Species Guild Composition: Relationship to Stand Age, Vegetation Structure, Landscape Composition and Configuration, and Spatial Structure of Data

The age of successional forests along with the related physiognomic and floristic changes in vegetation had perceptible effects on the feeding guilds. Interestingly, our results show that, in most cases, comparable patterns in bird feeding guild- environment relationships were found when using either the number of individuals per guild or number of species per guild. Other studies have found, when using abundance or presence/absence coded data to measure changes in bird communi­ties, that the results for both approaches do not show strong differences and both data sets are highly correlated, indicating high similarity among their predictions (Bart and Klosiewski 1989).

In our study feeding guilds varied in their response and susceptibility to these factors, which suggests that this ecological trait is useful to detect differential responses to habitat disturbance, as has been found in other studies in wet tropical regions (Raman et al. 1998; Gray et al. 2007). Also, responses of bird species are presumably affected by their ecology and evolutionary history (Bennet and Owens 2002), such as phylogeny, body size, local population size, and geographic range, which were not taken into consideration in this study.

Overall, most of the variation in number of individuals and number of species per feeding guild was accounted for by the spatial structure of data and vegetation variables (including stand age), whereas landscape variables had a particular marginal influence in both analyses. The low marginal effect of landscape compo­sition and configuration variables on bird feeding guilds suggests that, in this study, habitat characteristics play a greater role than landscape structure. In a tropical region, a comparable study found evidence that the bird community composition was more similar between the same age classes than with the adjacent vegetation regardless of distance between similar age class-vegetation patches (Raman et al. 1998), which stresses the importance of specific habitat attributes required by different bird feeding guilds regardless of the landscape structure. Also, the marginal effects of landscape structure found in this study were possibly affected by the landscape matrix, which is dominated by late succession (nearly 60 % corre­sponds to > 15-year-old vegetation), suggesting that the avian community perceives the landscape as fairly homogeneous.

The most important set of variables influencing bird feeding guilds was the spatial structure of sampling plots; this result could imply that species composition is highly associated with avian dispersal (Haas 1995). However, it is possible that variation in space likely had a considerable environmental or biotic component that was not detected because some relevant environmental or biotic variables, such predation and competition, were not considered in our analysis

Another important element influencing bird feeding guilds is the relatively small influence of the landscape matrix that differentially affects the capacity of species to move among patches of dissimilar habitats, which in turn may provide foraging or breeding habitat and ultimately influence the local persistence and abundances of species (Renjifo 2001). In our study, the landscape matrix is relatively homoge­neous, possibly providing a nonhostile landscape for dispersal that allows general­ists or species capable of thriving in a variety of habitats to persist.

As expected, bark insectivores (BI), constituting mostly forest-dependent spe­cies or specialists in their foraging strategy (e. g., army ant-following species such as Dendrocincla, or those feeding often at bromeliads such as Xiphorhynchus) in this study, were positively correlated with stand age and its associated changes in vegetation structure (e. g., increase in basal area and tree height). Dependency of BI on higher trees and thicker stems has been previously documented (Arriaga-Weiss et al. 2008), giving support to the hypothesis that forest birds follow the recovery of forest vegetation (Raman et al. 1998). In contrast, BI were negatively correlated to the abundance and density of saplings, which offer a comparatively lower surface area for feeding; such negative effects have been reported in other studies about forest succession (Raman et al. 1998) and other tropical forests (Thiollay 1994). In addition, canopy insectivores showed consistent patterns for both individuals-based and number of species-based analyses in which tree basal area (i. e., total basal area and basal area of adults) and abundance of adult plants were important predictors that showed a positive correlation. For insectivores, the structural complexity in vegetation, associated with older succession, has been proposed as being influential in providing opportunities for foraging (Willson 1974; Holmes et al. 1979; Terborgh 1985). Thus, an increase in vertical strata, tree height, and canopy cover provides potential new foraging opportunities (Bowman et al. 1990; Brady and Noske 2009). On the other hand, nectarivores (a feeding guild represented only by hummingbirds in this study) showed contrasting results between individuals – based and number of species-based analyses. The number of individuals showed a positive correlation with abundance and basal area of adult individuals, whereas the number of species showed a positive correlation with abundance, density, and basal area of saplings. The response in abundance suggests the use of secondary resources as supplementary food; especially, Amazilia Candida and A. yucatanensis, the most abundant species in the older succession, are reported as facultative feeders con­suming insects as a secondary resource (Arizmendi et al. 2010). Additionally, higher abundance of insects has been reported as related to forest complexity, which is typically associated with older succession (Sekercioglu et al. 2002). In contrast, the positive correlation of nectarivore species richness with abundance and density of saplings (associated with early succession) could be explained by the low sensitivity to disturbance of the hummingbird species grouped in this feeding guild; moreover, nectarivore species have been reported as frequent visitors of earlier successional stages (Robinson and Terborgh 1997) where flowering herba­ceous and scrubby vegetation is more abundant. However, the observed trend could also be attributed to sampling bias, as in older successional stages these humming­bird species could be difficult to detect because of their small size and also the weak vocalizations of some of the species.

