- Email: [email protected]
Disentangling elevational and vegetational effects on ant diversity patterns
Acta Oecologica 102 (2020) 103489
Contents lists available at ScienceDirect
Acta Oecologica journal homepage: www.elsevier.com/locate/actoec
Original article
Disentangling elevational and vegetational effects on ant diversity patterns a,∗
b
c
a
T
d
Chaim J. Lasmar , Carla R. Ribas , Julio Louzada , Antônio C.M. Queiroz , Rodrigo M. Feitosa , Mayara M.G. Imatab, Guilherme P. Alvesa, Gabriela B. Nascimentob, Frederico S. Nevese, Daniel Q. Domingosf a Programa de Pós–Graduação em Ecologia Aplicada, Setor de Ecologia e Conservação, Laboratório de Ecologia de Formigas, Departamento de Biologia, Universidade Federal de Lavras, PO Box 3037, Lavras, MG, 37200-000, Brazil b Laboratório de Ecologia de Formigas, Setor de Ecologia e Conservação, Departamento de Biologia, Universidade Federal de Lavras, PO Box 3037, Lavras, MG, 37200000, Brazil c Laboratório de Ecologia de Invertebrados, Setor de Ecologia e Conservação, Departamento de Biologia, Universidade Federal de Lavras, Lavras, MG, 37200-000, Brazil d Departamento de Zoologia, Universidade Federal do Paraná, CP 19020, Curitiba, PR, 81531-980, Brazil e Departamento de Genética, Ecologia e Evolução, Universidade Federal de Minas Gerais, Belo Horizonte, Minas Gerais, Brazil f Laboratório de Sistemática de Espermatófitas, Setor de Botânica, Departamento de Biologia, Universidade Federal de Lavras, MG, 31270-901, Brazil
A R T I C LE I N FO
A B S T R A C T
Keywords: Altitudinal gradient Effect size Elevational gradient Sampling bias Tropical mountain
When aiming to assess the effects of elevation on animal diversity, many studies have been carried out in different vegetation types occurring across elevational gradients. Thus, it remains unclear if any changes observed in species richness are caused by factors directly associated with elevation or are caused by vegetation change across the gradient. Here, we disentangled the effects of elevation from changes in vegetation by assessing ant diversity patterns along an elevational gradient. We analyzed patterns of ant diversity utilizing two different sampling approaches across the elevational gradient: (1) a standardized sampling including only forest formations and (2) a non–standardized sampling including forest (low elevational bands) and grasslands (high elevational bands). We sampled ants at eight elevational bands of Atlantic Forest in Brazil, and the highest three bands were sampled at both forest and grassland habitat. We found that the two approaches produce contrasting patterns of alpha and beta diversity, but the same pattern of gamma diversity. However, in the non–standardized sampling approach, the regression analysis produced a reduced explanation of the species richness gradient and a decrease in the elevational effect size. Different patterns found in the two approaches could be due to distinct environmental conditions in these habitats. In conclusion, our results highlight the potential bias of non–standardizing vegetation type across elevational gradients when assessing elevational patterns of species diversity.
1. Introduction
relatively small spatial scale (Körner, 2007; Sundqvist et al., 2013). They also present well established ecological patterns based on both historical and current factors, that shape species distribution (Colwell and Rangel, 2010; Rahbek, 2005) and mirror ecological patterns of latitudinal gradients (Rahbek, 2005). There are many factors intrinsically linked to elevational gradients such as temperature, area availability, atmospheric pressure, and UV–B radiation (Körner, 2007). The change in these factors across environmental gradients can affect the physiology of plants, metabolic processes, body size, and distribution, due to different selection pressures (Körner, 1998, 2007). Such adaptations to these pressures can be notably dramatic, especially when we consider the drastic change in vegetation type at the treeline of mountains. The treeline was
Two centuries have passed since von Humboldt's seminal study on diversity patterns across elevational gradients (von Humboldt, 1849); yet the subject is still being explored by ecologists and biogeographers around the world. Aiming to improve ecological theories, which underpin biodiversity conservation strategy, elevational gradients have been used to assess species richness patterns (Peters et al., 2016; Rahbek, 2005), species distributions ranges (Stevens, 1992; Wen et al., 2018) interspecific interactions (Roslin et al., 2017), community phylogenetic structure (Smith et al., 2014) and climate change impacts on biodiversity (Colwell et al., 2008). Elevational gradients are interesting because they present a high variability of environmental factors on a
∗
Corresponding author. E-mail address: [email protected] (C.J. Lasmar).
https://doi.org/10.1016/j.actao.2019.103489 Received 17 May 2019; Received in revised form 15 October 2019; Accepted 2 November 2019 1146-609X/ © 2019 Elsevier Masson SAS. All rights reserved.
Acta Oecologica 102 (2020) 103489
C.J. Lasmar, et al.
characterized by Harsch and Bader (2011) as “an ecotone delimited at the upper end by the tree species limit, the uppermost elevation or latitude at which tree species occur as trees, or krummholz, regardless of height, and at the lower end by continuous forest > 3 m tall”; for example, the shift from forest habitat to grassland habitats formations in tropical mountains. High levels of turnover of plant species occur throughout the elevation gradient (Sundqvist et al., 2013; Kraft et al., 2011) that could also form zonations of multilayered forests in tropical mountains (Hemp, 2006). However, such changes in vegetation type, in terms of forests and grassland or bushland formations, are not necessarily directly linked to the elevational gradient, and therefore should not be treated as an intrinsic elevational variable (unlike temperature, atmospheric pressure, etc.). Indeed, studies have shown that the occurrence of the treeline, and the vegetation types found above it, are not driven by climatic factors that are usually linked to elevation (Pompeu et al., 2018; Safford, 2001). Even though vegetation type has been demonstrated to affect diversity on elevational gradients (e.g. Axmancher and Fielder, 2008; Axmancher et al., 2009; Carneiro et al., 2014), many studies designed to assess elevation effects have still been carried out in different vegetation types located at distinct altitudes (e.g. Classen et al., 2015; Nakamura et al., 2015; Peters, 2016; Tello et al., 2014; Wen et al., 2018). Körner (2007) highlighted the importance of separating factors that are directly linked to elevational gradients from those that vary regionally but are not specific to elevation (i.e. precipitation, seasonality and wind velocity), in order to reduce confounding bias and obtain results that reflect only the elevation gradient. In this sense, it is possible that studies on the effects of elevational gradients conducted on different vegetation types are reflecting not only elevation patterns but also the inherent differences in vegetation. Therefore, without standardizing a survey to the same vegetation type, shifting sampling points among different vegetation types along an elevational gradient may influence the resulting patterns found in elevational gradient studies. Here, we disentangle the effects of elevation from differences in vegetation type on ant diversity patterns along an elevational gradient. We used ants as a study model because they are well distributed in tropical areas (Moreau and Bell, 2013), sensitive to changes in the environment (Philpott et al., 2010), on vegetation types (Andersen et al., 2007) and habitat features (Mezger Pfeiffer, 2011), and have been used successfully for assessing elevational gradients patterns (e.g. Bishop et al., 2017; Szewczyk and McCain, 2016). We tested whether sampling in different vegetation types may produce different patterns on ant species richness (α and ɣ diversities) and beta diversity (β–diversity). We hypothesized that vegetation type can influence the assessment of elevational gradients on biodiversity by biasing some of the observed patterns, due to confounding factors not necessarily directly linked to elevation. Specifically, through this study, we show that the influence of distinct vegetation types can inhibit clear interpretations of elevational patterns on ant species diversity.
hereafter, when referring to vegetation types in this study, we refer to the habitat formation defined simply as closed forest or open grassland. In this sense, until the treeline, there is a continuous dense Atlantic Forest formation that here we call “low elevation forests”. Above the treeline, the vegetation type changes to open grasslands, which we term “high elevation grasslands”. However, there are also some natural forest remnants that exist above the treeline that we characterize as “high elevation forests”. The low elevation forests present diverse plant families such as Arecaceae, Rubiaceae, Lauraceae, Burseraceae, Rutaceae, Fabaceae, Erythroxylaceae, Myrtaceae, Salicaceae, Euphorbiaceae, Araucariaceae, Bignoniaceae, and Piperaceae. The high elevation forests are composed of plants families such as Myrtaceae, Melastomataceae, Asteraceae, Fabaceae, Proteaceae, Aquifoliaceae, and Solanaceae. The high elevation grasslands present exposed rocks and are dominated by Poaceae species (grasses), also showing some elements of herbaceous plant families as Asteraceae, Apiaceae, and Alstroemeriaceae. It also presents bogs with plant families such as Cyperaceae, Lentibulariaceae, Xyridaceae, and bryophytes. 2.2. Ant sampling We sampled ants during the rainy season in 2015. We selected eight elevational bands that were situated at 600, 848, 1134, 1515, 1810, 2000, 2200 and 2457 m.a.s.l. As well as the elevation difference, each elevational band was located at least 1.4 km apart from each other. At each elevation, we installed a 200 m transect containing 10 sampling points spaced 20 m apart. In each sampling point, we placed four epigaeic pitfall traps in a square grid of 1.5 m × 1.5 m, resulting in a total of 40 traps at each elevation. The pitfalls were 11 cm in diameter and 11 cm in depth, filled with a 200 ml solution of water, salt (0.4%) and liquid soap (0.6%) (Bestelmeyer et al., 2000; Canedo–Júnior et al., 2016). Each trap remained operating for 48 h. We set one transect for each of the low elevation forest points, located from the first (660 m) to the fifth elevation band (1810 m) (i.e. below the treeline transects). At the treeline and onwards, which extended from the sixth elevation band (2000 m.a.s.l.) to the highest (2457 m.a.s.l.), we set two transects per elevation (one in each of the two habitat types), in order to sample both high elevation forests and high elevation grasslands. The transects in each vegetation type above the treeline (forest and grassland) were spatially separated by at least 160 m while maintaining both transects in the same elevation band. Ants were identified to genera following Baccaro et al. (2015), and whenever possible, to species level, through relevant literature and by matching collected specimens with identified material in the Laboratório de Ecologia de Formigas’ ant collection at Universidade Federal de Lavras (UFLA). Voucher specimens of all collected species/morphospecies were deposited at UFLA and at the Entomological Collection Padre Jesus Santiago Moure of the Universidade Federal do Paraná (DZUP), in Curitiba, Brazil. All applicable institutional and national guidelines for the care and use of animals were followed. We had a license for all samplings, which was conceded by Sistema de Autorização e Informação em Biodiversidade (SISBIO); license number 46564–1, date 17/10/2014, authentication code: 73877888; http:// www.icmbio.gov.br/sisbio/verificar_autenticidade.
