SEARCH

SEARCH BY CITATION

Keywords:

  • 4th-corner method;
  • agricultural landscapes;
  • avifauna;
  • habitat heterogeneity;
  • redundancy analysis

ABSTRACT

  1. Top of page
  2. ABSTRACTRESUMEN
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGMENTS
  7. LITERATURE CITED
  8. Supporting Information

This study examined the importance of habitat heterogeneity on the avian community composition, and investigated the scale at which species abundances respond to habitat variables. The study was conducted within a diverse landscape matrix of a shaded coffee region in Mexico. To detect at which characteristic spatial scale different species and foraging guilds respond most strongly we analyzed the effect of plot-, patch- and landscape-level variables at different spatial extent (i.e., different kilometer radii) on species composition and foraging guilds. We used redundancy analysis to identify species–environment correlations, and to identify predictor variables that best explained the bird community structure, quantified the influence of plot-, patch- and landscape-level variables on the bird community composition. In addition, we used the 4th-corner method to detect significant relationships between the dietary guilds and plot-, patch- and landscape-level variables. We recorded 12,335 individuals of 181 bird species; 105 bird species were recorded foraging within the shaded coffee plantations. We found that plot- and landscape-level variables significantly explained the bird community composition best across all scales, and were significantly correlated with the abundance of the dietary guilds. In contrast, patch-level variables were less important. Habitat composition variables (i.e., coffee, forest and agricultural area) were among the most important predictors. Canopy structure was more important than other vegetation structure variables in explaining dietary guild structure. Hence, the maintenance of a heterogeneous landscape with a high-quality matrix within an agro-ecological region enhances bird conservation.

RESUMEN

Este estudio examinó la importancia de la heterogeneidad del hábitat en la composición de la comunidad de aves, también investigó la escala a la cual las abundancias de las especies responden a variables del hábitat. Este estudio se llevo a cabo en una matriz del paisaje diversificada en una región con cafetalera con café bajo sombra. Para detectar cuál es la escala característica en la que diferentes especies y sus gremios de alimentación responden más fuertemente, nosotros analizamos el efecto de variables al nivel de parcela, parche y paisaje en diferentes extensiones espaciales (i.e., diferentes radios en kilómetros) en la composición de especies y sus gremios de alimentación. Nosotros utilizamos un análisis de redundancia (RDA por sus siglas en inglés) para identificar las correlaciones entre especies-ambiente, y para identificar variables predictivas que explicaran mejor y significativamente la estructura de la comunidad de aves. Además, usamos el método de 4th-corner para detectar las relaciones significativas entre los gremios de alimentación y variables a al nivel de parcela, parche y paisaje. Nosotros registramos 12,335 individuos de 181 especies de aves; de las cuales 105 especies de aves fueron registradas forrajeando en las plantaciones de café bajo sombra. Nosotros encontramos que las variables al nivel de parcela y paisaje explicaron significativamente la composición de la comunidad de aves a través de todas las escalas, y estuvieron significativamente correlacionadas con la abundancia de los gremios de alimentación. En contraste, las variables a nivel de parche fueron menos importantes. Las variables de composición de hábitat (i.e., área de café, bosque y agricultura) estuvieron entre las variables predictivas más importantes. La estructura del dosel fue la variable más importante en comparación a otras variables estructurales de la vegetación para explicar la estructura de los gremios de alimentación. Por lo tanto, el mantenimiento de un paisaje heterogéneo con una matriz de alta calidad dentro de una región agro-ecológica apoya la conservación de la avifauna.

The contribution of agricultural landscapes to the conservation of biodiversity has only been recognized recently (Tscharntke et al. 2005, Greenberg et al. 2008, Perfecto & Vandermeer 2008). This understanding increases the need for a land use planning strategy that incorporates the management and diversification of the anthropogenic matrix in which natural areas are embedded (Rosenzweig 2001, Bengtsson et al. 2003, Rosenzweig 2005, Toledo 2005). Many scientific studies focus on bird species richness (e.g., Wunderle & Latta 1996, Greenberg et al. 1997, Petit et al. 1999, Tejeda-Cruz & Sutherland 2004), but the relative importance of the spatial composition and configuration of the landscape on the bird community composition is poorly understood (Tscharntke et al. 2005, Komar 2006). There is therefore an urgent need to understand how the landscape mosaic, under influence of human impacts, shapes the avian community composition, such as by creating differences in the dietary guild structure.

A good example of such a landscape mosaic with different landscape elements is shaded coffee plantations in Latin America. The potential of shaded coffee plantations to conserve and maintain resident and Nearctic migrant avifauna has been debated in several studies (Rappole et al. 2003, Komar 2006), despite numerous scientific studies showing the large avian species richness in shaded coffee plantations (review in Moguel & Toledo 1999, Leyequien 2006, Perfecto et al. 2007, Philpott et al. 2008).

Landscape heterogeneity influences species assemblages, but species respond at spatial scales beyond the local plot level (Tscharntke et al. 2008). Ecological processes are influenced by factors acting across a range of scales, and so conclusions about the importance of factors that structure the assemblages must be based on observations at multiple scales (Turner 1989, Wiens 1989, Turner & Gardner 1991, Cushman & McGarigal 2002). However, studies investigating the impact of environmental factors on the composition of communities are mostly carried out at a single spatial scale (e.g., McGarigal & McComb 1995, Holland & Fahrig 2000) despite our knowledge that certain species or functional groups respond to the environment at a different spatial scale (Holland et al. 2004). Hence, there is a lack of research on how much of the variation in species composition or abundance of foraging guilds is a response of differences in structural characteristics of their environment (e.g., structural and floristic vegetation variables, patch and landscape variables), or differences in the spatial extent of these variables. One approach of estimating both the effect of environmental variables and their spatial scale on species composition is to analyze the species–environment correlations, and determine at which spatial scale the variation in species abundance can be best explained (e.g., Holland et al. 2004). We used this approach to explore the changes in the community assemblage and the link between diet and habitat type.

Our first research objective was to determine the significant influence of plot-, patch- and landscape-level habitat variables on the bird community by quantifying the conditional (or partial) effect, and the marginal (independent) effect of each habitat variables. We also aimed at detecting the characteristic scale at which species abundances respond to habitat variables, for that we used multiple radii: 1, 3, 5 and 10 km. To direct our analyses, we formulated the following hypotheses about bird community–habitat relationships from previous knowledge on the impact of habitat heterogeneity (Wiens 1989, Forman 1995, Freemark et al. 1995, Berg 1997, Schmiegelow et al. 1997, Saab 1999, Atauri & de Lucio 2001, Rickets 2001). First, we hypothesized that the structure of bird communities is significantly influenced by plot-, patch- and landscape-level variables, and that species abundance responses depend on the scale at which the variables are being measured. We expected that landscape-level variables have a larger explanatory power at all measured scales, whereas this would be lower for plot- and patch-level variables (Saab 1999). Second, we expected that the abundance of species with specialized dietary guilds or foraging strategies, e.g., bark insectivores, understory bark insectivores and large canopy frugivores increases with increasing tree-canopy cover, with highest densities in natural forest (Komar 2006) and shaded coffee areas (Tejeda-Cruz & Sutherland 2004). Third, generalist birds will be negatively correlated with increasing tree-canopy cover, and therefore found in higher abundance in areas with secondary vegetation and in agricultural areas (Komar 2006). We also expect that the abundance of birds in all dietary guilds will be greater in areas with shaded coffee (Perfecto et al. 1996, Dietsch et al. 2007), and that the abundance of insectivores (Thiollay 1995), frugivores, nectarivores (Carlo et al. 2004) and granivores will be higher in forests, whereas for omnivores a lower abundance in forested areas is expected (Calvo & Blake 1998). Omnivores are more found in agricultural and secondary vegetation areas (Calvo & Blake 1998). In addition, following Aguilar-Ortiz (1982), Greenberg et al. (1997) and Foster (2007), we expect that floristic and tree-canopy variables are positively correlated with insectivores, frugivores, nectarivores and granivores.

