Effect of overstory tree species diversity and composition on ground foraging ants (Hymenoptera: Formicidae) in timber plantations in Ghana

ABSTRACT

Plantation forests are becoming an increasingly important component of the world’s forested ecosystem. However, relatively little is known about how forest plantation management, overstory tree species composition and diversity impact biodiversity of nontree components of the forest. We assessed changes in ant functional group composition as related to changes in overstory tree diversity (monocultures vs. polycultures), species composition (native African species vs. exotic teak), and time (one and two years after planting). A pitfall trapping scheme was implemented during the summer months of 2006 and 2007. A total of 7473 specimens were collected representing six subfamilies, 22 genera, and 65 species. We found no significant differences in traditional diversity measures or functional group composition between treatments one year after planting. Two  years after planting, we found that species richness of ground foraging ants had significantly increased (F = 4.60, d.f. = 4, 15, p = 0.01). Several observed trends may have indicated that these ant communities were in transition and will likely become more distinct over time as the different plantation types recover from disturbance and diverge from each other in overstory structure.

EDITED BY Sheila Ward

KEYWORDS:

• Formicidae
• timber plantations
• Khaya ivorensis
• Tectona grandis
• biodiversity

Introduction

Tropical forests are under numerous economic, political, and social pressures including land conversion for agricultural and infrastructure development and tropical timber production. Africa has seen the highest rate of change in forest cover with an annual loss rate of just under 1 million hectares per year between 1990 and 2010 (FAO 2011). Decreasing tropical forest area, a decline in native timber stock, and increasing demand for timber products have led to plantation forests becoming an increasingly important component of the world’s forested lands. Between 2000 and 2010, the rate of planted forest establishment rose to 115,000 ha per year (FAO 2011). Planted forest area represented approximately 187 million hectares worldwide in 2000 (FAO 2001) and rose to 260 million hectares worldwide by 2010 (FAO 2011). Plantation forests are depended on for wood biomass production, soil and water conservation, and wind protection (Carnus et al. 2003) and are key sources of fuel wood and nontimber forest products (FAO 2015).

In 2010, planted forests represented almost 2.5 million hectares in West Africa (FAO 2011). In 2005, Ghana had an estimated 160,000 ha of forest plantations (FAO 2005), an increase in 85,000 ha since 2001 (FAO 2001). Of this approximately 90% was estimated to be exotic species, predominantly Tectona grandis monocultures. In spite of their high productivity and ability to meet some timber production goals, exotic species cannot satisfactorily fill the ecological and socioeconomic roles of native species. Little research has been undertaken to understand the relationship between forest management and biodiversity (Hartley 2002). In addition, the majority of assessments made have not satisfactorily evaluated a variety of alternative land uses such as native species plantations, polycultures, and nonforested land uses (Stephens & Wagner 2007).

With the increasing loss of habitats and biodiversity around the world, there is an urgent need for biodiversity assessment (Agosti et al. 2000). There is considerable interest in the identification of robust bioindicators for use in land monitoring and assessment programs (Noss 1990). Living organisms can integrate a variety of effects over time that short-term physical and chemical measures cannot, making them suitable indicators of environmental conditions (Danks 1992). Traditionally, bioindicators have been used to assess ecosystem response to human-induced environmental perturbation (Agosti et al. 2000). Terrestrial invertebrates are an important component of forest ecosystems and comprise a major part of their biological diversity (Beattie et al. 1992). In contrast to larger more mobile animals, terrestrial invertebrates are less likely to move between treatment units, so their presence is a better indication of and more strongly related to site condition (Bromham et al. 1999). Terrestrial invertebrates dominate the biomass, are highly diverse, occur ubiquitously, and are fundamentally important in ecosystem function (Samways 1994; Andersen 1995). Recent work on bioindicators identification has focused on soil invertebrates and microfauna (Cluzeaua et al. 2012). Specific insect guilds are highly sensitive to habitat disturbance (Day et al. 1993; Samways 1994; Mendez et al. 1995). For example, carabids (Kromp 1990; Beaudry et al. 1997; Villa-Castillo & Wagner 2002), Lepidoptera (Holloway & Stork 1991; Kremen 1992), Odonata (Samways 1996), and Formicidae (Buffington 1967; Majer 1982; Andersen 1990, 1991, 1995, 1997a; Roth et al. 1994; York 1994, 1999, 2000; Brown 1997; King et al. 1998; Peck et al. 1998; Bromham et al. 1999; Andrew et al. 2000; Vanderwoude et al. 2000; Stephens & Wagner 2006), have all been used successfully as bioindicators (Peck et al. 1998).

