suggest that the secondary forests experienced greater daily temperature fluctuations.

Bird community overview

During our survey, we recorded 31,177 bird individuals corresponding to 123 species (Appendix 1, Table S4), of which 113 entered our analyses after carnivorous raptors and strictly aerial insectivores were excluded. The most abundant species was the Mimic Honeyeater (Meliphaga analoga, 5.4% of all bird observations), followed by Helmeted Friarbird (Philemon buceroides, 5%) and Yellow-bellied Longbill (Toxorhamphus novaeguineae, 4.8%). Most species in our study were insectivores (N = 52), followed by frugivores (N = 50), nectarivores (N = 7) and omnivores (N = 5). Insectivores were also the most prevalent guild in terms of assemblage density (46% of individuals), followed by frugivores (35%), nectarivores (13%), and omnivores (6%). Our mist-netting detected only one additional bird species (to point counts), Black-winged Monarch (Monarcha frater), caught once.

Variation in bird community metrics

We found significant differences between forest types for overall bird species richness, but not for assemblage density (Table 1). We found that P1 forest had similar species richness (median SR = 59) as S1 forest (SR = 56), but these numbers dropped in P2 forests (SR = 44) and S2 forests (SR = 42, Fig. 3a), showing that connectivity with continuous primary rainforest was important for bird SR.

Insectivores

Overall, the most common insectivore species was the Mimic Honeyeater (12% of insectivore observations), followed by Yellow-bellied Longbill (10%), and Northern Variable Pitohui (Pitohui kirhocephalus, 8%). We identified insectivores as the bird group most sensitive to rainforest modification. Their species richness dropped from P1 forest to S1 and P2 forest (median SR = 27, 24, 22 respectively), and further in S2 forest (median SR = 18, Fig. 4a). In contrast, insectivore abundances were comparable in P1, S1, and P2 forest (median N/ha = 12, 11, 11), but collapsed in S2 forest (median N/ha = 7, Fig. 4b). Insectivore species composition changed among forest types, as is shown by our PCA results (Fig. 4c, variation explained by supplementary variables: 24.69%, variation explained by first two PCA axes: 30.35%). P1 and S1 forests were often occupied by the same bird species, such as Northern Variable Pitohui or Black-sided Robin (Poecilodryas hypoleuca). In contrast, P2 forest featured a distinct assemblage of birds, including e.g., Yellow-bellied Longbill or Spangled Drongo (Dicrurus bracteatus), as did the S2 forest, which was preferred by e.g., Mimic Honeyeater and Tawny-breasted Honeyeater (Xanthotis flaviventer).

Frugivores

The most abundant frugivore species was the Metallic Starling (Aplonis metallica, 10% of frugivore observations), followed by the Red-capped Flowerpecker (Dicaeum geelvinkianum, 7%) and Orange-bellied Fruit Dove (Ptilinopus iozonus, 6%). Frugivore assemblages were comparably species rich in P1 and S1 forest (median SR = 23, 24 respectively), but they showed a significant decrease in species richness in P2 and S2 forest (median SR = 17, 16 respectively, Fig. 5a, Table 1). However, mean frugivore abundances only differed marginally between forest types in pairwise post hoc tests. Frugivore species composition also varied between forest types (Fig. 5b, variation explained by supplementary variables: 15.3%, variation explained by first two PCA axes: 35.53%). P1 and S1 forests shared many species, such as Beautiful Fruit Dove (Ptilinopus pulchellus, Fig. 5) or Palm Cockatoo (Probosciger aterrimus). While some species preferred both P1 and P2 forests, e.g., King Bird-of-paradise (Cicinnurus regius) or Black Berrypecker (Melanocharis nigra), others strongly preferred P2 forests, such as Lesser Bird-of-paradise (Paradisaea minor) or Wompoo Fruit Dove (Ptilinopus magnificus). S2 forests were occupied by a distinct assemblage of frugivores not commonly present in other forest types, such as Orange-bellied Fruit Dove, Papuan King Parrot (Alisterus chloropterus) or Red-capped Flowerpecker. Despite P2 and S2 forests were comparable in total bird species richness and assemblage density, they were inhabited by different frugivore species, implying that habitat influenced community structure if not diversity.

Nectarivores

Nectarivore species richness only differed marginally among forest types, as shown by mostly non-significant post hoc tests (Fig. 6a). However, their density showed a highly positive response to habitat modification (Table 1, Fig. 6b). While P1 and S1 forests contained similar density (median N/ha = 1 and 1 respectively), these numbers increased in P2 forest (median N/ha = 2) and further in S2 (median N/ha = 3). The most abundant nectarivore was the Helmeted Friarbird (40 % of nectarivore observations), followed by the Black Sunbird (Leptocoma aspasia, 36%) and Coconut Lorikeet (Trichoglossus haematodus, 10%, illustrated in Fig. 5). Individual nectarivore species showed abundance patterns reflected in our model results, increasing in P2 and S2 forest, with the exception of Stella’s Lorikeet (Charmosyna stellae), which occurred in P1, S1 and P2 forest, but disappeared in S2 forest (Fig. 6c, variation explained by supplementary variables: 16.94%, variation explained by first two PCA axes: 87.55%).

