Current address: Institute of Ecology, Friedrich-Schiller-University, Dornburger Str. 159, 07743 Jena, Germany.
Ecosystem engineering by leaf-cutting ants: nests of Atta cephalotes drastically alter forest structure and microclimate
Article first published online: 15 NOV 2010
© 2010 The Authors. Ecological Entomology © 2010 The Royal Entomological Society
Volume 36, Issue 1, pages 14–24, February 2011
How to Cite
MEYER, S. T., LEAL, I. R., TABARELLI, M. and WIRTH, R. (2011), Ecosystem engineering by leaf-cutting ants: nests of Atta cephalotes drastically alter forest structure and microclimate. Ecological Entomology, 36: 14–24. doi: 10.1111/j.1365-2311.2010.01241.x
- Issue published online: 7 JAN 2011
- Article first published online: 15 NOV 2010
- Accepted 13 September 2010First published online 15 November 2010
- Canopy opening;
- nest effects;
- soil moisture;
- understorey gap
- Top of page
- Materials and methods
- Supporting Information
1. The role played by Atta species as ecosystem engineers remains poorly investigated despite previous evidence that their nests can impact plant assemblages.
2. In a large remnant of Atlantic forest, we compared forest structure at 36 Atta cephalotes nests to control sites and assessed shifts in microclimate along transects from nests up to 24 m into the forest (11 representative colonies).
3. Nests (average size: 55 m2) were virtually free of understorey vegetation with a high proportion of dead stems (up to 70%).
4. Canopy openness above colonies increased by roughly 40% compared with controls (5.3% at colony vs. 3.7% at control sites).
5. At nest centres, about 6% of the total radiation penetrated through the sparse canopy. Light levels declined exponentially, reaching a third (2%) in the unaffected forest understorey.
6. Likewise, maximum soil temperatures and daily amplitudes declined exponentially from 25 to 23 °C and 1.6 to 0.8 °C, respectively. Soil moisture increased significantly along transects, yet the effect was small and no differences were detected for air temperature and humidity.
7. We extrapolated that individual A. cephalotes nests modify the microclimate in an area of almost 200 m2 on average. For the population, this amounts to 6% of the area along forest edges, where colonies are strongly aggregated, compared with only 0.6% in the forest interior.
8. Nests changed microclimate to an extent that has been reported to impact seed germination, plant growth, and survival of seedlings, conclusively demonstrating that leaf-cutting ants act as ecosystem engineers.
- Top of page
- Materials and methods
- Supporting Information
Leaf-cutting ants (LCA) of the genus Atta have been called ecosystem engineers (e.g. Wirth et al., 2003; Rico-Gray & Oliveira, 2007), which is an organism modifying the availability of resources for other organisms by altering the physical environment (Jones et al., 1994). To act as ecosystem engineers would add to the ecological importance of LCA as dominant herbivores in the Neotropics (Wilson, 1986).Yet, environmental variables modified by LCA have hardly been explored in empirical studies. In Neotropical forests, ecosystem engineering by LCA can occur (1) in their foraging area where the removal of up to 15% of the standing leaf crop (Wirth et al., 2003; Urbas et al., 2007) affects the light microenvironment within and below the canopy (Wirth et al., 2003) and (2) at nest sites which are considered disturbances with the potential to influence plant regeneration in tropical systems (Garrettson et al., 1998). LCA nests frequently reach more than 100 m2 in surface area and 6 m or more in depth of subterranean chambers (Hölldobler & Wilson, 1990). Published effects of these conspicuous structures include (2a) soil disturbances (Alvarado et al., 1981; Perfecto & Vander Meer, 1993), (2b) the reduced resistance of soil to penetration (Moutinho et al., 2003), (2c) removal of litter resulting in bare ground (Weber, 1972), and (2d) enrichment of soil nutrients by the year-long harvesting activity (Haines, 1978; Moutinho et al., 2003; Verchot et al., 2003; Sternberg et al., 2007; but see Passlack, 2007).
Moreover, forest stands around LCA nests might be structurally altered. Farji-Brener and Illes (2000) proposed that LCA establish ‘bottom-up’ gaps (i.e. openings in the forest understorey) at nest sites by constantly clearing vegetation growing on or overhanging the immediate nest surface. In addition, first evidence for openings in the canopy above nests (sun flecks on nest surfaces) has been recently published (Corrêa et al., 2010). Despite the above evidence, a detailed comparison of forest structure at nest sites and undisturbed forest is lacking at present, but of importance to understand environmental effects of LCA nests. Regarding those impacts an increased light availability was documented most often by indirect means (Farji-Brener & Illes, 2000; Hull-Sanders & Howard, 2003; Corrêa et al., 2010), but never measured directly in a well replicated study. Furthermore, no investigation of climatic parameters, other than light, that are of importance for plant regeneration (such as temperature and humidity of air and soil) has been carried out. In addition, LCA nests have been viewed largely as punctual disturbances and their potential impact on the nest vicinity has been completely ignored at present, despite the demonstration of a micro-environmental continuum across the physical edge of forest gaps (Popma et al., 1988), with alterations that can penetrate several metres into the forest understorey (Brown, 1993).
Here, we assessed colonies of Atta cephalotes (L.) (Hymenoptera: Formicidae) in order to document nest-related shifts in forest structure and microclimate in a large piece of Atlantic forest in northeast Brazil. The aims of the current study were to (1) give a detailed description of the structure of forest understorey and canopy at nest sites of A. cephalotes compared with undisturbed control sites in the forest in order to (1a) document changes in the forest understorey due the ants' clearing activity and (1b) check for the presence of canopy gaps, (2) assess nature, magnitude, and spatial extent of nest-induced microclimatic alterations along gradients penetrating from the nest into the forest, and (3) estimate the proportion of forest area directly altered by Atta nests within the interior and the edge zone of the forest. Uncovered patterns are examined for their potential impact on plant regeneration with special reference to the role of LCA as an ecosystem engineer in human-modified landscapes.