We selected ants as potential bioindicators of land-use type, because of their high diversity and biomass, their ecological importance at all trophic levels, their ease of sampling, and their well-understood community dynamics (Andersen & Sparling 1997). Furthermore, ants consistently show strong successional patterns in ecosystems and their functional diversity and composition are related to land management practices and disturbances (Andersen & Sparling 1997), including forest management (Andersen & Sparling 1997; Vanderwoude et al. 2000; Willett 2001; Stephens & Wagner 2006). Majer (1983) showed that ants could be used to monitor ecosystem recovery and disturbance and have been used extensively in Australia to observe rehabilitation of mine sites. Additionally, ants have high spatial and temporal fidelity to sites (Nakamura et al. 2015).

The objectives of this study were to experimentally evaluate changes in ant species diversity and functional group composition under different levels of overstory tree diversity (a monoculture and six and 10 species polycultures of native species) and in plantations of native African species and an exotic species (Tectona grandis, or teak). We also looked for indicator species and functional groups for each type of plantation. We tested the hypothesis that changes in ant species diversity and ant functional group composition would be related to differences in overstory tree diversity and/or native vs. exotic species.

Materials and methods

Study sites

This study was conducted in Ghana, West Africa in the moist semi-deciduous forest zone between 6°34ʹ33ʺ N and 7°10ʹ49ʺ N. The moist semi-deciduous forest zone is the most extensive closed canopy forest type in Ghana, characterized by multistory canopy layers and approximately equal proportions of evergreen and deciduous species. This is the major timber-producing area in Ghana and is suitable for the production of agricultural crops, orchard crops, and exotic timber plantations (Wagner et al. 2008).

During July of 2005, experimental plantations were established on degraded forest sites within the South Fomangsu Forest Reserve (6°35ʹN, 0° 57ʹW) and Afram Headwaters Forest Reserve (7°10ʹN, 1°40ʹW). Both reserves are located within the moist semi-deciduous forest zone and are approximately 90 km apart. Degraded areas were identified within the reserves as lacking overstory trees and being dominated by weedy grass species and some woody shrubs. The degradation within the reserves was initially due to encroachment of illegal agriculture and chainsaw felling of timber. The degraded areas were within the bounds of the forest reserves and approximately 25–40 ha in size. We anticipated that the neighboring forests would provide a source for colonizing ant fauna. These plantations were established as part of a larger study examining the impacts of overstory tree diversity and species composition on ecosystem functions.

Experimental design

The treatments selected included (1) a native species monoculture, of Khaya ivorensis (an African mahogany), (2) exotic species monoculture, Tectona grandis (teak), and two polyculture conditions of (3) a six native timber species mixture, and (4) a 10 native timber species mixture (Table 1). The experimental plantations included tree species that represent different successional stages, growth forms and economic values ranging from timber to non-timber forest products (Table 2). Two replicates of each treatment were installed at each Forest Reserve. Plantations were set in complete randomized blocks with each block containing 1 replicate of each of the four treatments, for two replicates of each treatment at each forest reserve. Each individual plantation was 40 m × 40 m with an initial planting distance of 1 m. Within each mixed native species plantation, trees of each species were randomly distributed. All seedlings were of local provenance and reared in the FORIG nursery until approximately 3 months in age. A 4 m buffer was left between treatments.

Table 1. Species diversity and composition of treatments.

Description	Species
Treatment 1 Native species Monoculture	Khaya ivorensis
Treatment 2 Exotic species Monoculture	Tectona grandis
Treatment 3 Six native timber Species mixture	Albizia zygia, Ceiba pentandra, Khaya ivorensis, Mansonia altísima, Milicia excelsa, Pericopsis elata
Treatment 4 Ten native timber Species mixture	Albizia zygia, Ceiba pentandra, Khaya ivorensis, Mansonia altísima, Milicia excelsa, Pericopsis elata, Tieghemella heckelii, Terminila superba, Tetrapleura tetraptera

Table 2. Tree species characteristics.

Species	Common Name	Native vs. Exotic	Pioneer	Nonpioneering light demanding	Non-pioneering shade bearing	Nitrogen Fixer	Economic Importance	Threatened
Khaya ivorensis	African Mahogony	N		X			X	X
Milicia excesla	Iroko	N	X
Pericopsis elata	Afromosia	N		X			X
Albiza zygia	Okuro	N				X
Ceiba pentandra	Onyina	N	X
Mansonia altissima	Oprono	N		X			X
Terminalia superba	Ofram	N	X				X
Tetraplura tetraptera	Prekese	N
Tieghenella heckelli	Baku	N			X
Tectona grandis	Teak	E	X				X

Ant sampling

Sampling of ants via pitfall traps was conducted during July/August in 2006 and 2007, after the onset of the short rainy season, at one and two years after plantation establishment. Five pitfall traps were placed on each of two parallel 25-m transects at 5-m spacing, for a total of ten pitfall traps per treatment replica. Transects were placed at the center of each replicate plantation with a random start and direction within the stand. The traps were placed in two parallel transects instead of a single transect to reduce edge effects.