2. Material and methods 2.1. Study area We carried out the study in Itatiaia National Park (INP), in the southeast of Brazil (22°16’ – 22°28′ S and 44°34’ – 44°42’ W). The oldest national park in Brazil, the INP was established in 1937 to protect important remnants of the Atlantic Forest biome. The park is located in the Itatiaia Massif, on the highest portion of the Mantiqueira mountain range. The protected area falls within 600 m.a.s.l to the peak at 2878 m.a.s.l. The treeline is situated at around 2000 m.a.s.l. There are many vegetation types found on the Mantiqueira Mountains, ranging from lower montane forest (at 0–500 m.a.s.l.), montane forest (500–1500 m.a.s.l.), and upper montane forest (1500–2000 m.a.s.l.), to altitudinal grasslands (2000–2500 m.a.s.l.) (Safford, 1999). As we were interested in drastic changes in vegetation,
2.3. Data analysis To test whether sampling in different vegetation types in an elevational gradient may bias patterns on ant species diversity, we performed generalized linear models (GLMs) with and without mixed effects for the diversity components alpha, beta and gamma (α, β and ɣ, respectively) extracted from two different approaches. In the first approach, we used only closed forest transects from all elevations (low and high elevation forests) that we called single–habitat gradient. In the second approach, we used the five closed forest transects below the treeline (low elevation forests), plus the three open grassland transects above 2
Acta Oecologica 102 (2020) 103489
C.J. Lasmar, et al.
Fig. 1. Vegetation types sampled at the same elevation above the treeline (2000 m.a.s.l.) side by side. At the right, we have high elevation grasslands and at the left, we have high elevation forests. From top to down, images of areas located at 2000, 2200 and 2457 m.a.s.l.
(using only single or mixed habitat gradient), produced the same pattern, we performed a resampling application from 1000 bootstrap samples, replacing it in the ‘boot.ci()’ function from ‘boot’ package (Canty and Ripley, 2012). With this function, we can assess the 95% confidence intervals of a regression's R2 and slope, and the median precision of the GLMs of the two approaches. Thus, we could determine if there was a difference in model precision and the effect size of the regressions. We performed all analysis above at R 3.01 software (R Core Team, 2013).
the treeline (high elevation grasslands). Termed mixed–habitat gradient, this approach allowed us to assess the elevational gradient across different vegetation types in terms of habitat formation (Fig. 1). As response variables for both approaches, we considered α–diversity as the number of ant species richness per pitfall traps (four pitfall traps in a 1.5 m2 grid), ɣ–diversity the overall ant species richness per elevational band (transect) and β–diversity as a proxy of species composition changes within each elevational band (transects). For β–diversity, we extracted the total beta diversity (βsor), using ‘beta.part’ package, which results of turnover and nestedness mechanisms (Baselga and Orme, 2012). For α–diversity, at both approaches, we performed a generalized linear model with mixed effects (GLMM) with transect as the random and the elevation as the explanatory variable. Elevation values represent the elevation band, since there were no significant differences among sampling points in the same transect. We used the GLMM to reduce the pseudoreplication effects caused by the non–independence of pitfall traps placed in the same transects (Pinheiro and Bates, 2000). First, to investigate if ɣ diversities from two approaches are relate to elevation by a monotonic decline or hump–shaped function, we modified the Nagai (2011) function to generate linear and quadratic models for comparison from those diversity components. We verified which model would be more appropriated using the Aikaike's criteria (AICc), considering as the best model the one with lowest AIC value. After that, for β and ɣ diversities, we performed a generalized linear model (GLM) with the elevation also as an explanatory variable for both approaches. All GLMs and GLMM were performed using the ‘lme4’ package (Bates et al., 2013) and were assessed under residual analyses to obtain the adequacy of error distribution (Crawley, 2002). If both approaches
3. Results In total, we collected 148 ant morpho–species belonging to 37 genera and nine subfamilies, with 119 species recorded in low elevation forests, 21 at high elevation forests and 46 at high elevation grasslands (see Appendix A Table A). In addition, low elevation forests shared more species with high elevation grasslands than with high elevation forests (Fig. 2). Patterns of α–diversity differed between the two sampling approaches (Fig. 3a and b). Using the single–habitat gradient, elevation was found to influence α–diversity, presenting an inverse relationship (Chi = 12.05; p < 0.01), however in a mixed–habitat gradient the effect of elevation on α–diversity was not found to be significant (Chi = 2.23; p = 0.13). Regarding β–diversity, the two approaches also produced different patterns, although with contrasting results (Fig. 3c e d); in the single–habitat gradient β–diversity was not influenced by elevation (F = 1.62; p = 0.25), while in the mixed–habitat gradient elevation negatively influenced β–diversity (F = 11.32; p = 0.01). Regarding ɣ–diversity both approaches followed better a linear than a 3
Acta Oecologica 102 (2020) 103489
C.J. Lasmar, et al.
high elevation grasslands than with high elevation forests (Fig. 2), even considering that habitats from low and high elevation bands differ considerably in species composition. Ants are sensitive to the level of habitat openness, which could result in different species inhabiting these areas and consequently in the different values of species richness (Andersen et al., 2007; Botes et al., 2006; van Ingen et al., 2008). If we compare high elevation forest and grassland sites, forests are cooler and more humid than the low stature grassland habitat (Aparecido et al., 2018; Körner, 2007; Safford, 1999). Consequently, the warmer conditions and low levels of moisture in high elevation grasslands may permit more species to occur than in high forest habitats. Indeed, in this sense, despite inherent and notable differences, high elevation grassland habitat is more similar (in terms of temperature and moisture) to forests from low elevation bands which might improve the supporting of ant colonies in comparison to high elevational forests. Differences in β–diversity patterns (Fig. 3c and d) could be caused by different degrees of habitat heterogeneity between sites in high elevation grasslands and forests. The decrease of β–diversity according to elevation when analyzing the mixed–habitat gradient is likely due to high elevation grasslands sites being more homogenous in comparison to both low and high elevation forests. Such homogenization in terms of habitat conditions (related to temperature and moisture) would potentially permit ants in grass lands to colonize almost all sampling points in the transect; while this would likely not occur in more heterogeneous forest sites, leading to different patterns. We hypothesize that β–diversity within a single–habitat gradient may be driven by its own habitat heterogeneity (Jankowski et al., 2009; Nunes et al., 2016), which may not necessarily vary across the elevational gradient, at least in the extent of elevational gradient assessed in this study. Regarding ɣ–diversity we observed the same monotonic decline in diversity in both approaches (Fig. 3e and f). Elevational gradients have been found to produce a strong pattern on species diversity for other insect assemblages, resulting from the influence of temperature (Malsch et al., 2008; Sanders et al., 2007). Thus, even changes in vegetation type would likely not change this pattern. According to our results, when analyzing ɣ–diversity on a mixed–habitat gradient there was a decrease in model robustness in terms of the proportion of explained variance and effect size of elevation. In fact, species diversity is not influenced by elevation per se but by factors that covary with it. Sampling in different vegetation types likely means that diversity is dependent on factors related to differences in the structure of habitats, and not only limited to elevation–related factors. As Gotelli and Ellison (2004) have already pointed out, when many confounding factors are included in a study's design, it is difficult to disentangle their effects. Thus, the higher proportion of explanation of the variation in diversity found in a single–habitat gradient is likely due to the similarity of other environmental factors, except those directly linked to variation in elevation. Similarly, the effect size of elevation on species richness in a mixed–habitat gradient is lower because, as well as elevation, there are other environmental dissimilarities (as previously mentioned) that may permit more species to survive at higher elevations in grasslands than in the high elevation forests.