Materials and methods
- Top of page
- Materials and methods
- Supporting Information
Study site and species
The study was conducted in the Coimbra Forest (9°S, 35° 52′W), a privately owned fragment of the Atlantic forest in northeast Brazil. Coimbra forest is 3500 ha in size and consists of a largely well-conserved patch of lowland forest (300–400 m above sea level) that is completely surrounded by a homogeneous matrix of sugar cane fields along its 40 km of relatively old (at least 60 years) and stable edges (Santos et al., 2008). Such a human-modified landscape is typical across the Atlantic forest region and exhibits a tropical climate characterised by a 5-month dry season (<110 mm per month) lasting from September to January (annual precipitation about 2000 mm); the prevailing soils are latosols and podzols (IBGE, 1985). The edge zone (0–100 m into the forest) is largely dominated by pioneer plant species (Oliveira et al., 2004), which represent the preferred food source of LCA (Farji-Brener, 2001).
This study analyses forest structure and resulting microclimate at nests of Atta cephalotes, one of the two co-occurring Atta species in NE Brazil (Correa et al., 2005) that is widespread throughout the Neotropics from Mexico to Bolivia (Kempf, 1972). Colonies of A. cephalotes are long lived (life expectancy of 8 years; Meyer et al., 2009) and construct nests that are characterised by a large central nest mound sometimes accompanied by smaller lateral excavations. Nests of A. cephalotes are about an order of magnitude more frequent in the first 50 m of Coimbra forest than in the forest interior (2.8 vs. 0.3 colonies ha−1; Wirth et al., 2007). Thus, the Coimbra forest offered an interesting opportunity to document nest-related environmental changes in the predominant habitat of tropical human-modified landscapes, such as edges of a heavily fragmented forest.
Forest structure at nest sites
Forest structure at ant nests was examined across a set of 36 representative colonies of A. cephalotes, which were all the colonies encountered in a systematic trail-based census along 28 km of small footpaths (Wirth et al., 2007), the centre of the trail lay within the forest. Colonies occurred from the edge of Coimbra forest up to 800 m into forest, but were strongly aggregated at the edges (Wirth et al., 2007). Ant nest size was conventionally estimated as the ground surface area using ellipsoids that included all soil excavations carried out by the colony (Hernández et al., 1999; Wirth et al., 2003). Likewise, the bare area of understorey vegetation above nests (i.e. the forest understorey gaps) was estimated as an ellipsoid. Forest structure was examined across the 36 ant nests and 20 understorey control plots (5 × 5 m in size), which were haphazardly located within the edge zone (first 50 m) of Coimbra and showed no ant influence or disturbance, for example tree-fall gaps. To describe forest structure we studied understorey and canopy. Stem density of all stems >0.5 cm basal diameter was estimated, assigned into four size categories (0.5–2, 2–5, 5–10, >10 cm), ranging from small saplings (0.5–2 cm) to adult trees (>10 cm); stems were also recorded to be either living or dead to infer the ant-related mechanisms driving changes in forest structure. As measures of canopy structure we adopted canopy gap presence, gap size distribution (canopy gaps were assigned into three categories; small: 1–5, medium: 6–10 or big: >10 sky regions) and canopy openness estimated from fisheye photographs (for details on the canopy structure see Appendix S1A). Additionally, forest structure was indirectly characterised via measures of light availability (direct, diffuse, and total light transmission) also obtained by fisheye photographs (Appendix S1A). Measures like these are better gap size estimates than the physical gap size per se (Whitmore et al., 1993), because plants respond to microclimatic differences and physical measures of gap size (e.g. drip line projection sensuBrokaw, 1982) do not include small canopy holes, which are important sources of radiation in small gaps and under closed canopy (Mitchell & Whitmore, 1993). All the structural variables measured have been widely adopted to examine the role played by ant colonies, and other disturbances, as drivers of both forest microclimate and vegetation dynamics in tropical forests (see e.g. Denslow, 1980; Brown, 1993; Hull-Sanders & Howard, 2003; Corrêa et al., 2010).
Changes in microclimate
To characterise microclimatic alterations on and around nest sites, 11 adult colonies of A. cephalotes were selected that were located 13–295 m away from the nearest forest border and surrounded by homogeneous forest without openings other than the colony (e.g. no tree-falls or other gaps) as we aimed to document those changes in microclimate exclusively induced by ant nests and their associated vegetation gaps. These 11 colonies (four of which were part of the previous set of 36) differed neither in nest size (t = 1.34, d.f. = 14.7, P = 0.199), nor (analysed from fisheye photographs) in canopy openness (t = 0.98, d.f. = 17.7, P = 0.341), or transmittance of light (e.g. total transmittance: t = 1.33, d.f. = 15.9, P = 0.201) from the set of 36 colonies recorded in the representative census of Coimbra forest. Measurements of microclimatic changes at nest sites focused on three plant-relevant variables: (1) light (relative light interception and daily sum of photosynthetically active radiation), (2) soil and air water content, and (3) temperature, as detailed in Appendix S1B. Changes of soil parameters along the transect (nutrient availability, content of organic material) have been documented in a complementary study (Passlack, 2007) and will be presented elsewhere. Together these variables have been proposed to affect seed germination and seedling survivorship and growth (e.g. Everham et al., 1996; Kobe, 1999; Godoi & Takaki, 2004; de Gouvenain et al., 2007), thus, by altering them via nest-driven changes in forest structure, ants are likely to affect resource availability for plants. To account for the gradual change in microclimatic conditions from nests towards surrounding forest understorey and to determine how far nest-driven changes reach into the forest understorey (hereafter penetration distance), all variables were measured simultaneously at 3 m intervals along transects from the edge of the nest up to 24 m into the forest, with an additional measuring point at nest centres (i.e. 10 transect points). This was because we have previously observed that nest-related changes on forest structure do not penetrate farther than the vicinity of nest mounds (Corrêa et al., 2010). To control for confounding effects in case of nearby linear disturbances (forest edge, stream, valley), transects were run parallel to these structures.