Omnivores

Omnivores represented only a small portion of bird community. All five species of omnivores were comparable in their occurrence, with Long-Billed honeyeater (Melilestes megarhynchus) the most represented (22% of omnivore observations), followed by Rufous-bellied Kookaburra (Dacelo gaudichaud, 21%) and Hooded Butcherbird (Cracticus cassicus, 20%). While omnivore species richness was significantly affected by forest type (Table 1), there was only one significant pairwise comparison, marginalizing our results (Fig. S1a). Furthermore, our PCA analysis with supplementary variables showed that most omnivores were not particularly restricted to one forest type, with the exception of Hooded Butcherbird, which preferred S2 forests (Fig. S1).

DISCUSSION

Within our study areas, we show that deforestation leads to impoverished bird communities after decades of forest succession, showing that key forest characteristics for many species are still missing in human-modified forests (Watson et al. 2004). Bird communities in human-modified forests are also strongly influenced by the landscape context. Secondary forests after subsistence agriculture have retained connectivity to continuous primary rainforest, resulting in similar species richness and community structure. Conversely, fragments of primary forests encountered significant loss of species richness and strong changes in community composition due to habitat isolation. The suitability of human-modified forests as habitat for birds is significantly influenced by the adjacent land use systems (Saab 1999, Banks-Leite et al. 2010). As a result, bird biodiversity is shaped by the interplay of local vegetation characteristics and forest connectivity (Powell et al. 2013, Ramos et al. 2020, Salgueiro et al. 2021). This relationship is best explained by deconstructing total species richness and assemblage density into separate functional groups, such as feeding guilds (Sekercioglu 2012, Matuoka et al. 2020).

Insectivores

Insectivores were the dominant bird group in our study, and were the most sensitive to anthropogenic modification. Each modified forest type featured a significant decrease in species richness from the primary forest. Species richness in the S1 forests decreased similarly as in the P2 forest. In S1 forests, there were changes in important habitat characteristics that some insectivores depended on, such as proportion of plant litter and bare groundcover, and decreased canopy height and density (Tvardíková 2010). However, the effects of habitat loss were likely mediated by proximity to primary forest (Mayhew et al. 2019), with disappearance of species prevented by colonization (Bradfer-Lawrence et al. 2018) and community composition influenced by dispersal from the surrounding forest matrix (Johns 1991).

As for the primary forest fragments, our interpretation differs somewhat from Sam et al. (2014). They attributed the loss of insectivore diversity mainly to the decline of important microhabitats. However, our results show that the primary forest fragments were affected by a significant loss of species, although they retained the vegetation structure typical of primary forests. Dispersal limitation is considered by many to be a limiting factor for insectivore occupancy in fragmented forests (Sekercioglu 2002, Moore et al. 2008, Powell et al. 2013). While primary forest fragments are not strictly isolated from other forest habitats, our results showed that the surrounding secondary forest matrix produced by clear-cutting was unable to harbor many insectivore species, likely hindering their dispersal (Stratford and Stouffer 1999, Powell et al. 2015).

The post-clearcut secondary forests were not only species-poor, but also had significantly lower insectivore densities. We hypothesize that the decline in key vegetation features was exacerbated by the loss of connectivity to a contiguous primary forest, leading to a collapse of insectivore assemblages. The reduction in litterfall and bare ground cover was followed by creation of dense herb and shrub overgrowth, making secondary forests unsuitable for terrestrial species such as the Blue Jewel-babbler (Ptilorrhoa caerulescens), while the loss of vertical vegetation density impacted flycatchers (e.g., Golden Monarch Symposiachrus guttula) and midstory foragers (Sam et al. 2014), such as the Arafura Shrikethrush (Colluricincla megarhyncha, Billerman et al. 2022). Furthermore, greater daily temperature fluctuations may have had an additional detrimental effect on sedentary insectivores (Didham and Lawton 1999, Davies-Colley et al. 2000). Observed decrease of insectivore species richness is somewhat contrasting with Indonesia, where secondary forests functioned as suitable alternative habitats (Waltert et al. 2005, Imron et al. 2022). In fact, the loss of species we observed in secondary forests was comparably severe as in Indonesian industrial plantations (Prabowo et al. 2016), highlighting the sensitivity of New Guinean insectivores to severe habitat changes.

Frugivores

Forest connectivity was particularly important for frugivore diversity. Rainforest frugivores are a group that is generally highly affected by large-scale anthropogenic disturbance (Sodhi 2008, Newbold et al. 2013). They are highly sensitive to forest fragmentation (Farwig et al. 2017), as many frugivorous species are unable to survive in smaller forest remnants due to loss of key habitat features and large fruiting trees (Ferger et al. 2014, Mueller et al. 2014). In addition, hunting pressure and nest predation typically increase in fragmented forest landscapes, further increasing the vulnerability of large sedentary species (Riegert et al. 2021). We stress the necessity of focusing on all frugivore taxa, as our results contrast somewhat with Marsden and Pilgrim (2003), who may show slightly misleading results by studying opportunistic species benefiting from disturbances. Findings from Sumatra show a decrease in frugivore occurrence in secondary forests that supports our interpretation (Prabowo et al. 2016).