Fig. 2. The number of exclusive and shared ant species recorded in the three habitat types across an elevational gradient in Itatiaia National Park. The numbers within the overlapping parts of the circles represent the shared species found in the corresponding habitats, while the numbers outside of the overlap show the number of species exclusive to that habitat.
quadratic function (Appendix B Fig. B). Furthermore, both approaches exhibited the same pattern (Fig. 3e and f), with both single (F = 35.16; p < 0.01) and mixed–habitat gradient (F = 17.27; p < 0.01). Despite obtaining the same ɣ–diversity pattern for both transect approaches, we can observe a greater variance explained on single–habitat gradient than in mixed–habitat gradient regression (Fig. 4a). The single–habitat gradient also demonstrated a higher effect size based on the regression slope values (Fig. 4b). 4. Discussion The standardization of vegetation type across an elevational gradient has been empirically tested in this study. Our results unequivocally show that non–standardization of vegetation type can be a source of bias in the parameters evaluating community responses, which is in accord with Axmancher and Fielder (2008), who verified the vegetation type influence on geometrid moths’ diversity across an elevational gradient. The patterns observed for α–diversity and β–diversity clearly differed between the two approaches. Moreover, despite a monotonic decline observed for ɣ–diversity in both approaches, our analyses demonstrate that sampling in the same vegetation type leads to stronger results in order to explain the elevational gradient effects on species richness. We suggest that the environmental conditions imposed by the two vegetation types (grasslands and forests) are the main drivers of the contrasting elevation gradient patterns observed for α–diversity. Higher elevations present lower temperatures and high levels of moisture, which can affect ant species richness, likely producing the decrease of alpha diversity on a single–habitat gradient. Low temperatures and high levels of moisture affect the survival of ant colonies (for example decreasing foraging activity), and some species are unable to survive or colonize areas with such conditions (Bishop et al., 2017; Bruhl et al., 1999; Fisher, 1998; Malsch et al., 2008; Sanders et al., 2007). When analyzing the mixed–habitat gradient, we did not detect the same decrease of alpha diversity, as high elevation grasslands reached almost the same α–diversity of the most speciose low elevation forests (Fig. 3a and b). Furthermore, low elevation forests share more ant species with
5. Conclusions This study demonstrates that sampling in different vegetation types across an elevational gradient produces distinct and likely biased diversity patterns, compared to sampling in standardized habitats when aiming to assess exclusively elevational effects. As highlighted by Körner (2007), there is a risk of confounding geophysical factors not directly related to elevation, which may lead to assessments of other environmental gradients. With this in mind, we suggest that studies solely addressing elevational gradient effects must perform their sampling in the same vegetation type, in terms of closed forest or grasslands formation. However, when assessing only one vegetation type is not an option, we suggest authors test overall diversity (ɣ–diversity), instead 4
Acta Oecologica 102 (2020) 103489
C.J. Lasmar, et al.
Fig. 3. Relationship between ant diversity measures and elevation gradient with samples in only one vegetation type (single–habitat gradient) and with samples in two vegetation types (mixed–habitat gradient). Both single and mixed–habitat gradients present the same five first elevations (low elevation forest). The other three last elevations present only high elevation forests transects for single–habitat gradient and high elevation grasslands transects for mixed–habitat gradient. a) Ant species richness per pitfall traps (α–diversity) per sampling point of single–habitat gradient (Chi = 12.05; p < 0.01). b) Ant species richness per pitfall traps (α–diversity) of mixed–habitat gradient (Chi = 2.23; p = 0.13). c) Beta diversity (β–diversity) of only single–habitat gradient (F = 1.62; p = 0.25). d) Beta diversity (β–diversity) of mixed–habitat gradient (F = 11.32; p = 0.01) e) Overall ant species richness (ɣ–diversity) of single–habitat gradient (F = 35.16; p = 0.001). f) Overall ant species richness (ɣ–diversity) of mixed–habitat gradient (F = 17.27; p = 0.006).
of α–diversity and β–diversity, as the effects of elevation may be stronger at that scale. Furthermore, it seems that α–diversity and β–diversity responses are more sensitive to habitat changes. We also recommend that at the very least, vegetation types should be included in statistical models. The data presented here also indicate that many of published studies on elevational gradients performed on different vegetation types that aimed to assess only elevational effects, possibly confounded both elevational and vegetational effects leading to
potentially biased results. In order to reinforce this argument, future studies should address the influence of other vegetation types that present intermediate differences (e.g. shrublands, open woodland) on elevational gradient diversity patterns. Finally, we conclude that removing the non–standardized sampling bias is essential to fully understand elevational gradient patterns on biodiversity.
5
Acta Oecologica 102 (2020) 103489
C.J. Lasmar, et al.
Fig. 4. Comparison of GLM regression attributes between single and mixed–habitat elevation gradients on overall ant species richness (ɣ diversity). a) Proportion of variance explained (R2) and b) Effect size (slope of the regression line). Vertical dashed lines are bootstrapped confidence intervals based on 1000 bootstrap samples with replacement. Notch areas on boxplots mark the 95% confidence intervals of the median value (shown as a black horizontal line). The horizontal dotted line represent 0 value in y-axis. Black dots represent data outliers.
Author contributions
et al., 2018. Ecohydrological drivers of Neotropical vegetation in montane ecosystems. Ecohydrology 11 (3), e1932. https://doi.org/10.1002/eco.1932. Axmacher, J.C., Fiedler, K., 2008. Habitat type modifies geometryof elevational diversity gradients in geometrid moths (Lepidoptera Geometridae) on Mt Kilimanjaro, Tanzania. Trop. Zool. 21, 243–251. Axmacher, J.C., Brehm, G., Hemp, A., Tünte, H., Lyaruu, H.H.V., Müller–Hohenstein, K., Fiedler, K., 2009. Determinants of diversity in afrotropical herbivorous insects (Lepidoptera: geometridae): plant diversity, vegetation structure or abiotic factors? J. Biogeogr. 36, 337–349. https://doi.org/10.1111/j.1365-2699.2008.01997.x. Baccaro, F.B., Feitosa, R.M., Fernandez, F., Fernandes, I.O., Izzo, T.J., Souza, J.L.P., et al., 2015. Guia para gêneros de formigas do Brasil. Editora INPA, Manaus. https://doi. org/10.5281/zenodo.32912. Baselga, A., Orme, C.D.L., 2012. betapart: an R package for the study of beta diversity. Methods Ecol E 3, 808–812. https://doi.org/10.1111/j.2041–210X.2012.00224.x. Bates, D., Maechler, M., Bolker, B., Walker, S., 2013. lme4: linear mixed effects models using Eigen and S4. Cran.R - project. https://cran.r-project.org/web/packages/ lme4/index.html, version 1, Accessed date: 1 March 2017. Bestelmeyer, B.T., Agosti, D., Leeanne, F., Alonso, T., Brandão, C.R.F., Brown, W.L., et al., 2000. Field techniques for the study of ground–living ants, an overview, description, and evaluation. In: Agosti, D., Majer, J.D., Tennant, A., Schultz, T. (Eds.), Ants, Standard Methods for Measuring and Monitoring Biodiversity. Smithsonian Institution Press, Washington, pp. 122–144. https://doi.org/10.5281/zenodo.11736. Bishop, T.R., Robertson, M.P., van Rensburg, B.J., Parr, C.L., 2017. Coping with the cold: minimum temperatures and thermal tolerances dominate the ecology of mountain ants. Ecol. Entomol. 42, 105–114. https://doi.org/10.1111/een.12364. Botes, A., McGeoch, M.A., Robertson, H.G., van Niekerk, A., Davids, H.P., Chown, S.L., 2006. Ants, altitude and change in the northern cape floristic region. J. Biogeogr. 33, 71–90. https://doi.org/10.1111/j.1365–2699.2005.01336.x. Brühl, A., Mohamed, M., Linsenmair, K.E., 1999. Altitudinal distribution of leaf litter ants along a transect in primary forests on Mount Kinabalu, Sabah, Malaysia. J. Trop. Ecol. 15, 265–277. https://doi.org/10.1017/S0266467499000802. Canedo–Júnior, E.O., Cuissi, R.C., Curi, N.H.A., Demetrio, G.R., Lasmar, C.J., Malves, K., Ribas, C.R., 2016. Can anthropic fires affect epigaeic and hypogaeic Cerrado ant (Hymenoptera, Formicidae) communities in the same way? Rev. Biol. Trop. 64, 95–104. https://doi.org/10.15517/rbt.v64i1.18239. Canty, A., Ripley, B., 2012. Boot, Bootstrap R (S–Plus) Functions. R Package version1.3–4. https://CRAN.R_project.org/package=boot, Accessed date: 1 February 2017. Carneiro, E., Mielkel, O.H.H., Casagrande, M.M., Fiedler, K., 2014. Community structure of skipper butterflies (Lepidoptera, hesperiidae) along elevational gradients in Brazilian atlantic forest reflects vegetation type rather than altitude. PLoS One 9 (10), e108207. https://doi.org/10.1371/journal.pone.0108207. Classen, A., Peters, M.K., Kindeketa, W.J., Appelhans, T., Eardley, C.D., Gikungu, M.W., et al., 2015. Temperature versus resource constraints, which factors determine bee diversity on Mount Kilimanjaro, Tanzania? Glob. Ecol. Biogeogr. 24, 642–652. https://doi.org/10.1111/geb.12286. Colwell, R.K., Brehm, G., Cardelús, C.L., Gilman, A.C., Longino, J.T., 2008. Global warming, elevational range shifts, and lowland biotic attrition in the wet tropics. Science 322, 258–261. https://doi.org/10.1126/science.1162547. Colwell, R.K., Rangel, T.F., 2010. A stochastic, evolutionary model for range shifts and richness on tropical elevational gradients under Quaternary glacial cycles. Philos. T. R. Soc. B. 365, 3695–3707. https://doi.org/10.1098/rstb.2010.0293. Crawley, M.J., 2002. Statistical Computing, an Introduction to Data Analysis Using S–Plus, first ed. John Wiley Sons, London.