Parameters describing forest structure and light climate, derived from fisheye photographs at colonies and control plots, were square-root transformed prior to comparison in t-tests in order to stabilise variance, improve data normality and consequently increase the explanatory power of statistical models. Back-transformed 95% confidence intervals are presented along with the mean of the factors. Parameters of understorey and canopy gaps were analysed in detail by comparing colony and control site values of stem densities, the proportion of dead stems, number of gaps, and gap area using U-tests.
Changes of microclimatic parameters with distance to Atta colonies were analysed by building non-linear mixed effect (NLME) models as described by Pinheiro and Bates (2002). Mixed effect models have been frequently used to describe continuous data nested within individual samples that were randomly drawn from a population (e.g. Barrowman et al., 2003; Gillies et al., 2006). Using this approach, we can calculate a model representing the average of colonies in the population (fixed effects) and at the same time estimate parameters describing deviations of measurements at individual colonies from the mean microclimatic model (random effects). Models allowed for exponential and/or linear changes along transects (for details see Appendix S1D), such as those expected from microclimatic variables in response to the occurrence of forest canopy gaps (e.g. Brown, 1993). Moreover, random effects permitted the incorporation of variation due to different locations in the forest and macroclimatic conditions resulting from different measuring days, since obvious limitations in the amount of measuring equipment did not allow us to measure colonies simultaneously. Measuring days included various weather conditions typical for the dry season in Serra Grande (hot sunny days, different degrees of cloud cover and even some rain showers) and were spread over 2 months during the peak of the dry season 2005/2006.
Statistics were computed in R (version 2.6.0, R Development Core Team, 2007). Mixed effect models were built using the nlme library (version 3.1-85, Pinheiro et al., 2007). U-tests for the comparison of understorey and canopy gap characteristics were performed using the exact-Rank Tests library (version 0.8-16, Hothorn & Hornik, 2006).
Estimating impacted area
Light availability was chosen as the parameter to analyse the area of microclimatic nest alterations (i.e. the nest spatial impact), because relative irradiance is generally low in the understorey of tropical forests and therefore believed to be a limiting factor for plant growth (Chazdon, 1988). Moreover, it is directly changed by alterations in forest structure causing further microclimatic impacts (Whitmore et al., 1993). To calculate the area impacted per colony the continuous relationship between microclimatic alterations and distance from nests had to be converted into a radius up to which Atta nests alter relative irradiance. We based radius estimates on a conservative threshold value of 0.5% additional relative irradiance compared with the undisturbed understorey. Even smaller difference in light might induce biological effects, as for example, Montgomery and Chazdon (2002) showed plant growth and survival of different species to increase linearly with light availability in low-light environments between 0.2% and 6% relative irradiance. The distance from the nest edge up to which alterations exceeding the threshold occurred was calculated based on the model for relative irradiance (formula in Table 2). Adding this penetration distance to the radius of the average understorey gap allowed us to calculate an ellipsoid representing the area impacted by an average colony.
|Parameter||Model||Fixed effects (standard error)||Necessary random effects as SD||Residuals as SD|
|n (nest edge)||s1 (exponential)||s2 (linear)||f (forest)||n||S1||S2||f|
|Relative irradiance (RI)||y = n es1.x + f||1.33* (0.297)||−0.253 (0.042)||1.90 (0.312)||0.684||—||0.968||0.383|
|Sum of radiation (daily PPFD)||In(y) = n es1.x + f||0.800* (0.190)||−0.144 (0.036)||−0.596 (0.182)||0.040||—||0.443||0.539|
|Soil temperature maximum||y = n es1.x + f||0.668* (0.176)||−0.238 (0.017)||23.3 (0.128)||0.550||—||0.410||0.213|
|Soil temperature minimum||y = n es1.x + f||0.106* (0.055)||−0.243 (0.084)||22.5 (0.124)||—||—||0.397||0.206|
|Soil temperature amplitude||y = n es1.x + f||0.555* (0.095)||−0.228 (0.032)||0.835 (0.085)||0.252||—||0.265||0.212|
|Soil temperature mean||y = n es1.x + f||0.296* (0.069)||−0.252 (0.027)||22.9 (0.118)||0.193||—||0.384||0.140|
|Soil water content||y = s2 x + n||15.6(2.25)||—||0.111 (0.048)||—||5.36||—||0.102||—||1.60|
|Soil matric potential||y = s2 x + n||−1.23 (0.418)||—||0.040 (0.016)||—||0.968||—||0.036||—||0.179|
|Air temperature maximum||y = n es1.x + f||0.247* (0.102)||−0.182 (0.039)||28.0 (0.484)||0.285||—||1.58||0.235|
|Air temperature minimum||y = f||—||—||21.2 (0.261)||—||—||0.861||0.081|
|Air temperature amplitude||y = n es1.x + f||0.268* (0.103)||−0.197 (0.036)||6.78 (0.500)||0.290||—||1.632||0.230|
|Air temperature mean||y = f||—||—||24.0 (0.077)||—||—||—||0.812|
|Air humidity maximum||y = f||—||—||95.4 (0.282)||—||—||0.568||0.825|
|Air humidity minimum||y = f||—||—||55.3 (0.798)||—||—||—||5.65|
|Air humidity amplitude||y = f||—||—||40.0 (0.775)||—||—||—||5.48|
|Air humidity mean||y = f||—||—||81.9 (0.348)||—||—||—||2.46|
In order to compare the forest area influenced by A. cephalotes nests between the interior and edge of Coimbra on stand level, the impacted area per colony was multiplied with available colony densities for the two forest zones at the study site (Wirth et al., 2007). This extrapolation assumes per colony impacts do not differ between habitats. We checked this assumption by testing (1) the independence of colony size from distance to the forest edge in a Pearson correlation and (2) evaluating a possible effect of forest zone on microclimate by checking the explanatory power of the model for relative irradiance when including habitat as a covariate [edge zone: six colonies (13–66 m from edge) and forest interior: five colonies (138–295 m)].