We report that the species richness of frugivores decreased similarly in primary fragments and clear-cut secondary forests, when compared to continuous primary forest. However, the assemblages of frugivores in secondary forests often contained species that do not prefer the canopy, such as the Papuan King Parrot, or that inhabit forest edges and regrowth vegetation, e.g., Orange-bellied Fruit Dove (Billerman et al. 2022), in contrast to birds in forest fragments, which shared many species with primary forest. These results suggest that while vegetation characteristics influenced the species composition of frugivore groups, overall species richness was determined by the lack of forest connectivity (Moran et al. 2004), possibly associated with hunting or predation (Chmel et al. 2018). The relatively low importance of vegetation complexity for frugivore diversity appears to be region-specific for PNG (Sam et al. 2019), in contrast to other areas where it is a major determinant of frugivore species richness (Morante-Filho et al. 2018). Proximity to primary forest is probably the reason for the high species richness in secondary forests after subsistence agriculture (Mayhew et al. 2019). While these frugivores may show avoidance of large areas of fragmented forests, they are capable of dispersing for large distances within the boundaries of primary forest (Mack and Wright 2005).

Nectarivores

Nectarivores were the only feeding guild that increased in density in fragmented primary and post clear-cut secondary forests. In PNG, obligatory nectarivores form a species-poor assemblage, with their diversity mostly constant in natural forest ecosystems (Sam et al. 2019). However, previous research in PNG showed that both the landscape context and vegetation structure influences nectarivore abundances in human-modified areas (Marsden et al. 2006). We suggest that this is the case for our study. In secondary forests adjacent to forest fragments, the lower canopy density, together with increased shrub and herb cover, likely offered an increased supply of flowering plants (Blake and Loiselle 2001).

Simultaneously, nectarivores often have large home ranges, moving around in search of flowers (e.g., Helmeted Friarbird, Billerman et al. 2022). They can, therefore, disperse more easily and find more food sources in the open secondary forest matrix than in continuous primary forest. In addition, secondary forests are often in the vicinity of small village settlements with home-gardens (Sam et al. 2014), and gardens typically contain a higher density of flowering plants than continuous rainforests (Hagen and Kraemer 2010), suitable for generalist nectarivores. These and other characteristics of human-modified forests likely drive their suitability for nectarivores both in New Guinea and neighboring islands (Imron et al. 2022).

Research and conservation implications

In this study, we give a broad, guild-based overview of bird communities in a human-modified lowland rainforest landscape. While our findings offer valuable insights, they do not cover the regional variability of bird diversity (Sam et al. 2019). In PNG, environmental gradients create a wide range of forest ecosystems, from lowland alluvial rainforests to montane and cloud forests (Paijmans 1976) to semi-open woodlands and mangroves (Havel 1972). Human pressure on these forests, driven by subsistence-level agriculture and increasingly by large-scale commercial logging (Bryan and Shearman 2015) is significant and varies based on province and historical context (Haberle 2007). To fully understand the impact of human activities on PNG's avian biodiversity, future research should extend toward other areas within this unique landscape.

Guidelines for biodiversity conservation are dependent on baseline knowledge stemming from rigorous case study research (Matuoka et al. 2020, Sodhi 2008). In PNG, this knowledge is almost completely lacking. Our study seeks to address this, by examining birds in human-modified lowland rainforests, which in their natural state host greatest bird diversity across all elevations in PNG (Sam et al. 2019). Within our study areas, we show that large-scale deforestation has strong negative consequences on bird biodiversity. Moreover, we highlight that even large primary forest fragments (up to 1200 ha) only retain slightly greater biodiversity compared to surrounding secondary forests and only for some guilds. Despite this, we agree with Sam et al. (2014), that forest fragments are necessary for future bird conservation. The adjacent secondary forests are subject to regular anthropogenic disturbance that will only increase in magnitude with growing human population, as majority of people in PNG are dependent on home-gardens (Allen and Bourke 2009). Therefore, fragments of primary forest retained by villagers may thus become the only remaining forests in a landscape mostly converted to agricultural use.

Our results also show that the only way to fully protect lowland bird biodiversity is to rigorously maintain connectivity to large reserves of continuous primary forests. Research in other regions has shown that financial incentives for local communities can be the most effective means of preventing disturbance in large-scale forest reserves (Ramsdell et al. 2016). Such incentives may also discourage the sale of large forest areas to predatory logging companies (Novotny 2010). As a developing country, PNG could profit from funding by extra-national organizations for biodiversity conservation and monitoring projects (Turia et al. 2022). However, there remain many issues associated with external funding, such as emphasis on short-term results, poor execution of conservation actions, and corruption (Laurance et al. 2012).