CJL, CRR, JL and ACMQ originally formulated the idea. CJL, RMF, MMGI, GPA, GBN and DQD conducted fieldwork and collected the data. CJL, CRR, JL, ACMQ and FSN performed statistical analyses. CJL wrote the article with substantial collaboration from all co-authors. Declaration of competing interest None. Acknowledgements This work was funded by Fundação de Amparo à Pesquisa de Minas Gerais (FAPEMIG, grant PPM 00243/14). We are thankful to the staff of National Itatiaia Park, especially to Leo Nascimento, who permitted the sampling in the park area. We are also indebted to Maria Regina de Souza, Tobias R. Silva, Luiza Santiago, Edson Guilherme de Souza, Ernesto O. Canedo–Júnior, Graziele Santiago and Luana Zurlo for their help with logistics and fieldwork. We also thank Filipe França for his help and advice on statistical analyses. Thanks to Mariana Rabelo, Ícaro Carvalho and Felipe Lopes for helping with laboratory work. We are grateful to the staff at the Laboratório de Sistemática de Formigas da Universidade Federal do Paraná for confirming ant identification, especially to Alexandre Ferreira. This study was part of CJ Lasmar's MSc at UFLA that was supported by Coordenação de Aperfeiçoamento Pessoal (CAPES, Finance code: 001). RMF was supported by the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq grant 302462/2016–3). Appendix A. Supplementary data Supplementary data related to this article can be found at https:// doi.org/10.1016/j.actao.2019.103489. References Andersen, A.N., van Ingen, L.T., Campos, R.I., 2007. Contrasting rainforest and savanna ant faunas in monsoonal northern Australia, a rainforest patch in a tropical savanna landscape. Aust. J. Zool. 55, 363–369. https://doi.org/10.1071/ZO07066. Aparecido, L.M.T., Teodoro, G.S., Mosquera, G., Brum, M., Barros, F.V., Pompeu, P.V.,
6
Acta Oecologica 102 (2020) 103489
C.J. Lasmar, et al.
acprof:oso/9780199544639.001.0001. Pinheiro, J.C., Bates, D.M., 2000. In: Pinheiro, J.C. (Ed.), Mixed-effects models in S and SPLUS. Springer, New York. Pompeu, P.V., Fontes, M.A.L., Mulligan, M., Bueno, I.T., Siqueira, M.F., Acerbi–Júnior, F.W., et al., 2018. Assessing Atlantic cloud forest extent and protection status in southeastern Brazil. J. Nat. Conserv. 43, 146–155. https://doi.org/10.1016/j.jnc. 2018.04.003. R Core Team, 2013. R: A Language and Environment for StatisticalComputing. R Foundation for Statistical Computing, Vienna, Austria. Rahbek, C., 2005. The role of spatial scale and the perception of large–scale species–richness patterns. Ecol. Lett. 8, 224–239. https://doi.org/10.1111/j.1461–0248. 2004.00701.x. Roslin, T., Hardwick, B., Novotny, V., Petry, W.K., Andrew, N.R., Asmus, A., et al., 2017. Higher predation risk for insect prey at low latitudes and elevations. Science 356, 742–744. https://doi.org/10.1126/science.aaj1631. Safford, H.D., 1999. Brazilian Páramos I. An introduction to the physical environment and vegetation of the campos de altitude. J. Biogeogr. 26, 693–712. https://doi.org/10. 1046/j.1365–2699.1999.00313.x. Safford, H.F., 2001. Brazilian páramos III. Patterns and rates of post fire regeneration in the campos de Altitude. Biotropica 33, 282–302. https://doi.org/10.1111/j. 1744–7429.2001.tb00179.x. Sanders, N.J., Lessard, J.P., Fitzpatrick, M.C., Dunn, R.R., 2007. Temperature, but not productivity or geometry, predicts elevational diversity gradients in ants across spatial grains. Glob. Ecol. Biogeogr. 16, 640–649. https://doi.org/10.1111/j. 1466–8238.2007.00316.x. Smith, M.A., Hallwachs, W., Janzen, D.H., 2014. Diversity and phylogenetic community structure of ants along a Costa Rican elevational gradient. Ecography 37, 720–731. https://doi.org/10.1111/j.1600–0587.2013.00631.x. Stevens, G.C., 1992. The elevational gradient in altitudinal range, an extension of Rapoport's latitudinal rule to altitude. Am. Nat. 140, 893–911. https://doi.org/10. 1086/285447. Sundqvist, M.K., Sanders, N.J., Wardle, D.A., 2013. Community and ecosystem responses to elevational gradients, processes, mechanisms, and insights for global change. Annu. Rev. Ecol. Evol. Syst. 44, 261–280. https://doi.org/10.1146/ annurev–ecolsys–110512–135750. Szewczyk, T., McCain, C.M., 2016. A systematic review of global drivers of ant elevational diversity. PLoS One 11 (5), e0155404. https://doi.org/10.1371/journal.pone. 0155404. Tello, J.S., Myers, J.A., Macía, M.J., Fuentes, A.F., Cayola, L., Arellano, G., Loza, M.I., Torrez, V., Cornejo, M., Miranda, T.B., Jørgensen, P.M., 2014. Elevational gradients in β–diversity reflect variation in the strength of local community assembly mechanisms across spatial scales. PLoS One 10 (3), e0121458. https://doi.org/10.1371/ journal.pone.0121458. van Ingen, L.T., Campos, R.I., Andersen, A.N., 2008. At community structure along an extended rain forest–savanna gradient in tropical Australia. J. Trop. Ecol. 24, 445–455. https://doi.org/10.1017/S0266467408005166. von Humboldt, A., 1849. Aspects of nature. In: Different Lands and Different Climates, with Scientific Elucidations, second ed. Longman, London. Wen, Z., Wu, Y., Cheng, J., Cai, T., Du, Y., Ge, D., et al., 2018. Abundance of small mammals correlates with their elevational range sizes and elevational distributions in the subtropics. Ecography 41, 1–11. https://doi.org/10.1111/ecog.03558.
Fisher, B.L., 1998. Ant diversity patterns along an elevational gradient in the réserve spéciale d'Anjanaharibe–Sud and on the western Masoala Peninsula, Madagascar. Fieldiana, Life earth sci. 90, 39–67. Gotelli, N.J., Ellison, A.M., 2004. Designing successful field studies. In: Gotelli, N.J., Ellison, A.M. (Eds.), A Primer of Ecological Statistics. Sinauer Associates, Sunderland, Massachusetts, U.S.A, pp. 155–179. Harsch, M.A., Bader, M.Y., 2011. Treeline form – a potential key to understanding treeline dynamics. Glob. Ecol. Biogeogr. 20, 582–596. https://doi.org/10.1111/j.1466–8238. 2010.00622.x. Hemp, A., 2006. Continuum or zonation? Altitudinal gradients in the forest vegetation of Mt. Kilimanjaro. Plant. Ecol. 184, 27–42. https://doi.org/10.1007/ s11258–005–9049–4. Jankowski, J.E., Ciecka, A.L., Meyer, N.Y., Rabenold, K.N., 2009. Beta diversity along environmental gradients, implications of habitat specialization in tropical montane landscapes. J. Anim. Ecol. 78, 315–327. https://doi.org/10.1111/j.1365–2656.2008. 01487.x. Körner, C., 1998. A re–assessment of high elevation treeline positions and their explanation. Oecologia 115, 445–459. https://doi.org/10.1007/s004420050540. Körner, C., 2007. The use of ‘altitude’ in ecological research. Trends Ecol. Evol. 22, 569–574. https://doi.org/10.1016/j.tree.2007.09.006. Kraft, N.J.B., Comita, L.S., Chase, J.M., Sanders, N.J., Swenson, N.G., Crist, T.O., et al., 2011. Disentangling the drivers of β diversity along latitudinal and elevational gradients. Science 333, 1755–1758. https://doi.org/10.1126/science.1208584. Malsch, A.K.F., Fiala, B., Maschwitz, U., Mohamed, M., Nais, J., Linsenmair, K.E., 2008. An analysis of declining ant species richness with increasing elevation at Mount Kinabalu, Sabah, Borneo. Asian. Myrmecol. 2, 33–49. https://doi.org/10.20362/am. 002004. Mezger, D., Pfeiffer, M., 2011. Partitioning the impact of abiotic factors and spatial patterns on species richness and community structure of ground ant assemblages in four Bornean Rainforests. Ecography 34, 39–48. https://doi.org/10.1111/j. 1600–0587.2010.06538.x. Moreau, C.S., Bell, C.D., 2013. Testing the museum versus cradle tropical biological diversity hypothesis, phylogeny, diversification, and ancestral biogeographic range evolution of the ants. Evolution 67, 2240–2257. https://doi.org/10.1111/evo.12105. Nagai, M.E., 2011. Eco R: using R [WWW document]. Exerc. Do micael. http://ecologia. ib.usp.br/bie5782/doku.php?id=bie5782:05_curso_antigo:r2011:alunos:trabalho_ final:micael:start, Accessed date: 25 January 2018. Nakamura, A., Burwell, C.J., Ashton, L.A., Laidlaw, M.J., Katabuchi, M., Kitching, R.L., 2015. Identifying indicator species of elevation: comparing the utility of woody plants, ants and moths for long–term monitoring. Austral Ecol. 41, 179–188. https:// doi.org/10.1111/aec.12291. Nunes, C.A., Braga, F.R., Figueira, J.E.C., Neves, F.S., Fernandes, G.W., 2016. Dung beetles along a tropical altitudinal gradient, environmental filtering on taxonomic and functional diversity. PLoS One 11 (6), e0157442. https://doi.org/10.1371/ journal.pone.0157442. Peters, M.K., Hemp, A., Appelhans, T., Behler, C., Classen, A., Detsch, F., et al., 2016. Predictors of elevational biodiversity gradients change from single taxa to the multi–taxa community level. Nat. Commun. 7, 13736. https://doi.org/10.1038/ ncomms13736. Philpott, S.M., Perfecto, I., Armbrecht, I., Parr, C.L., 2010. Ant diversity and function in disturbed and changing habitats. In: Lach, L., Parr, C.L., Abbott, K.L. (Eds.), Ant Ecology. Oxford University Press, Oxford, pp. 137–156. https://doi.org/10.1093/
7
Contents lists available at ScienceDirect
Acta Oecologica journal homepage: www.elsevier.com/locate/actoec
Original article
Disentangling elevational and vegetational effects on ant diversity patterns a,∗
b
c
a
T
d
Chaim J. Lasmar , Carla R. Ribas , Julio Louzada , Antônio C.M. Queiroz , Rodrigo M. Feitosa , Mayara M.G. Imatab, Guilherme P. Alvesa, Gabriela B. Nascimentob, Frederico S. Nevese, Daniel Q. Domingosf a Programa de Pós–Graduação em Ecologia Aplicada, Setor de Ecologia e Conservação, Laboratório de Ecologia de Formigas, Departamento de Biologia, Universidade Federal de Lavras, PO Box 3037, Lavras, MG, 37200-000, Brazil b Laboratório de Ecologia de Formigas, Setor de Ecologia e Conservação, Departamento de Biologia, Universidade Federal de Lavras, PO Box 3037, Lavras, MG, 37200000, Brazil c Laboratório de Ecologia de Invertebrados, Setor de Ecologia e Conservação, Departamento de Biologia, Universidade Federal de Lavras, Lavras, MG, 37200-000, Brazil d Departamento de Zoologia, Universidade Federal do Paraná, CP 19020, Curitiba, PR, 81531-980, Brazil e Departamento de Genética, Ecologia e Evolução, Universidade Federal de Minas Gerais, Belo Horizonte, Minas Gerais, Brazil f Laboratório de Sistemática de Espermatófitas, Setor de Botânica, Departamento de Biologia, Universidade Federal de Lavras, MG, 31270-901, Brazil
A R T I C LE I N FO
A B S T R A C T
Keywords: Altitudinal gradient Effect size Elevational gradient Sampling bias Tropical mountain
