Realizing a synergy between research and education: how participation in ant monitoring helps raise biodiversity awareness in a resource-poor country

Abstract

Biodiversity-rich, resource-poor countries need to allocate scarce resources to the competing goals of identifying and monitoring their biodiversity and educating their populace about it. Often only relatively wealthy individuals participate in biodiversity-related volunteering, while the poor are left on the margins. We present a case study that shows how monitoring and education can be combined. South African high school scholars from mostly disadvantaged communities participated in ant monitoring in transformed sites and received lessons using their own data. The project provides baseline data on an important insect group in a region where invertebrate monitoring is rare. Participation in a real study enhances the scholars’ interest in science and direct interaction with scientists allows them to enquire about careers they might not otherwise consider. Here we outline how the project works, what participants learnt, and demonstrate that the data provide insights into ant diversity and the effects of landscape transformation.

Appendix

Box 1. Ant sampling

Ants are collected in 13 transformed sites associated with schools, in 13 natural sites close to the schools that serve as controls, and in a further 7 sites located in habitat types that were otherwise under-represented (Fig. 3). Two grids of 10 pitfall traps each are set in each site in a way that captures the local habitat diversity (Fig. 3). Traps within a grid are 10 m apart. The traps are plastic beakers of 7 cm diameter that are buried so that the rim is flush with the ground. Traps are partly filled with 50% propylene glycol solution. Propylene glycol is a low toxicity catching fluid which prevents ants from escaping and reduces evaporation. Traps are left open for 5 days at a time. Ant collections are done biannually in spring and autumn when ants are most active in this region. Collected samples were rinsed, sorted by trained staff and stored in 70% ethyl alcohol. By-catch with the exception of beetles—which will be processed at our own institution—has been made available to the Agricultural Research Council: South African National Survey of Arachnids (16 February 2009; www.arc.agric.za/home.asp?pid=3235) to avoid wasting specimens.

Fig. 3

Survey design. The map shows the locations of sites within the Western Cape Province of South Africa. School sites ranged from novel habitat types like lawns (shown left) to heavily trampled and otherwise impacted natural vegetation (right). Control sites were located in nearby nature reserves (middle) resulting in a matched-pair design. Some additional sites were selected to cover additional habitat types. Two grids with ten pitfall traps each are set in each site

Box 2. Environmental data

Additional environmental data were collected to examine what factors affect ant diversity patterns:

Topographic data: Scholars used clinometers, global positioning system (GPS) handsets, and compasses to measure, slope, location, elevation, and aspect.

Soil structure and chemical composition: Soil samples were collected once in all sites by taking 20 subsamples randomly in the area of each grid. Soil samples were then analysed in the laboratory for structural factors like sand, clay and stone content as well as for their content of nitrogen, phosphorous and other key nutrients.

Soil surface temperature: Soil surface temperature was measured hourly in pristine sites using small loggers (Thermocron i-Buttons, Semiconductor Corporation, Dallas/Maxim, TX, USA).

Vegetation structure surveys: These formed a key element of the field work done together with the scholars and were conducted each time traps were collected. Measurements were taken every 1 m along transects over the traps, starting five meters before and ending five meters after the last trap. At each point a measurement pole was set and maximum height at which different groups (e.g., grasses or woody shrubs) touched the pole was recorded. Additionally, the total number of intersections of the pole with dead or live vegetation, respectively was recorded every 5 m. The presence of large stones or exposed bedrock was also noted. The data thus collected were used to calculate descriptors of vegetation structure; e.g., vegetation cover was calculated as percentage of measurements taken where at least one plant touched the pole, while maximum vegetation height was the highest point at which any plant touched the pole. While the surveys had to be simple and thus provided only relatively coarse information, the data agreed well with productivity data gained from remote sensing when used to explain ant diversity patterns (data not shown). Ants are in the main not direct users of primary productivity and when using plant materials the interactions tend not to be highly specific. Thus a species-specific vegetation survey seems unlikely to be necessary to explain ant diversity patterns.

Box 3. Effects of habitat transformation on ant species richness and density

In the first 2 years 87,762 individuals belonging to 160 species and morpho-species (from now on also called species) in 32 genera were collected in all sites of the project. Of these 70 species were found in the 13 transformed school sites combined and 149 species were found in the other 21 sites. Transformed sites associated with schools had reduced species and genus richness but similar density to their control sites (Fig. 4; paired t-tests: species richness: t ₁₂ = 4.80, P < .001; density: t ₁₂ = 0.80, P = .44).

Fig. 4

Ant species richness and density in thirteen transformed sites associated with schools and thirteen control sites. Bars show means and 95% confidence intervals from pooled data from the two collections in 2007 in which all sites were sampled

The identification of the morpho-species is ongoing. However, it is already clear that new knowledge about the ranges of some South African species has been added. For example we found numerous Diplomorium workers in two sites in the Western Cape representing a Renosterveld site and an ecotonal site between Succulent Karoo and Fynbos. Diplomorium was previously only reported from the Eastern Cape Province at the eastern end of the Cape Floristic Region.