We suggest that additional benefits to particularly avian conservation could be in promotion of citizen science. For example, many other tropical countries already apply the popularity of keystone bird species to foster citizen engagement through banding and bird-watching organizations (Greenwood 2007). PNG is well-regarded for exotic bird-watching prospects, yet amateur ornithology is mostly practiced by foreign visitors (Newsome 2015). Yet local landowners are not opposed to conservation actions, and forest reserves with local fauna are viewed by many as places of cultural significance (Novotny 2010). The main obstacle to citizen science and engagement with amateur ornithology is represented by low standard of living that is joined with increasing criminality and exploitation of native land by foreign agents (Lakhani and Willman 2014). Combating such country-wide issues is mostly beyond the reach of expatriate researchers.

CONCLUSIONS

In response to ongoing deforestation, many tropical regions have seen extensive scientific coverage of human-modified forest ecosystems. There is a need to ensure that other regions, which are at earlier stages of a similar trajectory of forest loss, are not neglected. Furthermore, regional specifics of forest loss progress need to be considered. Our study on bird communities presents one of the few forays into secondary forests in PNG, and highlights the importance of continuous rainforest preservation in the face of increasing deforestation pressure. However, as we focus on only one lowland region of PNG, the elevational and geographical variation in these ecosystems must be investigated for effective countrywide conservation. Bird communities are the ideal study group for monitoring biodiversity in large-scale anthropogenic land use systems replacing natural forests. Therefore, expanding the scope of bird disturbance ecology to other human-modified landscapes in PNG is likely to yield highly valuable information for biodiversity conservation and ecosystem functioning in one of world’s least explored and largest rainforests.

LITERATURE CITED

Abdi, H., and L. J. Williams. 2010. Principal component analysis. WIREs Computational Statistics 2(4):433-459. https://doi.org/10.1002/wics.101

Alin, A. 2010. Multicollinearity. WIREs Computational Statistics 2(3):370-374. https://doi.org/10.1002/wics.84

Allen, B., and R. M. Bourke. 2009. People, land and environment. Food and Agriculture in Papua New Guinea. https://doi.org/10.22459/FAPNG.08.2009.01

Banks-Leite, C., R. M. Ewers, and J.-P. Metzger. 2010. Edge effects as the principal cause of area effects on birds in fragmented secondary forest. Oikos 119(6):918-926. https://doi.org/10.1111/j.1600-0706.2009.18061.x

Bartlett, A. G. 2018. Factors affecting the success of collaborative forestry research in Papua New Guinea. Australian Forestry 81(2):116-128. https://doi.org/10.1080/00049158.2018.1462546

Basset, Y., V. Novotny, S. E. Miller, and R. Pyle. 2000. Quantifying biodiversity: experience with parataxonomists and digital photography in Papua New Guinea and Guyana. BioScience 50(10):899-908. https://doi.org/10.1641/0006-3568(2000)050[0899:QBEWPA]2.0.CO;2

Bates, D., M. Maechler, B. Bolker, S. Walker, R. H. B. Christensen, H. Singmann, B. Dai, G. Grothendieck, P. Green, and M. B. Bolker. 2015. Package ‘lme4’. Convergence 12(1).

Beehler, B. M., and T. K. Pratt 2016. Birds of New Guinea. Princeton University Press, Princeton, New Jersey, USA. https://doi.org/10.1515/9781400880713

Billerman, M., B. K. Keeney, P. G. Rodewald, and T. S. Schulenberg. 2022. Birds of the world. Cornell Lab of Ornithology, Ithaca, New York, USA. https://doi.org/10.2173/bow

Blake, J. G., and B. A. Loiselle. 2001. Bird assemblages in second-growth and old-growth forests, Costa Rica: perspectives from mist nets and point counts. The Auk 118(2):304-326. https://doi.org/10.1093/auk/118.2.304

Blankespoor, G. W. 1991. Slash-and-burn shifting agriculture and bird communities in Liberia, West Africa. Biological Conservation 57(1):41-71. https://doi.org/10.1016/0006-3207(91)90107-K

Borges, S. H. 2007. Bird assemblages in secondary forests developing after slash-and-burn agriculture in the Brazilian Amazon. Journal of Tropical Ecology 23(4):469-477. Cambridge University Press, Cambridge, UK. https://doi.org/10.1017/S0266467407004105

Bradfer-Lawrence, T., N. Gardner, and D. H. Dent. 2018. Canopy bird assemblages are less influenced by habitat age and isolation than understory bird assemblages in Neotropical secondary forest. Ecology and Evolution 8(11):5586-5597. https://doi.org/10.1002/ece3.4086

Brooks, M. E., K. Kristensen, K. J. van Benthem, A. Magnusson, C. W. Berg, A. Nielsen, H. J. Skaug, M. Mächler, and B. M. Bolker. 2017. glmmTMB balances speed and flexibility among packages for zero-inflated generalized linear mixed modeling. R Journal 9:378-400. https://doi.org/10.32614/RJ-2017-066

Brooks, T. M., R. A. Mittermeier, G. A. B. Da Fonseca, J. Gerlach, M. Hoffmann, J. F. Lamoreux, C. G. Mittermeier, J. D. Pilgrim, and A. S. Rodrigues. 2006. Global biodiversity conservation priorities. Science 313(5783):58–61. https://doi.org/10.1126/science.1127609

Bryan, J. and P. L. Shearman. 2015. The state of the forests of Papua New Guinea 2014: measuring change over period 2002-2014. University of Tasmania, Australia.