METHODS

  1. Top of page
  2. ABSTRACTRESUMEN
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGMENTS
  7. LITERATURE CITED
  8. Supporting Information

Study site.— Coffee plantation areas overlap with regions of high biological diversity and high level of endemism (Hardner & Rice 2002). In northern Latin America, coffee grows in areas with little remaining original forest (Philpott et al. 2007) ranging in five coffee agricultural systems differentiated according to management level (Fuentes-Flores 1979). In Mexico there are >800,000 ha in coffee cultivation (SAGARPA 2007), and the majority of these coffee areas are traditional shaded coffee plantations, cultivated by small-scale community-based growers (Moguel & Toledo 1999). These traditional shaded coffee plantations are managed as segments of a larger multiple-use systems (Toledo et al. 2003), which combine several land use practices, creating diverse landscape mosaics. Coffee plantations in Mexico are located in strategic biogeographical and ecological belts, comprising tropical and temperate forests, and holding at least 14 priority conservation areas in Mexico (Moguel & Toledo 1999). We studied the birds in the North Eastern mountain range of Puebla in the Cuetzalan region (Fig. S1), an area dominated by traditional shaded coffee plantations. In this study, we refer to traditional shaded coffee plantations as plantations with some forest trees and some planted timber and fruit trees (for details see the classification of coffee plantations in Philpott et al. 2008). This coffee region is an ideal model landscape to study the effects of landscape composition measured at different scales and the impact of shaded coffee plantations on the bird community, as it provides a diverse and intricate mosaic of different land uses in a fragmented landscape. The area is a coffee agro-ecological zone, with shaded coffee plantations. Other important land uses are seasonal crops (i.e., milpa) and grazing lands. Small patches of forests are remnants of tropical evergreen forest, with low mountain rain forest in the higher areas (700–2500 m) and lowland rain forest in the lower areas (200–700 m). The study area (19°53′24″–20°09′36″ N, 97°22′12″–36′36″ W) covers an altitudinal range of 300–1200 m, the mean annual rainfall is 3500 mm without a defined dry season and the mean annual temperature is 22–26°C. The region is situated on the eastern side of the flyway of Nearctic migrant birds.

Sampling.— We assessed bird species composition on nine sites located within a coffee agro-ecological zone. The sites varied in landscape composition, and we selected them based on (1) minimum variability in slope, exposition, farming management and age of shaded coffee plantations; (2) relative proportion of the different land covers (i.e., shaded coffee plantations, agricultural area, forest, secondary vegetation, urban area and water bodies); and (3) logistic issues such as accessibility and land tenure (for description of sites see Table S1). The area of each site was 10 × 10 km. The bird survey was carried out from November 2002 to November 2003 for a total of 121 d, and we sampled each habitat during 12 d (minimum) or 13 d (maximum), in which sample days were regularly distributed over the 12 mo, sampling at least once a month. On each sampling day we carried out 12 point-counts at each site. We used a modified double-observer method (modified from Nichols et al. 2000; for a full description of the method see Leyequien 2007), using fixed point-counts with a 10-min observation period, recording all birds seen or heard within a 25 m radius. The small size of the point-count plots minimized double counting of individuals and reduced bias due to detectability of birds in plots with dissimilar vegetation (Borghesio 2007). To avoid hour-counting bias in sampling, we alternated the order of the point-counts at every visit, and we used a distance of ca 200 m between point-counts to guarantee data independency.

We measured independent variables within the nine sites, either through remote-sensing or in situ measurements. We considered three main levels to measure the variables: plot-level (i.e., structural and compositional vegetation variables), patch-level (i.e., patch metrics) and landscape-level (i.e., composition metrics: proportion of different land cover types) yielding three sub-sets of biophysical variables (Table 1). To detect the characteristic scale at which the species abundances respond to the selected biophysical variables, we measured them at multiple radii: 1, 3, 5 and 10 km. In this study, a difference in scale refers to a difference in the size of the radii (km) at which the variables were measured; other studies have used a similar approach and found ecological meaningful results (e.g., Cushman & McGarigal 2002, 2004; Holland et al. 2004). These radii were positioned with respect to the geometric center of the sites' area. The selection of the size of the radii was based on the following criteria: the species home ranges (e.g., Terborgh et al. 1990, Baillon et al. 1992, Freemark et al. 1995) and the area of potential influence of landscape features on the bird community (for a review, see Fahrig 2003). The first predictor subset contained floristic, stand structure and vegetation cover variables at plot level, measured in situ from 0.025 ha plots sampled using standard quadrants (Kent & Coker 1992), with four replicas in each of the nine sampled landscapes. We selected the vegetation plots randomly to ensure that they captured the variation in site heterogeneity and yielded a representative sample. We calculated plot-level variables for all the scales. For the calculation of the floristic variables for the different radii, we used Effort predictor (V 1.0) to calculate the predicted species number for both tree and herb species. We estimated the Smax richness estimator, using an exponential model where the maximum number of iterations was achieved when the model improvement was <0.0001. Thereafter, we extrapolated the curve for the number of samples needed to cover the area of the maximum radius (i.e., 10 km). The Pearson r2's between the fitted curves and species accumulation curves were >0.9. We then used the predicted number of species calculated for the corresponding number of samples, which was dependent on the area under each radius. Likewise, we calculated the stand structure variables for all the scales. We used the standard deviation of the trees' height and the average height of coffee shrubs in each site as a measure of the variability in these descriptors, and used it for all the radii. For the variable ‘density of trees’ we calculated the weighted mean density, taking into account the differences in area of each land use type covered by each radius. Also, for ‘number of coffee shrubs’ we calculated an average coffee shrub density in the area covered by each radius. Additionally, we calculated the average canopy crown spread and estimated the average crown spread occurring in the area covered by each radius.

