SEARCH

SEARCH BY CITATION

Keywords:

  • Atherton Tablelands;
  • Competition;
  • Edge effects;
  • Forest fragmentation;
  • Plant–plant interactions;
  • Point-pattern analysis

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Question

In fragmented forests, edge effects lead to changes in the distribution of plant species. In particular, tropical forest edges are increasingly dominated by lianas. Will this increase in lianas lead to changes in their interactions with other plant morphological groups? If so, will this alter the local distributions and abundance of other species?

Location

Plots located at increasing distances from the nearest forest edge and in remnant fragments of rain forest in the Atherton Tablelands, far northeast Queensland, Australia.

Methods

We mapped the distribution of trees, lianas and epiphytic ferns to better understand the role of forest disturbance in shaping their competitive and facilitative interactions. We then used specific spatial point-process analyses to examine the effects of the spatial distribution of trees on the presence and abundance of lianas and epiphytic ferns.

Results

Tree aggregation near forest edges was lower than that in the interior. The higher abundance of lianas near edges was associated with increased spatial segregation between lianas and epiphytic ferns. This segregation suggests there is competition between these two functional groups, and that lianas, being much more abundant, probably outcompete epiphytic ferns.

Conclusions

The ability of lianas to thrive in disturbed tropical rain forests appears to reduce the abundance of epiphytic ferns, probably via direct competition for space. Epiphytic ferns provide unique microclimates and harbour much biodiversity, and their decline could negatively affect many animals and plants that rely upon them.


Nomenclature
Tracey &

Webb (1975)

Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Tropical forests worldwide are being cleared and disturbed (Wright 2005, 2010; Laurance et al. 2012), leading to rapid changes in the distribution of many species (Fahrig 2003). Among the most studied effects of forest disturbance are those related to forest edges, which are artificial, often abrupt boundaries between remnant forest fragments and the modified habitats around them (Murcia 1995; Laurance et al. 2002, 2011; Ries et al. 2004; Harper et al. 2005; Lindenmayer & Fischer 2006). The creation of these new edges leads to a proliferation of pioneer plant species in detriment of long-lived ones that might have been common before disturbance (Laurance et al. 2000, 2006). These changes in the identity of the species in the edge community invariably lead to modifications in the persistence and distribution of many organisms as well as to alterations in plant demography (Murcia 1995; Jules 1998). This in turn can lead to important changes in ecological interactions between species (Fagan et al. 1999), to the extinction of some species (Laurance et al. 2000) and eventually to the formation of new ecological communities near forest edges (Ries et al. 2004). In some cases, edge effects can lead to a marked homogenization of edge communities (Tabarelli et al. 2008), with important consequences for functional diversity and ecosystem functioning across the landscape.

In tropical plant communities, one of the most conspicuous edge effects is an increase in the abundance of lianas (woody vines; Laurance et al. 2001; Schnitzer & Bongers 2011). Lianas are a widespread feature of tropical forests globally (Hegarty & Caballe 1991) and become particularly abundant in tree-fall gaps and sites affected by disturbance (Putz 1984; Schnitzer & Bongers 2002). The proliferation of lianas in fragmented and disturbed forests (Schnitzer & Bongers 2011) has been demonstrated in tropical Australia (Laurance 1991), south-eastern Brazil (Oliveira-Filho et al. 1997), the Amazon Basin (Laurance et al. 2001; Phillips et al. 2002; Benítez-Malvido & Martínez-Ramos 2003; Foster et al. 2008), Panama (Wright et al. 2004; Wright & Calderon 2006), French Guiana (Chave et al. 2008) and Costa Rica (Rutishauser 2011). Many of these studies evaluate the impact that lianas have on their host trees (Schnitzer & Bongers 2002, 2011; Paul & Yavitt 2011). However, abundant lianas could also cause other ecological changes in tropical forests. Vegetative reproduction and facilitation processes between different species of liana (Pinard & Putz 1994) lead to clumped spatial patterns and clustering, especially near forest edges (Schnitzer & Bongers 2011). Given the relative paucity of places to establish within tropical forests, such clumping could lead to patchy interspecific competition with other plant functional groups, such as epiphytes, that also rely on trees for support.

Epiphytes (those that root on the surface of tree trunks and branches but do not harm the host, sensu Benzing 2004) are major contributors to vascular plant diversity and biomass (Gentry & Dodson 1987), and play key roles in nutrient cycling (Benzing 1998; Muñoz et al. 2003). Of special relevance in tropical Australia are epiphytic ferns (Cummings et al. 2006), which support diverse communities of animals and plants, including a variety of bird (Cruz-Angon & Greenberg 2005), reptile (Freeman et al. 2005) and invertebrate (Ellwood et al. 2002) species, some of which have obligate relationships with their host epiphytes. These ‘islands’ of unique habitat might be sensitive to forest disturbance (Wenzhang et al. 2008), but the mechanisms responsible for this sensitivity are uncertain. One possible explanation is that epiphytes suffer from competition with other co-occurring plants, such as lianas. If this were the case, we might be able to detect this interspecific competition between both functional groups via increased spatial segregation at small scales. In addition, given the increased presence of lianas near forest edges, we would expect this spatial segregation to be more frequent there.

