Birds both promote and prosper from forest restoration. The ecosystem functions birds perform can increase the pace of forest regeneration and, correspondingly, increase the available habitat for birds and other forest-dependent species. The aim of this study was to learn how tropical forest restoration treatments interact with landscape tree cover to affect the structure and composition of a diverse bird assemblage. We sampled bird communities over two years in 13 restoration sites and two old-growth forests in southern Costa Rica. Restoration sites were established on degraded farmlands in a variety of landscape contexts, and each included a 0.25-ha plantation, island treatment (trees planted in patches), and unplanted control. We analyzed four attributes of bird communities including frugivore abundance, nectarivore abundance, migrant insectivore richness, and compositional similarity of bird communities in restoration plots to bird communities in old-growth forests. All four bird community variables were greater in plantations and/or islands than in control treatments. Frugivore and nectarivore abundance decreased with increasing tree cover in the landscape surrounding restoration plots, whereas compositional similarity to old-growth forests was greatest in plantations embedded in landscapes with high tree cover. Migrant insectivore richness was unaffected by landscape tree cover. Our results agree with previous studies showing that increasing levels of investment in active restoration are positively related to bird richness and abundance, but differences in the effects of landscape tree cover on foraging guilds and community composition suggest that trade-offs between biodiversity conservation and bird-mediated ecosystem functioning may be important for prioritizing restoration sites.
Citation: Reid JL, Mendenhall CD, Rosales JA, Zahawi RA, Holl KD (2014) Landscape Context Mediates Avian Habitat Choice in Tropical Forest Restoration. PLoS ONE 9(3): e90573. doi:10.1371/journal.pone.0090573
Editor: Nina Farwig, University of Marburg, Germany
Received: November 6, 2013; Accepted: February 3, 2014; Published: March 4, 2014
Copyright: © 2014 Reid et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: Financial support was provided by the US National Science Foundation (NSF-DEB 0918112 and graduate fellowships to J.L.R. and C.D.M.), Peter and Helen Bing, Ralph and Louise Haberfield, the Moore Family Foundation, the Koret Foundation, the Mertz Gilmore Foundation, and the Winslow Foundation. Funding for Open Access provided by the University of California Santa Cruz Open Access Fund. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Ecological restoration is the process of assisting the recovery of degraded ecosystems to their historic trajectories . Interventions such as tree planting are effective for restoring biodiversity (e.g., species diversity, abundances, or biomass) and ecosystem services (e.g., nutrient cycling, soil stabilization, or climate regulation) in tropical terrestrial ecosystems . Tropical restoration efforts help offset the impacts of ongoing deforestation  that threaten to exacerbate climate change and drive extinctions in the world's richest biological communities .
Restoration projects are spatially explicit, but rarely replicated across landscapes due to high implementation costs . As such, our understanding of the importance of landscape context on the restoration of communities and their associated ecosystem functions and societal benefits is limited to a relatively small number of studies –. Nonetheless, funding for tropical forest restoration is increasingly available from national payment for ecosystem services programs and climate change mitigation initiatives such as Reducing Emissions from Deforestation and Forest Degradation (REDD+) . Better understanding of landscape effects in tropical forest restoration will help inform sub-national prioritization criteria to effectively allocate limited funding and conform to environmental safeguards , .
We chose birds for this study because they are both beneficiaries and benefactors of forest restoration. Many tropical bird populations are in decline because of habitat loss , , but forest restoration aims to reverse this trend through gains in habitat . Correspondingly, birds provide ecosystem functions that reduce biotic barriers to forest succession –, including, seed dispersal –, increased germination via gut passage , herbivorous arthropod control , and pollination , .
Early research on birds and tropical forest restoration showed that habitat structures such as isolated trees – and artificial perches – increased bird visitation and seed dispersal in degraded habitats, but isolated trees are not always available and seedling recruitment rarely increased below artificial perches . Additional research demonstrated that mixed-species tree plantations both attracted seed dispersers and improved conditions for seed germination and seedling survival –. Seed dispersal trends were driven by small omnivores , , and the effects of tree planting varied by feeding guild , ,  and habitat associations , . Still, knowledge gaps remain . In particular, few tropical forest restoration studies have been sufficiently replicated across the landscape to assess interactions between local restoration treatments and landscape context , .
The aim of this research was to learn how tropical forest restoration treatments interact with landscape tree cover to affect bird visitation and community composition in regenerating farmlands. We sought to identify differences in explanatory factors for avian foraging guilds versus birds associated with old-growth forest. We addressed these questions in a replicated restoration experiment in southern Costa Rica and found that both local treatments and landscape context were important predictors of bird community structure and composition.
Materials and Methods
Permission to conduct this research was granted by the Costa Rican Ministry of the Environment and Energy. Permission to work on private lands was granted by all landowners. This research did not require approval for animal care and use because it was an observational field study that did not involve the capture or handling of wild animals nor their maintenance in captivity.
We sampled bird communities in 13 restoration sites and two old-growth forests in southern Costa Rica (canton of Coto Brus). Restoration sites were located between the Las Cruces Biological Reserve (8° 47′ N, 82° 57′ W) and the town of Agua Buena (8° 44′ N, 82° 56′ W). Study sites were 1100–1400 m a.s.l., and the dominant natural ecosystem was premontane moist forest . Most study sites were on the Fila Cruces, but one old-growth site was at similar elevation in the Talamanca Mountains. Precipitation across the study areas varies with topography but is ~3.4 m y−1, with a mean annual temperature of 21°C at the Las Cruces Biological Station.
The study region was settled in the 1950s–80s by farmers from the Central Valley of Costa Rica and colonists from southern Italy . Pioneers cleared the contiguous forest for coffee production, but many converted their coffee plantations to cattle pastures when coffee prices sunk in the 1990s . The study area is currently a diverse mix of agricultural fields, coffee plantations, cattle pastures, small urban centers, and several types of forest elements including forest fragments, riparian forests, fencerows, and isolated trees. Forest elements comprise ~35% of the study region .
