Elsevier, Geoderma, Volume 323, 1 August 2018
It is commonly acknowledged that ants improve the hydraulic properties of soils in which they build their nests. To date, however, most studies of such soil modifications have focused on one ant species and one type of ecosystem, rather than investigating how different ant species affect different types of land cover within the same landscape. Our study focused on modifications to water infiltration and surface texture of Haplic Luvisols by two ant species—one of them present only in a forest and the other present only in a pasture. These two sites are located near each other in the Caatinga, a tropical dry forest region of Brazil, and have similar climatic conditions and soil type. We found that in the forest site, soils undisturbed by ant activity already showed water infiltration 13 times higher than the similarly undisturbed soils of the pasture site; and that the ant Dinoponera quadriceps more than doubled water infiltration in their nest annular zones compared with the surrounding forest areas. In addition, D. quadriceps increased clay content in their mound soils by 50% compared with the soils of the surrounding forest. In contrast, at the pasture site, the leaf-cutter ant Atta laevigata had a mixed effect on water infiltration: compared with pasture matrix soils, their nest-building activities led to a threefold increase in infiltration of the mound soils, but a threefold decrease in the nest annular zones. In addition, compared with the pasture matrix, A. laevigata almost doubled sand content in their mounds—which explains the comparatively high infiltration rates. These results suggest that the contrasting soil modifications by the two species could lead to amplifying feedbacks in opposite directions: an ant species that depends on healthy forests is improving water infiltration and soil texture (and thus increasing water available to plants) in its environment, while another ant species that is known to profit from disturbances is having the opposite effect.