Here we assess whether and how forest fragmentation alters the ecological interactions between lianas and epiphytic ferns in tropical Queensland, where lianas thrive under disturbed conditions (Laurance 1997). Our aim is to determine whether fern–liana–tree interactions have a strong spatial component (see Raventós et al. 2011) and to address three specific questions: (1) do lianas and epiphytic ferns respond to edge effects; (2) do they interact in space; and (3) do they compete for the use of trees? Our findings may help to reveal how competition for supporting trees helps to structure rain forest plant communities.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Study site

The study was carried out on the Atherton Tableland in NE Queensland, Australia (Fig. 1). Mean annual precipitation ranges from 2000 to 3000 mm at different locations, and rainfall is highly seasonal, peaking from January to April. The rain forests in the study area are complex notophyll vine-forests (Tracey & Webb 1975), ranging in height from 20 to 40 m, and are dominated by the following tree species: Agathis microstachya (Araucariaceae), Ficus spp. (Moraceae), Flindersia brayleyana (Rutaceae), Geissois biagiana (Cunoniaceae), Cryptocarya onoprienkoana (Lauraceae), Argyrodendron peralatum (Malvaceae), Castanospermum australe (Fabaceae) and Cardwellia sublimis (Proteaceae). Lianas and epiphytes are common throughout these forests (Tracey 1987; Cummings et al. 2006), with lianas known to increase in disturbed areas (Laurance 1997).

image

Figure 1. (a) Map showing study area location in northeast Queensland, Australia. (b) Location of forests in Atherton Tablelands. Dark grey areas indicate primary forest, black areas represent secondary forest and light grey areas show the fragments sampled in the study. (c,d) Example of one of the plots measured inside one of the forests. (c) Location of five plots in relation to forest edge. (d) Trees measured inside a particular plot (located 80–100 m from the nearest forest edge). In light grey are trees with lianas, dark grey is trees with no epiphytes or climbers and black is trees with epiphytic ferns. Different sized circles indicate different loads of lianas and epiphytic ferns.

Download figure to PowerPoint

Queensland tropical forests are listed among 200 ecoregions that should be prioritized for global conservation due to their vulnerable status (Olson & Dinerstein 2002) and biologically unique nature. These forests have historically suffered heavy logging and land clearing for agriculture (Goosem et al. 1999). Clearing in the study area began with European colonization around 1909 and proceeded rapidly until the 1940s, then slowed in the following decades (Winter et al. 1987; Eacham Historical Society 1995). Currently, most primary forest in the study area comprises forest fragments ranging from 1 to 600 ha in area, surrounded by cattle pastures and secondary forests, most of which are under some form of protection. The main threats currently faced by forests are due to isolation, edge effects and invasive species, although present plant distributions could reflect past disturbances (Dupouey et al. 2002; Lindborg & Eriksson 2004). Steeper, more nutrient-poor areas to the south, west and east of the Atherton Tablelands sustain larger, continuous forests.

The forests of the Atherton Tablelands harbour several species of epiphytic fern, among the most common of which are those in the genera Asplenium (Aspleniaceae) and Platycerium (Polypodiaceae). Asplenium creates a unique microclimate in the forest (Cummings et al. 2006) and harbours important invertebrate communities (Ellwood et al. 2002), whereas Platycerium has important functions in capturing and retaining moisture in the forest canopy (Cummings et al. 2006). There are >390 species of liana known in the study area, belonging to four main families, Apocynaceae, Convolvulaceae, Fabaceae and Vitaceae (Centre for Australian National Biodiversity Research 2010). These lianas have an array of climbing strategies that include mainstem twiners, branch twiners, tendril climbers, root climbers and scramblers (see Putz 1984). Like other lianas worldwide, they frequently reproduce vegetatively. However, whereas lianas in the New World contain many wind-dispersed species (Gentry 1991), those in Australian tropical forests, similar to others found in Central African and some Asian regions, often have fleshy fruits (Centre for Australian National Biodiversity Research 2010) presumably adapted for animal dispersal.

Plot measures

From April to June 2012 we selected five primary rain forest fragments ranging from ca. 18 to 9500 ha in area (Table S1). Within each fragment we located five 20 × 20-m plots at increasing distances from the nearest forest edge (0–20 m, 20–40 m, 40–60 m, 60–80 m, 80–100 m; see Fig. 1), following a linear transect perpendicular to the edge. To avoid possible interactions due to the presence of other neighbouring edges within a forest fragment (Porensky & Young 2013), all plots were located >150 m from any second edge. In addition, we tried to avoid the eastern faces of the fragments, which are more exposed to cyclonic activity. We measured canopy cover in the four corners and centre of each plot, by averaging at each point four spherical densitometer readings taken at cardinal directions.

For each tree, liana and epiphytic fern in our plots, we recorded their spatial location with a handheld GPS (Garmin GPSMAP 62). To increase precision, we averaged 50–100 GPS measurements for each plant, and corrected GPS measurements using known waypoints. We also measured the DBH of each tree ≥10 cm and recorded the number of lianas and epiphytic ferns using that tree for support. Fern abundance was evaluated by counting individual ferns from the forest floor using binoculars; individual species can be distinguished once their reproductive fronds have developed. However, we counted only ferns with a basal diameter of ≥10 cm (where basal diameter is the area of contact between fern and host tree; Cummings et al. 2006). We also measured the DBH of all lianas ≥1 cm in DBH at 1.3 m from the rooting point, following recent protocols (Gerwing et al. 2006; Schnitzer et al. 2006, 2008), and recorded all trees (≥10 cm DBH) that they used as support (although lianas can also climb over trees with DBH < 10 cm). Each liana was assigned to one of five climbing guilds: mainstem twiner, branch twiner, tendril climber, root climber and scrambler (Putz 1984). For each plot, we calculated the ‘morphological diversity’ of lianas and epiphytic ferns using Shannon's diversity index (H), based on the proportions of plants in the two fern genera and five liana climbing guilds.