A pitfall trap consisted of a plastic deli style container (11 cm diameter by 8 cm deep) buried flush to ground level. During the 48-hour sampling periods, traps were set flush with the soil surface and with approximately 250 ml of soapy water in the trap. Any debris inside the trap was carefully removed when the traps were set. Preliminary trap testing conducting in the field ranged from 24 hours to 10 days. Analysis of these trap catches showed a 48-hour trapping period allowed for sufficient ant capture and reduced the risk of trap disturbance and flooding, especially during the rainy season. In addition, species richness curves were optimized at between six and eight traps (unpublished data).

Pitfall trapping has been widely used in studies of Australian ant communities (Andersen 1995) and has been shown to provide a reliable estimate of species composition (Andersen 1991; Agosti et al. 2000). Andersen (1991) showed that pitfall trapping for ants was an adequate collection method for habitats with dense litter and vegetation as occurred at the sites in this study, in comparison to more time-intensive quadrant counts, which are better suited to open habitat types. The use of pitfall traps may underrepresent cryptic, soil-litter-dwelling ant species, although they do occur in small numbers.

The contents of each trap were collected individually and transported back to the lab for sorting, preservation, and preparation of museum vouchers. Ants were identified to species and placed into functional groups as adapted from Andersen (2007 and personal communication) and described in Table 3; identifications were based on Taylor (2007). Members of the genus Dorylus were excluded from this study because they are highly mobile and do not establish long-term stationary colonies, and therefore are not associated with long-term site characteristics. This behavior precludes them from being suitable bioindicators. Specimens were vouched by Dr. Brian Taylor and have been deposited into the Northern Arizona University School of Forestry voucher collection and in the Forestry Research Institution of Ghana insectarium.

Table 3. Classification of ant functional groups collected from the semi-deciduous forest zone of Ghana, West Africa.

Functional Groups (code)	Characteristics	Species	Species Code	Species	Species Code
Subordinate Camponotini (SC)	Ecologically separated due to large body size and nocturnal foraging. Several arboreal nesters.	Camponotus sp. (Myrmacrhaphe)	CAMY	Camponotus furvus	CAFU
		Camponotus acvapimenis	CAAC	Camponotus maculatus	CAMA
		Camponotus flavomarginatus	CAFL
Tropical Climate Specialist (TCS)	Distribution centered on tropical climatic zone. Unspecialized ants.	Aenictus guineensis	AEGU	Pristomyrmex orbiceps	PROR
		Aenictus species 2	AESP
Cryptic Species (CS)	Forage and nest primarily within soil and litter, having relatively little interaction with epigaetic ants	Cerapachys nitidulus	CENI	Pyramica cryptura	PYCR
		Hypoponera cammerunensis	HYCA	Pyramica membranifera	PYME
		Oligomyrmex diabolus	OLDI	Strumigenys petiolata	STPE
		Oligomyrmex vorax	OLVO	Strumigenys rufobrunea	STRU
Opportunist (O)	Unspecialized, poorly competitive species, characteristic of disturbed sites	Cardiocondyla emeryi	CAEM	Tetramorium angulinode	TEAN
		Cardiocondyla neferka	CANE	Tetramorium calinum	TECA
		Decamorium decem	DEDE	Tetramorium flavithorax	TEFL
		Odontomachus troglodytes	ODTR	Tetramorium intonsum	TEIN
		Lepisiota cacozela	LECA	Tetramorium minimum	TEMI
		Lepisiota laevis	LELA	Tetramorium sericeiventre	TESE
		Paratrechina albipes	PAAL	Tetramorium sillimimum	TESI
		Paratrechina arlesi	PAAR	Tetramorium quadridentatum	TEQU
		Tapinoma lugubre	TALU
Generalist (G)	Ubiquitous species occurring in most habitats. Rapid recruitment to, and successful defense of food resources. Predominate moderately productive habitats.	Crematogaster censor	CRCE	Pheidole concinna	PHCO
		Crematogaster zavattarii	CRZA	Pheidole costauriensis	PHCOS
		Monomorium bicolor	MOBI	Pheidole impressifrons	PHIM
		Monomorium exiguum	MOEX	Pheidole megacephala	PHME
		Monomorium pharaonis	MOPH	Pheidole melancholica	PHMEL
		Monomorium rosae	MORO	Pheidole nigritella	PHNIG
		Monomorium tanysum	MOTA	Pheidole occipitalis	PHOC
		Monomorium trake	MOTR	Pheidole pusilla	PHPU
		Monomorium vaguum	MOVA	Pheidole retronitens	PHRE
		Pheidole aeberlii	PHAE	Pheidole schoutedeni	PHSC
		Pheidole aurivillii	PHAU	Pheidole speculifera	PHSPE
		Pheidole bayeri	PHBA	Pheidole weissi	PHWE
		Pheidole caffra	PHCA
Specialized Predators (SP)	Medium to large bodied strictly predaceous species. Well-developed sight, highly active foragers mostly solitary. Little competitive interactions with other ants behaving more like other predatory arthropods.	Anochetus bequaerti	ANBE	Pachycondyla silvestrii	PASI
		Cerapachys sudanensis	CESU	Pachycondyla caffraria	PACA
		Leptogenys conradti	LECO	Pachycondyla tarsata	PATA
		Pachycondyla pachyderma	PAPA