Buckland, S. T., D. R. Anderson, K. P. Burnham, and J. L. Laake. 2005. Distance sampling. In: Encyclopedia of biostatistics. John Wiley and Sons, Hoboken, New Jersey, USA. https://doi.org/10.1002/0470011815.b2a16019

Burivalova, Z., D. Miteva, N. Salafsky, R. A. Butler, and D. S. Wilcove. 2019. Evidence types and trends in tropical forest conservation literature. Trends in Ecology & Evolution 34(7):669–679. https://doi.org/10.1016/j.tree.2019.03.002

Chen, Y. 2016. Spatial autocorrelation approaches to testing residuals from least squares regression. PLoS ONE 11:e0146865. https://doi.org/10.1371/journal.pone.0146865

Chmel, K., J. Riegert, L. Paul, M. Mulau, K. Sam, and V. Novotny. 2018. Predation on artificial and natural nests in the lowland rainforest of Papua New Guinea. Bird Study 65(1):114-122. https://doi.org/10.1080/00063657.2017.1420751

Corlett, R. T. 1994. What is secondary forest? Journal of Tropical Ecology 10(3):445–447. https://doi.org/10.1017/S0266467400008129

Corlett, R. T. and R. B. Primack. 2006. Tropical rainforests and the need for cross-continental comparisons. Trends in Ecology & Evolution 21:104–110. https://doi.org/10.1016/j.tree.2005.12.002

Creagh, A., J. Fawcett, D. Jones. 2004. Shrub density correlates with densities of white-browed scrubwrens (Sericornis frontalis) in three forest types in south-eastern Queensland. Corella 28(1):20-21.

Davies-Colley, R. J., G. W. Payne, and M. van Elswijk. 2000. Microclimate gradients across a forest edge. New Zealand Journal of Ecology 24(2):111-121.

Dean, C. B., and J. D. Nielsen. 2007. Generalized linear mixed models: a review and some extensions. Lifetime Data Analysis 13:497-512. https://doi.org/10.1007/s10985-007-9065-x

Didham, R. K., and J. H. Lawton. 1999. Edge structure determines the magnitude of changes in microclimate and vegetation structure in tropical forest fragments. Biotropica 31(1):17-30. https://doi.org/10.1111/j.1744-7429.1999.tb00113.x

Dormann, C. F., S. J. Schymanski, J. Cabral, I. Chuine, C. Graham, F. Hartig, M. Kearney, X. Morin, C. Römermann, B. Schröder, A. Singer. 2012. Correlation and process in species distribution models: bridging a dichotomy. Journal of Biogeography 39(12):2119-2131. https://doi.org/10.1111/j.1365-2699.2011.02659.x

ESRI. 2024. ArcGIS Pro Desktop: Version 3.3. Environmental Systems Research Institute, Redlands, CA, USA.

Farwig, N., D. G. Schabo, and J. Albrecht. 2017. Trait-associated loss of frugivores in fragmented forest does not affect seed removal rates. Journal of Ecology 105(1):20-28. https://doi.org/10.1111/1365-2745.12669

Ferger, S.W., M. Schleuning, A. Hemp, K.M. Howell, and K. Böhning-Gaese. 2014. Food resources and vegetation structure mediate climatic effects on species richness of birds. Global Ecology and Biogeography 23(5):541-549. https://doi.org/10.1111/geb.12151

Fox, J. C. 2011. Native forest management in Papua New Guinea: advances in assessment, modelling and decision-making. Canberra (Australia) ACIAR.

Fox, J. 2015. Applied regression analysis and generalized linear models. Sage, Thousand Oaks, California, USA.

Gamoga, G., R. Turia, H. Abe, M. Haraguchi, and O. Iuda. 2021. The forest extent in 2015 and the drivers of forest change between 2000 and 2015 in Papua New Guinea: deforestation and forest degradation in Papua New Guinea. Case Studies in the Environment 5(1):1442018. https://doi.org/10.1525/cse.2021.1442018

Gill F., D. Donsker, and P. Rasmussen, editors. 2024. IOC World Bird List (v14.1).

Greenwood, J. J. D. 2007. Citizens, science and bird conservation. Journal of Ornithology 148:77-124. https://doi.org/10.1007/s10336-007-0239-9

Haberle, S. G. 2007. Prehistoric human impact on rainforest biodiversity in highland New Guinea. Philosophical Transactions of the Royal Society B: Biological Sciences 362(1478):219-228. https://doi.org/10.1098/rstb.2006.1981

Hagen, M., and M. Kraemer. 2010. Agricultural surroundings support flower-visitor networks in an Afrotropical rainforest. Biological Conservation 143(7):1654-1663. https://doi.org/10.1016/j.biocon.2010.03.036

Hartig, F. 2018. DHARMa: residual diagnostics for hierarchical (multi-level / mixed) regression models. R Package version 020. https://doi.org/10.32614/CRAN.package.DHARMa