Data analysis (between plots)

We analysed our data (including tree abundance, and liana and fern presence and abundance per plot) using generalized linear mixed-effect models (GLMMs) with package glmmADMB in R 2.15.0 (R Foundation for Statistical Computing, Vienna, AT; see Table S2 for fixed effects included, and error distributions and link functions fitted for each model). ‘Fragment’ was always included as a random variable, given the dependence of plots within the same fragment. We created all possible model combinations using the dredge function in the MuMIn package and then selected all models with ΔAIC < 2 and calculated averaged confidence intervals for all variables included in the subset of best-performing models (Burnham & Anderson 2002).

Data analysis (within plots)

We used techniques of spatial point-pattern analysis (Wiegand & Moloney 2004; Law et al. 2009) to describe the locations of host trees, and the presence/absence and abundance of lianas and epiphytic ferns. The spatial explicit distribution of plants carries information about the processes that are currently operating or that occurred in the past, and hence we can infer ecological processes with the analysis of the spatially explicit distribution of points (Law et al. 2009). Analyses were performed using the software PROGRAMITA (Wiegand & Moloney 2004). For each plot, we performed spatial point-pattern analyses of host trees at three levels: (1) the spatial structure of trees within the forest, which gives us the spatial scale over which lianas and ferns interact; (2) the spatial distribution of host trees with lianas or ferns; and (3) the spatial pattern of the presence and abundance of ferns on host trees occupied by lianas (see Raventós et al. 2011). The former analyses were based on the theory of ‘unmarked’ (i.e. the points only depend on the spatial location of host trees) and ‘marked’ point patterns (i.e. apart from location, each tree carries marks, such as species identity, epiphyte type colonizing it or size; Illian et al. 2008).

For each analysis and plot, we calculated the summary statistics for the observed and simulated (null) point-process models. We used simulations (199 replicates) to estimate the envelopes encircling the 95% range of values of summary statistics under a given point-process model (i.e. the 5th lowest and highest values). Departures from the point-process models occurred if the observed summary statistics were outside the simulation envelopes (Wiegand & Moloney 2004; Illian et al. 2008). We also used Diggle's (2003) goodness of fit (GoF) test to assess the overall fit of null models for a given test statistic (Loosmore & Ford 2006). If the rank of the test was >200, the data showed a significant departure from the null model, with an error rate α = 0.05.

Given that we were interested in obtaining global results while using the full statistical power of our data, we combined the summary statistics of different plots using techniques for replicated patterns (Law et al. 2009). These techniques allow for the combination of all summary statistics into a single ‘master’ statistical test that basically represents the average of the summary statistics of each plot, weighted by the number of focal points in the focal pattern (Illian et al. 2008; Raventós et al. 2011). In other words, we focused on the average process, rather than on the potential variability for each individual plot (for a similar study following the same methodology, see Raventós et al. 2011). We thus pooled our plots into interior (60–100 m from edge) and edge (0–60 m from edge) categories, as these were relatively distinct. For ferns, however, we grouped our results by fragment as there was little difference between forest edges and interiors (see 'Results').

Spatial distribution of trees

To describe the pattern of tree aggregation with distance from the forest edge we carried out ‘unmarked’ point-pattern analyses (Diggle 2003; Wiegand & Moloney 2004; Illian et al. 2008) using the univariate pair-correlation function g(r) (Diggle 2003; Illian et al. 2008). For each distance, the significance of the observed g(r) describing tree aggregation was evaluated against a random pattern (null model) generated by randomly drawing an identical tree number for each plot. Analyses were carried out separating edge and interior plots.

Spatial distribution of ferns and lianas

In this case we analysed whether the spatial pattern of host trees with and without lianas or ferns was randomly distributed. The latter analyses are based on the use of null models of random labelling: we randomly assigned the pattern of trees with the labels (a) with and without lianas, and (b) with and without ferns, conditional on the observed tree distribution. Specifically, we used the univariate mark-connection function p11(r), where trees with lianas or ferns (type 1) were compared to the distribution of trees without lianas or ferns (type 2; Diggle 2003; Illian et al. 2008). If p11(r) of the observed pattern showed a positive departure from that of the null model [i.e. p11(r) p1p1(r)], the presence of either lianas or ferns was significantly clustered at scale r, conditional on tree distribution. We used the bivariate function p12(r) to further explore the spatial relationship between trees with (type 1) or without (type 2) lianas or ferns. Attraction (or segregation) between type 1 and type 2 groups occurs if observed p12(r) shows positive (or negative) departures from that of the null model. We did not detect differences between edge and interior plots (results not shown), and we thus carried out these analyses combining all plots.

Spatial distribution of ferns dependent on lianas

We used a trivariate random labelling (e.g. Raventós et al. 2011) to test whether trees with and without lianas differed in their probability of bearing ferns. We contrasted the observed point pattern to an expected null model of random fern distribution conditional on host trees (Raventós et al. 2011). With this null model, ‘fern presence’ was randomly shuffled among host trees, leaving the observed liana presence on trees fixed. Analyses were carried out for both edge and interior plots. We did not analyse the effect of ferns on lianas because ferns were much less abundant than lianas (see 'Results').