The abundance and species richness of mid-canopy insectivores (MCI) showed a positive correlation to basal area and abundance of saplings and the Simpson diversity index (dominance). A previous study in our research area found that the early successional stages are related to higher abundance and basal area of saplings where plant communities are dominated by few dominant species (Dupuy et al. 2011). The species composition of the MCI in this study includes mostly habitat generalists that reside or feed in open woodlands, plantations, clearings, scrubby vegetation, forest edge, or secondary forests (Shulenberg 2010); consequently, it is not surprising that MCI were associated with early succession. Also, understory insectivores (UI) presented similar patterns as the MCI, for both abundance and species richness, and similarly most of the species grouped as UI can inhabit second-growth forest and scrubby vegetation where the foliage is concentrated in the first meters above the ground and where food and microclimate conditions are suitable (Karr and Freemark 1983; Johns 1991).

Canopy frugivores (CF) showed a positive correlation with stand age and basal area, both for abundance and for species richness. The CF guild is composed of a wide array of species with unique natural history and ecology (Wiens 1989; Simberloff 1994); however, some species attributes appear to be related to vulner­ability to habitat alteration. For example, body size may influence the response of CF, because large-bodied species (e. g., Amazona albifrons, and Patagioenas flavirostris with more than 130 g) tend to be more susceptible to habitat alteration (Raman and Sukumar 2002) as they may have smaller population sizes, lower reproductive rates, and larger home or geographic range requirements than small­bodied species (Gaston and Blackburn 1995). In addition, some species have specific reproduction strategies that link them to older succession; for example, species such as Amazona albifrons, Aratinga nana, and Tityra semifasciata require thicker and higher stems to construct their nests although they can forage in a wide range of habitats. In contrast, understory frugivores (UF) were positively correlated with early succession structural characteristics, which is not surprising as this guild was basically represented by Crypturellus cinnamomeus, a habitat generalist that inhabits secondary forests, thickets, and shrublands, where it usually hides under the dense understory (Bribiesca-Formisano et al. 2010). Moreover, the marginal influence of landscape structure in terms of patch density and higher fragmentation (TECI) could be partially explained by the low sensitivity to disturbance of C. cinnamomeus. In addition, understory granivores (UG) comprised mostly spe­cies that inhabit scrub or shrub vegetation, early successional forests, dense under­growth, and for some species even open habitats. The majority of UG species also have low sensitivity to disturbance, which explains their positive correlation with early succession structural characteristics (e. g., abundance and density of saplings and a plant community dominated by one or few species), and the marginal positive correlation with patch density and fragmentation (TECI). It is noteworthy that the only two species with high sensitivity to disturbance, Dactylortyx thoracicus and Meleagris ocellata, occurred in low numbers in early succession or were very rare during the sampling. Finally, MCG were also composed of species that inhabit scrub vegetation, second-growth scrub, shrubby areas, and open woodlands, and commonly inhabit disturbed habitat; thus, these species are habitat generalists with no reported negative effect of habitat fragmentation. Two MCG species, Zenaida asiatica and Cardinalis cardinalis, have actually expanded their geographic range northward in past centuries, correlating with human habitation and agricultural activities (Shulenberg 2010).

The different responses among feeding guilds indicate that this trait may be a suitable predictor of avian susceptibility to habitat disturbance. Secondary forests are ubiquitous in the tropics and specifically in the TDF (Wright 2005; Sanchez – Azofeifa et al. 2005; Miles et al. 2006), and thus potentially play a key role in the conservation of animal communities as the bulk of tropical species will need to persist in degraded and secondary habitats (Bhagwat et al. 2008). However, it is important to remain cautious if sound conservation and management are to be done, because forest specialist species require pristine or late secondary forests to persist (Laurence 2007), species that may occur in secondary forests but would otherwise use mature forests may decline in secondary forests (Graham and Blake 2001), and species that possibly feed in a variety of habitats may not necessarily reproduce in all habitat types. Moreover, in coming decades the tendency in changes of forest cover is toward clearing, logging, fragmentation, or degradation (Rappole et al. 1992; Wright and Muller-Landau 2006), in which the type and rate of recovery could strongly affect species prone to extinction (Laurence 2007). Similarly, secondary forests are typically transient elements of the landscapes that undergo frequent reclearing, and which could go beyond survival thresholds of vulnerable species to disturbance. Finally, it is of great importance to account for the syner­getic effect of both the temporal variability of secondary forests as a potential habitat and the landscape matrix in which secondary forests are immersed, espe­cially in highly heterogeneous and fragmented landscapes.