Restoration sites were established on degraded farmlands (mostly pastures) in 2004–2006. Each site included three 50×50 m plots (N = 39 plots), which were cleared of vegetation and randomly assigned to one of three treatments (Fig. 1). Controls were allowed to regenerate naturally; islands (i.e., applied nucleation treatments) were planted with six tree seedling patches, or “islands”, of various sizes; and contiguous plantations were planted with rows of seedlings across the entire plot. The restoration treatments spanned a gradient of intervention intensity; tree seedling density (seedlings 0.25 ha−1) ranged from zero seedlings in controls, 86 in islands, to 313 in plantations. Seedlings planted in islands and plantations were a mix of two native timber species, Terminalia amazonica (J.F. Gmel.) Exell (Combretaceae) and Vochysia guatemalensis Donn. Sm. (Vochysiaceae), and two naturalized legumes, Erythrina poeppigiana Walp. Skeels and Inga edulis Mart. (Fabaceae). All restoration plots were cleared with machetes at ~3-mo intervals for 2.5 y to allow planted seedlings to grow above existing grasses and forbs. Treatments had been in place for 3–7 y when bird counts took place at which time there were already large differences in vegetation structure and plant composition between treatments. See Cole et al.  for detailed descriptions of the restoration treatments and planted tree species selection.
Figure 1. Local restoration treatments.
Restoration treatments randomly applied to plots of heavily degraded pasture at 13 sites in 2004–2006 in southern Costa Rica. Green denotes areas planted with seedlings. Control plots were cleared of vegetation and allowed to regenerate naturally; islands were cleared and planted with 86 seedlings of four species in six patches (two each 4×4, 8×8, 12×12 m); plantations were cleared and planted in uniform rows throughout the plot (313 seedlings).doi:10.1371/journal.pone.0090573.g001
Bird communities in restoration treatments were compared to a reference community sampled from two old-growth forests. The first old-growth forest site (Las Cruces: 8° 47′ N, 82° 57′ W) was located 1.2–8.9 km from restoration sites, and the second (La Amistad: 8° 56′ N, 82° 50′ W) was located 21.5–27.7 km from restoration sites. Each old-growth forest hosted six plots to sample the reference bird community (N = 12 plots).
Birds in restoration sites were surveyed seven times by a single observer, J.A.R., in Dec 2009 and Apr, Jul, and Nov 2010 and 2011. Each experimental plot at each site was actively searched in a random order for 20 min per observation, and all birds seen or heard within the plot were recorded. Old-growth forests were surveyed using point counts in May-Sep 2010. Each point count was conducted for 30 min by J. Figueroa-Sandí within a maximum radius of 50 m. Surveying methods differed because old-growth forests were initially sampled for a separate study . We used these data to calculate an index of community similarity, but we did not directly compare bird communities in restoration and old-growth. All surveys were conducted from sunrise (~5:30 AM) until 9:00 AM. Birds flying over sites without using them were excluded from analyses. Bird taxonomy follows the American Ornithologists' Union Checklist of the Birds of North America and its supplements .
We assigned bird species to three guilds (frugivores, migrant insectivores, nectarivores) based on published dietary descriptions . We selected frugivores, migrant insectivores, and nectarivores based on their effects on ecosystem functions that benefit society (Table S1 in File S1). Frugivore abundance was a strong predictor of seed rain richness (r2 = 0.95) and abundance (r2 = 0.71) in the study area  (but see ), and migrant insectivore richness was the best community predictor of arthropod reduction in a coffee agroforestry system in southern Mexico (r2 = 0.64) . We assumed that nectarivore abundance would have greater power than species richness for predicting pollination because vertebrate pollination networks are characterized by low dependency . Many species were omnivores, and some were classified in more than one guild. Because we were interested in guilds as they relate to ecosystem functioning, frugivores were limited to fruit-eating birds that also disperse viable seeds. Thus, seed predators (i.e., Psittacids) and species that eat fruit but rarely defecate viable seeds in the study region were not included .
Observational data from restoration sites and trait data are publicly archived at https://merritt.cdlib.org/m/ucsc_lib_hollzahawi.
Landscape tree cover
We used landscape tree cover as a measure of landscape context because it has performed well in previous studies , , . Landscape tree cover was classified by manually digitizing aerial photographs from 2003 and 2005 with 2-m resolution . Tree cover includes primary and secondary forest fragments of all sizes, single trees, early secondary growth, live fences, hedgerows, non-native timber and fruit tree plantations, and nonnative garden ornamentals. We calculated the percentage of landscape tree cover within 36 concentric rings around each restoration treatment (every 10 m from 10–200 m, every 50 m from 250–1000 m). This range of spatial scales includes scales that have been found to be relevant to bird communities in other studies , , .
To compare community composition of birds in restoration plots and old-growth forests, we used a presence-absence matrix to calculate a Sørensen similarity index: QSij = 2Cij/Si+Sj, where C is the number of species in common between sites i and j and S is the total number of species at a given site . We selected a presence-absence similarity index rather than one based on abundance in order to account for differences in survey methods in old-growth forests and restoration sites. Migratory songbirds were excluded from the similarity analysis due to different sampling seasons for restoration sites and old-growth forests. No other seasonal trends were evident. We used a Mantel test of a similarity matrix and distance matrix to evaluate potential for spatial autocorrelation. After removing an outlying old-growth forest site (La Amistad) that was >21 km from all restoration sites, similarity values between sites were not explained by proximity (r = 0.19, p = 0.116, 9999 permutations).
We analyzed bird communities using linear mixed-effects models and maximum likelihood model selection , . Response variables included frugivore and nectarivore abundance (detections 20-min observation−1), migrant insectivore richness (total number of species observed over two years of sampling), and compositional similarity of birds in restoration plots to birds in old-growth forest (QS).
General model structure was y = β0+β1x1+/×β2x2+si+εi where y is one of four bird community response variables, β0 is the y-intercept, β1 and β2 are vectors of fixed-effect coefficients for the three restoration treatments (x1) and landscape tree cover (x2), si is the random effect for the ith site, and εi is the error term. Responses were modeled using log-link for Poisson-distributed responses (frugivore and nectarivore abundance and migrant insectivore richness) and identity-link for normally-distributed responses (compositional similarity to old-growth forest). Spatial and temporal non-independence of treatments within sites were modeled as nested random effects.