Adapted from Andersen (1997b, 2007).

Dominant Dolichoderine, Cold and Hot Climate Specialist functional groups removed, none were expected or observed.

Dorlyus spp. were excluded from the study because of their characteristic nesting and foraging behavior.

Ant species diversity and functional groups

The traditional diversity measures of species richness (R), the Shannon-Wiener Diversity Index (H), and the Simpson Dominance Index (D) were calculated for each treatment. Species richness is the number of species observed. The Shannon-Wiener Diversity Index was calculated as a measure of relative diversity per treatment type per year. The Simpson Dominance Index was used to determine the strength of species numerical abundance. These traditional diversity measures fail to assess species assemblages or groups with particular traits, so species were also grouped into functional groups for further analysis. Functional groups were determined according to behaviors, habitat preferences, and trophic level interactions, using existing functional group assignments developed by Andersen (2007 and personal communication).

Statistical analyses

Prior to any statistical interpretation of the ant data, we decided to pool the trap collections across treatment replicate to obtain one value per treatment replicate per year, because we were primarily interested in observing indicator species and guild responses to treatment. Effect of treatment was evaluated in two separate analyses: (1) comparing plantations with different levels of diversity of native species (1, 6, and 10 native species; Table 1), and (2) comparing native African species (all plantations) with exotic teak monocultures. We used data from each year to determine if changes in communities occurred over time.

We applied standard analysis of variance (ANOVA) tests to determine differences between traditional diversity indices in overstory tree diversity treatments using JMP In 5.1.2 (SAS 2004) after meeting assumptions of normality and equal variance. Nonparametric Welch ANOVA was applied when standard ANOVA assumptions were not met (SAS 2004). To determine differences between species diversity indices for the native vs. exotic treatments, we applied a two-sample t-test after meeting assumptions of normality and equal variance (Hays 1963).

Functional group analysis

Multiple response permutation analyses (MRPP) was performed for the relative abundance measures of each ant functional group to test the null hypothesis of no difference in functional group composition among treatments (McCune & Mefford 2006). Indicator species value analysis (referred to as indicator value) by the Dufrene and Legendre (1997) method was performed to determine the specific contributions of each species to the functional group and to determine the strength of indication of treatment provided by either a functional group or an individual species (Stephens & Wagner 2006).

Results

A total of 7473 specimens were collected representing six subfamilies, 22 genera and 65 species (Table 4). The number of specimens collected per treatment was similar and ranged from 746 to 1262. The greatest number of species was trapped in the genus Pheidole (16 species). The most commonly trapped ants were members of Pheidole, represented by three to six species per treatment. The most frequently observed ant was Odontomachus troglodytes Santschi, which represented 17–55% of ants collected per treatment. Composition of ant communities by functional group varied little between treatments. Results from two years after planting showed increases in ant species richness and nonsignificant differences in ant community composition.

Table 4. Composition of ant genera by functional group. Data are number of species per genera, with percent total ants per treatment per year (column) in brackets.