Havel, J. J. 1972. New Guinea forests: structure, composition and management. Australian Forestry 36(1):24–37. https://doi.org/10.1080/00049158.1972.10675567

Heffner, R. A., M. J. Butler, and C. K. Reilly. 1996. Pseudoreplication revisited. Ecology 77(8):2558-2562. https://doi.org/10.2307/2265754

Huang, A. 2017. Mean-parametrized Conway-Maxwell-Poisson regression models for dispersed counts. Statistical Modelling 17(6):359-380. SAGE Publications, India. https://doi.org/10.1177/1471082X17697749

Imron, M. A., M. Campera, D. Al Bihad, F. D. Rachmawati, F. E. Nugroho, B. Budiadi, K. F. Wianti, E. Suprapto, V. Nijman, and K. A. I. Nekaris. 2022. Bird assemblages in coffee agroforestry systems and other human modified habitats in Indonesia. Biology 11(2):310. https://doi.org/10.3390/biology11020310

Johns, A. D. 1991. Responses of Amazonian rain forest birds to habitat modification. Journal of Tropical Ecology 7(4):417-437. https://doi.org/10.1017/S0266467400005812

Lakhani, S. and A. M. Willman 2014. Drivers of crime and violence in Papua New Guinea. Research and dialogue series no. 2. The World Bank Group, Washington, D.C., USA.

Laurance, W. F., T. Kakul, M. Tom, R. Wahya, and S. G. Laurance. 2012. Defeating the ‘resource curse’: key priorities for conserving Papua New Guinea’s native forests. Biological Conservation 151(1):35-40. https://doi.org/10.1016/j.biocon.2011.10.037

Laurance, W. F. 2015. Emerging threats to tropical forests. BioOne Digital Library 100(3):159–169. https://doi.org/10.3417/2011087

Lenth, R. 2022. emmeans: Estimated marginal means, aka least-squares means. R package version 1.7.2. https://doi.org/10.32614/CRAN.package.emmeans

Mack, A. L., and D. D. Wright 2005. The frugivore community and the fruiting plant flora in a New Guinea rainforest: identifying keystone frugivores. Pages 185-203 in J. L. Dew and J. P. Boubli, editors. Tropical fruits and frugivores: the search for strong interactors. Springer Netherlands, Dordrecht. https://doi.org/10.1007/1-4020-3833-X_10

Marsden, S. J., and J. D. Pilgrim. 2003. Factors influencing the abundance of parrots and hornbills in pristine and disturbed forests on New Britain, PNG. Ibis 145(1):45-53. https://doi.org/10.1046/j.1474-919X.2003.00107.x

Marsden, S. J., C. T. Symes, and A. L. Mack. 2006. The response of a New Guinean avifauna to conversion of forest to small-scale agriculture. Ibis 148(4):629-640. https://doi.org/10.1111/j.1474-919X.2006.00577.x

Marsden, S. J., and C. T. Symes. 2008. Bird richness and composition along an agricultural gradient in New Guinea: the influence of land use, habitat heterogeneity and proximity to intact forest. Austral Ecology 33(6):784–793. https://doi.org/10.1111/j.1442-9993.2008.01850.x

Martin, T. E., and G. A. Blackburn. 2014. Conservation value of secondary forest habitats for endemic birds, a perspective from two widely separated tropical ecosystems. Ecography 37:250-260. https://doi.org/10.1111/j.1600-0587.2013.00234.x

Matuoka, M. A., M. Benchimol, J. M. de Almeida-Rocha, and J. C. Morante-Filho. 2020. Effects of anthropogenic disturbances on bird functional diversity: a global meta-analysis. Ecological Indicators 116:106471. https://doi.org/10.1016/j.ecolind.2020.106471

Mayhew, R. J., J. A. Tobias, L. Bunnefeld, and D. H. Dent. 2019. Connectivity with primary forest determines the value of secondary tropical forests for bird conservation. Biotropica 51(2):219-233. https://doi.org/10.1111/btp.12629

Mcdonald, T., J. Carlisle and A. Mcdonald. 2019. Rdistance: distance-sampling analyses for density and abundance estimation. R package version 2.1.3. https://doi.org/10.32614/CRAN.package.Rdistance

Moore, R. P., W. D. Robinson, I. J. Lovette, and T. R. Robinson. 2008. Experimental evidence for extreme dispersal limitation in tropical forest birds. Ecology Letters 11(9):960-968. https://doi.org/10.1111/j.1461-0248.2008.01196.x

Moran, C., C. Catterall, R. Green, and M. Olsen. 2004. Functional variation among frugivorous birds: implications for rainforest seed dispersal in a fragmented subtropical landscape. Oecologia 141:584-595. https://doi.org/10.1007/s00442-004-1685-1

Morante-Filho, J. C., V. Arroyo-Rodríguez, M. de S. Pessoa, E. Cazetta and D. Faria. 2018. Direct and cascading effects of landscape structure on tropical forest and non-forest frugivorous birds. Ecological Applications 28(8):2024-2032. https://doi.org/10.1002/eap.1791