For each response variable, we tested models that included fixed effects for: (1) restoration treatment only, (2) restoration treatment + tree cover, and (3) restoration treatment × tree cover. For model types that included tree cover, we compared 36 individual models with tree cover calculated within different-sized concentric rings (10–1000 m radius) around each experimental plot at each site. To avoid collinearity, we used a separate model for each buffer scale. We used Akaike Information Criterion (AIC) scores and weights corrected for small sample sizes (AICc) to select the best model from each set. We report the model with the lowest AICc score as well as the range of buffer scales that resulted in models with ΔAICc<2. Effects of local restoration treatments on avian communities were also analyzed using non-parametric Kruskal-Wallis and Wilcoxon rank sum tests with Bonferroni corrections. For the frugivore abundance analysis, we excluded one site from the model selection procedure because it had a disproportionate influence for models with small tree cover radii. Significance of individual fixed factors was assessed by removing one factor from the model and comparing AICc scores. To assess whether patterns observed at the community level made good predictions at the individual species level, we inspected plots of the most abundant species in each group across sites and tree cover gradients. Analyses were conducted in R 2.15.0  using the lme4 package .
Our analytical approach addresses several common criticisms of bird community studies in conservation biology. First, species richness metrics and forest dependency indices are superficial measures of biodiversity and its responses to change , . We avoided this issue by using similarity to old-growth forest as a response variable rather than abundance or richness of forest-dependent birds. Second, recent studies have highlighted heterogeneous bird detectability in different habitats, which may confound cross-habitat comparisons . We addressed this problem by intensively surveying small areas, where detection probability was likely high and assumptions of occupancy models could not be met. Implications of this decision are considered in the discussion. Third, many studies compare birds across small spatial scales and are biased by spatial autocorrelation . Sites in this study were sufficiently spaced (>700 m separation), but treatments within sites were separated by only ~10–200 m. This spatial arrangement was ideal for our study because we were interested in relative habitat visitation by birds presented with a choice of three restoration treatments at each site. Spatial non-independence was therefore modeled as a nested random effect. Finally, studies evaluating vertebrate responses to small habitat manipulations have unique challenges and should be explicit about the inferences that can be made . We do not assume that any birds complete their life cycle within the restoration sites that we studied or that these interventions have restored bird communities per se. Rather, we infer that bird visitation denotes that restored habitat supports one or more aspects of a bird's ecology.
We observed 3852 bird visitations to restoration sites representing 125 species from 29 families (Figures S1–S2 in File S1). Avian guilds and compositional similarity to old-growth forest differed among local restoration treatments (all Χ2≥14; p<0.001). Plantations had significantly greater frugivore and nectarivore abundance, migrant insectivore richness, and compositional similarity to old-growth forest birds than controls (all p<0.01) and islands were intermediate (Table S2 in File S1).
Supported models included local restoration treatments and landscape tree cover for frugivore abundance (range of tree cover buffer scales with ΔAICc<2: 300–900 m; best-fit buffer scale: 350 m), nectarivore abundance (range: 250–950 m; best: 450 m), and compositional similarity to old-growth forest (range: 170–250, 350–750 m; best: 550 m) (Fig. 2, Table S3–4 in File S1). Migrant insectivore richness was only predicted by local restoration treatment (Table 1).
Figure 2. Landscape tree cover model comparison.
Model comparison for community similarity to old-growth forest (diamonds) and foraging guilds (frugivores = circles; nectarivores = squares; migrant insectivores = triangles) predicted by tree cover at varying buffer distances around restoration sites. AICc represents an Akaike Information Criterion score corrected for small sample sizes. ΔAICc represents the difference in AICc scores between a given model and the model with the lowest AICc.doi:10.1371/journal.pone.0090573.g002
Table 1. Model estimates for bird community attributes in tropical forest restoration.doi:10.1371/journal.pone.0090573.t001
The best-fit model for predicting compositional similarity to old-growth forest included an interaction between local restoration treatment and landscape tree cover (Fig. 3A). Bird communities in plantations more closely resembled bird communities in old-growth forest plots when there was greater landscape tree cover within a radius of 550 m (adj. r2 = 0.83, p<0.001; Table 1, Tables S3-S4 in File S1). An outlying island plot with low tree cover and relatively high compositional similarity appeared to drive results for island plots (Fig. 3A). When we analyzed the data with this outlier removed, the coefficient for tree cover effect on island plots was greater (0.122±0.079 SE) than with the outlier included (0.023±0.077), however this difference had little impact over the range of possible tree cover values. Results were not substantially different when we ran the analysis excluding two sites that were immediately adjacent to old-growth forest patches.
Figure 3. Compositional and functional attributes of bird communities predicted by restoration treatments and landscape tree cover.
Error bars represent ±1 s.e.m. (A–C) Controls are represented by squares; islands by triangles, and plantations by circles. (A) Sørensen similarity of bird communities in restoration sites to bird communities in reference forest (550 m tree cover buffer); (B) frugivore abundance per observation (7 observations per point; 350 m buffer); (C) nectarivore abundance per observation (450 m buffer); (D) migrant insectivore richness (equal sampling intensity). Significance calculated using pairwise Wilcoxon rank sum tests with Bonferroni corrections (all P<0.004).doi:10.1371/journal.pone.0090573.g003
Frugivore and nectarivore relative abundance declined as landscape tree cover increased within radii of 350 m and 450 m, respectively (Fig. 3B–C), but local treatment × landscape tree cover interactions were not significant. The effect of landscape tree cover was not apparent in the trends of individual species (Table S5 in File S1). Of the ten most abundant species in each analysis group, eight frugivores, seven nectarivores, nine migratory insectivores, and eight reference forest species were detected more frequently in plantations than controls (Table S5 in File S1). Five of the most frequently detected species in old-growth forests increased in plantations relative to controls as tree cover increased within a radius of 550 m. No individual species had a negative relationship with landscape tree cover.