- Top of page
- Materials and methods
- Supporting Information
Forest structure at nest sites
Nests of A. cephalotes represented major disturbances in the forest understorey of Coimbra. The average size of 36 representative nests (main mound and lateral excavations) was 55 m2 (95% CI: 22–136) and showed no correlation with distance to the forest edge (Pearson correlation r = 0.134, d.f. = 34, P = 0.438); therefore, no confounding effects of edge distance on nest size are to be expected in our study. Nests were associated with zones of strongly reduced understorey vegetation concentric with the main mound of the colony. These understorey gaps encompassed on average 48 m2 (95% CI: 18–130) and their size correlated to a high degree with colony size (Person correlation r = 0.637, d.f. = 34, P < 0.001). Densities of stems up to diameters of 10 cm were reduced in the nest areas compared with control sites in the forest (Fig. 1a). Stem density in the smallest diameter class (0.5–2 cm) decreased from about 2 m−2 to less than 1 m−2. In addition, about 70% of the stems on nests in this size class were dead, compared with less than 10% at control sites. Generally, nest sites had a much higher proportion of dead stems with highly significant differences for diameters up to 5 cm (Fig. 1b). Nest sites were de facto free of small living vegetation and contrasted strongly with the surrounding forest, creating an understorey gap that was well demarcated from the adjacent understorey by a rather distinct edge.
Analysis of fisheye photographs taken from colony centres confirmed our visual observations that alterations of forest structure were not restricted to the understorey, but extended into the canopy (Table 1). Canopy openness measured above colonies increased by ca 40% compared with control sites (5.3% at colony vs. 3.7% at control sites) allowing about 1.5 times as much light to penetrate the canopy (total transmittance of 7.9% vs. 5.4%). In the canopy above nests, the number of small, medium, and large gaps was markedly higher compared with control sites (Fig. 2a). While the absolute difference in gap numbers between colonies and controls was greatest for small gaps, this resulted in only a relatively small, yet highly significant, difference in total gap area (Fig. 2b). Big gaps, although not as frequent as smaller ones, caused a much stronger increase in total gap area at nest sites, highlighting their importance for the increased canopy openness above nests. On average, the largest opening in the canopy at nest sites encompassed about 7 sky regions (roughly 12 m2) – almost three times the size measured at control sites (2.5 sky regions; Table 1). In addition, the right side of the 95% confidence interval for gap sizes above colonies extended up to 75 sky regions (roughly 135 m2), while 95% of all control sites had a maximum opening in the canopy equal to or less than 10 sky regions (Table 1).
|Canopy openness (%)||5.3 (2.4–11.7)||3.7 (2.4–5.7)||4.3||54||<0.001|
|Direct transmittance (%)||8.5 (3.2–22.8)||5.9 (3.4–10.3)||3.5||54||<0.001|
|Diffuse transmittance (%)||7.2 (3.4–15.3)||4.9 (3.3–7.3)||5.0||54||<0.001|
|Total transmittance (%)||7.9 (3.5–17.9)||5.4 (3.6–8.1)||4.5||54||<0.001|
|Largest single opening (sky regions)||6.7 (0.6–75.1)||2.5 (0.6–10.0)||3.8||54||<0.001|
Alterations in forest structure at nest sites of A. cephalotes caused changes in forest microclimate that were measured in linear transects extending from the edge of the understorey gap 24 m into the forest with an additional measurement point at the centre of nests. Nest centres received about three times as much light as the forest understorey with light levels declining exponentially with distance from nests (Fig. 3a,b, Table 2). This pattern was observed for relative irradiance (6 vs. 2% relative irradiance) as well as for the daily sum of photosynthetically active radiation (2.6 vs. 0.6 mol m−2 day−1 PPFD). Neither at nest sites nor in the surrounding forest was relative irradiance statistically different between colonies from forest edge or interior (including habitat as a covariate into the model for relative irradiance did not increase the explanatory power of the model), yet, there was considerable variation between the light levels at individual colonies (random effects; Table 2). Paralleling light levels reaching the forest floor, maximum soil temperatures declined exponentially along transects from 25 to 23 °C (Fig. 3d). Minimum soil temperatures were virtually constant at 22.5 °C (Fig. 3d). Thus, the course of daily amplitude and mean soil temperature along transects closely resembled maximum soil temperatures (Table 2). Mean soil temperature was 1 °C higher with daily temperature amplitudes twice as large in nest centres compared with the forest understorey (1.6 vs. 0.8 °C). Gravimetric soil water content significantly increased with distance to the nest (Table 2), but the effect was small (increase from 15% at nest centres to 18% at 24 m from the nest edge; Fig. 3f). Stronger effects became apparent when soil water content was converted into soil matric potential to express plant water availability (see Appendix S1C for conversion). Matric potential increased linearly from less than −1 MPa at the nest centres to values close to 0 MPa at the end of transects (Fig. 3f), with considerable variation in slope and offset of the curves among colonies (random effects; Table 2). Air temperature (Fig. 3c) and air humidity (Fig. 3e) changed very little or not at all along transects compared to diurnal fluctuations visualised as minimum (21 °C; 55% RH) and maximum values (28 °C; 95% RH) in the graphs. Also, variation between colonies was high (Table 2).