Spatial abundance of lianas and ferns

To examine the spatial pattern of the abundance of lianas or ferns on host trees, we used the univariate r-mark correlation function km1(r) (Stoyan & Stoyan 1994; Getzin et al. 2008). In each plot null models were generated randomly, shuffling the mark m (i.e. liana or fern abundance on host trees) over the observed tree distribution. If the analysis of liana abundance on trees departed negatively from the null model [i.e. km1(r) < 1], then this means that host trees with high loads of lianas are free of trees with low loads of lianas (i.e. inhibition of lianas). By contrast, a positive departure from the null model of the liana abundance [i.e. km1(r) > 1] would indicate that host trees with high loads of lianas are also surrounded by trees with low loads of lianas (i.e. mutual facilitation of lianas). The same procedure could be extrapolated to fern abundance. Analyses for liana abundance were carried out for edge and interior plots, while all plots were combined for fern abundance.

Furthermore, the bivariate mark-correlation function km1, m2(r) gives the mean mark product of the number of lianas (m1) and ferns (m2) on trees at distance r. If the host trees have smaller than average marks m1 and m2 when they are closer together, there is inhibition; if they have larger than average marks when they are closer together, there is mutual facilitation. To test for a significant correlation of the m2 marks dependent on the position of m1 marks (i.e. spatial abundance of ferns dependent on that of lianas; see above for a similar analysis for lianas or fern presence), we used a null model where m1 marks were fixed and m2 marks were randomly shuffled among all trees (with and without lianas or ferns). The former analyses were carried out combining all plots.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Analysis of lianas and ferns between plots

Across the 25 plots in our five forest fragments, we recorded 889 trees, 882 lianas and 98 epiphytic ferns. Tendril climbers were the most abundant liana climbing guild (40% of all lianas), followed by mainstem twiners (29%), branch twiners (18%) and root climbers and scramblers (13%). Asplenium spp. were the most abundant of the ferns sampled (73%). Comparing edge and interior plots, we found that lianas were twice as abundant in general in edge plots as in interior ones (mean ± SE = 24.5 ± 5.72 and 12.38 ± 3.14, respectively), and that all of the liana climbing guilds were more abundant near forest edges (Table 1). They were also morphologically more diverse on edges than interiors based on the Shannon diversity index (Table 1).

Table 1. Values of abundance for the different liana guilds and Shannon's diversity index for the edge and interior plots
 EdgeInteriorχ2
  1. ***< 0.001, *< 0.05.

Branch Twiners21 ± 9.74.5 ± 3.518.69***
Mainstem Twiners26 ± 4.517.5 ± 0.53.79*
Root Climbers12.33 ± 7.336.5 ± 1.53.95*
Tendril Climbers38.66 ± 19.2221 ± 811.49***
Shannon Diversity0.95 ± 0.090.58 ± 0.125.94*

The abundance of trees per plot was not explained by any of the variables included in our GLMM models (Table 2; see Table S2 for original models and fixed effects included). However, lianas were more frequently present and abundant near forest edges and on larger trees (Table 2). Fern presence and abundance were both higher in plots with less canopy cover and larger trees. Also, ferns declined near forest edges (Table 2).

Table 2. Results for the average of the subset of best performing models selected. Numbers indicate 95% confidence intervals
 Tree AbundanceLiana PresenceLiana LoadFern PresenceFern Abundance
  1. Bold letters indicate significant effects. n.s., non-significant effects. **< 0.001.

Fragment Size−0.003 ± 0.0014n.s.−0.003, 0.02n.s.−0.003, 0.002n.s.−0.01, 0004n.s.−0.01, 0.002n.s.
Distance to Edge−0.004 ± 0.0014n.s.0.34,0.04**0.32,0.06** 0.016, 0.36 ** −0.06, 0.32
Canopy Cover −0.06, 0.02n.s.−0.06, 0.01n.s.0.15,0.05**0.14,0.03**
Log Tree Size 0.18, 0.75 ** 0.27, 0.82 ** 1.24, 1.93 ** 1.57, 2.39 **
Liana Presence   −0.51, 0.33n.s.−0.24, 0.69n.s.
Fragment Size × Distance to Edge −0.006, 0.0008n.s. −0.004, 0.005n.s. 

Distribution of lianas and ferns within plots

Point-pattern analyses revealed that trees were unevenly distributed within plots, occurring at nearest-neighbour distances (r) of roughly 2–3 m (Fig. 2a,b). At local distances of < 0.5 m, trees in interior plots showed stronger clustering than did trees in edge plots.

image

Figure 2. Analyses of the spatial distribution of trees, ferns and lianas. The pair correlation functions g(r) were used to estimate the tree distribution in edge (a) and internal plots (b). The univariate mark-connection functions p11(r) were used to explore the spatial structure of trees occupied by ferns (c) and lianas (d) over all trees. The bivariate mark-connection functions p12(r) was used to test if there was spatial differentiation between host trees with and without ferns (e) or lianas (f). The functions estimated from the observed point pattern (lines with dots) were contrasted to the simulation envelopes (grey polygons) derived from 199 runs of the null model chosen in each analysis (see further details in 'Methods'). In (a) and (b), the expected results of the functions under a random pattern are shown as black lines. For a given analysis, we further computed the goodness of fit (GoF) test for the intervals of distances of observed data departing from simulation envelopes (distances between brackets).