Our results show that local restoration and landscape context interacted to affect bird community composition, but not foraging guilds. Compared to less intensive restoration practices, tree plantations had: (1) greater abundance of frugivores and nectarivores, (2) greater migratory insectivore species richness; and (3) greater compositional similarity to old-growth forest. Similarity to old-growth bird communities increased with greater tree cover in the surrounding landscape, but only in plantations. In contrast, frugivore and nectarivore abundance and migrant insectivore richness were greatest in tree plantations regardless of landscape tree cover, and frugivore and nectarivore abundance were lower in landscapes with high tree cover.
Spatial scaling of landscape effects was similar (best models: 0.35–0.55 km radius) for old-growth similarity, frugivores, and nectarivores, but the direction of these effects was surprisingly different. Variance in local × landscape interactions for old-growth similarity and foraging guilds may be best explained by dispersal limitation and niche complementarity. Guilds are delineated on the basis of species traits, such as diet , but species identity is central to measures of community composition  and associated conservation values. High similarity to old-growth bird communities in plantations embedded within well-forested landscapes are contingent upon birds dispersing from pre-existing forest into restoration sites . Strong evidence shows that some forest birds are unable to cross even small distances through unusable habitat , and that the most dispersal-limited species are typically also the most prone to extinction in fragmented landscapes . Fragmentation studies have often highlighted that terrestrial insectivores are among the most extinction-prone birds . Likewise, we found that the most frequently detected species in old-growth forest was a terrestrial insectivore, Formicarius analis, which was only recorded in two restoration sites, one of which was adjacent to an old-growth forest fragment. Intensive local restoration efforts may thus provision suitable habitat for forest-dependent species, but their colonization depends upon matrix permeability and the composition of regional species pools , .
In contrast to old-growth forest birds, frugivores, nectarivores, and migrant insectivores were more abundant or speciose in tree plantations than in less-intensive restoration treatments regardless of landscape context. This observation could result from niche complementarity – the tendency of species similar on one niche axis to differ along another. In our study area, frugivores, nectarivores, and insectivores that are otherwise similar (sometimes congeneric) are separated by habitat affinity into partially overlapping agricultural and forest communities . The result is a portfolio effect, where reductions in forest-affiliated frugivores, for example, are balanced by increased abundance of agriculture-affiliated frugivores, potentially maintaining a constant level of bird-mediated ecosystem functions despite high species turnover , . However, guild classifications simplify functional heterogeneity. For example, frugivore trends in this study were driven by small omnivores, whereas wide-gaped species that could disperse large seeds were virtually absent. In this context, intensive local restoration in sites with low landscape tree cover may attract a subset of agriculture-affiliated birds already present in the surrounding matrix by provisioning food resources, favorable microclimate, or cover from diurnal predators , .
Subtle community-wide increases in frugivore and nectarivore detections in sites with lower tree cover may be explained by the marginal value theorem of optimal foraging . If lower tree cover in the surrounding landscape indicates greater travel distances between patches, then birds may spend more time and potentially be detected more frequently exploiting food resources in restoration sites with little tree cover in the surrounding landscape. This observation suggests that smaller forest elements become more valuable (to a subset of the regional bird species pool) when they make up larger proportions of local forest cover. A lack of landscape tree cover effect on migrants compared to other groups could be due to territorial exclusion if individuals are commonly relegated to low quality patches  or from a preference for early-successional habitats . Alternately, hierarchical landscape selection by migrants could occur at a spatial scale larger than the maximum buffer of 1 km used here .
Differences in landscape effects on old-growth forest species and foraging guilds support the hypothesis that restoration does not necessarily optimize biodiversity conservation and ecosystem functioning simultaneously . A growing body of literature is establishing connections between biodiversity and ecosystem functioning (BEF) – (but see ), but trade-offs between ecosystem services like carbon sequestration and conservation-relevant biodiversity outcomes are evident for ecological restoration at national and global scales , . Terminology is a primary hang-up. Biodiversity in the BEF conversation is taken to include taxonomic, phylogenetic, genetic, functional, spatial, temporal, interaction, and landscape diversity , but conservation priorities are typically designated using other biodiversity concepts, such as the richness or abundance of threatened, endemic, and forest-affiliated species, population diversity, and community composition , , . Our data suggest that this divide between basic and applied biodiversity-ecosystem function science may extend to bird communities if bird-mediated functions are redundant across species or do not align with species-specific conservation priorities.
The local × landscape interaction that we detected for predicting old-growth species composition has implications for allocating restoration funds at sub-national scales. Large-scale tropical forest restoration is on the rise, due in large part to increased funding from REDD+ and various payments for ecosystem services programs , , . Our data suggest that new restoration projects in areas with high forest cover will likely promote greater colonization by species representative of reference communities – and thus safeguard biodiversity conservation – than similar projects in habitat-poor landscapes . Also, more intensive local interventions are likely to benefit birds over a 5–7 yr period more than less intensive or passive restoration techniques , , . Given time, we expect that effects of local restoration treatments will converge as low-intensity control plots increasingly resemble closed-canopy forest , , but landscape effects are likely to endure.
An alternative explanation for differences in observed bird visitations between local restoration treatments could be that detectability varied among treatments , . We were unable to evaluate detection probabilities, but we expect that a habitat-specific bias would favor increased bird detections in open control plots relative to closed-canopy plantations. Such a bias would strengthen our conclusion that relative abundance is greater in plantations, however, this result does not denote a successful restoration of the bird community per se in any particular treatment . Many habitat manipulation studies, including this one, are too small to reliably detect differences in population or demographic variables needed to infer community restoration, but they still contain useful information .
We have demonstrated experimentally that intensive local restoration of degraded pastures promotes three functionally-relevant foraging guilds regardless of surrounding tree cover. Also, intensive restoration coupled with high amounts of surrounding tree cover increases habitat for species affiliated with old-growth forest within a few years of the intervention. We note that while ecological restoration holds great hope for slowing or reversing the tide of biodiversity loss , restoration cannot be considered a substitute for the preservation of existing forest .