When aiming to assess the effects of elevation on animal diversity, many studies have been carried out in different vegetation types occurring across elevational gradients. Thus, it remains unclear if any changes observed in species richness are caused by factors directly associated with elevation or are caused by vegetation change across the gradient. Here, we disentangled the effects of elevation from changes in vegetation by assessing ant diversity patterns along an elevational gradient. We analyzed patterns of ant diversity utilizing two different sampling approaches across the elevational gradient: (1) a standardized sampling including only forest formations and (2) a non–standardized sampling including forest (low elevational bands) and grasslands (high elevational bands). We sampled ants at eight elevational bands of Atlantic Forest in Brazil, and the highest three bands were sampled at both forest and grassland habitat. We found that the two approaches produce contrasting patterns of alpha and beta diversity, but the same pattern of gamma diversity. However, in the non–standardized sampling approach, the regression analysis produced a reduced explanation of the species richness gradient and a decrease in the elevational effect size. Different patterns found in the two approaches could be due to distinct environmental conditions in these habitats. In conclusion, our results highlight the potential bias of non–standardizing vegetation type across elevational gradients when assessing elevational patterns of species diversity.
1. Introduction
relatively small spatial scale (Körner, 2007; Sundqvist et al., 2013). They also present well established ecological patterns based on both historical and current factors, that shape species distribution (Colwell and Rangel, 2010; Rahbek, 2005) and mirror ecological patterns of latitudinal gradients (Rahbek, 2005). There are many factors intrinsically linked to elevational gradients such as temperature, area availability, atmospheric pressure, and UV–B radiation (Körner, 2007). The change in these factors across environmental gradients can affect the physiology of plants, metabolic processes, body size, and distribution, due to different selection pressures (Körner, 1998, 2007). Such adaptations to these pressures can be notably dramatic, especially when we consider the drastic change in vegetation type at the treeline of mountains. The treeline was
Two centuries have passed since von Humboldt's seminal study on diversity patterns across elevational gradients (von Humboldt, 1849); yet the subject is still being explored by ecologists and biogeographers around the world. Aiming to improve ecological theories, which underpin biodiversity conservation strategy, elevational gradients have been used to assess species richness patterns (Peters et al., 2016; Rahbek, 2005), species distributions ranges (Stevens, 1992; Wen et al., 2018) interspecific interactions (Roslin et al., 2017), community phylogenetic structure (Smith et al., 2014) and climate change impacts on biodiversity (Colwell et al., 2008). Elevational gradients are interesting because they present a high variability of environmental factors on a
∗
Corresponding author. E-mail address: [email protected] (C.J. Lasmar).
https://doi.org/10.1016/j.actao.2019.103489 Received 17 May 2019; Received in revised form 15 October 2019; Accepted 2 November 2019 1146-609X/ © 2019 Elsevier Masson SAS. All rights reserved.
Acta Oecologica 102 (2020) 103489
C.J. Lasmar, et al.
characterized by Harsch and Bader (2011) as “an ecotone delimited at the upper end by the tree species limit, the uppermost elevation or latitude at which tree species occur as trees, or krummholz, regardless of height, and at the lower end by continuous forest > 3 m tall”; for example, the shift from forest habitat to grassland habitats formations in tropical mountains. High levels of turnover of plant species occur throughout the elevation gradient (Sundqvist et al., 2013; Kraft et al., 2011) that could also form zonations of multilayered forests in tropical mountains (Hemp, 2006). However, such changes in vegetation type, in terms of forests and grassland or bushland formations, are not necessarily directly linked to the elevational gradient, and therefore should not be treated as an intrinsic elevational variable (unlike temperature, atmospheric pressure, etc.). Indeed, studies have shown that the occurrence of the treeline, and the vegetation types found above it, are not driven by climatic factors that are usually linked to elevation (Pompeu et al., 2018; Safford, 2001). Even though vegetation type has been demonstrated to affect diversity on elevational gradients (e.g. Axmancher and Fielder, 2008; Axmancher et al., 2009; Carneiro et al., 2014), many studies designed to assess elevation effects have still been carried out in different vegetation types located at distinct altitudes (e.g. Classen et al., 2015; Nakamura et al., 2015; Peters, 2016; Tello et al., 2014; Wen et al., 2018). Körner (2007) highlighted the importance of separating factors that are directly linked to elevational gradients from those that vary regionally but are not specific to elevation (i.e. precipitation, seasonality and wind velocity), in order to reduce confounding bias and obtain results that reflect only the elevation gradient. In this sense, it is possible that studies on the effects of elevational gradients conducted on different vegetation types are reflecting not only elevation patterns but also the inherent differences in vegetation. Therefore, without standardizing a survey to the same vegetation type, shifting sampling points among different vegetation types along an elevational gradient may influence the resulting patterns found in elevational gradient studies. Here, we disentangle the effects of elevation from differences in vegetation type on ant diversity patterns along an elevational gradient. We used ants as a study model because they are well distributed in tropical areas (Moreau and Bell, 2013), sensitive to changes in the environment (Philpott et al., 2010), on vegetation types (Andersen et al., 2007) and habitat features (Mezger Pfeiffer, 2011), and have been used successfully for assessing elevational gradients patterns (e.g. Bishop et al., 2017; Szewczyk and McCain, 2016). We tested whether sampling in different vegetation types may produce different patterns on ant species richness (α and ɣ diversities) and beta diversity (β–diversity). We hypothesized that vegetation type can influence the assessment of elevational gradients on biodiversity by biasing some of the observed patterns, due to confounding factors not necessarily directly linked to elevation. Specifically, through this study, we show that the influence of distinct vegetation types can inhibit clear interpretations of elevational patterns on ant species diversity.
hereafter, when referring to vegetation types in this study, we refer to the habitat formation defined simply as closed forest or open grassland. In this sense, until the treeline, there is a continuous dense Atlantic Forest formation that here we call “low elevation forests”. Above the treeline, the vegetation type changes to open grasslands, which we term “high elevation grasslands”. However, there are also some natural forest remnants that exist above the treeline that we characterize as “high elevation forests”. The low elevation forests present diverse plant families such as Arecaceae, Rubiaceae, Lauraceae, Burseraceae, Rutaceae, Fabaceae, Erythroxylaceae, Myrtaceae, Salicaceae, Euphorbiaceae, Araucariaceae, Bignoniaceae, and Piperaceae. The high elevation forests are composed of plants families such as Myrtaceae, Melastomataceae, Asteraceae, Fabaceae, Proteaceae, Aquifoliaceae, and Solanaceae. The high elevation grasslands present exposed rocks and are dominated by Poaceae species (grasses), also showing some elements of herbaceous plant families as Asteraceae, Apiaceae, and Alstroemeriaceae. It also presents bogs with plant families such as Cyperaceae, Lentibulariaceae, Xyridaceae, and bryophytes. 2.2. Ant sampling We sampled ants during the rainy season in 2015. We selected eight elevational bands that were situated at 600, 848, 1134, 1515, 1810, 2000, 2200 and 2457 m.a.s.l. As well as the elevation difference, each elevational band was located at least 1.4 km apart from each other. At each elevation, we installed a 200 m transect containing 10 sampling points spaced 20 m apart. In each sampling point, we placed four epigaeic pitfall traps in a square grid of 1.5 m × 1.5 m, resulting in a total of 40 traps at each elevation. The pitfalls were 11 cm in diameter and 11 cm in depth, filled with a 200 ml solution of water, salt (0.4%) and liquid soap (0.6%) (Bestelmeyer et al., 2000; Canedo–Júnior et al., 2016). Each trap remained operating for 48 h. We set one transect for each of the low elevation forest points, located from the first (660 m) to the fifth elevation band (1810 m) (i.e. below the treeline transects). At the treeline and onwards, which extended from the sixth elevation band (2000 m.a.s.l.) to the highest (2457 m.a.s.l.), we set two transects per elevation (one in each of the two habitat types), in order to sample both high elevation forests and high elevation grasslands. The transects in each vegetation type above the treeline (forest and grassland) were spatially separated by at least 160 m while maintaining both transects in the same elevation band. Ants were identified to genera following Baccaro et al. (2015), and whenever possible, to species level, through relevant literature and by matching collected specimens with identified material in the Laboratório de Ecologia de Formigas’ ant collection at Universidade Federal de Lavras (UFLA). Voucher specimens of all collected species/morphospecies were deposited at UFLA and at the Entomological Collection Padre Jesus Santiago Moure of the Universidade Federal do Paraná (DZUP), in Curitiba, Brazil. All applicable institutional and national guidelines for the care and use of animals were followed. We had a license for all samplings, which was conceded by Sistema de Autorização e Informação em Biodiversidade (SISBIO); license number 46564–1, date 17/10/2014, authentication code: 73877888; http:// www.icmbio.gov.br/sisbio/verificar_autenticidade.