Download figure to PowerPoint

In all plots, host trees with ferns were aggregated within 5–6 m of each other (Fig. 2c). Trees with lianas were spatially segregated beyond 8 m (Fig. 2d), which may be attributed to an uneven distribution of trees within plots. Furthermore, trees with ferns and lianas were clustered and separated from those without ferns and lianas at distances of 2.1 m (Fig. 2e) and 0.6–1.5 m (Fig. 2f), respectively. In other words, host trees with ferns were clustered and separated from those without ferns. However, we found a finer aggregation for host trees with and without lianas than for those with or without ferns (i.e. segregation between trees with and without lianas is higher as it appears at distances under 2 m, coinciding with the distance at which tree aggregation stabilizes; see above).

Additionally, we found a significant lower probability of ferns in the vicinity of trees with lianas, but only at distances beyond 1.2 m in edge plots (Fig. 3a). For interior plots, the observed data did not depart from the null model, except for distances >9 m (Fig. 3b), probably due to an uneven distribution of trees within plots (see Fig. 2d for a similar result). We observed a random pattern of fern abundance on host trees (Fig. 3c, joining all plots). We observed mutual exclusion of trees with high and low loads of lianas at small distances (between 0.6 and 1.5 m), but only for edge plots (Fig. 3d,e); in other words, edge plots have clusters of host trees with high liana loads, and that probably favours high competition with ferns. Finally, we detected a random pattern in the association between the abundance of lianas and ferns (Fig. 3f). That is, competitive interactions are only apparent when analysing the presence of lianas and ferns.

image

Figure 3. Analyses of spatial distributions of presence and abundance of ferns and lianas. Probability pa,2(r) that a fern (subscript 2) was present at distance r from a tree occupied by a liana (subscript a) in border (a) and internal plots (b). The univariate r-mark correlation functions km1 and km2 give the mean fern abundance in all plots (c) and the mean liana abundances in border (d) and internal (e) plots, at distance r from a host tree. (e) The bivariate mark-correlation function km1,m2 tests the mean product of fern abundances dependent on the liana abundance (ind m−2). For further conventions, see Fig. 2.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Our study is the first to suggest that lianas may have an impact on plant species other than trees, namely epiphytic ferns. Previous studies have shown that lianas can negatively affect their tree hosts, especially in disturbed forests where they become hyper-abundant (e.g. Putz 1984; Laurance et al. 2001; Schnitzer & Bongers 2002, 2011; Paul & Yavitt 2011). In disturbed environments, lianas may affect trees negatively by loading heavily on their crowns (Putz & Mooney 1991), competing for light, water (Pérez-Salicrup & Barker 2000; Pérez-Salicrup et al. 2001) and soil nutrients (Schnitzer et al. 2005), reducing tree growth and reproduction (Lowe & Walker 1977; Putz 1984; Schnitzer & Bongers 2002), suppressing tree regeneration in canopy gaps (Schnitzer et al. 2000) and accelerating tree turnover rates (Laurance et al. 2001). Our analysis suggests that following disturbance, trees become an increasingly aggregated resource for climbing and epiphytic plants, grouped in clusters of roughly 2–3 m in diameter, particularly dense in forest interiors. The lower numbers of trees and the large increases in liana abundances in edge environments (twice as abundant as in the interior), together with the probable increase in liana re-sprouting near edges in response to increased light levels, is apparently translated into clumped distributions of lianas and epiphytic ferns. The lower abundances of epiphytic ferns in edge environments may very well reflect a response to abiotic drivers such as humidity, which is lower in these areas. However, the presence of ferns within plots with less canopy cover suggests this is not the only feature determining their distribution, and that a lack of suitable hosts for colonization is also limiting their distribution. Indeed, our finding that in edge plots, epiphytic ferns were less abundant near lianas (within ca. 1 m), suggests the existence of a high interspecific competition between them. Certainly, the prodigious re-sprouting ability of lianas enhances the creation of massive clumps of them around small groups of trees (Benítez-Malvido & Martínez-Ramos 2003). This, coupled with their high specific leaf area, allows them to allocate large amounts of canopy leaves above their hosts, competing aggressively with trees (Schnitzer & Bongers 2011) and leaving relatively little space available for other lianas and epiphytic plants. And such a pattern is supported by our data at distances <2 m. Specifically, we found a clustering of host trees with high liana loads at edge plots, which is probably leading to a high intra- and interspecific competition at local scales. However, although epiphytic ferns were largely excluded from using trees colonized by lianas in our study area, the reverse could actually be true as in certain other systems, where ants inhabiting epiphytic ferns have been found to actively exclude lianas from attaching to their host trees (Tanaka & Itioka 2011; see also Fayle et al. 2011).

Our results suggest that lianas with tendril climbing mechanisms are more abundant than the other guilds. Although tendril climbers are generally thought to require smaller (<10 cm diameter) supports (Putz 1984), many species found in our study area initially use their tendrils to attach to small trees but, when fully grown, drop their tendrils and sometimes twine completely around the branches or trunks of trees. Lianas bearing these kinds of climbing mechanisms are probably the first to colonize small trees, acting then as facilitators for other lianas to further colonize host trees and contributing to the creation of clumps of lianas (Pinard & Putz 1994). Therefore, re-sprouting and potential facilitation processes between types of liana seems to create groups of trees with large amounts of lianas in close vicinity, hindering the colonization of epiphytic ferns, especially so near forest edges.

A key caveat of our study is that we measured morphological plant guilds rather than individual species (see Fayle et al. 2009 for differences in the distribution of species of Asplenium ferns). However, we observed a clear pattern of fern displacement in relation to trees with lianas, despite the large spatial heterogeneity of lianas in edge plots. This strong response (especially apparent when analysing the presence of epiphytic ferns in host trees with lianas) would probably be even stronger if species were analysed separately. Larger sample sizes for both ferns and lianas would be needed to assess species-specific analyses, and such studies with a taxonomic component will help in understanding the mechanisms behind the patterns observed in this paper.