Functional group	Exotic teak monoculture		Native African mahogany monoculture		Six native timber species mixture		Ten native timber species mixture
Genera	2006	2007	2006	2007	2006	2007	2006	2007
Subordinate Camponotini
Camponotus	3 (15.67)	1 (2.39)	2 (5.74)	1 (4.77)	4 (5.14)	1 (1.27)	3 (7.91)	1 (4.83)
Tropical Climate Specialist
Aenictus		1 (0.13)
Pristomyrmex			2 (0.22)
Cryptic Species
Cerapachys		1 (0.26)		1 (0.10)		1 (0.18)		1 (0.08)
Hyperponera								1 (0.16)
Oligomyrmex			1 (0.11)	1 (0.10)		1 (0.51)
Pyramica				1 (0.10)		1 (0.36)		1 (0.16)
Strumigenys		1 0.13)		1 (0.20)	1 (0.12)	1 (0.73)	1 (0.13)	1 (0.24)
Oppurtunist
Tapinoma		1 (15.52)	1 (4.27)	1 (9.44)		1 (8.45)		1 (12.60)
Lepisiota	2 (5.41)	1 (1.33)	2 (7.42)	1 (1.32)	2 (8.30)	2 (6.09)	2 (5.23)	2 (2.46)
Paratrechina	1 (4.96)	2 (9.81)	2 (3.04)	1 (1.42)	1 (1.64)	2 (3.00)	1 (3.89)	2 (1.35)
Cardiocondyla		1 (3.15)	1 (0.56)	1 (0.79)	2 (0.35)	1 (4.00)		1 (3.25)
Decamorium			1 (0.11)
Tetramorium	1 (1.35)	2 (1.86)	3 (3.26)	3 (0.81)		7 (5.36)		3 (2.06)
Odontomachus	1 (55.02)	1 (24.80)	1 (17.32)	1 (46.90)	1 (41.82)	1 (25.18)	1 (45.04)	1 (23.53)
Generalist
Crematogaster				1 (1.62)		1 (0.64)	1 (0.13)	1 (1.23)
Monomurium	2 (3.16)	3 (1.59)	1 (1.46)	2 (0.71)	1 (1.99)	4 (2.65)	4 (2.68)	5 (2.38)
Pheidole	5 (2.37)	3 (27.19)	6 (34.53)	2 (19.19)	5 (22.43)	4 (24.45)	4 (19.97)	3 (34.15)
Specialized Predators
Cerapachys				1 (0.10)
Anochetus			1 (0.34)	1 (0.20)		1 (0.09)		1 (0.08)
Leptogenys	1 (0.11)			1 (0.10)
Pachycondyla	2 (11.95)	2 (14.19)	2 (21.60)	2 (9.64)	2 (12.97)	3 (16.82)	2 (15.01)	3 (11.33)
Total Individuals	887	754	889	985	856	1100	746	1262
Total # of species	18	20	25	24	18	33	19	29
Mean R (SE)	8.75 (1.25)	11.75 (1.48)	11.25 (1.25)	13.25 (1.48)	9.50 (1.25)	19.75 (1.48)	10.25 (1.25)	17.00 (1.48)
Mean H (SE)	1.36 (0.27)	1.84 (0.24)	1.76 (0.27)	1.38 (0.24)	1.47 (0.27)	2.22 (0.24)	1.67 (0.27)	1.94 (0.24)
Mean D (SE)	0.58 (0.11)	0.78 (0.09)	0.75 (0.11)	0.54 (0.09)	0.63 (0.11)	0.83 (0.09)	0.70 (0.11)	0.78 (0.09)

Ant species diversity and overstory tree species diversity

At one year after planting, there were no significant differences in measures of ant species diversity in relation to overstory diversity of native tree species (Table 4). Insignificant measures included species richness (F = 0.96; d.f. = 4, 15; p = 0.5), Shannon-Weiner Diversity Index (F = 0.82; d.f. = 4, 15; p = 0.5), and Simpson’s Dominance Index (F = 0.76; d.f. = 4, 15; p = 0.6) (Table 4). At two  years after planting species richness was significantly different between treatments (F = 4.60; d.f. = 4, 15; p = 0.01). Species richness was highest in the plantations of six native timber species (R = 19.75 ±SE 1.48) and lowest in the native African mahogany monoculture (R = 13.25 ±SE 1.48) and (Figure 1). Shannon-Weiner Diversity Index (F = 2.28; d.f. = 4, 15; p = 0.1) was highest in the plantations of six native timber species (H = 2.22 ±SE 0.24) and lowest in native African mahogany monocultures (H = 1.38 ±SE 0.24) (Figure 2). Simpson’s Dominance Index (F = 1.64; d.f. = 4, 15; p = 0.2) was highest in the plantations of six native timber species (D = 0.83 ±SE 0.09) and lowest in native African mahogany monocultures (D = 0.54 ±SE 0.09) (Figure 3).

Figure 1. Ant species richness (R) comparisons with standard error bars in 2-year-old plantations for (a) overstory tree species diversity, and (b) teak monocultures s. all native species plantations. Significance shown at α = 0.05. Bars with different letters are statistically different based on a standard mean separation test.

Figure 2. Shannon Weiner Index (H) comparisons with standard error bars for ant species in 2-year-old plantations for (a) overstory tree species diversity, and (b) teak monocultures vs. all native species plantations. Significance shown at α = 0.05. Bars with different letters are statistically different based on a standard mean separation test.