2. Material and methods 2.1. Study area We carried out the study in Itatiaia National Park (INP), in the southeast of Brazil (22°16’ – 22°28′ S and 44°34’ – 44°42’ W). The oldest national park in Brazil, the INP was established in 1937 to protect important remnants of the Atlantic Forest biome. The park is located in the Itatiaia Massif, on the highest portion of the Mantiqueira mountain range. The protected area falls within 600 m.a.s.l to the peak at 2878 m.a.s.l. The treeline is situated at around 2000 m.a.s.l. There are many vegetation types found on the Mantiqueira Mountains, ranging from lower montane forest (at 0–500 m.a.s.l.), montane forest (500–1500 m.a.s.l.), and upper montane forest (1500–2000 m.a.s.l.), to altitudinal grasslands (2000–2500 m.a.s.l.) (Safford, 1999). As we were interested in drastic changes in vegetation,
2.3. Data analysis To test whether sampling in different vegetation types in an elevational gradient may bias patterns on ant species diversity, we performed generalized linear models (GLMs) with and without mixed effects for the diversity components alpha, beta and gamma (α, β and ɣ, respectively) extracted from two different approaches. In the first approach, we used only closed forest transects from all elevations (low and high elevation forests) that we called single–habitat gradient. In the second approach, we used the five closed forest transects below the treeline (low elevation forests), plus the three open grassland transects above 2
Acta Oecologica 102 (2020) 103489
C.J. Lasmar, et al.
Fig. 1. Vegetation types sampled at the same elevation above the treeline (2000 m.a.s.l.) side by side. At the right, we have high elevation grasslands and at the left, we have high elevation forests. From top to down, images of areas located at 2000, 2200 and 2457 m.a.s.l.
(using only single or mixed habitat gradient), produced the same pattern, we performed a resampling application from 1000 bootstrap samples, replacing it in the ‘boot.ci()’ function from ‘boot’ package (Canty and Ripley, 2012). With this function, we can assess the 95% confidence intervals of a regression's R2 and slope, and the median precision of the GLMs of the two approaches. Thus, we could determine if there was a difference in model precision and the effect size of the regressions. We performed all analysis above at R 3.01 software (R Core Team, 2013).
the treeline (high elevation grasslands). Termed mixed–habitat gradient, this approach allowed us to assess the elevational gradient across different vegetation types in terms of habitat formation (Fig. 1). As response variables for both approaches, we considered α–diversity as the number of ant species richness per pitfall traps (four pitfall traps in a 1.5 m2 grid), ɣ–diversity the overall ant species richness per elevational band (transect) and β–diversity as a proxy of species composition changes within each elevational band (transects). For β–diversity, we extracted the total beta diversity (βsor), using ‘beta.part’ package, which results of turnover and nestedness mechanisms (Baselga and Orme, 2012). For α–diversity, at both approaches, we performed a generalized linear model with mixed effects (GLMM) with transect as the random and the elevation as the explanatory variable. Elevation values represent the elevation band, since there were no significant differences among sampling points in the same transect. We used the GLMM to reduce the pseudoreplication effects caused by the non–independence of pitfall traps placed in the same transects (Pinheiro and Bates, 2000). First, to investigate if ɣ diversities from two approaches are relate to elevation by a monotonic decline or hump–shaped function, we modified the Nagai (2011) function to generate linear and quadratic models for comparison from those diversity components. We verified which model would be more appropriated using the Aikaike's criteria (AICc), considering as the best model the one with lowest AIC value. After that, for β and ɣ diversities, we performed a generalized linear model (GLM) with the elevation also as an explanatory variable for both approaches. All GLMs and GLMM were performed using the ‘lme4’ package (Bates et al., 2013) and were assessed under residual analyses to obtain the adequacy of error distribution (Crawley, 2002). If both approaches
3. Results In total, we collected 148 ant morpho–species belonging to 37 genera and nine subfamilies, with 119 species recorded in low elevation forests, 21 at high elevation forests and 46 at high elevation grasslands (see Appendix A Table A). In addition, low elevation forests shared more species with high elevation grasslands than with high elevation forests (Fig. 2). Patterns of α–diversity differed between the two sampling approaches (Fig. 3a and b). Using the single–habitat gradient, elevation was found to influence α–diversity, presenting an inverse relationship (Chi = 12.05; p < 0.01), however in a mixed–habitat gradient the effect of elevation on α–diversity was not found to be significant (Chi = 2.23; p = 0.13). Regarding β–diversity, the two approaches also produced different patterns, although with contrasting results (Fig. 3c e d); in the single–habitat gradient β–diversity was not influenced by elevation (F = 1.62; p = 0.25), while in the mixed–habitat gradient elevation negatively influenced β–diversity (F = 11.32; p = 0.01). Regarding ɣ–diversity both approaches followed better a linear than a 3
Acta Oecologica 102 (2020) 103489
C.J. Lasmar, et al.
high elevation grasslands than with high elevation forests (Fig. 2), even considering that habitats from low and high elevation bands differ considerably in species composition. Ants are sensitive to the level of habitat openness, which could result in different species inhabiting these areas and consequently in the different values of species richness (Andersen et al., 2007; Botes et al., 2006; van Ingen et al., 2008). If we compare high elevation forest and grassland sites, forests are cooler and more humid than the low stature grassland habitat (Aparecido et al., 2018; Körner, 2007; Safford, 1999). Consequently, the warmer conditions and low levels of moisture in high elevation grasslands may permit more species to occur than in high forest habitats. Indeed, in this sense, despite inherent and notable differences, high elevation grassland habitat is more similar (in terms of temperature and moisture) to forests from low elevation bands which might improve the supporting of ant colonies in comparison to high elevational forests. Differences in β–diversity patterns (Fig. 3c and d) could be caused by different degrees of habitat heterogeneity between sites in high elevation grasslands and forests. The decrease of β–diversity according to elevation when analyzing the mixed–habitat gradient is likely due to high elevation grasslands sites being more homogenous in comparison to both low and high elevation forests. Such homogenization in terms of habitat conditions (related to temperature and moisture) would potentially permit ants in grass lands to colonize almost all sampling points in the transect; while this would likely not occur in more heterogeneous forest sites, leading to different patterns. We hypothesize that β–diversity within a single–habitat gradient may be driven by its own habitat heterogeneity (Jankowski et al., 2009; Nunes et al., 2016), which may not necessarily vary across the elevational gradient, at least in the extent of elevational gradient assessed in this study. Regarding ɣ–diversity we observed the same monotonic decline in diversity in both approaches (Fig. 3e and f). Elevational gradients have been found to produce a strong pattern on species diversity for other insect assemblages, resulting from the influence of temperature (Malsch et al., 2008; Sanders et al., 2007). Thus, even changes in vegetation type would likely not change this pattern. According to our results, when analyzing ɣ–diversity on a mixed–habitat gradient there was a decrease in model robustness in terms of the proportion of explained variance and effect size of elevation. In fact, species diversity is not influenced by elevation per se but by factors that covary with it. Sampling in different vegetation types likely means that diversity is dependent on factors related to differences in the structure of habitats, and not only limited to elevation–related factors. As Gotelli and Ellison (2004) have already pointed out, when many confounding factors are included in a study's design, it is difficult to disentangle their effects. Thus, the higher proportion of explanation of the variation in diversity found in a single–habitat gradient is likely due to the similarity of other environmental factors, except those directly linked to variation in elevation. Similarly, the effect size of elevation on species richness in a mixed–habitat gradient is lower because, as well as elevation, there are other environmental dissimilarities (as previously mentioned) that may permit more species to survive at higher elevations in grasslands than in the high elevation forests.