Table 1. Description of the explanatory variables at plot, patch and landscape level, including the explanatory variable type.
Explanatory setExplanatory variable typeDescriptionCode
PlotFloristicEstimated density of tree species in each radii per site within coffee plantationsTree species
 Estimated density of herb species in each radii per site within coffee plantationsHerb species
Stand structureStandard deviation of all tree heights per siteTree height
 Average height of coffee shrubs (m) per siteCoffee height
 Average density of trees estimated for the area under each radiusDensity trees
 Average number of coffee shrubs estimated for the area under each radiusCoffee shrubs
Vegetation coverAverage of crown spread of canopy estimated for the area under each radiusCrown spread
PatchConfigurationArea (ha) of coffee plantations in which each plot was locatedArea patch
 Perimeter of coffee plantation patch in which each plot was locatedEdge patch
LandscapeLandscape compositionProportion of shaded coffee plantation, forest, secondary vegetation, urban areas, agricultural landsArea coffee, area forest, area secondary vegetation, area urban, area agricultural

Additionally, we generated a land-cover map using an ETM satellite image (year 2003; 30 m spatial resolution, 70% accuracy). After image rectification and restoration, and image enhancement, we applied a Gaussian maximum likelihood classifier method, including prior probabilities to generate a supervised classification, which contained seven land-cover classes (agricultural land, shaded coffee plantations, forest, secondary vegetation, urban areas, water bodies and clouds). These land-cover classes were verified by means of a previous field sampling carried out by the Coffee Census (Consejo Mexicano del Café 2001) that provided us with digital images (point and polygon data) and tabular data of all the coffee plots existing in Cuetzalan region, including data on abandoned plots. In addition, the National Institute of Statistic and Geographic Information (INEGI 2001) provided us with a land use/land cover digital map (1:10,000). We used half of this data set to create a training set to develop a numerical description of the spectral attributes of each land cover type that were used as inputs in the supervised classification. Once we developed the spectral signatures and run the classification, we used the other half of the data to validate the classification, which was verified by ground truthing. We extracted patch area and perimeter, and landscape composition variables subsets from the supervised classification using Fragstats (version 3.3, McGarigal et al. 2002) and Arc Info (version 8.3).

We classified birds into five dietary guilds: frugivores, nectarivores, granivores, insectivores and omnivores. We categorized the species into one dietary guild, and in the case of facultative feeders we grouped them based on the primary food resource reported for the species (Leyequien 2006, Bird Life International 2008). We categorized the bird species on only one dietary guild to enable an analysis with sufficiently large sample sizes for each guild class (see review Philpott et al. 2008; e.g., Şekercioğlu et al. 2004), although we are fully aware that facultative species may shift to secondary resources under specific circumstances. This approach is in line with the recommendations of Philpott et al. (2008) to provide additional knowledge on the effects of coffee systems on species assemblages and dietary guilds. Raptors and full aerial birds were excluded from the analyses due to low sample size. The foraging strata was classified into terrestrial birds (T) that forage primarily on the ground; understory birds (U) that forage mostly within 5 m of the ground in shrubs and small trees in closed canopy forest; undergrowth birds (UG) that forage up to 3 m above ground; midstory birds (M) that generally forage above 5 m but below the canopy; and canopy birds (C) that forage in the tree canopy (Cohn-Haft et al. 1997, Henriques et al. 2003). Additionally, we assigned classes of high, medium and low sensitivity to disturbance to each of the species using the classification of Bird Life International (2008).

Statistical analysis.— First we examined the strength of species–environment relationships for the community as a whole. We tested for significant effects of the plot-, patch- and landscape-level variables upon the bird community composition, for each of the four different radii. A redundancy analysis (RDA) provides robust modeling of multivariate response data from reported species–environment correlations, and the associations of plot-, patch- and landscape-level variables could thereby be quantified. We did not intend to test specific hypothesis about cause–effect relationships of the plot-, patch- and landscape-level variables on the bird community composition. First, we used a detrended correspondence analysis to test for unimodal or linear species responses to the underlying environmental gradient; this analysis confirmed a linear response (length of gradient <3.0) (Lepš & Šmilauer 2003). We therefore used an RDA to find significant predictors that ‘best’ explained the species ordination. For the RDA we log-transformed the data and used a stepwise forward selection procedure with a Monte Carlo permutation test (using the reduced model to minimize Type I error; 999 permutations) to test for the statistical significance of each environmental variable. In the stepwise forward selection we included only those environmental variables that proved significant (P<0.05) in the final ordination. Thereafter, we used the variance partitioning method (Borcard et al. 1992) to separate the conditional (or partial) effect, and the marginal (independent) effect of each selected variable sets (i.e., plot-, patch- or landscape-level variables) to explain the observed variation in the bird community composition. The conditional effect is the variance explained by a given set of factors after removing the variance that is jointly explainable by the one or both of the other factors, while the marginal effect is the total variation explained by a given set (Cushman & McGarigal 2004). Finally, for display purposes, we only plotted the species with a fit range ≥50 percent (i.e., the proportion of explained variance) with the axes into the ordination space, the relative size of the symbols (circles) corresponds to the fraction of explained variance of a species accounted for by the model (Ter Braak & Šmilauer 1998). The species fit represents the percentage of variance of the species' abundance that is explained by the ordination (for details of used calculations, see Ter Braak & Šmilauer 1998, Lps & Šmilauer 2003). We conducted all multivariate analyses in CANOCO 4.5 (Ter Braak 1987).

For the analysis of the dietary guilds, we included only those species fulfilling the following two criteria in the analysis: (1) birds were actively foraging during point-counts, and (2) species with ≥50 percent of variance explained. We used the 4th-corner method to examine whether the different variables in the plot-, patch- and landscape-level variables that were measured at different scales (radii) were predictive of the dietary assemblage (P≤0.05) (Legendre et al. 1997). The 4th-corner method develops a matrix that relates the distinct kinds of habitat to biological traits (i.e., a habitat vs. traits matrix) using three matrices: the A matrix containing the taxon distribution (taxa vs. sites), the B matrix including the behavioral traits (taxa vs. traits) and the C matrix containing the environmental characteristics (sites vs. environment or habitat) (Bonada et al. 2007). In this study, traits refer to the dietary guilds included in the analysis (i.e., nectarivores, frugivores, insectivores, granivores and omnivores), and taxa represent the species. The species were categorized into each dietary guild based on the primary food resource reported for the species (Leyequien 2006, Bird Life International 2008). We constructed a matrix A ( p×n) containing data on the presence or absence of p species at n sites; a second matrix B ( p×q) describing q behavioral traits (i.e., dietary guilds) of the same p species; and a third table C (m×n) containing information about m habitat characteristics (plot-, patch- and landscape-level variables) at the n sites. In our case, the A matrix was the bird species matrix transformed to presence/absence (species × sites: 33 × 9, 57 × 9, 49 × 9, 104 × 9 from smaller to larger radii, respectively). The behavioral matrix comprised the dietary guilds of each bird species (species × trait categories): 33 × 5, 57 × 5, 49 × 5, 104 × 5 from smaller to larger radii, respectively; and the environmental-habitat matrix included 14 environmental variables reflecting plot-, patch- and landscape-level variables that were measured at different radii (scales) (habitat variables × sites: 14 × 9). We carried out the 4th corner analysis for each radius (scale). The Holm's procedure was used to adjust individual probabilities (Holm 1979) at the 5 percent significance level after 999 permutations (to minimize a Type I error). For the 4th-corner statistics, we used a qualitative behavioral matrix and a quantitative habitat matrix. The Pearson product–moment correlation coefficient (r) was used to indicate the strength and direction of association between the behavioral characteristics of species (i.e., dietary guilds) and the habitat characteristics (plot-, patch- and landscape-level variables). We used Model 2 (environmental control over species assemblages), because we assumed that species assemblages are dependent on the habitat characteristics of the sites where they are actually found (Legendre & Legendre 1998). We carried out all analyses using the 4th corner program (DOS version, Legendre et al. 1997).