Figure 3. Simpsons Dominance Index (D) with standard error bars for ant species in 2-year-old plantations for (a) overstory tree species diversity, and (b) teak monocultures vs. all native species plantations. Significance shown at α = 0.05. Bars with different letters are statistically different based on a standard mean separation test.

Ant species diversity and native species vs. teak overstories

At one  year after planting, there were no significant differences between the plantations of native African species and teak in measures of ant species diversity, including species richness (native species R = 11.25; exotic teak R = 8.75) (t = −1.43, d.f. = 18, p = 0.1); Shannon-Weiner Diversity Index (t = −1.24, d.f. = 18, p = 0.1); or Simpson’s Dominance Index (t = −1.26, d.f. = 18, p = 0.2) (Table 4). However, at two years after planting, species richness (t = −2.54, d.f. = 18, p = 0.02) was significantly higher in the plantations of native species (R = 13.25) than in the exotic teak plantations (R = 11.75) (Figure 1).

Functional group analysis

During both the 2006 and 2007 collection periods, neither ant functional group composition nor individual ant species composition differed significantly at among the one, six, and 10 species treatment, nor between the native African species and teak plantations (α = 0.05, Table 5; Figures 4 and 5). The A statistic for the MRPP comparisons of both overstory diversity and native vs. teak plantations was very close to zero for both ant functional groups and individual species in both years (Table 5), indicating that the heterogeneity within groups was equal to that expected by chance.

Table 5. Multiple response permutation procedure (MRPP) to determine ant functional group and individual species relative abundance response to treatment per year.

	Individual species				Functional groups
	Teak vs. all native plantations		Overstory diversity		Teak vs. all native plantations		Overstory diversity
	2006	2007	2006	2007	2006	2007	2006	2007
Test statistic T	1.32	−0.01	1.51	0.45	−0.36	−0.96	0.81	−1.12
Observed Delta	138.72	120.43	143.53	123.67	114.49	119.31	119.57	114.55
Expected Delta	135.08	120.48	135.08	120.48	115.41	122.76	115.41	122.76
A*	−0.03	0.0003	−0.06	−0.02	0.007	0.02	−0.04	0.06
p (α = 0.05)	1.00	0.38	0.97	0.62	0.29	0.15	0.78	0.13

*A statistic chance corrected within group agreement, A(max) = 1 when all items are identical within groups. A = 0 when heterogeneity within groups is equal to expected by chance. A < 0 when heterogeneity is greater within groups than expected by chance. Small value for A indicates that groups are appropriately assigned, members within groups are more similar to each other than to members of other groups. p-value indicates probability of smaller or equal delta.

Figure 4. Ant functional group composition 1 year after planting for (a) overstory tree species diversity, and (b) teak monocultures vs. all native species plantations. Different ratios of column fills indicate differences in functional groups composition – greater dissimilarity between columns equals greater dissimilarity between ant communities. TCS – tropical climate specialist, SP – specialized predators, SC – subordinate camponotini, O – opportunist, G – generalist, CS – cryptic species.

Figure 5. Ant functional group composition 2 years after planting for (a) overstory tree species diversity, and (b) teak monocultures vs. all native species plantations. Different ratios of column fills indicate differences in functional groups composition – greater dissimilarity between columns equals greater dissimilarity between ant communities. TCS – tropical climate specialist, SP – specialized predators, SC – subordinate camponotini, O – opportunist, G – generalist, CS – cryptic species.

While no indicator species or groups were identified for any treatments, the Opportunist functional group dominated all of the treatments in both years, representing 36–67% of the ants collected in each treatment. The Generalist functional group was typically the second most common functional group in both years representing an average of 26% of ants collected in each treatment. Specialized predators and Subordinate Camponotini averaged 13% and 11%, respectively. Tropical climate specialist and Cryptic species were very seldom observed and contributed to less than 2% of the ants collected from any treatment (Table 4, Figures 4 and 5).

Changes over time

Over the 2 years of the study, we observed a significant change in species richness (F = 53.22, d.f. = 8, 30, p = < 0.0001) in all treatments (Figures 1 and 6). The exotic teak and native African mahogany monocultures gained two and lost two species between years. The mixed native species plantations gained 10–15 species between years. We observed only small, nonsignificant changes in the proportions of functional group in each treatment (Figure 7). Subordinate Camponotini species declined in all treatments, Cryptic species increased in all treatments, and Specialized Predators increased in the plantations of one, six, and 10 native tree species and declined in exotic teak monocultures. Opportunists increased in the exotic teak monocultures and native mixed species plantations and decreased in native African mahogany monocultures. Generalists increased or remained unchanged in mixed native species plantations and decreased in the teak and African mahogany monocultures.