Fig. 2. The number of exclusive and shared ant species recorded in the three habitat types across an elevational gradient in Itatiaia National Park. The numbers within the overlapping parts of the circles represent the shared species found in the corresponding habitats, while the numbers outside of the overlap show the number of species exclusive to that habitat.
quadratic function (Appendix B Fig. B). Furthermore, both approaches exhibited the same pattern (Fig. 3e and f), with both single (F = 35.16; p < 0.01) and mixed–habitat gradient (F = 17.27; p < 0.01). Despite obtaining the same ɣ–diversity pattern for both transect approaches, we can observe a greater variance explained on single–habitat gradient than in mixed–habitat gradient regression (Fig. 4a). The single–habitat gradient also demonstrated a higher effect size based on the regression slope values (Fig. 4b). 4. Discussion The standardization of vegetation type across an elevational gradient has been empirically tested in this study. Our results unequivocally show that non–standardization of vegetation type can be a source of bias in the parameters evaluating community responses, which is in accord with Axmancher and Fielder (2008), who verified the vegetation type influence on geometrid moths’ diversity across an elevational gradient. The patterns observed for α–diversity and β–diversity clearly differed between the two approaches. Moreover, despite a monotonic decline observed for ɣ–diversity in both approaches, our analyses demonstrate that sampling in the same vegetation type leads to stronger results in order to explain the elevational gradient effects on species richness. We suggest that the environmental conditions imposed by the two vegetation types (grasslands and forests) are the main drivers of the contrasting elevation gradient patterns observed for α–diversity. Higher elevations present lower temperatures and high levels of moisture, which can affect ant species richness, likely producing the decrease of alpha diversity on a single–habitat gradient. Low temperatures and high levels of moisture affect the survival of ant colonies (for example decreasing foraging activity), and some species are unable to survive or colonize areas with such conditions (Bishop et al., 2017; Bruhl et al., 1999; Fisher, 1998; Malsch et al., 2008; Sanders et al., 2007). When analyzing the mixed–habitat gradient, we did not detect the same decrease of alpha diversity, as high elevation grasslands reached almost the same α–diversity of the most speciose low elevation forests (Fig. 3a and b). Furthermore, low elevation forests share more ant species with
5. Conclusions This study demonstrates that sampling in different vegetation types across an elevational gradient produces distinct and likely biased diversity patterns, compared to sampling in standardized habitats when aiming to assess exclusively elevational effects. As highlighted by Körner (2007), there is a risk of confounding geophysical factors not directly related to elevation, which may lead to assessments of other environmental gradients. With this in mind, we suggest that studies solely addressing elevational gradient effects must perform their sampling in the same vegetation type, in terms of closed forest or grasslands formation. However, when assessing only one vegetation type is not an option, we suggest authors test overall diversity (ɣ–diversity), instead 4
Acta Oecologica 102 (2020) 103489
C.J. Lasmar, et al.
Fig. 3. Relationship between ant diversity measures and elevation gradient with samples in only one vegetation type (single–habitat gradient) and with samples in two vegetation types (mixed–habitat gradient). Both single and mixed–habitat gradients present the same five first elevations (low elevation forest). The other three last elevations present only high elevation forests transects for single–habitat gradient and high elevation grasslands transects for mixed–habitat gradient. a) Ant species richness per pitfall traps (α–diversity) per sampling point of single–habitat gradient (Chi = 12.05; p < 0.01). b) Ant species richness per pitfall traps (α–diversity) of mixed–habitat gradient (Chi = 2.23; p = 0.13). c) Beta diversity (β–diversity) of only single–habitat gradient (F = 1.62; p = 0.25). d) Beta diversity (β–diversity) of mixed–habitat gradient (F = 11.32; p = 0.01) e) Overall ant species richness (ɣ–diversity) of single–habitat gradient (F = 35.16; p = 0.001). f) Overall ant species richness (ɣ–diversity) of mixed–habitat gradient (F = 17.27; p = 0.006).
of α–diversity and β–diversity, as the effects of elevation may be stronger at that scale. Furthermore, it seems that α–diversity and β–diversity responses are more sensitive to habitat changes. We also recommend that at the very least, vegetation types should be included in statistical models. The data presented here also indicate that many of published studies on elevational gradients performed on different vegetation types that aimed to assess only elevational effects, possibly confounded both elevational and vegetational effects leading to
potentially biased results. In order to reinforce this argument, future studies should address the influence of other vegetation types that present intermediate differences (e.g. shrublands, open woodland) on elevational gradient diversity patterns. Finally, we conclude that removing the non–standardized sampling bias is essential to fully understand elevational gradient patterns on biodiversity.
5
Acta Oecologica 102 (2020) 103489
C.J. Lasmar, et al.
Fig. 4. Comparison of GLM regression attributes between single and mixed–habitat elevation gradients on overall ant species richness (ɣ diversity). a) Proportion of variance explained (R2) and b) Effect size (slope of the regression line). Vertical dashed lines are bootstrapped confidence intervals based on 1000 bootstrap samples with replacement. Notch areas on boxplots mark the 95% confidence intervals of the median value (shown as a black horizontal line). The horizontal dotted line represent 0 value in y-axis. Black dots represent data outliers.
Author contributions
et al., 2018. Ecohydrological drivers of Neotropical vegetation in montane ecosystems. Ecohydrology 11 (3), e1932. https://doi.org/10.1002/eco.1932. Axmacher, J.C., Fiedler, K., 2008. Habitat type modifies geometryof elevational diversity gradients in geometrid moths (Lepidoptera Geometridae) on Mt Kilimanjaro, Tanzania. Trop. Zool. 21, 243–251. Axmacher, J.C., Brehm, G., Hemp, A., Tünte, H., Lyaruu, H.H.V., Müller–Hohenstein, K., Fiedler, K., 2009. Determinants of diversity in afrotropical herbivorous insects (Lepidoptera: geometridae): plant diversity, vegetation structure or abiotic factors? J. Biogeogr. 36, 337–349. https://doi.org/10.1111/j.1365-2699.2008.01997.x. Baccaro, F.B., Feitosa, R.M., Fernandez, F., Fernandes, I.O., Izzo, T.J., Souza, J.L.P., et al., 2015. Guia para gêneros de formigas do Brasil. Editora INPA, Manaus. https://doi. org/10.5281/zenodo.32912. Baselga, A., Orme, C.D.L., 2012. betapart: an R package for the study of beta diversity. Methods Ecol E 3, 808–812. https://doi.org/10.1111/j.2041–210X.2012.00224.x. Bates, D., Maechler, M., Bolker, B., Walker, S., 2013. lme4: linear mixed effects models using Eigen and S4. Cran.R - project. https://cran.r-project.org/web/packages/ lme4/index.html, version 1, Accessed date: 1 March 2017. Bestelmeyer, B.T., Agosti, D., Leeanne, F., Alonso, T., Brandão, C.R.F., Brown, W.L., et al., 2000. Field techniques for the study of ground–living ants, an overview, description, and evaluation. In: Agosti, D., Majer, J.D., Tennant, A., Schultz, T. (Eds.), Ants, Standard Methods for Measuring and Monitoring Biodiversity. Smithsonian Institution Press, Washington, pp. 122–144. https://doi.org/10.5281/zenodo.11736. Bishop, T.R., Robertson, M.P., van Rensburg, B.J., Parr, C.L., 2017. Coping with the cold: minimum temperatures and thermal tolerances dominate the ecology of mountain ants. Ecol. Entomol. 42, 105–114. https://doi.org/10.1111/een.12364. Botes, A., McGeoch, M.A., Robertson, H.G., van Niekerk, A., Davids, H.P., Chown, S.L., 2006. Ants, altitude and change in the northern cape floristic region. J. Biogeogr. 33, 71–90. https://doi.org/10.1111/j.1365–2699.2005.01336.x. Brühl, A., Mohamed, M., Linsenmair, K.E., 1999. Altitudinal distribution of leaf litter ants along a transect in primary forests on Mount Kinabalu, Sabah, Malaysia. J. Trop. Ecol. 15, 265–277. https://doi.org/10.1017/S0266467499000802. Canedo–Júnior, E.O., Cuissi, R.C., Curi, N.H.A., Demetrio, G.R., Lasmar, C.J., Malves, K., Ribas, C.R., 2016. Can anthropic fires affect epigaeic and hypogaeic Cerrado ant (Hymenoptera, Formicidae) communities in the same way? Rev. Biol. Trop. 64, 95–104. https://doi.org/10.15517/rbt.v64i1.18239. Canty, A., Ripley, B., 2012. Boot, Bootstrap R (S–Plus) Functions. R Package version1.3–4. https://CRAN.R_project.org/package=boot, Accessed date: 1 February 2017. Carneiro, E., Mielkel, O.H.H., Casagrande, M.M., Fiedler, K., 2014. Community structure of skipper butterflies (Lepidoptera, hesperiidae) along elevational gradients in Brazilian atlantic forest reflects vegetation type rather than altitude. PLoS One 9 (10), e108207. https://doi.org/10.1371/journal.pone.0108207. Classen, A., Peters, M.K., Kindeketa, W.J., Appelhans, T., Eardley, C.D., Gikungu, M.W., et al., 2015. Temperature versus resource constraints, which factors determine bee diversity on Mount Kilimanjaro, Tanzania? Glob. Ecol. Biogeogr. 24, 642–652. https://doi.org/10.1111/geb.12286. Colwell, R.K., Brehm, G., Cardelús, C.L., Gilman, A.C., Longino, J.T., 2008. Global warming, elevational range shifts, and lowland biotic attrition in the wet tropics. Science 322, 258–261. https://doi.org/10.1126/science.1162547. Colwell, R.K., Rangel, T.F., 2010. A stochastic, evolutionary model for range shifts and richness on tropical elevational gradients under Quaternary glacial cycles. Philos. T. R. Soc. B. 365, 3695–3707. https://doi.org/10.1098/rstb.2010.0293. Crawley, M.J., 2002. Statistical Computing, an Introduction to Data Analysis Using S–Plus, first ed. John Wiley Sons, London.