Morris, R. J. 2010. Anthropogenic impacts on tropical forest biodiversity: a network structure and ecosystem functioning perspective. Philosophical Transactions of the Royal Society B: Biological Sciences 365(1558):3709–3718. https://doi.org/10.1098/rstb.2010.0273

Mueller, T., J. Lenz, T. Caprano, W. Fiedler, and K. Böhning-Gaese. 2014. Large frugivorous birds facilitate functional connectivity of fragmented landscapes. Journal of Applied Ecology 51(3):684-692. https://doi.org/10.1111/1365-2664.12247

Mulvaney, J. M., and M. I. Cherry. 2020. The effectiveness of point counts and mist-netting in surveying Afrotemperate forest bird community structure in South Africa. Emu - Austral Ornithology 120(2):103–113. https://doi.org/10.1080/01584197.2020.1726186

Newbold, T., J. P. W. Scharlemann, S. H. M. Butchart, Ç.H. Şekercioğlu, R. Alkemade, H. Booth, and D. W. Purves. 2013. Ecological traits affect the response of tropical forest bird species to land-use intensity. Proceedings of the Royal Society B: Biological Sciences 280:20122131. https://doi.org/10.1098/rspb.2012.2131

Newsome, D. 2015. Conflicts between cultural attitudes, development and ecotourism: the case of bird watching tours in Papua New Guinea. Pages 194-210 in K. Markwell, editor. Animals and tourism: understanding diverse relationships. Channel View Publications, Bristol, UK. https://doi.org/10.21832/9781845415051-016

Novotny, V. 2010. Rain forest conservation in a tribal world: why forest dwellers prefer loggers to conservationists. Biotropica 42(5):546-549. https://doi.org/10.1111/j.1744-7429.2010.00658.x

Paijmans, K. 1976. New Guinea vegetation. Elsevier Scientific, Amsterdam, Netherlands.

Pangau-Adam, M., M. Mühlenberg, and M. Waltert. 2015. Rainforest disturbance affects population density of the northern cassowary Casuarius unappendiculatus in Papua, Indonesia. Oryx 49(4):735–742. https://doi.org/10.1017/S0030605313001464

Powell, L. L., J. D. Wolfe, E. I. Johnson, J. E. Hines, J. D. Nichols, and P. C. Stouffer. 2015. Heterogeneous movement of insectivorous Amazonian birds through primary and secondary forest: a case study using multistate models with radiotelemetry data. Biological Conservation 188:100-108. https://doi.org/10.1016/j.biocon.2015.01.028

Powell, L. L., P. C. Stouffer, and E. I. Johnson. 2013. Recovery of understory bird movement across the interface of primary and secondary Amazon rainforest. Auk 130(3):459-468. https://doi.org/10.1525/auk.2013.12202

Prabowo, W. E., K. Darras, Y. Clough, M. Toledo-Hernandez, R. Arlettaz, Y. A. Mulyani, and T. Tscharntke. 2016. Bird responses to lowland rainforest conversion in Sumatran smallholder landscapes, Indonesia. PLoS ONE 11:e0154876. https://doi.org/10.1371/journal.pone.0154876

R Core Team. 2024. R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria.

Ramos, D. L., M. A. Pizo, M. C. Ribeiro, R. S. Cruz, J. M. Morales, and O. Ovaskainen. 2020. Forest and connectivity loss drive changes in movement behavior of bird species. Ecography 43(8):1203-1214. https://doi.org/10.1111/ecog.04888

Ramsdell, C. P., M. G. Sorice, and A. M. Dwyer. 2016. Using financial incentives to motivate conservation of an at-risk species on private lands. Environmental Conservation 43(1):34-44. https://doi.org/10.1017/S0376892915000302

Renjifo, L. M. 1999. Composition changes in a Subandean avifauna after long-term forest fragmentation. Conservation Biology 13(5):1124–1139. https://doi.org/10.1046/j.1523-1739.1999.98311.x

Riegert, J., K. Chmel, J. Vlček, Z. Hrázský, O. Sedláček, S. Grill, J. Reif, and D. Horák 2021. Alarming declines in bird abundance in an Afromontane global biodiversity hotspot. Biodiversity and Conservation 30:3385-3408. https://doi.org/10.1007/s10531-021-02252-1

Saab, V. 1999. Importance of spatial scale to habitat use by breeding birds in riparian forests: a hierarchical analysis. Ecological Applications 9(1):135-151. https://doi.org/10.1890/1051-0761(1999)009[0135:IOSSTH]2.0.CO;2

Salgueiro, P.A., F. Valerio, C. Silva, A. Mira, J.E. Rabaça and S.M. Santos. 2021. Multispecies landscape functional connectivity enhances local bird species’ diversity in a highly fragmented landscape. Journal of Environmental Management 284:112066. https://doi.org/10.1016/j.jenvman.2021.112066

Sam, K., L. R. Jorge, B. Koane, R. J. Hazell, P. L. Shearman, and V. Novotny. 2024. Reorganization of bird communities along a rainforest elevation gradient during a strong El Niño event in Papua New Guinea. Ecology and Evolution 14(5):e10955. https://doi.org/10.1002/ece3.10955