Figure 6. Change in ant species richness one and two years after planting for (a) overstory tree species diversity, and (b) teak monocultures vs. all native species plantations. One year after planting there were no differences in species richness among treatments. 2 years after planting all treatments showed an increase in species richness which differed for both overstory diversity and between teak monoculture and the native species plantations (see also Figure 1).

Figure 7. Changes in number of ant species representing each functional group over time for exotic teak and native African mahogany monocultures, 6 and 10 native species mixed plantations. Positive bars indicate an increase in species per functional group, negative bars indicate a decrease in the number of species per functional group. TCS – tropical climate specialist, SP – specialized predators, SC – subordinate camponotini, O – opportunist, G – generalist, CS – cryptic species.

Discussion

Over the two  years of the study, we observed a 20% increase in species richness across all treatments (Figure 6). If current trends persist, we predict species diversity will become stable over time. As supported by other observations (Stephens et al. 2008), we expect distinct ant communities will develop in these plantation types, which may include specific indicator species or functional groups.

This study highlights the potential of ants as bioindicators of habitats conditions in tropical forested systems, where they have been generally underutilized (Piper et al. 2009). Anderson (1991) indicated that ants are associated with structural components of their habitats. Peck et al. (1998) saw significant changes in ant communities based on vegetation community diversity and structure. Disturbance resulting from fire and timber management has been observed to affect ant communities (Neumann 1992; Andrew et al. 2000; York 2000).

Changes in ant diversity and overstory composition

Overstory tree diversity had a positive impact on ant species diversity at two years after plantation establishment (Figures 1–3, and Table 4), although not at one year after planting. Increasing overstory tree diversity from one to six species resulted in a 30% increase in ant species richness. The species present in the six and ten native timber species plantations increased by 15 and 10 species respectively between the first and second years (Table 4). This suggests that, within plantation forests, moderate increases in overstory tree diversity can have a significant effect on ant biodiversity, and perhaps on other taxa.

Native environments typically support higher number of insect species than non-native environments (DeGomez & Wagner 2001). In this study, the native species plantations had approximately 20% more ant species than the exotic teak plantation over both years (Table 4). At two years, the native species plantations had significantly higher species richness teak plantations. A nonsignificant difference in species composition was also observed (Table 4) with six genera being unique to the native African species overstories and only one genera unique to teak plantations. Since teak is a major plantation species in West Africa and elsewhere (FAO 20001), the lower associated ant diversity in this study is of concern.

Ant functional groups and succession

Whereas traditional diversity measures are based on numerical counts, functional group analysis is based on differences in characteristics among species groups. Unlike traditional diversity measures, functional group analysis permits examination of how species habitat requirements and competitive interactions may impact diversity patterns (Anderson 1991). In this study, the relative abundances of each functional group were similar among treatments especially at one year after planting (Figures 4 and 5), with MRRP A values indicating high homogeneity between treatments (Table 5). However, by Year 2, we see a trend of increasing heterogeneity in ant functional groups among treatments in MRRP A values (although still insignificant) (Table 5).

This study may have presented the early stages of establishment of distinct ant communities under differing overstory composition. As the treatment areas recover from forest degradation, disturbance, and plantation establishment (site prep, planting, and early site maintenance), and their composition and structure become more distinct, more distinct ant communities may emerge. Significant differences among plantation types in ant diversity indicators were detectable at only twoyears after plantation establishment. Other work (Stephens et al. 2008) assessed ants as indicators of land use in Ghana at 10–13 years after planting and found distinct patterns in communities of ants and other organisms among agriculture, citrus orchards, exotic species plantations, mixed native species plantations, and native forests. The plantations were similar in species composition to the 10 native timber species plantation and teak monocultures used in this study. The agricultural areas and citrus orchards were also similar to the degraded conditions prior to plantation establishment for this study at the Forest Reserve sites. Furthermore, indicator species and functional groups were identified for the more mature land uses (Stephens et al. 2008). These observations support the conjecture that the newly established plantations of this study are in the early stages of establishment for distinct ant communities.

In Years 1 and 2 under different overstory types, the ant communities were generally dominated by Opportunists and Generalists (Figures 4 and 5). We have observed that Opportunists tend to move into disturbed habitats (Stephen et al. unpublished), whereas Generalists seem indifferent to successional stage. In this study, between Years 1 and 2, Opportunists increased (except for the native African mahogany monoculture) and Generalists showed no pattern (Figure 7). In another study (Stephen et al. unpublished), a similar relationship between Opportunist and Generalist functional groups and different land uses was observed, with Opportunists dominating agricultural fields. We anticipate future changes from Opportunist-dominated ant communities to more functional group diversity and stronger links to overstory tree diversity and composition.