In summary, our results suggest important differences in the composition of epiphytic and climbing plants between the edges and interiors of forest fragments in our study area. Edge ecotones seem to be the domain of lianas with detrimental effects upon epiphytic ferns. The vital importance of ferns in tropical forests in Australia (Cummings et al. 2006) and their apparent decline in fragmented forests might have cascading impacts on animals and plants that depend on them.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

AM was funded by a Basque Government post-doctoral fellowship. JRP was funded by the Spanish Government through the project CGL2011-28430 (MICECO). WFL was supported by an Australian Laureate Fellowship and an Australian Research Council Discovery grant. MC was supported by an Australian Postgraduate Award scholarship. We thank Eduardo Velázquez, Andy Gillison and Mike Lawes and two anonymous reviewers for comments in the early version of the manuscript.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information
  • Benítez-Malvido, J. & Martínez-Ramos, M. 2003. Impact of forest fragmentation on understory plant species richness in Amazonia. Conservation Biology 17: 389400.
  • Benzing, D.H. 1998. Vulnerabilities of tropical forests to climate change: the significance of resident epiphytes. Climatic Change 39: 519540.
  • Benzing, D.H. 2004. Vascular epiphytes. In: Lowman, M.D. & Rinker, H.B. (eds.) Forest canopies, pp. 175211. Elsevier, Sydney, AU.
  • Burnham, K.P. & Anderson, D.R. 2002. Model selection and multi-model inference. A practical information theoretic approach, Springer, New York, NY, US.
  • Centre for Australian National Biodiversity Research. 2010. Australian Tropical Rainforest Plants Edition 6 [online version] Available at http://www.anbg.gov.au/cpbr/cd-keys/rfk/index.html Accessed January 2011.
  • Chave, J., Condit, R., Muller-Landau, H.C., Thomas, S.C., Ashton, P.S., Bunyavejchewin, S., Co, L.L., Dattaraja, H.S., Davies, S.J., (…) & Losos, E.C. 2008. Assessing evidence for a pervasive alteration in tropical tree communities. PLoS Biology 6: 04550462.
  • Cruz-Angon, A. & Greenberg, R. 2005. Are epiphytes important for birds in coffee plantations? An experimental assessment. Journal of Applied Ecology 42: 150159.
  • Cummings, J., Martin, M. & Rogers, A. 2006. Quantifying the abundance of four large epiphytic fern species in remnant complex notophyll vine forest on the Atherton Tableland, north Queensland, Australia. Cunninghamia 9: 521527.
  • Diggle, P.J. 2003. Statistical analysis of spatial point patterns. Edward Arnold, London, UK.
  • Dupouey, J.L., Dambrine, E., Laffite, J.D. & Moares, C. 2002. Irreversible impact of past land use on forest soils and biodiversity. Ecology 83: 29782984.
  • Eacham Historical Society 1995. Malanda: in the shadow of Bartle Frere. Eacham Historical Society, Malanda, AU.
  • Ellwood, M.D.F., Jones, D.F. & Foster, W.A. 2002. Canopy ferns in lowland dipterocarp forest support a prolific abundance of ants, termites, and other invertebrates. Biotropica 34: 575583.
  • Fagan, W.F., Cantrell, R.S. & Cosner, C. 1999. How habitat edges change species interactions. The American Naturalist 153: 165182.
  • Fahrig, L. 2003. Effects of habitat fragmentation on biodiversity. Annual Review of Ecology Evolution and Systematics 34: 487515.
  • Fayle, T.M., Chung, A.Y.C., Dumbrell, A.J., Eggleton, P. & Foster, W.A. 2009. The effect of rain forest canopy architecture on the distribution of epiphytic ferns (Asplenium spp.) in Sabah. Malaysia. Biotropica 41: 676681.
  • Fayle, T.M., Edwards, D.P., Turner, E.C., Dumbrell, A.J., Eggleton, P. & Foster, W.A. 2011. Public goods, public services and by-product mutualism in an ant–fern symbiosis. Oikos 121: 12791286.
  • Foster, J.R., Townsend, P.A. & Zganjar, C.E. 2008. Spatial and temporal patterns of gap dominance by low-canopy lianas detected using EO-1 Hyperion and Landsat Thematic Mapper. Remote Sensing of Environment 112: 21042117.
  • Freeman, A.B., Freeman, A.N.D. & Krockenberger, A. 2005. Predators on the prowl: how does Australia's largest snake survive in a fragmented landscape?. Joint Meeting of Ichthyologists and Herpetologists, Tampa, FL, US.
  • Gentry, A.H. 1991. The distribution and evolution of climbing plants. In: Putz, F.E. & Mooney, H.A. (eds.) The biology of vines, pp. 349. Cambridge University Press, Cambridge, UK.
  • Gentry, A.H. & Dodson, C. 1987. Contribution of non-trees to species richness of a tropical rain forest. Biotropica 19: 149156.
  • Gerwing, J.J., Schnitzer, S.A., Burnham, R.J., Bongers, F., Chave, J., Dewalt, S., Ewango, C.E.N., Foster, R., Kenfack, D., (…) & Thomas, D.W. 2006. A standard protocol for liana censuses. Biotropica 38: 256261.