CJL, CRR, JL and ACMQ originally formulated the idea. CJL, RMF, MMGI, GPA, GBN and DQD conducted fieldwork and collected the data. CJL, CRR, JL, ACMQ and FSN performed statistical analyses. CJL wrote the article with substantial collaboration from all co-authors. Declaration of competing interest None. Acknowledgements This work was funded by Fundação de Amparo à Pesquisa de Minas Gerais (FAPEMIG, grant PPM 00243/14). We are thankful to the staff of National Itatiaia Park, especially to Leo Nascimento, who permitted the sampling in the park area. We are also indebted to Maria Regina de Souza, Tobias R. Silva, Luiza Santiago, Edson Guilherme de Souza, Ernesto O. Canedo–Júnior, Graziele Santiago and Luana Zurlo for their help with logistics and fieldwork. We also thank Filipe França for his help and advice on statistical analyses. Thanks to Mariana Rabelo, Ícaro Carvalho and Felipe Lopes for helping with laboratory work. We are grateful to the staff at the Laboratório de Sistemática de Formigas da Universidade Federal do Paraná for confirming ant identification, especially to Alexandre Ferreira. This study was part of CJ Lasmar's MSc at UFLA that was supported by Coordenação de Aperfeiçoamento Pessoal (CAPES, Finance code: 001). RMF was supported by the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq grant 302462/2016–3). Appendix A. Supplementary data Supplementary data related to this article can be found at https:// doi.org/10.1016/j.actao.2019.103489. References Andersen, A.N., van Ingen, L.T., Campos, R.I., 2007. Contrasting rainforest and savanna ant faunas in monsoonal northern Australia, a rainforest patch in a tropical savanna landscape. Aust. J. Zool. 55, 363–369. https://doi.org/10.1071/ZO07066. Aparecido, L.M.T., Teodoro, G.S., Mosquera, G., Brum, M., Barros, F.V., Pompeu, P.V.,
6
Acta Oecologica 102 (2020) 103489
C.J. Lasmar, et al.
acprof:oso/9780199544639.001.0001. Pinheiro, J.C., Bates, D.M., 2000. In: Pinheiro, J.C. (Ed.), Mixed-effects models in S and SPLUS. Springer, New York. Pompeu, P.V., Fontes, M.A.L., Mulligan, M., Bueno, I.T., Siqueira, M.F., Acerbi–Júnior, F.W., et al., 2018. Assessing Atlantic cloud forest extent and protection status in southeastern Brazil. J. Nat. Conserv. 43, 146–155. https://doi.org/10.1016/j.jnc. 2018.04.003. R Core Team, 2013. R: A Language and Environment for StatisticalComputing. R Foundation for Statistical Computing, Vienna, Austria. Rahbek, C., 2005. The role of spatial scale and the perception of large–scale species–richness patterns. Ecol. Lett. 8, 224–239. https://doi.org/10.1111/j.1461–0248. 2004.00701.x. Roslin, T., Hardwick, B., Novotny, V., Petry, W.K., Andrew, N.R., Asmus, A., et al., 2017. Higher predation risk for insect prey at low latitudes and elevations. Science 356, 742–744. https://doi.org/10.1126/science.aaj1631. Safford, H.D., 1999. Brazilian Páramos I. An introduction to the physical environment and vegetation of the campos de altitude. J. Biogeogr. 26, 693–712. https://doi.org/10. 1046/j.1365–2699.1999.00313.x. Safford, H.F., 2001. Brazilian páramos III. Patterns and rates of post fire regeneration in the campos de Altitude. Biotropica 33, 282–302. https://doi.org/10.1111/j. 1744–7429.2001.tb00179.x. Sanders, N.J., Lessard, J.P., Fitzpatrick, M.C., Dunn, R.R., 2007. Temperature, but not productivity or geometry, predicts elevational diversity gradients in ants across spatial grains. Glob. Ecol. Biogeogr. 16, 640–649. https://doi.org/10.1111/j. 1466–8238.2007.00316.x. Smith, M.A., Hallwachs, W., Janzen, D.H., 2014. Diversity and phylogenetic community structure of ants along a Costa Rican elevational gradient. Ecography 37, 720–731. https://doi.org/10.1111/j.1600–0587.2013.00631.x. Stevens, G.C., 1992. The elevational gradient in altitudinal range, an extension of Rapoport's latitudinal rule to altitude. Am. Nat. 140, 893–911. https://doi.org/10. 1086/285447. Sundqvist, M.K., Sanders, N.J., Wardle, D.A., 2013. Community and ecosystem responses to elevational gradients, processes, mechanisms, and insights for global change. Annu. Rev. Ecol. Evol. Syst. 44, 261–280. https://doi.org/10.1146/ annurev–ecolsys–110512–135750. Szewczyk, T., McCain, C.M., 2016. A systematic review of global drivers of ant elevational diversity. PLoS One 11 (5), e0155404. https://doi.org/10.1371/journal.pone. 0155404. Tello, J.S., Myers, J.A., Macía, M.J., Fuentes, A.F., Cayola, L., Arellano, G., Loza, M.I., Torrez, V., Cornejo, M., Miranda, T.B., Jørgensen, P.M., 2014. Elevational gradients in β–diversity reflect variation in the strength of local community assembly mechanisms across spatial scales. PLoS One 10 (3), e0121458. https://doi.org/10.1371/ journal.pone.0121458. van Ingen, L.T., Campos, R.I., Andersen, A.N., 2008. At community structure along an extended rain forest–savanna gradient in tropical Australia. J. Trop. Ecol. 24, 445–455. https://doi.org/10.1017/S0266467408005166. von Humboldt, A., 1849. Aspects of nature. In: Different Lands and Different Climates, with Scientific Elucidations, second ed. Longman, London. Wen, Z., Wu, Y., Cheng, J., Cai, T., Du, Y., Ge, D., et al., 2018. Abundance of small mammals correlates with their elevational range sizes and elevational distributions in the subtropics. Ecography 41, 1–11. https://doi.org/10.1111/ecog.03558.
Fisher, B.L., 1998. Ant diversity patterns along an elevational gradient in the réserve spéciale d'Anjanaharibe–Sud and on the western Masoala Peninsula, Madagascar. Fieldiana, Life earth sci. 90, 39–67. Gotelli, N.J., Ellison, A.M., 2004. Designing successful field studies. In: Gotelli, N.J., Ellison, A.M. (Eds.), A Primer of Ecological Statistics. Sinauer Associates, Sunderland, Massachusetts, U.S.A, pp. 155–179. Harsch, M.A., Bader, M.Y., 2011. Treeline form – a potential key to understanding treeline dynamics. Glob. Ecol. Biogeogr. 20, 582–596. https://doi.org/10.1111/j.1466–8238. 2010.00622.x. Hemp, A., 2006. Continuum or zonation? Altitudinal gradients in the forest vegetation of Mt. Kilimanjaro. Plant. Ecol. 184, 27–42. https://doi.org/10.1007/ s11258–005–9049–4. Jankowski, J.E., Ciecka, A.L., Meyer, N.Y., Rabenold, K.N., 2009. Beta diversity along environmental gradients, implications of habitat specialization in tropical montane landscapes. J. Anim. Ecol. 78, 315–327. https://doi.org/10.1111/j.1365–2656.2008. 01487.x. Körner, C., 1998. A re–assessment of high elevation treeline positions and their explanation. Oecologia 115, 445–459. https://doi.org/10.1007/s004420050540. Körner, C., 2007. The use of ‘altitude’ in ecological research. Trends Ecol. Evol. 22, 569–574. https://doi.org/10.1016/j.tree.2007.09.006. Kraft, N.J.B., Comita, L.S., Chase, J.M., Sanders, N.J., Swenson, N.G., Crist, T.O., et al., 2011. Disentangling the drivers of β diversity along latitudinal and elevational gradients. Science 333, 1755–1758. https://doi.org/10.1126/science.1208584. Malsch, A.K.F., Fiala, B., Maschwitz, U., Mohamed, M., Nais, J., Linsenmair, K.E., 2008. An analysis of declining ant species richness with increasing elevation at Mount Kinabalu, Sabah, Borneo. Asian. Myrmecol. 2, 33–49. https://doi.org/10.20362/am. 002004. Mezger, D., Pfeiffer, M., 2011. Partitioning the impact of abiotic factors and spatial patterns on species richness and community structure of ground ant assemblages in four Bornean Rainforests. Ecography 34, 39–48. https://doi.org/10.1111/j. 1600–0587.2010.06538.x. Moreau, C.S., Bell, C.D., 2013. Testing the museum versus cradle tropical biological diversity hypothesis, phylogeny, diversification, and ancestral biogeographic range evolution of the ants. Evolution 67, 2240–2257. https://doi.org/10.1111/evo.12105. Nagai, M.E., 2011. Eco R: using R [WWW document]. Exerc. Do micael. http://ecologia. ib.usp.br/bie5782/doku.php?id=bie5782:05_curso_antigo:r2011:alunos:trabalho_ final:micael:start, Accessed date: 25 January 2018. Nakamura, A., Burwell, C.J., Ashton, L.A., Laidlaw, M.J., Katabuchi, M., Kitching, R.L., 2015. Identifying indicator species of elevation: comparing the utility of woody plants, ants and moths for long–term monitoring. Austral Ecol. 41, 179–188. https:// doi.org/10.1111/aec.12291. Nunes, C.A., Braga, F.R., Figueira, J.E.C., Neves, F.S., Fernandes, G.W., 2016. Dung beetles along a tropical altitudinal gradient, environmental filtering on taxonomic and functional diversity. PLoS One 11 (6), e0157442. https://doi.org/10.1371/ journal.pone.0157442. Peters, M.K., Hemp, A., Appelhans, T., Behler, C., Classen, A., Detsch, F., et al., 2016. Predictors of elevational biodiversity gradients change from single taxa to the multi–taxa community level. Nat. Commun. 7, 13736. https://doi.org/10.1038/ ncomms13736. Philpott, S.M., Perfecto, I., Armbrecht, I., Parr, C.L., 2010. Ant diversity and function in disturbed and changing habitats. In: Lach, L., Parr, C.L., Abbott, K.L. (Eds.), Ant Ecology. Oxford University Press, Oxford, pp. 137–156. https://doi.org/10.1093/
7