In general, bark insectivores (BI) increased in species density and abundance with stand age. Although BI have been reported to adapt to microclimatic changes linked to disturbance (Johns 1991), our results and those of other studies show an increase in species richness in BI with tree height (Arriaga-Weiss et al. 2008), and

consequently if the surface area of feeding sites increases (Thiollay 1994; Raman et al. 1998), which are features associated with late succession. Also, BI are associated with habitats that are structurally complex, which in turn provides a higher abundance of insects (Johnson 2000; Brady and Noske 2009). Our results highlight the importance of maintaining large forest tracts within the landscape matrix, if more specialized feeding guilds such as the BI are to be conserved.

11.4 Conclusion

Insofar as possible, when developing a national carbon sequestration strategy, the costs and benefits of different land use/land cover alternatives should be taken into account, which includes the potential effects on biodiversity. In this study, we provided important elements to assess the impact of a priori identified factors influencing bird diversity that may derive from carbon sequestration policies in a tropical dry forest: these include the variation in species diversity and composition of the bird communities in secondary forests of different ages. Cost-benefit analysis can be done when all relevant gains in a given land use/land cover scenario are included, such as the impacts on biodiversity.

Acknowledgments We thank James Callaghan and Kaxil Kiuic A. C. for logistic support. We also thank Rosalina Rodriguez Roman, Filogonio May Pat, Fernando Tun Dzul, Victor Marin Perez, Ramiro Lara Castillo, Feliciano Pech PinziSn, Evelio Uc Uc, Mario Evelio Uc Uc, and Santos Armin Uc Uc for fieldwork and technical assistance. Funding for this research was provided by CICY, FOMIX-Yucatan (project YUC-2008-C06-108863) and CONACYT (CB-127800)

Appendix 11.1 Bird feeding guilds based on their primary food source and foraging strata, and presence of bird species in four forest classes

Age of secondary forest Feeding >15 years >15 years Species guild 3-8 years old 9-15 years old old, flat land old, hill Crypturellus cinnamomeus UF P P P P Ortalis vetula MCG P P P P Meleagris ocellata UG A P P A Colinus nigrogularis UG A P P A Dactylortyx thoracicus UG P P P P Patagioenas cayennensis CF A A A P Patagioenas flavirostris CF P P P P Zenaida asiatica MCG P P P P Columbina passerina UG P P P P Columbina talpacoti UG P P A A Claravis pretiosa UF A P P P Leptotila verreauxi UG P P P P

Species	guild	3-8 years old	9-15 years old	old, flat land	old, hill
Leptotila jamaicensis	UG	P	P	P	P
Amazona albifrons	CF	P	P	P	P
Aratinga nana	CF	P	P	P	P
Piaya cayana	CI	P	P	P	P
Dromococcyx phasianellus	UI	P	P	P	P
Chordeiles acutipennis	UI	A	A	P	A
Campylopterus curvipennis	N	A	A	P	P
Anthracothorax prevostii	N	P	A	P	P
Chlorostilbon canivetii	N	P	P	P	P
Amazilia candida	N	P	P	P	P
Amazilia rutila	N	P	P	P	P
Amazilia yucatanensis	N	P	P	P	P
Trogon melanocephalus	MCI	P	P	P	P
Trogon collaris	MCI	A	P	A	P
Trogon caligatus	MCI	P	P	P	P
Momotus momota	MCI	P	P	P	P
Eumomota superciliosa	MCI	P	P	P	P
Melanerpes pygmaeus	BI	P	P	P	P
Melanerpes aurifrons	BI	P	P	P	P
Picoides scalaris	BI	A	P	P	A
Veniliornis fumigatus	BI	P	P	P	P
Colaptes rubiginosus	BI	P	P	P	P
Dryocopus lineatus	BI	P	P	P	P
Campephilus guatemalensis	BI	A	P	P	A
Dendrocincla anabatina	BI	A	A	P	P
Dendrocincla homochroa	BI	A	P	P	P
Sittasomus griseicapillus	BI	P	P	P	P
Xiphorhynchus flavigaster	BI	P	P	P	P
Thamnophilus doliatus	UI	P	P	P	P
Camptostoma imberbe	CI	P	P	P	P
Myiopagis viridicata	MCI	P	P	P	P
Elaenia flavogaster	CI	P	P	P	P
Oncostoma cinereigulare	MCI	P	P	P	P
Tolmomyias sulphurescens	CI	P	P	P	P
Platyrinchus cancrominus	UI	A	A	P	P
Contopus virens	MCI	P	A	P	A
Contopus cinereus	MCI	P	P	P	P
Attila spadiceus	CI	P	P	P	P
Empidonax flaviventris	MCI	P	P	P	P
Empidonax virescens	MCI	A	P	A	P
Empidonax minimus	MCI	P	P	P	P
Myiarchus yucatanensis	CI	P	P	P	P
Myiarchus tuberculifer	CI	P	P	P	P