Sam, K., B. Koane, S. Jeppy, and V. Novotny. 2014. Effect of forest fragmentation on bird species richness in Papua New Guinea. Journal of Field Ornithology 85(2):152–167. https://doi.org/10.1111/jofo.12057

Sam, K., B. Koane, D.C. Bardos, S. Jeppy, and V. Novotny. 2019. Species richness of birds along a complete rain forest elevational gradient in the tropics: habitat complexity and food resources matter. Journal of Biogeography 46(2):279-290. https://doi.org/10.1111/jbi.13482

Sam, K., B. Koane, S. Jeppy, J. Sykorova and V. Novotny. 2017. Diet of land birds along an elevational gradient in Papua New Guinea. Scientific Reports 7:44018. https://doi.org/10.1038/srep44018

Sekercioglu, C. H. 2002. Forest fragmentation hits insectivorous birds hard. Directions in Science 1:62-64. https://www.researchgate.net/publication/228764669_Forest_Fragmentation_Hits_Insectivorous_Birds_Hard

Sekercioglu, C. H. 2012. Bird functional diversity and ecosystem services in tropical forests, agroforests and agricultural areas. Journal of Ornithology 153:153-161. https://doi.org/10.1007/s10336-012-0869-4

Sekhran, N., and S. E. Miller, editors. 1995. Papua New Guinea country study on biological diversity. Dept. of Environment and Conservation, Conservation Resource Centre; Centre for Resources and Environment, Waigani, Papua New Guinea.

Shearman, P., and J. Bryan. 2011. A bioregional analysis of the distribution of rainforest cover, deforestation and degradation in Papua New Guinea. Austral Ecology 36(1):9–24. https://doi.org/10.1111/j.1442-9993.2010.02111.x

Silva, R. R. Da, B. Vilela, D. P. Silva, A. F. A. De Andrade, P. V. Cerqueira, G. S. R. Gonçalves, and M. P. D. Santos. 2020. The role of ecological niche evolution on diversification patterns of birds distinctly distributed between the Amazonia and Atlantic rainforests. PLoS ONE 15:e0238729. https://doi.org/10.1371/journal.pone.0238729

Šmilauer, P. and J. Lepš. 2014. Multivariate analysis of ecological data using CANOCO 5. Cambridge University Press, Cambridge, UK. https://doi.org/10.1017/CBO9781139627061

Sodhi, N. S., M. R. C. Posa, T. M. Lee, and I. G. Warkentin. 2008. Perspectives in ornithology: effects of disturbance or loss of tropical rainforest on birds. Auk 125(3):511–519. https://doi.org/10.1525/auk.2008.1708

Stratford, J. A., and P. C. Stouffer. 1999. Local extinctions of terrestrial insectivorous birds in a fragmented landscape near Manaus, Brazil. Conservation Biology 13(6):1416-1423. https://doi.org/10.1046/j.1523-1739.1999.98494.x

ter Braak, C. J. F., and P. Smilauer. 2012. Canoco reference manual and user’s guide: software for ordination, version 5.0. Microcomputer Power. Ithaca, New York, USA.

Thiollay, J.-M. 1997. Disturbance, selective logging and bird diversity: a Neotropical forest study. Biodiversity & Conservation 6:1155–1173. https://doi.org/10.1023/A:1018388202698

Turubanova, S., P. V. Potapov, A. Tyukavina, and M. C. Hansen. 2018. Ongoing primary forest loss in Brazil, Democratic Republic of the Congo, and Indonesia. Environmental Research Letters 13(7):074028. https://doi.org/10.1088/1748-9326/aacd1c

Turia R., G. Gamoga, H. Abe, V. Novotny, F. Attorre, and L. Vesa. 2022. Monitoring the multiple functions of tropical rainforest on a national scale: an overview from Papua New Guinea. Case Studies in the Environment 6(1):1547792. https://doi.org/10.1525/cse.2021.1547792

Tvardíková, K. 2010. Bird abundances in primary and secondary growths in Papua New Guinea: a preliminary assessment. Tropical Conservation Science 3(4):373–388. https://doi.org/10.1177/194008291000300403

Waltert, M., K. S. Bobo, N. M. Sainge, H. Fermon, and M. Mühlenberg. 2005. From forest to farmland: habitat effects on Afrotropical forest bird diversity. Ecological Applications 15(4):1351–1366 https://doi.org/10.1890/04-1002

Watson, J. E., R. J. Whittaker, and T. P. Dawson. 2004. Habitat structure and proximity to forest edge affect the abundance and distribution of forest-dependent birds in tropical coastal forests of southeastern Madagascar. Biological Conservation 120(3):311-327. https://doi.org/10.1016/j.biocon.2004.03.004

Whitworth, A., C. Beirne, E. Flatt, R. Pillco Huarcaya, J. C. Cruz Diaz, A. Forsyth, P. K. Molnár, J. S. Vargas Soto. 2018. Secondary forest is utilized by Great Curassows (Crax rubra) and Great Tinamous (Tinamus major) in the absence of hunting. Condor 120(4):852-862. https://doi.org/10.1650/CONDOR-18-57.1