  • Getzin, S., Wiegand, T., Wiegand, K. & He, F. 2008. Heterogeneity influences spatial patterns and demographics in forest stands. Journal of Ecology 96: 807820.
  • Goosem, S., Morgan, G. & Kemp, J.E. 1999. Wet tropics. In: Sattler, P.S. & Williams, R.D. (eds.) The conservation status of Queensland's Regional Ecosystems. Environmental Protection Agency, Brisbane, AU.
  • Harper, K., Macdonald, S., Burton, P., Chen, J., Brosofske, K. & Saunders, S. 2005. Edge influence on forest structure and composition in fragmented landscapes. Conservation Biology 19: 768782.
  • Hegarty, E.E. & Caballe, G. 1991. Distribution and abundance of vines in forest communities. In: Putz, F.E. & Mooney, H.A. (eds.) The biology of vines, pp. 313335. Cambridge University Press, Cambridge, UK.
  • Illian, J., Penttinen, A., Stoyan, H. & Stoyan, D. 2008. Statistical analysis and modelling of spatial point patterns. Wiley, Chichester, UK.
  • Jules, E.S. 1998. Habitat fragmentation and demographic change for a common plant: Trillium in old-growth forest. Ecology 79: 16451656.
  • Laurance, W.F. 1991. Ecological correlates of extinction proneness in Australian tropical rain forest mammals. Conservation Biology 5: 7989.
  • Laurance, W.F. 1997. Hyper-disturbed parks: edge effects and the ecology of isolated rainforest reserves in tropical Australia. In: Laurance, W.F. & Bierregaard, R.O. Jr (eds.) Tropical forest remnants: ecology, management, and conservation of fragmented communities, pp. 7183. University of Chicago Press, Chicago, IL, US.
  • Laurance, W.F., Delamonica, P., Laurance, S.G., Vasconcelos, H.L. & Lovejoy, T.E. 2000. Rainforest fragmentation kills big trees. Nature 404: 836.
  • Laurance, W.F., Pérez-Salicrup, D., Delamonica, P., Fearnside, P.M., D'Angelo, S., Jerozolinski, A., Pohl, L. & Lovejoy, T. 2001. Rain forest fragmentation and the structure of Amazonian liana communities. Ecology 82: 105116.
  • Laurance, W.F., Lovejoy, T.E., Vasconcelos, H.L., Bruna, E.M., Didham, R.K., Stouffer, P.C., Gascon, C., Bierregaard, R.O., Laurance, S.G. & Sampiao, E. 2002. Ecosystem decay of Amazonian forest fragments: a 22-year investigation. Conservation Biology 16: 605618.
  • Laurance, W.F., Nascimento, H., Laurance, S.G., Andrade, A., Fearnside, P.M. & Ribeiro, J. 2006. Rain forest fragmentation and the proliferation of successional trees. Ecology 87: 469482.
  • Laurance, W.F., Camargo, J., Luizao, R., Laurance, S.G., Pimm, S.L., Bruna, E., Stouffer, P., Williamson, G.B., Benitez-Malvido, J., (…) & Lovejoy, T.E. 2011. The fate of Amazonian forest fragments: a 32-year investigation. Biological Conservation 144: 5667.
  • Laurance, W.F., Useche, D.C., Rendeiro, J., Kalka, M., Bradshaw, C.J.A., Sloan, S.P., Laurance, S.G., Campbell, M., Abernethy, K., (…) & Zamzani, F. 2012. Averting biodiversity collapse in tropical forest protected areas. Nature 489: 290294.
  • Law, R., Illian, J., Burslem, D.F.R.P., Gratzer, G., Gunatilleke, C.V.S. & Gunatilleke, I.A.U.N. 2009. Ecological information from spatial patterns of plants: insights from point process theory. Journal of Ecology 97: 616628.
  • Lindborg, R. & Eriksson, O. 2004. Historical landscape connectivity affects present plant species diversity. Ecology 85: 18401845.
  • Lindenmayer, D.B. & Fischer, J. 2006. Habitat fragmentation and landscape change. Island Press, Washington, DC, US.
  • Loosmore, N.B. & Ford, E.D. 2006. Statistical inference using the G or K point pattern spatial statistics. Ecology 87: 19251931.
  • Lowe, R.G. & Walker, P. 1977. Classification of canopy, stem, crown status and climber infestation in natural tropical forest in Nigeria. Journal of Applied Ecology 14: 897903.
  • Muñoz, A.A., Chacón, P., Pérez, F., Barnert, E.S. & Armesto, J.J. 2003. Diversity and host preferences of vascular epiphytes and vines in a temperate rainforest in southern Chile. Australian Journal of Botany 51: 381391.
  • Murcia, C. 1995. Edge effects in fragmented forests: implications for conservation. Trends in Ecology & Evolution 10: 5862.
  • Oliveira-Filho, A.T., De Mello, J.M. & Scolforo, J.R.S. 1997. Effects of past disturbance and edges on tree community structure and dynamics within a fragment of tropical semideciduous forest in south-eastern Brazil over a five-year period (1987–1992). Plant Ecology 131: 4566.
  • Olson, D.M. & Dinerstein, E. 2002. The global 200: priority ecoregions for global conservation. Annals of the Missouri Botanical Garden 89: 199224.
  • Paul, G.S. & Yavitt, J.B. 2011. Tropical vine growth and the effects on forest succession: a review of the ecology and management of tropical climbing plants. Botanical Review 77: 1130.