Microclimatic alterations were not restricted to the nest site itself, but penetrated beyond the nest edge into the surrounding forest (Fig. 3). A difference of 0.5% in light availability (relative irradiance) was present up to about 4 m from the edge of the understorey gap around colonies. Therefore, on average, an area of 195 m2 (95% CI: 120–318) was influenced by every colony, which is roughly four times the area of the average understorey gap 48 m2 (95% CI: 18–130). Based on the densities of A. cephalotes in the interior and edge of the forest (0.3 and 3 colonies ha−1, respectively) the proportion of the forest area directly impacted by the presence of A. cephalotes nests thus increased from 0.6% in the forest interior to 6% at the forest edge.
- Top of page
- Materials and methods
- Supporting Information
The present study demonstrated that nest sites of A. cephalotes are associated with a striking reduction of understorey vegetation and with canopy gaps above nests. Not only was an elevated frequency of small and medium-sized canopy openings observed at nest sites, but large gaps as well, which were completely absent from the undisturbed canopy of control plots. Such marked changes in forest structure (canopy openness at nest sites was on average 40% higher) allowed for a three times higher light interception with increased soil temperatures and reduced water availability. Microclimatic alterations substantially penetrated into the adjacent forest understorey causing consistent patterns of variation relative to distance from nest centres, despite idiosyncrasies of colonies, individual gaps, the surrounding local forest stand, and the macroclimatic conditions on different measuring days. Finally, nest-driven changes can affect a considerable portion of the forest as every colony alters on average 200 m2 of forest habitat.
Previous studies had documented understorey gaps associated with nests of some Atta species (Garrettson et al., 1998; Hull-Sanders & Howard, 2003) and characterised them as ‘bottom-up gaps' due to a pronounced reduction in stem density in the forest understorey and a lack of canopy openings (Farji-Brener & Illes, 2000). For the first time, we documented a strongly elevated proportion of dead stems at nest sites compared with undisturbed forest, hinting at an elevated mortality of small plants (≤5 cm basal diameter) as an underlying mechanism for this reduction. Broadening the concept of ‘bottom up-gaps' at nest sites, our study demonstrates that nest-driven changes in forest structure extend to the canopy causing a 40% higher canopy openness above nests (see also Corrêa et al., 2010). Thus, contrary to the prior perception, gaps at LCA nests are not restricted to the forest understorey. Also, we are the first to document that resulting microclimatic alterations exceed light climate and include pronounced changes in soil temperature and impacts on water availability. Microclimatic impacts substantially penetrated into the adjacent forest understorey, thereby creating an ant-impacted area rather than a punctual disturbance.
We propose that different-sized gaps above colonies are created by different mechanisms that should be, at last briefly, presented here to drive further investigations. A high number of small and medium-sized openings in the canopy is typical for the patchy herbivory by LCA (Wirth et al., 2003) and may be accounted for by a diffusely elevated cutting probability at nest sites rather than a specialised behaviour, like nest clearing on the ground. On the other hand, large and very large gaps in the canopy at nest sites (which were the main source of elevated canopy openness) suggest additional mechanisms. We observed frequent tree falls at nest sites during 8 years of research on LCA in the study area (including year-long observations at several colonies) that seem to exceed the rate of tree falls in the surrounding forest. Elevated mortality for large trees might, as for seedlings and saplings, be caused by activities of the ants, since (1) large trees on nests can be repeatedly defoliated by the ants, which ultimately may lead to a depletion of stored reserves, (2) their roots can be cut in the area of the colony which might promote pathogen infections in addition to immediate effects of root loss, and (3) trees might lose their support when the soil at nest sites is destabilised by excavation activities of the colony. Data on forest structure at nest sites together with the presented putative mechanisms indicate that colonies of LCA maintain and even create (augment) gaps at their nest site. While we cannot completely exclude the possibility of an association of LCA with forest gaps purely due to site preferences of founding queens (documented by Vasconcelos, 1990), the temporal scales of the processes involved give a strong indication for a causal role played by LCA activities. Tree fall gaps, which are the main disturbance causing canopy openings that can serve as colony founding site, have been shown to be closed by regenerating vegetation in about 2 years (Fetcher et al., 1985). LCA colonies, in contrast, maintain gaps for about 8 years, which is the average life expectancy of A. cephalotes colonies in our study area (Meyer et al., 2009). Therefore, forest gaps at nest sites are maintained much longer than these gaps are expected to persist without the ants' activities.
Irrespective of the mechanisms leading to the emergence of nest canopy gaps, they had a large impact on forest microclimate. Microclimatic parameters responded consistently to the distance from colonies, despite considerable variation among and within colonies probably attributable to idiosyncrasies of individual gaps, the surrounding local forest stand, or the macroclimatic conditions on the measuring day. Nest centres received about three times as much light as the surrounding forest and maximum soil temperature closely followed light interception along transects as is typical for rain forest gaps (Whitmore et al., 1993). Since minimum soil temperatures were almost constant, nest centres experienced on average higher daily amplitudes and moderately higher mean temperatures. Vegetation temperatures, although not measured, are expected to follow the same pattern because, like soil, vegetation absorbs more energy with increasing light irradiance. Higher temperatures of soil and vegetation augment evaporation and transpiration, respectively. In combination, these effects promote water loss from the soil and might explain the observed lower water content and water availability at nests. Similarly, lower soil water contents were documented for nests of A. sexdens in the Brazilian Amazon by Moutinho et al. (2003), who observed higher root densities in nest soil and proposed increased water uptake by trees as a mechanism for reduced water content. Above the soil surface, no or only weak spatial patterns were observed in air temperature or humidity, suggesting sufficient air turbulence to ameliorate differences on these small spatial scales.