Table S1. Correlates of bird-mediated seed dispersal, arthropod control, and pollination in tropical ecosystems. Table S2. Effects of local restoration treatments on bird community attributes. Table S3. Maximum likelihood model selection for bird community attributes. Table S4. Maximum likelihood tests for significance of fixed factors explaining bird community attributes. Table S5. Detection trends for common species in each group. Figure S1. Weights and family composition of bird species detected in restoration sites. Figure S2. Detections and species composition of large frugivores (weight >100 g) in restoration sites.
We thank G. Daily, P. Ehrlich, J. Figueroa-Sandí, the Holl lab at UCSC, F. Lu, F. Oviedo Brenes, A. Wrona, J. Zook, many field assistants, landowners, and the staff of Las Cruces Biological Station for their support. We also thank two anonymous reviewers. Data for this manuscript are available at the Merritt data repository site (https://merritt.cdlib.org/m/ucsc_lib_hollzahawi).
Conceived and designed the experiments: JLR CDM RAZ KDH. Performed the experiments: JLR CDM JAR RAZ KDH. Analyzed the data: JLR CDM. Wrote the paper: JLR CDM JAR RAZ KDH.
- 1. SER (Society for Ecological Restoration) (2004) The SER primer on ecological restoration. Available: http://www.ser.org/resources/resources-detail-view/ser-international-primer-on-ecological-restoration. Accessed 05 November 2013.
- 2. Rey Benayas JM, Newton AC, Diaz AD, Bullock JM (2009) Enhancement of biodiversity and ecosystem services by ecological restoration: a meta-analysis. Science 325: 1121–1124. doi: 10.1126/science.1172460
- 3. Redo DJ, Grau HR, Aide TM, Clark ML (2012) Asymmetric forest transition driven by the interaction of socioeconomic development and environmental heterogeneity in Central America. Proc Natl Acad Sci USA 109: 8839–8844. doi: 10.1073/pnas.1201664109
- 4. Webb TJ, Gaston KJ, Hannah L, Woodward FI (2006) Coincident scales of forest feedback on climate and conservation in a diversity hot spot. Proc Roy Soc B-Biol Sci 273: 757–765. doi: 10.1098/rspb.2005.3364
- 5. Ruiz-Jaen MC, Aide TM (2005) Restoration success: how is it being measured? Restor Ecol 13: 569–577. doi: 10.1111/j.1526-100x.2005.00072.x
- 6. Vellend M (2003) Habitat loss inhibits recovery of plant diversity as forests regrow. Ecology 84: 1158–1164. doi: 10.1890/0012-9658(2003)084[1158:hlirop]2.0.co;2
- 7. Grainger MJ, van Aarde RJ, Wassenaar TD (2011) Landscape composition influences the restoration of subtropical coastal dune forest. Restor Ecol 19: 111–120. doi: 10.1111/j.1526-100x.2009.00630.x
- 8. Huxel GR, Hastings A (1999) Habitat loss, fragmentation, and restoration. Restor Ecol 7: 309–315. doi: 10.1046/j.1526-100x.1999.72024.x
- 9. Cole RJ, Holl KD, Zahawi RA (2010) Seed rain under tree islands planted to restore degraded lands in a tropical agricultural landscape. Ecol Appl 20: 1255–1269. doi: 10.1890/09-0714.1
- 10. Lindenmayer DB, Knight EJ, Crane MJ, Montague-Drake R, Michael DR, et al. (2010) What makes an effective restoration planting for woodland birds? Biol Conserv 143: 289–301. doi: 10.1016/j.biocon.2009.10.010
- 11. Edwards DP, Fisher B, Boyd E (2010) Protecting degraded rainforests: enhancement of forest carbon stocks under REDD+. Conserv Lett 3: 313–316. doi: 10.1111/j.1755-263x.2010.00143.x
- 12. Harvey CA, Dickson B, Kormos C (2010) Opportunities for achieving biodiversity conservation through REDD. Conserv Lett 3: 53–61. doi: 10.1111/j.1755-263x.2009.00086.x
- 13. Gardner TA, Burgess ND, Aguilar-Amuchastegui N, Barlow J, Berenguer E, et al. (2012) A framework for integrating biodiversity concerns into national REDD+ programmes. Biol Conserv 154: 61–71. doi: 10.1016/j.biocon.2011.11.018
- 14. Sodhi NS, Smith KG (2007) Conservation of tropical birds: mission possible? J Ornithol 148: S305–S309. doi: 10.1007/s10336-007-0180-y
- 15. Dirzo R, Raven PH (2003) Global state of biodiversity and loss. Ann Rev Env Resour 28: 137–167.
- 16. Young TP (2000) Restoration ecology and conservation biology. Biol Conserv 92: 73–83. doi: 10.1016/s0006-3207(99)00057-9
- 17. Aide TM, Cavelier J (1994) Barriers to lowland tropical forest restoration in the Sierra Nevada de Santa Marta, Colombia. Restor Ecol 2: 219–229. doi: 10.1111/j.1526-100x.1994.tb00054.x
- 18. Holl KD (1999) Factors limiting tropical rain forest regeneration in abandoned pasture: seed rain, seed germination, microclimate, and soil. Biotropica 31: 229–242. doi: 10.1111/j.1744-7429.1999.tb00135.x
- 19. Sekercioglu CH (2006) Increasing awareness of avian ecological function. Trends Ecol Evol 21: 464–471. doi: 10.1016/j.tree.2006.05.007
- 20. Ingle NR (2003) Seed dispersal by wind, birds, and bats between Philippine montane rainforest and successional vegetation. Oecologia 134: 251–261.
- 21. Wunderle JM (1997) The role of animal seed dispersal in accelerating native forest regeneration on degraded tropical lands. Forest Ecol Manag 99: 223–235. doi: 10.1016/s0378-1127(97)00208-9
- 22. Whittaker RJ, Jones SH (1994) The role of frugivorous bats and birds in the rebuilding of a tropical forest ecosystem, Krakatau, Indonesia. J Biogeogr 21: 245–258. doi: 10.2307/2845528
- 23. Traveset A, Verdú M (2002) A meta-analysis of the effect of gut treatment on seed germination. In: Levey DJ, Silva WR, Galetti M, editors. Seed dispersal and frugivory: ecology, evolution, and conservation. Wallingford: CAB International. pp. 339–350.