RESULTS

  1. Top of page
  2. ABSTRACTRESUMEN
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGMENTS
  7. LITERATURE CITED
  8. Supporting Information

Effects on bird community.— We recorded 12,335 individuals over 181 bird species. For the RDA analyses, the most important environmental variables predicting bird community composition in all the models were those at plot and landscape level, whereas patch-level variables had no significant contribution (Fig. 1). Table S2 shows only the significant plot- and landscapes-level variables in each considered scale (radii). At the scale of 1 km only plot-level variables (P<0.05) had a significant contribution, namely the density of the trees and the coffee height. For the 3 km scale only tree height and two landscape-level variables, i.e., area of shaded coffee plantations and the area of forest were significant. At the 5 km scale only landscape-level variables were significant, i.e., the area of shaded coffee plantations, the area of forest and the area of agricultural lands. Finally, for the 10 km scale also the same three landscape-level variables were significant (the area of shaded coffee plantations, the area of forest and the area of agricultural lands). The percentage of explained variance by plot- and landscape-level variables is shown in Figure 2 for all models. The total amount of explained variance is relatively high (≥50%), and the marginal effect is larger than the conditional effect of the variables, indicating that these variables individually contribute more in explaining the species data of the community assemblage than their shared part.

image

Figure 1.  Correlation biplot resulted from the redundancy analysis, where the symbol size corresponds to the values of the fit of species into the ordination space: (A) up to 1 km, (B) 3 km, (C) 5 km, and (D) 10 km. Species numbers, Group I–Group II and fit values are presented in Table S5.

Download figure to PowerPoint

image

Figure 2.  The cumulative amount of explained variance by each significant variable alone (marginal effect), and the conditional effect explained by all significant variables, as calculated from the variance partitioning method. The unexplained variance is also depicted.

Download figure to PowerPoint

The total number of species that were explained in each model with a fit range of ≥50 percent of explained variance was 29.3 percent (53) for the 1 km model, 50.3 percent (91) for the 3 km, 37.6 percent (68) for the 5 km model and 49.2 percent (89) for the 10 km models.

Effects on dietary guilds.— A total of 105 bird species were recorded foraging within the shaded coffee plantations over 1360 observations, of which 68.6 percent (72 species) were resident bird species and 31.4 percent (33) were migratory species (Table S3). The abundance of bark insectivores (Lepidocolaptes affinis, Xiphorhynchus flavigaster, Picoides fumigatus), understory bark insectivores (Sittasomus griseicapillus) and large canopy frugivores (Pionus senilis, Tytira semifasciata) was positively correlated with forest area across the medium and larger scales (i.e., 3, 5 and 10 km). Interestingly, all species with a near threatened IUCN (2007) status showed a positive correlation with forest area at medium and larger scales, i.e., Sporophila aurita (midstory granivore), Vireo belli (understory/undergrowth insectivore) and Contopus cooperi (canopy insectivore). Other species with medium sensitivity to disturbance were also positively correlated with forest area across the medium to larger spatial scales (e.g., Amazilia cyanocephala, canopy nectarivore; Aulacorhynchus prasinus, canopy omnivore; Euphonia elegantisima, canopy frugivore; Henicorhina leucosticta, underground/undergrowth insectivore; and Chlorospingus ophtalmicus, midstory granivore).

The abundance of bark insectivores, understory bark insectivores and large canopy frugivores species (e.g., L. affinis, X. flavigaster, S. griseicapillus, P. senilis, T. semifasciata) was negatively correlated with coffee. Overall, across medium to larger scales, the species positively correlated with coffee area were obligatory or facultative insectivores (e.g., Polioptila caerulea, Cyclarhis gujanensis, Mniotilta varia, canopy insectivore; Icterus gradacauda, Icterus gularis and Piranga rubra, canopy insectivore/frugivore), and omnivores (e.g., Cyanocorax yncas, canopy omnivore; and Momotus momota, midstory omnivore). Interestingly, Amazilia candida (canopy nectarivore/insectivore) with medium sensitivity to disturbance showed a positive correlation with coffee area across medium to larger scales.

Results from the 4th-corner method analyses showed significant relationships between landscape and plot variables and the dietary guilds across all spatial scales, whereas patch variables were not significant in any dietary guild across any spatial scale. Overall, although we found positive and negative significant correlations, the correlation coefficients were relatively low. Table S4 presents the significant relationships for each dietary guild per spatial scale. Specifically, the abundance of insectivores was positively correlated with floristic and structural variables of the arboreal layer at the 1 and 3 km radii, and with coffee area at the 3 km radii, whereas secondary vegetation and coffee height were negatively correlated. The abundance of granivores was positively correlated with only coffee area and coffee height at 1 and 10 km radii, respectively; in contrast, forest, secondary vegetation and agricultural areas were negatively correlated as well as coffee shrubs across all radii except the 1 km. The frugivores were positively correlated with coffee and forest areas across medium scales (radii of 3 and 5 km), and with structural variables of the coffee layer and vegetation across all scales, whereas agricultural area (at 1 and 10 km radii) and structural characteristic of trees (across all radii except at 5 km) were negatively correlated. Nectarivores presented a negative correlation with most of the vegetation structure and landscape composition variables across all scales, with the exception of coffee area (3 km radius), secondary vegetation (5 km) and herb species (10 km). Other vegetation structure variables such as tree density and coffee shrubs did not show consistent positive or negative correlations across scales. Finally, omnivores were positively correlated with tree density and coffee shrubs density over nearly all scales, and at larger scales with coffee area, tree height and crown spread and herb species. We also found that omnivores were positively correlated with agricultural area (across all radii except at 10 km), whereas negatively correlated with forest area (at 1 and 3 km radii).

DISCUSSION

  1. Top of page
  2. ABSTRACTRESUMEN
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGMENTS
  7. LITERATURE CITED
  8. Supporting Information

Effects on bird community.— Our results show that the bird community composition is significantly influenced by plot- and landscape-level variables. However, patch-level variables were not significant, independent of the scale (radii) of analysis. Habitat composition variables, rather than vegetation features, were the most important and frequent predictors of bird community composition. Habitat composition variables overall accounted for more than 50 percent of variability in community composition in the medium to larger scales. Specifically, forest, coffee and agricultural areas were significant at the medium and larger scales (i.e., 3, 5 and 10 km). Coffee and forest areas were consistently significant in explaining community composition. This underlines that the landscape matrix strongly influences the bird assemblage structure within the coffee agro-ecological zone, and suggests the necessity to consider the surrounding matrix for management and conservation plans for birds. It had been predicted that population persistence could be maintained if a low-quality matrix is converted to a high-quality matrix by maintaining heterogeneous landscapes with a diversity of vegetated features (Fahrig 2001). A high-quality matrix facilitates the individuals' influx maintaining source–sink populations (Vandermeer & Carvajal 2001). Moreover, our results support Fahrig's findings and confirm the significant influence of habitat heterogeneity on bird communities in fragmented agricultural landscapes (e.g., McGarigal & McComb 1995, Saab 1999, Atauri & de Lucio 2001). Moreover, the results indicate that in order to understand the bird community structure it is a prerequisite to incorporate observations measured at multiple scales (Cushman & McGarigal 2004). More specifically, the marginal effects of coffee, forest and agricultural area were significantly high in the medium to larger scales with relatively comparable contributions of forest and coffee area. However, the abundance of many specialist species positively correlated with the percentage of forest in the landscape, suggesting that maintaining forest or other habitats that maintain landscape connectivity (e.g., shaded coffee) may benefit conservation of these rare species (Petit et al. 1999, Roberts et al. 2000, Tejeda-Cruz & Sutherland 2004, Shankar-Raman 2006). In addition, it has been argued that the response in large-scale population processes (e.g., dispersal) to landscape heterogeneity should be assessed at large spatial scales; the benefit of any one habitat type is likely to be dependent at the regional population level (Benton et al. 2003). Additionally, at the smallest scale (1 km) only the structural variables of vegetation (plot-level variables) in shaded coffee plantations showed a significant association with the community composition. The latter may reflect how a group of species (community) perceive the landscape, in which the vegetation structure determines the heterogeneity at local scale (Benton et al. 2003). This local heterogeneity in vegetation structure is an important factor for birds as it affects, for example, the accessibility and visibility of both potential prey and potential predators (Lima & Dill 1990), or the reproductive success of specific species, e.g., thrush species, that depend on woods that have a thick, brushy understory, or tanager species that strongly rely on epiphytes' presence (Cruz-Angon et al. 2008). In contrast, Rotenberry (1985) had proposed that the distribution and abundance of bird species may be more related to plant taxonomic composition than with the structure and configuration of the vegetation; however, the flaw in Rotenberry's study is that the diversity indices proved insufficient to detect the aforementioned relationships. The abovementioned results stress the importance of ecological studies based on community structure and composition rather than solely on species richness or diversity.