  • Pérez-Salicrup, D.R. & Barker, M.G. 2000. Effect of liana cutting on water potential and growth of adult Senna multijuga (Caesalpinioideae) trees in a Bolivian tropical forest. Oecologia 124: 469475.
  • Pérez-Salicrup, D.R., Sork, V.L. & Putz, F.E. 2001. Lianas and trees in a liana forest of Amazonian Bolivia. Biotropica 33: 3447.
  • Phillips, O.L., Vasquez Martinez, R., Arroyo, L., Baker, T.R., Killeen, T., Lewis, S.L., Mahli, Y., Monteagudo Mendoza, A., Neill, D., (…) & Vinceti, B. 2002. Increasing dominance of large lianas in Amazonian forests. Nature 418: 770774.
  • Pinard, M.A. & Putz, F.E. 1994. Vine infestation of large remnant trees in logged forest in Sabah, Malaysia: biomechanical facilitation in vine succession. Journal of Tropical Forest Science 6: 302309.
  • Porensky, L.M. & Young, T.P. 2013. Edge-effect interactions in fragmented and patchy landscapes. Conservation Biology 27: 509519.
  • Putz, F.E. 1984. The natural history of lianas on Barro Colorado Island, Panama. Ecology 65: 17131724.
  • Putz, F.E. & Mooney, H.A. 1991. The biology of vines. Cambridge University Press, Cambridge, UK.
  • Raventós, J., Mujica, E., Wiegand, T. & Bonet, A. 2011. Analyzing the spatial structure of Broughtonia cubensis (Orchidaceae) populations in the dry forests of Guanahacabibes, Cuba. Biotropica 43: 173182.
  • Ries, L., Fletcher, R.J., Battin, J. & Sisk, T.D. 2004. Ecological responses to habitat edges: mechanisms, models, and variability explained. Annual Review of Ecology, Evolution and Systematics 35: 491522.
  • Rutishauser, S.E. 2011. Increasing liana abundance and biomass in tropical forests: testing mechanistic explanations. MSc Thesis, University of Wisconsin - Milwaukee, Milwaukee, WI, US.
  • Schnitzer, S.A. & Bongers, F. 2002. The ecology of lianas and their role in forests. Trends in Ecology & Evolution 17: 223230.
  • Schnitzer, S.A. & Bongers, F. 2011. Increasing liana abundance and biomass in tropical forests: emerging patterns and putative mechanisms. Ecology Letters 14: 397406.
  • Schnitzer, S.A., Dalling, J.W. & Carson, W.P. 2000. The impact of lianas on tree regeneration in tropical forest canopy gaps: evidence for an alternative pathway of gap-phase regeneration. Journal of Ecology 88: 655666.
  • Schnitzer, S.A., Kuzee, M.E. & Bongers, F. 2005. Disentangling above- and below-ground competition between lianas and trees in a tropical forest. Journal of Ecology 93: 11151125.
  • Schnitzer, S.A., Dewalt, S.J. & Chave, J. 2006. Censusing and measuring lianas: a quantitative comparison of the common methods. Biotropica 38: 581591.
  • Schnitzer, S.A., Rutishauser, S. & Aguilar, S. 2008. Supplemental protocol for liana censuses. Forest Ecology and Management 255: 10441049.
  • Stoyan, D. & Stoyan, H. 1994. Fractals, random shapes and point fields: methods of geometrical statistics. Wiley, Chichester, UK.
  • Tabarelli, M., Lopes, A.V. & Peres, C.A. 2008. Edge-effects drive tropical forest fragments towards and early-successional system. Biotropica 40: 657661.
  • Tanaka, H.O. & Itioka, T. 2011. Ants inhabiting myrmecophytic ferns regulate the distribution of lianas on emergent trees in a Bornean tropical rainforest. Biology Letters 7: 706709.
  • Tracey, J.G. 1987. The vegetation of the humid tropical region of north Queensland. CSIRO, Melbourne, AU.
  • Tracey, J.G. & Webb, L.J. 1975. Vegetation of the humid tropical region of north Queensland. CSIRO, Brisbane, AU.
  • Wenzhang, M., Wenyao, L., Lipan, Y. & Guoping, Y. 2008. Edge effects on epiphytes in montane moist evergreen broad-leaved forest. Biodiversity Science 16: 245254.
  • Wiegand, T. & Moloney, K.A. 2004. Rings, circles, and null-models for point pattern analysis in ecology. Oikos 104: 209229.
  • Winter, J.W., Bell, F.C., Pahl, L.I. & Atherton, R.G. 1987. Rainforest clearfelling in Northeastern Australia. Proceedings of the Royal Society of Queensland 98: 4157.
  • Wright, S.J. 2005. Tropical forests in a changing environment. Trends in Ecology & Evolution 20: 553560.
  • Wright, S.J. 2010. The future of tropical forests. Annals of the New York Academy of Sciences 1195: 127.
  • Wright, S.J. & Calderon, O. 2006. Seasonal El Niño and longer term changes in flower and seed production in a moist tropical forest. Ecology Letters 9: 3544.
  • Wright, S.J., Calderón, O., Hernández, A. & Paton, S. 2004. Are lianas increasing in importance in tropical forests? A 17-year record from Panama. Ecology 85: 484489.

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information
FilenameFormatSizeDescription
avsc12104-sup-0001-AppendixS1-S2.docxWord document20K

Appendix S1. Sizes of fragments included in the study.

Appendix S2. List of generalized mixed models fitted for the different response variables.

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.