- 24. Morrison EB, Lindell CA (2012) Birds and bats reduce insect biomass and leaf damage in tropical forest restoration sites. Ecol Appl 22: 1526–1534. doi: 10.1890/11-1118.1
- 25. Dixon KW (2009) Pollination and restoration. Science 325: 571–573. doi: 10.1126/science.1176295
- 26. Lindell CA, Thurston GM (2013) Bird pollinator visitation is equivalent in island and plantation planting designs in tropical forest restoration sites. Sustainability 5: 1177–1187. doi: 10.3390/su5031177
- 27. Guevara S, Laborde J (1993) Monitoring seed dispersal at isolated standing trees in tropical pastures: consequences for local species availability. Vegetatio 108: 319–338. doi: 10.1007/978-94-011-1749-4_22
- 28. Galindo-González J, Guevara S, Sosa VJ (2000) Bird- and bat-generated seed rains at isolated trees in pastures in a tropical rainforest. Conserv Biol 14: 1693–1703. doi: 10.1046/j.1523-1739.2000.99072.x
- 29. da Silva JMC, Uhl C, Murray G (1996) Plant succession, landscape management, and the ecology of frugivorous birds in abandoned Amazonian pastures. Conserv Biol 10: 491–503. doi: 10.1046/j.1523-1739.1996.10020491.x
- 30. Duncan RS, Chapman CA (1999) Seed dispersal and potential forest succession in abandoned agriculture in tropical Africa. Ecol Appl 9: 998–1008. doi: 10.1890/1051-0761(1999)009[0998:sdapfs]2.0.co;2
- 31. Holl KD (1998) Do bird perching structures elevate seed rain and seedling establishment in abandoned tropical pasture? Restor Ecol 6: 253–261. doi: 10.1046/j.1526-100x.1998.00638.x
- 32. Shiels AB, Walker LR (2003) Bird perches increase forest seeds on Puerto Rican landslides. Restor Ecol 11: 457–465. doi: 10.1046/j.1526-100x.2003.rec0269.x
- 33. Graham LLB, Page SE (2012) Artificial bird perches for the regeneration of degraded tropical peat swamp forest: a restoration tool with limited potential. Restor Ecol 20: 631–637. doi: 10.1111/j.1526-100x.2011.00805.x
- 34. Reid JL, Holl KD (2012) Arrival ≠ survival. Restor Ecol 21: 153–155. doi: 10.1111/j.1526-100x.2012.00922.x
- 35. Cole RJ (2009) Postdispersal seed fate of tropical montane trees in an agricultural landscape, southern Costa Rica. Biotropica 41: 319–327. doi: 10.1111/j.1744-7429.2009.00490.x
- 36. Zahawi RA, Augspurger CK (2006) Tropical forest restoration: tree islands as recruitment foci in degraded lands of Honduras. Ecol Appl 16: 464–478. doi: 10.1890/1051-0761(2006)016[0464:tfrtia]2.0.co;2
- 37. de la Peña-Domene M, Martínez-Garza C, Howe HF (2013) Early recruitment dynamics in tropical restoration. Ecol Appl 23: 1124–1134. doi: 10.1890/12-1728.1
- 38. Parrotta JA, Knowles OH, Wunderle Jr JM (1997) Development of floristic diversity in 10-year-old restoration forests on a bauxite mined site in Amazonia. Forest Ecol Manag 99: 21–42. doi: 10.1016/s0378-1127(97)00192-8
- 39. Jansen A (2005) Avian use of restoration plantings along a creek linking rainforest patches on the Atherton Tablelands, North Queensland. Restor Ecol 13: 275–283. doi: 10.1111/j.1526-100x.2005.00035.x
- 40. Reid JL, Harris JBC, Martin LJ, Barnett JR, Zahawi RA (2008) Distribution and abundance of nearctic-neotropical songbird migrants in a forest restoration site in southern Costa Rica. J Trop Ecol 24: 685–688. doi: 10.1017/s0266467408005415
- 41. Lindell CA, Reid JL, Cole RJ (2013) Planting design effects on avian seed dispersers in a tropical forest restoration experiment. Restor Ecol 21: 515–522. doi: 10.1111/j.1526-100x.2012.00905.x
- 42. Moran C, Catterall C, Green RJ, Olsen MF (2004) Functional variation among frugivorous birds: implications for rainforest seed dispersal in a fragmented subtropical landscape. Oecologia 141: 584–595. doi: 10.1007/s00442-004-1685-1
- 43. Reid JL, Harris JBC, Zahawi RA (2012) Avian habitat preference in tropical forest restoration in southern Costa Rica. Biotropica 44: 350–359. doi: 10.1111/j.1744-7429.2011.00814.x
- 44. Catterall CP, Freeman AND, Kanowski J, Freebody K (2012) Can active restoration of tropical rainforest rescue biodiversity? a case with bird community indicators. Biol Conserv 146: 53–61. doi: 10.1016/j.biocon.2011.10.033
- 45. Freeman AN, Freeman AB, Burchill S (2010) Bird use of revegetated sites along a creek connecting rainforest remnants. Emu 109: 331–338. doi: 10.1071/mu09089
- 46. Ortega-Álvarez R, Lindig-Cisneros R (2012) Feathering the scene: the effects of ecological restoration on birds and the role birds play in evaluating restoration outcomes. Ecol Restor 30: 116–127. doi: 10.3368/er.30.2.116
- 47. Munro NT, Lindenmayer DB, Fischer J (2007) Faunal response to revegetation in agricultural areas of Australia: a review. Ecol Manag Restor 8: 199–207. doi: 10.1111/j.1442-8903.2007.00368.x
- 48. Holdridge LR, Grenke WC, Hatheway WH, Liany T, Tosi Jr JA (1971) Forest environments in tropical life zones. New York: Pergamon Press. 747 p.
- 49. Seligson M (1980) Peasants of Costa Rica and the development of agrarian capitalism. Madison: University of Wisconsin Press. 220 p.