In our study, we did not detect any significant effects of patch area on the bird community, which is in agreement with previous studies (Andrén 1994, Mönkkönen & Reunanen 1999, Bhagwat et al. 2005). The effects of landscape composition appear to veil the influence of patch size on the community composition. The latter could be explained because habitat patches are components of the landscape mosaic and the occurrence of a particular species in a patch may be a function of the neighboring habitat and not only of patch size, especially in fragmented landscapes (Andrén 1994, Bender & Fahrig 2005). In contrast, Connor et al. (2000) found in a meta-analysis that birds generally showed moderately to large positive effects to differences in patch size; however, these relationships were stronger in temperate and boreal environments than in tropical ones. Additionally, we found no significant effect of patch edge on the community composition, while previous studies reported small effects of edge effects on specific species and dietary guilds (Saab 1999, Sallabanks et al. 2000). Villard et al. (1999) found that area-sensitive bird species do not avoid edges in most forested areas, and it has been hypothesized that this response is due to increased food availability at edges (Schmiegelow & Mönkkönen 2002). Helle (1983), for instance, showed that insectivorous birds clearly preferred edges.

Effects on dietary guilds.— The results suggest that bark insectivores, understory bark insectivores, large canopy frugivores and species with high sensitivity to disturbance are dependent on forest, whereas shaded coffee area benefits facultative and obligatory insectivores, omnivores and in lower proportion midstory and understory/undergrowth granivores. This dietary guild structure and forage strata suggest that the area of shaded coffee plantations increases the availability of food resources for a variety of dietary guilds. However, the structural characteristics in the canopy layer and the increase in the herb stratum may also explain the strong associations between shaded coffee and insectivores, omnivores and granivores. Moreover, the bark insectivores, understory bark insectivores and large canopy frugivores species that were detected as feeding in the shaded coffee plantations could make facultative use of shaded plantations. Tejeda-Cruz and Sutherland (2004) also found that pristine forest held higher numbers and abundance of disturbance-sensitive species, and a low abundance of frugivores. However, in contrast with our study they found lower abundances of insectivores. Specifically, it has been stressed that because of the relative sedentary nature of understory bark insectivores they may be less able than other bird species to travel from nesting areas to foraging areas that are unsuitable for nesting (Şekercioğlu et al. 2004). Other studies also reported that forest-associated species were reduced in shaded coffee plantations (Greenberg et al. 1997, Roberts et al. 2000, Perfecto & Armbrecht 2003). Nevertheless, various forest species with high and medium sensitivity have been reported in shaded coffee plantations (Tejeda-Cruz & Sutherland 2004), including our study. More research is certainly required to evaluate the reproduction success of different species from vulnerable dietary guilds within certain forage strata (e.g., bark insectivores, understory bark insectivores and large canopy frugivores), comparing shaded coffee plantations with natural forest.

The patch habitat variables that were measured in this study were not significantly correlated with any of the dietary guilds. In general, results showed that canopy structure (i.e., crown spread) could not explain the differences in densities of granivores and omnivores across all spatial scales. The latter results suggest that these dietary guilds are probably not affected by variation in canopy structure whereas insectivores were positively correlated at small and medium scales, suggesting that they benefit from the available food that canopy provides. Other studies carried out in shaded coffee plantations demonstrated that birds, specifically insectivorous, frequently forage in the tree canopies; thus, canopy structure is an important vegetation variable, influencing the dietary guild composition as it is rich in arthropods (Greenberg et al. 1997, Wunderle & Latta 1998, Jones et al. 2002). For nectarivores and frugivores there was no clear pattern across the spatial scales; however, crown spread was only significantly negatively correlated to the abundance of these guilds. Conversely, it has been argued that canopy structure is only an indirect variable as it is correlated with other important variables such as food availability (Komar 2006). Thus, there is a need for more specific studies about food supply, vegetation structure and composition in shaded coffee plantations, and the use by birds.

Other interesting correlations appeared between the amount of different habitats in the landscape matrix and the observed number of birds per dietary guilds within the shaded coffee plantations. Specifically, agricultural area was positively correlated with omnivores (generalists and opportunistic feeders), birds that probably benefit from the variety of food resources available in the milpas, the most frequent agricultural use in the study region. However, frugivores, nectarivores and granivores were negatively correlated with the agricultural area. Specifically, frugivores and nectarivores could be affected by the reduction in plant species and the replacement of the tree canopies by crop species. Greenberg et al. (1997) reported that the majority of the bird observations within the coffee plantations were registered in the canopy, as most of the fruit and nectar resources is located in the canopy (Komar 2006). Thus, the frugivore and nectarivore species that we registered as using the food resources within the shaded coffee plantations appeared to be negatively affected by the increase of the agricultural area. Interestingly, the coffee area was positively correlated with all dietary guilds at the medium scale, with the exception of the nectarivores that showed a negative correlation at this 5 km scale. The aforementioned results imply that at medium scales the increase in shaded coffee plantations' area could benefit virtually all the dietary guilds; however, at the larger scale the coffee area was not a significant predictor. It had been argued that shaded coffee plantations could facilitate bird dispersal across the landscape, in particular forest birds (Tejeda-Cruz & Sutherland 2004), and that they may function as transition areas among different habitat types, either natural or humanized (Komar 2006). However, the lack of explanatory power of coffee area at larger scales points out the necessity to consider the surrounding landscape matrix and the proportion of the different land covers surrounding the shaded coffee plantations more fully if we aim to develop prudent conservation measures in coffee regions. Besides foraging, other important factors need to be systematically evaluated such as survivorship and reproduction success. In addition, the forest area showed a negative correlation with omnivores at the smallest scales, and with nectarivores and granivores at the 3 km scale, whereas it was positively correlated with frugivores at the smaller scales, and with insectivores at the 3 km scale. Specifically, omnivores, nectarivores and granivores (in our case ground-feeding granivores) may prevail outside forest fragments as they could benefit from the food resources provided by the patchier canopy and denser shrubbery offered by the shaded coffee plantations, and by the high diversity of plant species that could provide food. Leyequien and Toledo (2009) reported higher resource density in coffee areas (i.e., 35–150 plant species per 0.5–2 ha), providing a wide variety of food resources for frugivores, nectarivores and insectivores. However, there were relatively few obligatory frugivores recorded in our study mostly from the families Turdidae, Mimidae and Fringilidae, and even less registered as feeding within the shaded coffee plantations. Moreover, the seasonal variation in plant phenology, which is an important factor influencing the dietary guild structure (Dietsch et al. 2007), was not included in our analyses. In specific, the insectivores could shift in the resource use from shaded coffee plantations to forest and vice versa over the seasons. In accordance, Dietsch et al. (2007) found that insectivores presented a variation in resource use originated from changes in flowering and fruiting.