It is already clear that ant nests drive changes in both forest structure and microclimate at nest sites, but is that an indication that ants are acting as ecosystems engineers by altering patterns of resource availability for plants? To answer this question we must examine the potential of these changes to impact plant performance and their spatial and temporal extent as follows. Soil temperature and especially temperature amplitudes, in the range resulting from LCA nests, have been demonstrated to be of importance for the induction of germination and resulting germination rates (Everham et al., 1996; Pearson et al., 2002; Godoi & Takaki, 2004). For example, an increase in soil temperature of as little as 0.5 °C has been shown to significantly impact germination rates of tropical trees (Everham et al., 1996), which is about half of the average increase of mean soil temperatures at nest centres compared with the forest understorey measured in our study. Also, drier soil at nest sites can delay or inhibit germination of seeds and, in fact, germination rates of Chrysophyllum viride seeds have been documented to be markedly reduced on and close to nests of A. cephalotes (Corrêa et al., 2010). Following germination, tree species-specific differences in establishment of seedlings, juvenile growth, and survivorship strongly influence forest dynamics, successional trajectory, and species composition and diversity (Kobe, 1999). Light is believed to be the principal limiting factor for seedling growth in the shaded understorey of tropical forests (Denslow, 1980; Chazdon, 1988) since only 1–2% of the photosynthetically active radiation hitting the canopy reaches the forest floor (Chazdon, 1988; 1.9% in our study area). Under such low light conditions all plant species tend to increase growth rates as light increases, although slope and shape of this relationship may be strongly species-specific (Agyeman et al., 1999; Poorter, 1999). Along the Atta nest forest understorey transect we documented a light gradient from 6% to 2% relative irradiance, for which a simple model of potential biomass gain predicts strong effects on seedling growth rates including rank changes between species, comparing nest area and forest understorey (Appendix S2). Generally, the model indicates that potential growth rates are highest on the nest surface and in fact, in a follow-up study, transplanted seedlings of six Atlantic forest species grew more new leaves on nests (Meyer et al., 2010). However, during the lifetime of the colony, seedlings will not realise potential growth rates on the nests due to the zealous nest-clearing behaviour of the ants (Haines, 1975; Farji-Brener & Silva, 1995; Garrettson et al., 1998; Meyer et al., 2010). Yet, microclimatic nest effects and predicted increased growth rates penetrated several metres into the forest understorey, while the high cutting pressure for seedlings ended at the nest edge (Garrettson et al., 1998; Meyer et al., 2010). Thus, we have demonstrated for the first time that the vicinity of living LCA nests has the potential to function as a regeneration site. On a spatial scale, the area impacted by a single colony is augmented therefore by no less than a factor of four considering the near nest zone with higher light availability. Concluding, there can be no doubt that microclimatic alterations at nest sites have potential impacts on plant performance and LCA thus act as ecosystem engineers.
In contrast to other prominent ecosystem engineers that have been substantially decimated by human activities (e.g. beavers and elephants: Syphard & Garcia, 2001; Dunham, 2008; Hood & Bayley, 2008), some species of LCA profit from anthropogenic disturbances and increase in abundance with increasing agricultural land-use, deforestation, and landscape alterations (Jonkman, 1979; Fowler et al., 1986; Jaffe & Vilela, 1989; Wirth et al., 2007). In such habitats as forest edges, the high number of LCA colonies is likely to locally increase LCA impact. At the edge of Coimbra, areas 10 times as big as in the interior receive higher light levels due to the ecosystem engineering of A. cephalotes, although per colony impact did not differ between the forest edge and the forest interior. This physical disturbance caused by nests of hyper-abundant (Wirth et al., 2007) and persisting (Meyer et al., 2009) populations of A. cephalotes complement their dramatic trophic effects in the edge zone of Coimbra, where A. cephalotes exerts an exceptional herbivory pressure removing 36% of the available foliage compared with only one-sixth (6%) in the forest interior (Urbas et al., 2007; Meyer et al., 2010). Both trophic and physical LCA effects are likely among the contributing factors that enable cycles of pioneer self-replacement at forest edges (Tabarelli et al., 2008). These cycles and the concomitant domination of closed edges in old fragmented landscapes by pioneer vegetation have been forecast to be the future scenario for tropical forests in general (Tabarelli et al., 2008). Thus, LCA are among the organisms profiting from anthropogenic forest fragmentation that, by their own activities, help to maintain these newly created habitats (Meyer et al., 2010; Wirth et al., 2008).
In synthesis, we demonstrate that living LCA nests drastically alter forest structure by opening understorey, but also forest canopy, causing a higher light irradiance on the nest surface and in its vicinity in addition to higher amplitudes in soil temperatures and reduced water availability. The magnitude exhibited by nest-driven microclimatic changes varied in a range that has been reported to impact seed germination, growth rates and survival of seedlings, conclusively demonstrating that LCA act as ecosystem engineers. The extensive spatial impacts in habitats where these ants are hyper-abundant, such as the edges of the Coimbra forest, highlights the need to understand the impact of ecosystem engineering by LCA in both pristine and fragmented forests.