- 50. Rickert E (2005) Environmental effects of the coffee crisis: a case study of land use and avian communities in Agua Buena, Costa Rica. Masters Thesis. Olympia: Evergreen State College. 219 p.
- 51. Mendenhall CD, Sekercioglu CH, Brenes FO, Ehrlich PR, Daily GC (2011) Predictive model for sustaining biodiversity in tropical countryside. Proc Nat Acad Sci USA 108: 16313–16316. doi: 10.1073/pnas.1111687108
- 52. American Ornithologists' Union (1998) Check-list of North American birds. Washington, D.C.: American Ornithologists' Union. 829 p.
- 53. Stiles FG, Skutch AF (1989) A guide to the birds of Costa Rica. Ithaca: Cornell University Press. 511 p.
- 54. Pejchar L, Pringle RM, Ranganathan J, Zook JR, Duran G, et al. (2008) Birds as agents of seed dispersal in a human-dominated landscape in southern Costa Rica. Biol Conserv 141: 536–544. doi: 10.1016/j.biocon.2007.11.008
- 55. García D, Martínez D (2012) Species richness matters for the quality of ecosystem services: a test using seed dispersal by frugivorous birds. Proc Roy Soc B-Biol Sci 279: 3106–3113. doi: 10.1098/rspb.2012.0175
- 56. Van Bael SA, Philpott SM, Greenberg R, Bichier P, Barber NA, et al. (2008) Birds as predators in tropical agroforestry systems. Ecology 89: 928–934. doi: 10.1890/06-1976.1
- 57. Bascompte J, Jordano P (2007) Plant-animal mutualistic networks: the architecture of biodiversity. Annu Rev Ecol Evol Syst 38: 567–593. doi: 10.1146/annurev.ecolsys.38.091206.095818
- 58. Mendenhall CD, Archer HM, Brenes FO, Sekercioglu CH, Sehgal RNM (2012) Balancing biodiversity with agriculture: land sharing mitigates avian malaria prevalence. Conserv Lett 6: 125–131. doi: 10.1111/j.1755-263x.2012.00302.x
- 59. Karp DS, Mendenhall CD, Sandí RF, Chaumont N, Ehrlich PR, et al. (2013) Forest bolsters bird abundance, pest control and coffee yield. Ecol Lett 16: 1339–1347. doi: 10.1111/ele.12173
- 60. Pearman PB (2002) The scale of community structure: habitat variation and avian guilds in tropical forest understory. Ecol Monogr 72: 19–39. doi: 10.2307/3100083
- 61. Montague-Drake RM, Lindenmayer DB, Cunningham RB (2009) Factors affecting site occupancy by woodland bird species of conservation concern. Biol Conserv 142: 2896–2903. doi: 10.1016/j.biocon.2009.07.009
- 62. Oksanen J, Blanchet FG, Kindt R, Legendre P, Minchin PR, et al.. (2012) Vegan: community ecology package. R package version 2.0-5. Available: http://cran.r-project.org/web/packages/vegan/index.html. Accessed 24 January 2013.
- 63. Burnham KP, Anderson DR (1998) Model selection and inference: a practical information-theoretic approach. New York: Springer. 353 p.
- 64. Gelman A, Hill J (2006) Data analysis using regression and multilevel/hierarchical models. Cambridge: Cambridge University Press. 648 p.
- 65. R Development Core Team (2012) R: a language and environment for statistical computing. Vienna: R Foundation for Statistical Computing. Available: http://www.R-project.org/. Accessed: 24 January 2013.
- 66. Bates D, Maechler M, Bolker B (2011) Lme4: linear mixed-effects models using S4 classes. Vienna: Institute for Statistics and Mathematics, University of Economics and Business. Available: http://CRAN.R-project.org/package=lme4. Accessed: 24 January 2013.
- 67. Ruiz-Gutiérrez V, Zipkin EF, Dhondt AA (2010) Occupancy dynamics in a tropical bird community: unexpectedly high forest use by birds classified as non-forest species. J Appl Ecol 47: 621–630. doi: 10.1111/j.1365-2664.2010.01811.x
- 68. Mendenhall CD, Daily GC, Ehrlich PR (2012) Improving estimates of biodiversity loss. Biol Conserv 151: 32–34. doi: 10.1016/j.biocon.2012.01.069
- 69. MacKenzie DI, Royle JA (2005) Designing occupancy studies: general advice and allocating survey effort. J Appl Ecol 42: 1105–1114. doi: 10.1111/j.1365-2664.2005.01098.x
- 70. Hill JK, Hamer KC (2004) Determining impacts of habitat modification on diversity of tropical forest fauna: the importance of spatial scale. J Appl Ecol 41: 744–754. doi: 10.1111/j.0021-8901.2004.00926.x
- 71. Robinson WD (2010) The challenges of studying vertebrates in habitat treatment plots. Open Environ Sci 4: 21–23. doi: 10.2174/1876325101004010021
- 72. Ricklefs RE (2004) A comprehensive framework for global patterns in biodiversity. Ecol Lett 7: 1–15. doi: 10.1046/j.1461-0248.2003.00554.x
- 73. Ibarra-Macias A, Robinson WD, Gaines MS (2011) Experimental evaluation of bird movements in a fragmented Neotropical landscape. Biol Conserv 144: 703–712. doi: 10.1016/j.biocon.2010.08.006
- 74. Moore RP, Robinson WD, Lovette IJ, Robinson TR (2008) Experimental evidence for extreme dispersal limitation in tropical forest birds. Ecol Lett 11: 960–968. doi: 10.1111/j.1461-0248.2008.01196.x
- 75. Stouffer PC, Strong C, Naka LN (2009) Twenty years of understorey bird extinctions from Amazonian rain forest fragments: consistent trends and landscape-mediated dynamics. Divers Distrib 15: 88–97. doi: 10.1111/j.1472-4642.2008.00497.x
- 76. Ricketts TH (2001) The matrix matters: effective isolation in fragmented landscapes. Am Nat 158: 87–99. doi: 10.1086/320863