Management implications.— Our results support the perception that plot- and landscape-level variables are important factors shaping the bird community composition. Specifically, habitat composition variables, such as coffee and forest area within the coffee agro-ecological zone, were important factors shaping the community composition. Hence, a high-quality heterogeneous landscape contributes to bird conservation, including some forest birds. However, it is important to stress that several studies have emphasized the need to identify and set a threshold for coverage of suitable habitat, below which there may be excessive fragmentation (e.g., Santos & Tellería 1997, Hughes et al. 2002). In addition, shaded coffee plantations provide food resources for all the dietary guilds, although we expect that the seasonal use by certain dietary guilds, e.g., frugivore or insectivore species may depend on the plant phenology, especially of the canopy species. As there is a clear dependence of many birds on natural forest (Daily et al. 2001, Komar 2006), it needs to be emphasized that the conversion of natural forest into any other land cover is not preferable, and that shaded coffee plantations cannot substitute them. However, the maintenance of canopy species within the shaded coffee plantations, and consequently the canopy plant diversity and the vertical vegetation heterogeneity, is an important factor to include in management plans (Hughes et al. 2002, Laube et al. 2008). Moreover, it has been suggested that most tropical species strongly depend on secondary habitats (that surrounds natural reserves) encompassed by the landscape matrix, such as shaded coffee plantations, and these plantations thus play a crucial role in the maintenance and conservation of biodiversity (Bhagwat et al. 2008). Finally, the preservation of a diverse landscape matrix and specifically agro-ecosystems such as shaded coffee plantations can offer more stable livelihoods to the farmers in terms of a diversified food supply and economic profits with lower risks (avoiding dependence on only one commodity), and the promotion of forested land use, improving overall forest connectivity and minimizing habitat loss and fragmentation (Philpott et al. 2008, Tscharntke et al. 2008).

ACKNOWLEDGMENTS

  1. Top of page
  2. ABSTRACTRESUMEN
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGMENTS
  7. LITERATURE CITED
  8. Supporting Information

We thank the Cooperativa Agropecuaria Tosepan Titataniske for facilitating the logistics and providing proper permits to carry out the fieldwork. We also thank S. López de Aquino for field assistance, F. Langevelde for GIS data generation and T. A. Essens for useful suggestions about statistical methods. We also thank two anonymous reviewers that provided valuable comments and helped us improve the manuscript. Funding for this research was provided by the CONACYT grant 149557 and the fellowship granted by the Foundation for Scientific Research in the Tropics to E. Leyequién.