- Top of page
- Materials and methods
- Supporting Information
We thank Gordon Frazer for helpful discussions on the analysis of hemispherical photographs, Markus Eichhorn for counsel on the use of NLME models and Flavio Roces for comments on the manuscript. Manoel Vieira Araújo Jr, Christoph Dohm and Simone Jürgens helped during data collection in the field. We also thank four anonymous referees whose comments helped to improve the manuscript. The study was supported by a Brazil–Germany collaboration project (PROBRAL CAPES/DAAD, projects D/06/33907 and 257/07), the ‘Deutsche Forschungsgemeinschaft' (DFG, process WI 1959/1-1, 1-2), a grant from the Schimper foundation, and ‘Conselho Nacional de Desenvolvimento Científico e Tecnológico’ (CNPq, processes 540322/01-6). We are grateful to Burkhard Büdel (University of Kaiserslautern) for the PhD position for S. Meyer, to the DAAD for a travel grant to Simone Jürgens, and to the ‘Usina Serra Grande’, the ‘Centro de Pesquisas Ambientais do Nordeste’ (CEPAN) and ‘Conservação Internacional do Brasil’ (CI-Brasil) for logistical support.
- Top of page
- Materials and methods
- Supporting Information
- 1999) Responses of tropical forest tree seedlings to irradiance and the derivation of a light response index. Journal of Ecology, 87, 815–827. , & (
- 1981) Leaf-cutter ant (Atta cephalotes) influence on the morphology of andepts in Costa Rica. Soil Science Society of America Journal, 45, 790–794. , & (
- 2003) The variability among populations of Coho salmon in the maximum reproductive rate and depensation. Ecological Applications, 13, 784–793. , , , & (
- 1982) The definition of treefall gap and its effect on measures of forest dynamics. Biotropica, 14, 158–160. (
- 1993) The implications of climate and gap microclimate for seedling growth-conditions in a Bornean lowland rain-forest. Journal of Tropical Ecology, 9, 153–168. (
- 1988) Sunflecks and their importance to forest understorey plants. Advances in Ecological Research, 18, 1–63. (
- 34, 695–698. , , & (2005) Occurence of Atta cephalotes (L.) (Hymenoptera: Formicidae) in Alagoas, Northeastern Brazil. Neotropical Entomology
- 2010) How leaf-cutting ants impact forests: drastic nest effects on light environment and plant assemblages. Oecologia, 162, 103–115. , , , & (
- 1980) Gap partitioning among tropical rainforest trees. Biotropica, 12, 47–55. (
- 2008) Detection of anthropogenic mortality in elephant Loxodonta africana populations: a long-term case study from the Sebungwe region of Zimbabwe. Oryx, 42, 36–48. (
- 1996) Effects of light, moisture, temperature, and litter on the regeneration of five tree species in the tropical montane wet forest of Puerto Rico. American Journal of Botany, 83, 1063–1068. , & (
- 2001) Why are leaf-cutting ants more common in early secondary forests than in old-growth tropical forests? An evaluation of the palatable forage hypothesis. Oikos, 92, 169–177. (
- 2000) Do leaf-cutting ant nests make ‘bottom-up’ gaps in neotropical rain forests?: a critical review of the evidence. Ecology Letters, 3, 219–227. & (
- 1995) Leaf-cutting ants and forest groves in a tropical parkland savanna of Venezuela: facilitated succession? Journal of Tropical Ecology, 11, 651–669. & (
- 1985) Vegetation effects on microclimate in lowland tropical forest in Costa-Rica. International Journal of Biometeorology, 29, 145–155. , & (
- 1986) Population dynamics of leaf cutting ants: a brief review. Fire Ants and Leaf-Cutting Ants (ed. by C. S.Lofgren and R. K. V.Meer), pp. 123–145. Westview Press, Boulder, Colorado. , , & (
- 1998) Diversity and abundance of understorey plants on active and abandoned nests of leaf-cutting ants (Atta cephalotes) in a Costa Rican rain forest. Journal of Tropical Ecology, 14, 17–26. , , , , & (
- 2006) Application of random effects to the study of resource selection by animals. Journal of Animal Ecology, 75, 887–898. , , , , , et al. (
- 2004) Effects of light and temperature on seed germination in Cecropia hololeuca Miq. (Cecropiaceae). Brazilian Archives of Biology and Technology, 47, 185–191. & (
- 2007) Partitioning of understorey light and dry-season soil moisture gradients among seedlings of four rain-forest tree species in Madagascar. Journal of Tropical Ecology, 23, 569–579. , & (
- 1975) Impact of leaf-cutting ants on vegetation development at Barro Colorado Island. Tropical Ecological Systems (ed. by F. G.Golley and E.Medina), pp. 99–111. Springer, New York, New York. (
- 1978) Element and energy flows through colonies of the leaf-cutting ant, Atta colombica in Panama. Biotropica, 10, 270–277. (
- 1999) Growth of Atta laevigata (Hymenoptera: Formicidae) nests in pine plantations. Florida Entomologist, 82, 97–103. , , & (
- 1990) The Ants. Springer, Berlin, Germany. & (
- 2008) Beaver (Castor canadensis) mitigate the effects of climate on the area of open water in boreal wetlands in western Canada. Biological Conservation, 141, 556–567. & (
- 2006) exactRankTests: exact distributions for rank and permutationtests [WWW document]. URL http://cran.r-project.org. & (
- 2003) Impact of Atta colombica colonies on understory vegetation and light availability in a neotropical forest. Biotropica, 35, 441–445. & (
- IBGE (1985) Atlas Nacional do Brasil: Região Nordeste. IBGE (Instituto Brasileiro de Geografia e Estatística), Rio de Janeiro, Brazil.