- 77. Zobel M (1997) The relative of species pools in determining plant species richness: an alternative explanation of species coexistence? Trends Ecol Evol 12: 266–269. doi: 10.1016/s0169-5347(97)01096-3
- 78. Doak DF, Bigger D, Harding E, Marvier M, O'Malley R, et al. (1998) The statistical inevitability of stability-diversity relationships in community ecology. Am Nat 151: 264–276. doi: 10.1086/286117
- 79. Karp DS, Ziv G, Zook J, Ehrlich PR, Daily GC (2011) Resilience and stability in bird guilds across tropical countryside. Proc Nat Acad Sci USA 108: 21134–21139. doi: 10.1073/pnas.1118276108
- 80. Fink RD, Lindell CA, Morrison EB, Zahawi RA, Holl KD (2009) Patch size and tree species influence the number and duration of bird visits in forest restoration plots in southern Costa Rica. Restor Ecol 17: 479–486. doi: 10.1111/j.1526-100x.2008.00383.x
- 81. Morrison EB, Lindell CA, Holl KD, Zahawi RA (2010) Patch size effects on avian foraging behaviour: implications for tropical forest restoration design. J Appl Ecol 47: 130–138. doi: 10.1111/j.1365-2664.2009.01743.x
- 82. Charnov EL (1976) Optimal foraging: the marginal value theorem. Theor Popul Biol 9: 129–136. doi: 10.1016/0040-5809(76)90040-x
- 83. Marra PP, Sherry TW, Holmes RT (1993) Territorial exclusion by a long-distance migrant warbler in Jamaica: a removal experiment with American Redstarts (Setophaga ruticilla). Auk 110: 565–572. doi: 10.2307/4088420
- 84. Blake JG, Loiselle BA (2001) Bird assemblages in second-growth and old-growth forests, Costa Rica: perspectives from mist nets and point counts. Auk 118: 304–326. doi: 10.2307/4089793
- 85. Hutto RL (1985) Habitat selection by nonbreeding, migratory land birds. In: Cody ML, editor. Habitat selection in birds. Orlando: Academic Press, Inc. pp. 455–476.
- 86. Tscharntke T, Tylianakis JM, Rand TA, Didham RK, Fahrig L, et al. (2012) Landscape moderation of biodiversity patterns and processes - eight hypotheses. Biol Rev 87: 661–685. doi: 10.1111/j.1469-185x.2011.00216.x
- 87. Hooper DU, Chapin FS, Ewel JJ, Hector A, Inchausti P, et al. (2005) Effects of biodiversity on ecosystem functioning: a consensus of current knowledge. Ecol Monogr 75: 3–35. doi: 10.1890/04-0922
- 88. Schwartz MW, Brigham CA, Hoeksema JD, Lyons KG, Mills MH, et al. (2000) Linking biodiversity to ecosystem function: implications for conservation ecology. Oecologia 122: 297–305. doi: 10.1007/s004420050035
- 89. Naeem S, Duffy JE, Zavaleta E (2012) The functions of biological diversity in an age of extinction. Science 336: 1401–1406. doi: 10.1126/science.1215855
- 90. Wardle DA, Bonner KI, Nicholson KS (1997) Biodiversity and plant litter: experimental evidence which does not support the view that enhanced species richness improves ecosystem function. Oikos 79: 247–258. doi: 10.2307/3546010
- 91. Mason NWH, Ausseil AE, Dymond JR, Overton JM, Price R, et al. (2012) Will use of non-biodiversity objectives to select areas for ecological restoration always compromise biodiversity gains? Biol Conserv 155: 157–168. doi: 10.1016/j.biocon.2012.05.019
- 92. Phelps J, Friess DA, Webb EL (2012) Win–win REDD+ approaches belie carbon–biodiversity trade-offs. Biol Conserv 154: 53–60. doi: 10.1016/j.biocon.2011.12.031
- 93. Myers N, Mittermeier RA, Mittermeier CG, da Fonseca GAB, Kent J (2000) Biodiversity hotspots for conservation priorities. Nature 403: 853–858. doi: 10.1038/35002501
- 94. Roberts CM, McClean CJ, Veron JEN, Hawkins JP, Allen GR, et al. (2002) Marine biodiversity hotspots and conservation priorities for tropical reefs. Science 295: 1280–1284. doi: 10.1126/science.1067728
- 95. Arriagada RA, Ferraro PJ, Sills EO, Pattanayak SK, Cordero-Sancho S (2012) Do payments for environmental services affect forest cover? a farm-level evaluation from Costa Rica. Land Econ 88: 382–399.
- 96. Brancalion PHS, Viani RAG, Strassburg BBN, Rodrigues RR (2012) Finding the money for tropical forest restoration. Unasylva 239: 41–50.
- 97. Tambosi LR, Martensen AC, Ribeiro MC, Metzger JP (2013) A framework to optimize biodiversity restoration efforts based on habitat amount and landscape connectivity. Restor Ecol. DOI 10.1111/rec.12049.
- 98. Lindell CA, Cole RJ, Holl KD, Zahawi RA (2012) Migratory bird species in young tropical forest restoration sites: effects of vegetation height, planting design, and season. Bird Conserv Int 22: 94–105. doi: 10.1017/s0959270911000177
- 99. O'Dea N, Whittaker RJ (2007) How resilient are Andean montane forest bird communities to habitat degradation? Biodivers Conserv 16: 1131–1159. doi: 10.1007/s10531-006-9095-9
- 100. Gu W, Swihart RK (2004) Absent or undetected? effects of non-detection of species occurrence on wildlife–habitat models. Biol Conserv 116: 195–203. doi: 10.1016/s0006-3207(03)00190-3
- 101. Munro NT, Fischer J, Barrett G, Wood J, Leavesley A, et al. (2011) Bird's response to revegetation of different structure and floristics—are “restoration plantings” restoring bird communities? Restor Ecol 19: 223–235. doi: 10.1111/j.1526-100x.2010.00703.x
- 102. Gibson L, Lee TM, Koh LP, Brook BW, Gardner TA, et al. (2011) Primary forests are irreplaceable for sustaining tropical biodiversity. Nature 478: 378–381. doi: 10.1038/nature10425