Box 4. Effects of habitat transformation on ant species composition

Species richness was decreased in most transformed sites when compared to natural sites in the same biome (Table 1). This reduced species richness was not due to random absences but rather to differences in how different genera coped with transformation. While some genera did very well in transformed areas others were completely absent from them (electronic supplementary Table S1). Large-bodied species from the genera Camponotus and Pachycondyla were absent or rare in transformed sites, while some small species of the genera Tetramorium and Monomorium reached high densities (Fig. 5). The alien invasive Argentine ant (Linepithema humile) was the most abundant species in many school sites but was only found in one nature reserve, where it had low abundance. Linepithema humile is known negatively to affect many though not all indigenous species (Christian 2001). Indeed, the most common species in transformed sites was the indigenous seed disperser Tetramorium quadrispinosum which has been previously shown to co-exist with L. humile (Christian 2001). This indicates that transformed sites retained some functionality, though it remains to be examined to what degree still present species can fulfil the roles of those that disappeared.

Table 1 Species richness in sites included in the Iimbovane project

Fig. 5

Abundance in transformed sites within the CFR pooled over the first 2 years and body size range for ant genera. Body size ranges were extracted from Brian Taylor’s “The ants of (sub-Saharan) Africa” (Taylor 2007) and taxonomic papers referenced therein for all species of each genus that were recorded from South Africa and Lesotho. For Brachymyrmex no sizes were reported for South African species and the size range given in Mayr (1868) and Santschi (1923) defining the genus was accepted instead. In polymorphic species all workers including majors were considered and the absolute minimum and maximum values for each genus are reported. The only worker caste excluded was repletes (a caste used as storage vessels) as they would not leave the nest. Bars for genera not represented in transformed sites are slightly offset for better visibility. Genus abbreviations are: Acr Acropyga, Ano1 Anochetus, Ano2 Anoplolepis, Bra Brachymyrmex, Cam Camponotus, Car Cardiocondyla, Cre Crematogaster, Dip Diplomorium, Dor Dorylus, Hyp Hypoponera, Lep1 Lepisiota, Lep2 Leptogenys, Lin Linepithema, Mer Meranoplus, Mes Messor, Mon Monomorium, Nes Nesomyrmex, Ocy Ocymyrmex, Oli Oligomyrmex, Pac Pachycondyla, Phe Pheidole, Pla Plagiolepis, Pol Polyrhachis, Pyr Pyramica, Sol Solenopsis, Tap1 Tapinolepis, Tap2 Tapinoma, Tec Technomyrmex, Tet1 Tetramorium, Tet2 Tetraponera

Changes in ant species composition depended on the degree of disturbance. Fig. 6 show the result of non-metric multidimensional scaling (MDS) using pooled data from the first 2 years. Primer 5 version 5.2.0 was used for analysis. In the MDS, sites are arranged in a way that reflects their ranked similarities, so that sites with a more similar species composition are closer together (Clarke and Warwick 1994). A finer division than for the preceding analyses was used. School sites were divided between fully transformed sites (novel habitats like lawns or thickets of alien invasive trees), partly transformed sites (novel habitats with remnants of disturbed but more natural vegetation), and heavily disturbed sites (original vegetation type but heavily impacted by a multitude of factors including trampling, refuse, alien plants, or livestock). Non-school sites were divided between pristine sites, slightly disturbed sites (low-level disturbance, or close to disturbed sites or roads), and moderately disturbed sites (previously heavily disturbed now restored sites or sites with some moderate disturbance). The non-school sites cluster together, indicating that low-level disturbance did not change ant species composition (Fig. 6). By contrast the school sites are clearly separated from the natural sites with the distance greatest for the worst transformation.

Fig. 6

Non-metric multidimensional scaling analysis of all sites in the Iimbovane project based on Bray-Curtis similarity. The analysis is based on the pooled abundance data for each species for the first 2 years of the project. In a few sites which joined the project later the actual period sampled is shorter. Data were 4th root-transformed prior to analysis to decrease the effect of very abundant species. Black symbols denote sites within the Cape Floristic Region (Fynbos and Renosterveld area) while open symbols denote sites in the Succulent Karoo. Grey symbols denote sites located within the Nama-Karoo biome. Degree of disturbance: P Pristine, SD Slightly disturbed, MD Moderately disturbed, HD Heavily disturbed, PT Partly transformed else heavily disturbed, T Fully transformed (see Box 4 for explanation of categories)

MDS also shows that sites within the Cape Floristic Region (Fynbos and Renosterveld) cluster separately from those in the more arid karoo biomes. The only exception was an ecotonal and extralimital site. In contrast the ant species composition of the two karoo biomes seems similar. These findings are interesting in the light of large shifts in biome boundaries predicted for the area as a consequence of climate change (e.g., Midgley et al. 2002; Hannah et al. 2005). Both these findings are also supported by cluster analysis (not shown).