- 1989) On nest densities of the leaf-cutting ant Atta cephalotes in tropical primary rainforest. Biotropica, 21, 234–236. & (
- 1994) Organisms as ecosystem engineers. Oikos, 69, 373–386. , & (
- 1979) Population dynamics of leaf-cutting ant nests in a Paraguayan pasture. Journal of Applied Entomology, 87, 281–293. (
- 1972) Catálogo abreviado das formigas da Região Neotropical (Hymenoptera: Formicidae). Studia Entomologica, 15, 3–344. (
- 1999) Light gradient partitioning among tropical tree species through differential seedling mortality and growth. Ecology, 80, 187–201. (
- 2010). Performance and fate of tree seedlings on and around nests of Atta cephalotes: ecological filters in a fragmented forest. Austral ecology, in press. , & (
- 2009) Persisting hyper-abundance of leaf-cutting ants (Atta spp.) at the edge of an old Atlantic Forest fragment. Biotropica, 41, 711–716. , & (
- 1993) Use of hemispherical photographs in forest ecology. Oxford Forestry Institute Occasional Papers 44. & (
- 2002) Light gradient partitioning by tropical tree seedlings in the absence of canopy gaps. Oecologia, 131, 165–174. & (
- 2003) Influence of leaf-cutting ant nests on secondary forest growth and soil properties in Amazonia. Ecology, 84, 1265–1276. , & (
- 2004) Forest edge in the Brazilian Atlantic forest: drastic changes in tree species assemblages. Oryx, 38, 389–394. , & (
- 2007) Effekte von Blattschneiderameisen auf die Regeneration fragmentierter Regenwälder– Einfluss von Nestbau auf Bodeneigenschaften und Nahrungspräferenz für trockengestresste Pflanzen. Diploma thesis, University of Kaiserslautern, Kaiserslautern, Germany. (
- 2002) Germination ecology of neotropical pioneers: interacting effects of environmental conditions and seed size. Ecology, 83, 2798–2807. , , & (
- 1993) Distribution and turnover rate of a population of Atta cephalotes in a tropical rain forest in Costa Rica. Biotropica, 25, 316–321. & (
- 2002) Mixed-Effects Models in S and S-Plus. Springer, Berlin, Germany. & (
- R Core Team (2007) nlme: linear and nonlinear mixed effects models [WWW document]. URL http://cran.r-project.org. , , , &
- 1999) Growth responses of 15 rain-forest tree species to a light gradient: the relative importance of morphological and physiological traits. Functional Ecology, 13, 396–410. (
- 1988) Pioneer species distribution in treefall gaps in Neotropical rain forest; a gap definition and its consequences. Journal of Tropical Ecology, 4, 77–88. , , & (
- R Development Core Team (2007) R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria.
- 2007) The Ecology and Evolution of Ant–Plant Interactions. University of Chicago Press, Chicago, Illinois. & (
- 2008) Drastic erosion in functional attributes of tree assemblages in Atlantic forest fragments of northeastern Brazil. Biological Conservation, 141, 249–260. , , , , & (
- 2007) Plants use macronutrients accumulated in leaf-cutting ant nests. Proceedings of the Royal Society B – Biological Sciences, 274, 315–321. , , , , & (
- 2001) Human- and beaver-induced wetland changes in the Chickahominy River watershed from 1953 to 1994. Wetlands, 21, 342–353. & (
- 2008) Edge-effects drive tropical forest fragments towards an early-successional system. Biotropica, 40, 657–665. , & (
- 2007) Cutting more from cut forests – edge effects on foraging and herbivory of leaf-cutting ants. Biotropica, 39, 489–495. , , & (
- 1990) Habitat selection by the queens of the leaf-cutting ant Atta sexdens L. in Brazil. Journal of Tropical Ecology, 6, 249–252. (
- 2003) Leaf-cutting ant (Atta sexdens) and nutrient cycling: deep soil inorganic nitrogen stocks, mineralization, and nitrification in Eastern Amazonia. Soil Biology and Biochemistry, 35, 1219–1222. , & (
- 1972) Gardening Ants: The Attines. The American Philosophical Society, Philadelphia, Pennsylvania. (
- 1993) Use of hemispherical photographs in forest ecology: measurement of gap size and radiation totals in a Bornean tropical rain forest. Journal of Tropical Ecology, 9, 131–151. , , , , & (
- 1986) The defining traits of fire ants and leaf-cutting ants. Fire Ants and Leaf-Cutting Ants: Biology and Management (ed. by C. S.Lofgren and R. K.Vander Meer), pp. 1–9. Westview Press, Boulder, Colorado. (
- 2003) Herbivory of Leaf-cutting Ants–A Case Study on Atta colombica in the Tropical Rainforest of Panama. Springer, Berlin, Germany. , , , & (
- 2007) Increasing densities of leaf-cutting ants (Atta spp.) with proximity to the edge in a Brazilian Atlantic forest. Journal of Tropical Ecology, 23, 501–505. , , , , & (
- 2008) Plant–herbivore interactions at the forest edge. Progress in Botany, 69, 423–448. , , & (
- Top of page
- Materials and methods
- Supporting Information
Additional Supporting Information may be found in the online version of this article under the DOI reference: DOI: 10.1111/j.1365-2311.2010.01241.x
Appendix S1. Details of methodologies used.
Appendix S2. Model of potential biomass gain of seedlings along transects from Atta nests into the forest understorey.
|EEN_1241_sm_AppendixS1.doc||196K||Supporting info item|
|EEN_1241_sm_AppendixS2.doc||64K||Supporting info item|
Please note: Neither the Editors nor Wiley Blackwell are responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.