LITERATURE CITED

  1. Top of page
  2. ABSTRACTRESUMEN
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGMENTS
  7. LITERATURE CITED
  8. Supporting Information
  • Aguilar-Ortiz, F. 1982. Estudio ecológico de las aves del cafetal. In E.Ávila-Jiménez and A.Gomez-Pompa (Eds.). Estudios ecológicos en el agroecosistema cafetalero, pp. 103128. Instituto Nacional de Investigaciones Sobre Recursos Bióticos, Xalapa, Veracruz, México.
  • Andrén, H. 1994. Effects of habitat fragmentation on birds and mammals in landscapes with different proportions of suitable habitat: A review. Oikos 71: 355366.
  • Atauri, J. A., and J. V. de Lucio. 2001. The role of landscape structure in species richness distribution of birds, amphibians, reptiles and lepidopterans in Mediterranean landscapes. J. Landscape Ecol. 16: 147159.
  • Baillon, F., S. Benvenuti, and P. Ioale. 1992. Fidelity to non breeding site in some species of birds in Senegal. Trop. Zool. 5: 3143.
  • Bender, D. J., and L. Fahrig. 2005. Matrix structure obscures the relationship between interpatch movement and patch size and isolation. Ecology 86: 10231033.
  • Bengtsson, J. P., Angelstam, and T. Elmqvis. 2003. Reserves, resilience, and dynamic landscapes. Ambio 32: 38996.
  • Benton, T. G., J. A. Vickery, and J. D. Wilson. 2003. Farmland biodiversity: Is habitat heterogeneity the key? Trends Ecol. Evol. 18: 182188.
  • Berg, Å. 1997. Diversity and abundance of birds in relation to forest fragmentation, habitat quality and heterogeneity. Bird Study 44: 355366.
  • Bhagwat, S. A., C. G. Kushalappa, P. H. Williams, and N. D. Brown. 2005. Landscape approach to biodiversity conservation of sacred groves in the Western Ghats of India. Conserv. Biol. 19: 18531862.
  • Bhagwat, S. A., K. J. Willis, H. J. B. Birks, and R. J. Whittaker. 2008. Agroforestry: A refuge for tropical biodiversity? Trends Ecol. Evol. 23: 261267.
  • Bird life international. 2008. Species factsheet. Available at http://www.birdlife.org/ (accessed January 3, 2008).
  • Bonada, N., M. Rieradevall, and N. Pratonada. 2007. Macroinvertebrate community structure and biological traits related to flow permanence in a Mediterranean river network. Hydrobiologia 589: 91106.
  • Borcard, D., P. Legendre, and P. Drapeau. 1992. Partialling out the spatial component of ecological variation. Ecology 73: 10451055.
  • Borghesio, L. 2007. Effects of human subsistence activities on forest birds in northern Kenya. Conserv. Biol. 22: 384394.
  • Calvo, L., and J. Blake. 1998. Bird diversity and abundance on two different shade coffee plantations in Guatemala. Bird Conserv. Int. 8: 297308.
  • Carlo, T. A., J. A. Collazo, and M. J. Groom. 2004. Influences of fruit diversity and abundance on bird use of two shaded coffee plantations. Biotropica 36: 602614.
  • Cohn-Haft, M., A. Whittaker, and P. C. Stouffer. 1997. A new look at the “species-poor” central Amazon: The avifauna north of Manaus, Brazil. Ornithol. Monogr. 48: 205235.
  • Connor, E. F., A. C. Courtney, and J. M. Yode. 2000. Individuals–area relationships: The relationship between animal population density and area. Ecology 81: 734748.
  • Consejo mexicano del cafe. 2001. Censo cafetalero. Consejo mexicano del cafe, Puebla, México.
  • Cruz-Angon, A., T. S. Sillett, and R. Greenberg. 2008. An experimental study of habitat selection by birds in a coffee plantation. Ecology 89: 921927.
  • Cushman, S. A., and K. McGarigal. 2002. Hierarchical, multiscale decomposition of species–environmental relationships. Landscape Ecol. 17: 637646.
  • Cushman, S. A., and K. McGarigal. 2004. Patterns in the species–environment relationship depend on both scales and choice of response variables. Oikos 105: 117124.
  • Daily, G. C., P. R. Ehrlich, and G. A. Sanchez-Azofeifa. 2001. Countryside biogeography: Use of human-dominated habitats by the avifauna of southern Costa Rica. Ecol. Appl. 11: 113.
  • Dietsch, T. V., I. Perfecto, and R. Greenberg. 2007. Avian foraging behaviour in two different types of coffee agroecosystem in Chiapas, Mexico. Biotropica 39: 232240.
  • Fahrig, L. 2001. How much habitat is enough? Biol. Conserv. 100: 6574.
  • Fahrig, L. 2003. Effects of habitat fragmentation on Biodiversity. Annu. Rev. Ecol. Evol. Syst. 34: 487515.
  • Forman, R. T. T. 1995. Land mosaics: The ecology of landscapes and regions. Cambridge University Press, New York, New York.
  • Foster, M. S. 2007. The potential of fruit trees to enhance converted habitats for migrating birds in southern Mexico. Bird Conserv. Int. 17: 4561.
  • Freemark, K. E., J. B. Dunning, S. J. Hejl, and J. R. Probst. 1995. A landscape ecology perspective for research, conservation and management. In T. E.Martin and D. M.Finch (Eds.). Population ecology and conservation of neotropical migrant birds, pp. 381427. Oxford University Press, New York, New York.
  • Fuentes-Flores, R. 1979. Coffee production systems in Mexico. In F.De las Salas (Ed.). Workshop on agroforestry systems in Latin America, pp. 6072. Centro Agronómico Tropical de Investigación y Enseñanza, Turrialba, Costa Rica.
  • Greenberg, R., P. Bichier, and J. Sterling. 1997. Bird populations in rustic and planted shade coffee plantations of eastern Chiapas, México. Biotropica 29: 501514.
  • Greenberg, R., I. Perfecto, and S. M. Philpott. 2008. Agroforests as model systems for tropical ecology. Ecology 89: 913914.
  • Hardner, J., and R. Rice. 2002. Rethinking green consumerism. Sci. Am. 286 (5): 8995.
  • Helle, P. 1983. Bird communities in open ground–climax forest edges in northeastern Finland. Oulanka Rep. 3: 3946.
  • Henriques, L. M. P., J. M. Jr. Wunderle, and M. R. Willig. 2003. Birds of the Tapajós National Forest, Brazilian Amazon: A preliminary assessment. Ornitol. Neotrop. 14: 307338.
  • Holland, J. D., D. G. Bert, and L. Fahrig. 2004. Determining the spatial scale of species' response to habitat. BioScience 54: 227233.
  • Holland, J. D., and L. Fahrig. 2000. Effect of woody borders on insect density and diversity in crop fields: A landscape-scale analysis. Agric., Ecosyst. Environ. 78: 115122.
  • Holm, S. 1979. A simple sequentially rejective multiple test procedure. Scand. J. Stat. 6: 6570.
  • Hughes, J. B., G. C. Daily, and P. R. Ehrlich. 2002. Conservation of tropical forest birds in countryside habitats. Ecol. Lett. 5: 121129.
  • INEGI. 2001. Land use/land cover digital map. Available at http://www.inegi.org.mx/inegi/default.aspx (accessed January 27, 2003).
  • IUCN. 2007. Species' red list status. Available at http://www.iucnredlist.org/info/programme (accessed January 3, 2008).
  • Jones, J., P. Ramoni-Perazzi, E. H. Carruthers, and R. J. Tobertson. 2002. Species composition of bird communities in shade coffee plantations in the Venezuelan andes. Ornithol. Neotrop. 13: 397412.
  • Kent, M., and P. Coker. 1992. Vegetation description and analysis. A practical approach. John Wiley & Sons, Chichester, UK.
  • Komar, O. 2006. Priority contribution. Ecology and conservation of birds in coffee plantations: A critical review. Bird Conserv. Int. 16: 123.
  • Laube, I., N. Breitbach, and K. Bohning-Gaese. 2008. Avian diversity in a Kenyan agroecosystem: Effects of habitat structure and proximity to forest. J. Ornithol. 149: 181191.
  • Legendre, P., R. Galzin, and M. L. Harmelin-Vivien. 1997. Relating behaviour to habitat: Solutions to the fourth-corner problem. Ecology 78: 547562.
  • Legendre, P., and L. Legendre. 1998. Numerical ecology. Developments in environmental modelling 20. Elsevier, Amsterdam, The Netherlands.
  • Lepš, J., and P. Šmilauer. 2003. Multivariate analysis of ecological data using CANOCO. Cambridge University Press, Cambridge, UK.
  • Leyequien, E. 2006. Birds, coffee and spatial complexity: The diversity puzzle. Wageningen University, Wageningen, The Netherlands.
  • Leyequien, E. 2007. Influence of body size on coexistence of bird species. Ecol. Res. 22: 735741.
  • Leyequien, E., and V. M. Toledo. 2009. Flores y aves de cafetales: Ensambles de biodiversidad en paisajes humanos. Biodiversitas 83: 710.
  • Lima, S. L., and L. M. Dill. 1990. Behavioral decisions made under the risk of predation: A review and prospectus. Can. J. Zool. 68: 619640.
  • McGarigal, K., S. A. Cushman, and M. C. Neel. 2002. FRAGSTATS: Spatial pattern analysis program for categorical maps. Computer software program. University of Massachusetts, Amherst, Massachusetts.
