Prairie flowers at the Fort Worth Nature Center, Fort Worth, Texas

Ants in the Grassland: their importance and potential as indicators of ecosystem health

This is the presentation I made at the America’s Grassland Conference recently  in Manhattan, Kansas (August 2013).

I’ve posted the power point presentation below with a few additions and included the questions asked.



While we often think of ants as annoying pests, ants are important members of nearly every terrestrial ecosystem, except Antarctica. There are  perhaps 40,000 species worldwide and we have good species descriptions for about 14,000 of them – all of which are in the same family, the Formicidae.

Ants are incredibly diverse: they vary in their morphology, their behavior, their physiology, and their ecology. Ants may engineer ecosystems through their nesting and foraging habits – greatly shaping the physical landscape and thereby impacting a variety of other organisms including the plant community. Ants also have a diversity of relationships. They are important prey items as well as predators; they have important mutualistic relationships with plants, fungi, and other arthropods; they have their own parasites; they have commensals and parasites that live in their nests.

Because of all this ecological diversity, ants may be good indicators of habitats and ecosystem health. If there are changes in any of these relationships, for any of these organisms, this change may affect ant presence, activity, and abundance. Because ants are small and live on a small scale, they may detect such changes earlier than larger monitored species, such as vertebrates. Ants are also good candidates for indicators because they are easy to collect and do not have the problems of monitoring vertebrate populations which may be difficult to track, endangered or threatened species sensitive to handling, etc. The possibility for such utility has been shown in previous research.

I investigated the potential for grassland ant assemblages to be used as bio-indicators in prairies in the Fort Worth Nature Center and Wildlife Refuge in Fort Worth, Texas, including to discern habitat type and response to disturbance. I collected the ground active ants of 17 sites monthly from March – September 2012 using pitfall traps.  The 17 sites constituted a natural experiment: 3 were wooded sites and 14 were prairie sites. The sites were paired according to soil and ecological unit (from the Natural Resources Conservation Service) for wooded vs. non-wooded (3 replicates); mowed (and soil disturbance) vs. non-mowed (4 replicates); and low intensity burn vs. non-burned (2 replicates).

For each site, I measured environmental variables which are known or thought to be important to ants in choosing their nesting areas,  including depth of sand, soil penetration (compaction), depth to the restrictive layer, percent slope, drainage, percent bare ground, percent litter cover, percent standing plant cover, percent total cover, latitude, ecological units (from the NRCS), and soil type.

I used the program CANOCO to do ordination analyses: principle components analysis (PCA) on the environmental variables only and redundancy analysis (RDA) combining the environmental variables with species presence. Ant species were characterized by functional groups following Andersen (1997). The PCA confirmed that the variables chosen could be used to distinguish among sites. The RDA revealed that some of the ant species were aligned with habitats but disturbance did not matter. The sites grouped into three sets which aligned with soil types and ecological units. Some species did not align with their habitats but this may be explained by the foraging of those species into habitats other than where they nest. The RDA showed a strong relationship between the ants and the environmental variables with the interaction between percent litter cover and drainage, percent litter cover by itself and drainage by itself being significant factors. However, these factors combined did not explain more than 20% of the variation so either there are other significant factors or many factors account for the local presence of ants with none being particularly significant.

Andersen’s functional groups are problematic for these sites because some of the species placed in the groups do not have similar ecological roles as the Australian species upon which this work is based. Species richness by functional group did not vary significantly among the sites. And although the functional group designations are problematic, there is a  pattern in the composition of these assemblages with general myrmicines contributing most, followed by hot climate specialists, cryptic and opportunistic species, then tropical climate specialists and dominant species. This suggests a shape to the assemblages that may transcend individual species.

In conclusion, this project indicated weak support for these assemblages s as bioindicators and only two ants could be considered indicators of habitat: the carpenter ant species (Camponotus americanus and Camponotus pennsylvanicus) occurring in the woodlands and the Comanche harvester ant (Pogonomyrmex comanche) occurring in the Aquilla prairie.