Nonsignificant shifts in functional group composition between Year 1 and Year 2 include decreases in the subordinate Camponotini (SC) and increases in the Cryptic species functional groups in all treatments. The SC group was composed largely of arboreal Camponotus species strongly associated with downed woody debris (Higgins & Lingren 2006; Stephens & Wagner 2006). The Camponotus species that were found in Year 1 may have been residual populations using coarse woody debris and the few small shrubs remaining from before overstory removal and plantation establishment. By Year 2, since woody material on the treatment sites had lessened and the young trees were not of sufficient size to provide suitable habitat, Camponotus species were probably being displaced by more pioneering species more adapted to disturbance and early successional conditions. Camponotus species may re-enter these areas when the overstory trees mature and provide habitat.

In contrast to the SC group, Cryptic species such as Strumigenys species, specialize in predation on leaf litter dwelling microinvertebrates (Holldobler & Wilson 1990). The increase in Cryptic species between Years 1 and 2 was likely a response to changes in understory and soil conditions after removal of agricultural crops and plantation establishment that improved microhabitats for their prey. The number of species in the Specialist Predators (SP) functional group also increased in all of the native overstory treatments (Figure 7) from Year 1 to Year 2, perhaps as a consequence of understory changes. For example, Cerapachys sp. are predators of termites and other ants, particularly Pheidole sp. (Holldobler & Wilson 1990). Pheidole is a Generalist, which was readily abundant in treatments where Specialist Predators were observed (Table 3, Figure 7).

As these young plantations transition through successional stages of stand development, it seems likely that distinct ant community patterns will emerge. Our experimental plantations with managed understories and open canopies may not yet have developed all of the characteristics that promote specific ant functional groups or species. Regular weeding was used to control the understory vegetation during the early years of plantation development, and the tree species used in these plantations have different growth characteristics (Table 2) and grew at significantly different rates, all of which may have impacted ant community composition and diversity. The cumulative differences in the plantation types as the stands develop will impact both the overstory and understory ant communities. As canopy structure matures, shading of the understory will increase and changes in understory vegetation and leaf litter will support different ant communities. For example, as individual trees increase in bole diameter and height during stand succession, several species of arboreal nesters would be expected to appear, including carton-nesting Crematogastor spp. and Camponotus spp. (Holldobler & Wilson 1990).

Future research is required to fully assess the stages of stand development and management that are most influential on ant communities. As these plantations mature and are managed over time, some tree species are likely to be thinned or harvested to meet management objectives. This would present the opportunity to assess how those changes in stand structure and diversity further impact ant communities. A longer period of observation over 5–10 years may reveal the full sequence of ant community assemblage establishment and the appearance of indicator species and functional groups associated with overstory tree diversity and composition. Additional work could be conducted to assess the association of ant communities with changes in understory vegetation, leaf litter, and soil characteristics.

Conclusion

Biodiversity is clearly an important component of ecosystems, supporting their function (Tilman et al. 1996) and their anthropocentric value (Wilson 1984). Conventional wisdom has long held that biodiversity and plantation forests are mutually exclusive (Stephens & Wagner 2007). Stephens and Wagner (2007) argued the importance of land use comparisons prior to passing judgment on plantation forest impacts on biodiversity. Observed cases in the literature indicate that plantation management, can impact biodiversity positively and negatively, depending on the strategies used. Based on the results of this study and other work (Stephens et al. 2008), the six native tree species mixture at only two years after planting recouped an astonishing 80% of the total number of ant species, and from 40% to 100% of the individual ant functional groups observed at native forest sites. The native African mahogany monoculture recouped 60% of native forest ant diversity. These numbers show that biodiversity and plantation forests are not necessarily at odds. Casual observation of soil, microclimate, understory plants, and birds, reptiles and mammals in the more mature mixed native species and exotic teak plantations suggest that strong community difference are likely occur across numerous floral and faunal groups in different kinds of plantations. Our findings suggest that moderate changes in management, such as focusing on mixed native species plantations, can recapture a significant level of native forest ant diversity and potentially other taxa as well.

Acknowledgements

The authors would like to thank Dr. Brian Taylor for his assistance with African ant taxonomy, Dr. Andersen for his assistance with ant functional group assessment, Dr. McGavin and the current staff of the Oxford University Museum of Natural History – entomology collections for lab space, and the Forestry Research Institute of Ghana staff for assistance with field work, logistics and a home away from home.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This study was funded in part by International Tropical Timber Organization grant PD 256/03 Rev. 1 (F) and a Fellowship from the International Tropical Timber Organization.