Species	guild	3-8 years old	9-15 years old	old, flat land	old, hill
Myiarchus tyrannulus	CI	P	P	P	P
Pitangus sulphuratus	CI	P	P	P	P
Myiodynastes luteiventris	CI	A	A	P	A
Megarynchus pitangua	CI	P	P	P	P
Myiozetetes similis	CI	P	P	P	P
Tyrannus melancholicus	CI	P	P	P	P
Tyrannus couchii	CI	P	P	P	P
Tityra semifasciata	CF	P	P	P	P
Tityra inquisitor	CI	A	P	A	A
Pachyramphus major	CI	P	P	P	P
Pachyramphus aglaiae	CI	P	P	P	P
Saltator coerulescens	CF	P	P	P	P
Saltator atriceps	CF	P	P	P	P
Vireo griseus	MCI	P	P	P	P
Vireo pallens	MCI	P	P	P	P
Vireo flavifrons	CI	A	A	P	P
Vireo olivaceus	CI	P	P	A	A
Vireo flavoviridis	CI	P	P	P	P
Hylophilus decurtatus	CI	P	P	P	P
Cyclarhis gujanensis	CI	P	P	P	P
Cyanocorax yncas	MCI	P	P	P	P
Psilorhinus morio	CI	P	P	P	P
Cyanocorax yucatanicus	MCI	P	P	P	P
Pheugopedius maculipectus	UI	P	P	P	P
Thryothorus ludovicianus	UI	P	P	P	P
Uropsila leucogastra	MCI	P	P	P	P
Ramphocaenus melanurus	MCI	P	P	P	P
Polioptila caerulea	CI	P	P	P	P
Polioptila plumbea	CI	P	P	P	P
Hylocichla mustelina	UI	P	A	P	P
Turdus grayi	CI	P	P	P	P
Dumetella carolinensis	UI	P	P	P	P
Melanoptila glabrirostris	MCI	P	P	P	P
Vermivora cyanoptera	CI	A	A	A	P
Oreothlypis peregrina	CI	A	A	P	A
Parula americana	CI	P	P	P	P
Setophaga petechia	MCI	A	A	P	P
Setophaga magnolia	MCI	P	P	P	P
Setophaga caerulescens	CI	A	A	P	A
Setophaga virens	CI	P	A	P	P
Setophaga dominica	MCI	P	P	A	A
Mniotilta varia	MCI	P	P	P	P
Setophaga ruticilla	CI	P	P	P	P
Seiurus aurocapilla	UI	P	P	P	A

Species	guild	3-8 years old	9-15 years old	old, flat land	old, hill
Geothlypis trichas	UI	P	A	A	A
Setophaga citrina	UI	P	P	P	P
Icteria virens	UI	P	A	A	A
Eucometis penicillata	UI	A	A	P	P
Cyanerpes cyaneus	CI	A	A	A	P
Volatinia jacarina	UG	P	P	A	A
Sporophila torqueola	UG	P	P	A	A
Tiaris olivaceus	UG	P	P	A	P
Arremonops rufivirgatus	UI	P	P	P	P
Arremonops chloronotus	UI	P	P	P	P
Habia fuscicauda	UI	P	P	P	P
Piranga roseogularis	CI	P	P	P	P
Piranga rubra	CI	A	P	A	P
Cardinalis cardinalis	MCG	P	P	P	P
Cyanocompsa parellina	UG	P	P	P	P
Passerina cyanea	UG	P	P	P	P
Granatellus sallaei	MCI	P	P	P	P
Dives dives	CI	P	P	P	P
Molothrus aeneus	UG	P	P	P	P
Icterus prosthemelas	CI	P	P	P	P
Icterus spurius	CI	A	A	P	A
Icterus cucullatus	CI	P	P	P	P
Icterus chrysater	CI	P	P	P	P
Icterus mesomelas	CI	P	P	A	A
Icterus auratus	CF	P	P	P	P
Icterus gularis	CI	P	P	P	P
Amblycercus holosericeus	UI	P	P	P	P
Euphonia affinis	CF	P	P	P	P
Euphonia hirundinacea	CF	P	P	P	P

P present, A absent Feeding guilds: BI bark insectivores, CI canopy insectivores, N nectarivores, MCI mid-canopy insectivores, UI understory insectivores, CF canopy frugivores, UF understory frugivores, UG understory granivores, MCG mid-canopy granivores