  • McGarigal, K., and W. McComb. 1995. Relationships between landscape structure and breeding birds in the Oregon coast range. Ecol. Monogr 65: 235260.
  • Moguel, P., and V. M. Toledo. 1999. Biodiversity conservation in traditional coffee systems in Mexico. Conserv. Biol. 13: 112.
  • Mönkkönen, M., and P. Reunanen. 1999. On critical thresholds in landscape connectivity: A management perspective. Oikos 84: 302305.
  • Nichols, J. D., J. E. Hines, J. R. Sauer, F. W. Fallon, J. E. Fallon, and P. J. Heglund. 2000. A double-observer approach for estimating detection probability and abundance from point counts. Auk 117: 393408.
  • Perfecto, I., and I. Armbrecht. 2003. The coffee agroecosystem in the Neotropics: Combining ecological and economic goals. In J. H.Vandermeer (Ed.). Tropical agroecosystems, pp. 159194. CRC Press, Boca Raton, Florida.
  • Perfecto, I., I. Armbrecht, S. M. Philpott, T. Dietsch, and L. Soto-Pinto. 2007. Shaded coffee and the stability of rainforest margins in Latin America. In T.Tscharntke, C.Leuschner, M.Zeller, E.Guhudja, and A.Bidin (Eds.). The stability of tropical rainforest margins: Linking ecological, economic, and social constraints of land use and conservation. Environmental Science Series, pp. 227264. Springer, Heidelberg, Germany.
  • Perfecto, I., R. Rice, R. Greenberg, and M. van der Voort. 1996. Shade coffee: A disappearing refuge for biodiversity. BioScience 46: 598603.
  • Perfecto, I., and J. Vandermeer. 2008. Biodiversity conservation in tropical ecosystems. A new conservation paradigm. Ann. N.Y. Acad. Sci. 1134: 173200.
  • Petit, L. J., D. R. Petit, D. G. Christian, and H. D. W. Powell. 1999. Bird communities of natural and modified habitats in Panama. Ecography 22: 292304.
  • Philpott, S. M., W. J. Arendt, I. Armbrecht, P. Bichier, T. V. Diestch, C. Gordon, R. Greenberg, I. Perfecto, R. Reynoso-Santos, L. Soto-Pinto, C. Tejeda-Cruz, G. Williams-Linera, J. Valenzuela, and J. M. Zolotoff. 2008. Biodiversity loss in Latin American coffee landscapes: Review of the evidence on ants, birds, and trees. Conserv. Biol. 22: 10931105.
  • Philpott, S. M., P. Bichier, R. Rice, and R. Greenberg. 2007. Field-testing ecological and economic benefits of coffee certification programs. Conserv. Biol. 21: 975985.
  • Rappole, J. H., D. I. King, and J. H. Vega-Rivera. 2003. Coffee and conservation. Conserv. Biol. 17: 334336.
  • Rickets, T. H. 2001. The matrix matters: Effective isolation in fragmented landscapes. Am. Nat. 158: 8799.
  • Roberts, D. L., R. J. Cooper, and L. J. Petit. 2000. Flock characteristics of ant-following birds in premontane moist forest and coffee agroecosystems. Ecol. Appl. 10: 14141425.
  • Rosenzweig, M. L. 2001. Loss of speciation rate will impoverish future diversity. Proc. Natl. Acad. Sci. USA 98: 54045410.
  • Rosenzweig, M. L. 2005. Avoiding mass extinction: Basic and applied challenges. 24th Annual Midwest Ecology and Evolution Conference, Department of Ecology and Evolutionary Biology, University of Arizona, Tucson, Arizona.
  • Rotenberry, J. T. 1985. The role of habitat in avian community composition: Physiognomy or floristics? Oecologia 67: 213217.
  • Saab, V. 1999. Importance of spatial scale to habitat use by breeding birds in riparian forests: A hierarchical analysis. Ecol. Appl. 9: 135151.
  • SAGARPA. 2007. Avance de siembras y cosechas perennes 2007, riego y temporal. Available at http://w2.siap.sagarpa.gob.mx/Brio/ihtml/OpenDoc?DocInstanceID=00000110a8ac0920-0000-111f-0a0b0236&autologin=yes&cultivo=5500&DocUUID=0000010f53bb0965-0000-0bf3-0a0b0236&pass=sisprro&DocVersion=1&ciclo=3 (accessed January 3, 2008).
  • Sallabanks, R., J. R. Walters, and J. A. Collazo. 2000. Breeding bird abundance in bottomland hardwood forests: Habitat, edge, and patch size effects. Condor 102: 748758.
  • Santos, T., and J. L. Tellería. 1997. Efectos de la fragmentación sobre la saves insectívoras forestales de dos localidades europeas. Ardeola 44: 113117.
  • Schmiegelow, F. K. A., C. S. Machtans, and S. J. Hannon. 1997. Area boreal birds resilient to forest fragmentation? An experimental study of short-term community responses. Ecology 78: 19141932.
  • Schmiegelow, F. K. A., and M. Mönkkönen. 2002. Habitat loss and fragmentation in dynamic landscapes: Avian perspectives from the boreal forest. Ecol. Appl. 12: 375389.
  • Şekercioğlu, C. H., G. C. Daily, and P. R. Ehrlich. 2004. Ecosystem consequences of bird declines. Proc. Natl. Acad. Sci. USA 101: 1804218047.
  • Shankar-Raman, T. R. 2006. Effects of habitat structure and adjacent habitats on birds in tropical rainforest fragments and shaded plantations in the Western Ghats, India. Biodivers. Conserv. 15: 15771607.
  • Tejeda-Cruz, C., and W. Sutherland. 2004. Bird responses to shade coffee production. Anim. Conserv. 7: 169179.
  • Ter Braak, C. J. F. 1987. The analysis of vegetation–environment relationships by canonical correspondence analysis. Plant Ecol. 69: 6977.
  • Ter Braak, C. J. F., and T. Šmilauer. 1998. CANOCO reference manual and user's guide to CANOCO for windows: software for canonical community ordination (version 4.5). Microcomputer Power, Ithaca, New York.
  • Terborgh, J., S. K. Robinson, T. A. Parker III, C. A. Munn, and N. Pierpont. 1990. Structure and organization of an Amazonian forest bird community. Ecol. Monogr. 60: 213238.
  • Thiollay, J. M. 1995. The role of traditional agroforests in the conservation of rain forest bird diversity in Sumatra. Conserv. Biol. 9: 335353.
  • Toledo, V. M. 2005. Repensar la conservación ¿Áreas naturales protegidas o estrategia bioregional? Gaceta Ecol. 77: 6782.
  • Toledo, V. M., B. Ortiz-Espejel, L. Cortés, P. Moguel, and X. Ordoñez. 2003. The multiple use of tropical forests by indigenous peoples in Mexico: Case of adaptive management. Conserv. Ecol. 7: 9. Available at http://www.consecol.org/vol7/iss3/art9 (accessed October 12, 2005).
  • Tscharntke, T., M. Klein, A. Kruess, I. Steffan-Dwenter, and C. Thies. 2005. Landscape perspectives on agricultural intensification and biodiversity—ecosystem service management. Ecol. Lett. 8: 857874.
  • Tscharntke, T., C. H. Şekercioğlu, T. V. Dietsch, N. S. Sodhi, P. Hoehn, and M. Tylianakis. 2008. Landscape constraints on functional diversity of birds and insects in tropical agroecosystems. Ecology 89: 944951.
  • Turner, M. G. 1989. Landscape ecology: The effect of pattern and process. Ann. Rev. Ecol. Syst. 20: 17197.
  • Turner, M. G., and R. H. Gardner. 1991. Quantitative methods in landscape ecology: The analysis and interpretation of landscape heterogeneity. Springer-Verlag, New York, New York.
  • Vandermeer, J., and R. Carvajal. 2001. Metapopulation dynamics and the quality of the matrix. Am. Na. 158: 211220.
  • Villard, M. A., M. K. Trzcinski, and G. Merriam. 1999. Fragmentation effects on forest birds: Relative influence of woodland cover and configuration on landscape occupancy. Conserv. Biol. 13: 774783.
  • Wiens, J. A. 1989. Spatial scaling in ecology. Funct. Ecol. 3: 385397.
  • Wunderle, J. M. Jr., and S. C. Latta. 1998. Avian resource use in Dominican shade coffee plantations. Wilson Bull. 110: 271281.
  • Wunderle, J. M. Jr., and S. C. Latta. 1996. Avian abundance in sun and shade coffee plantations and remnant pine forest in the Cordillera Central, Dominican Republic. Ornitol. Neotrop. 7: 1934.

Supporting Information

  1. Top of page
  2. ABSTRACTRESUMEN
  3. METHODS
  4. RESULTS
  5. DISCUSSION
  6. ACKNOWLEDGMENTS
  7. LITERATURE CITED
  8. Supporting Information

FIGURE S1. Study area in Cuetzalan region in the Northern Mountain Range of Puebla, México.

TABLE S1. Description of the nine sites, varying in landscape composition at the four radii (scales) of 1-km, 3- km, 5-km, 10-km.

TABLE S2. Plot, patch and landscape-level variables that showed significant relationship (F-ratio, P-value) with the bird community composition in at least one radii (scale).

TABLE S3. Bird species recorded foraging within the shaded coffee plantations, with guild, foraging strategy, sensitivity to disturbance, and IUCN status.

TABLE S4. Results of the 4th-corner analysis per dietary guild at each radius (scale).

TABLE S5. Species' codes and fit for RDA models, group I and II, and guild, foraging strategy, sensitivity to disturbance, and IUCN status.

Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.

FilenameFormatSizeDescription
BTP_553_sm_supplfig.pdf202KSupporting info item
BTP_553_sm_suppltable.doc689KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.