Pedogeomorphology and paleoenvironmental implications of large termite mounds at the Brazilian semiarid landscape

Abstract

Large termite mounds, regionally called Murundus, forming hillocks throughout the Brazilian territory, occur, with different names, across the entire tropical belt. Little is known about the genesis and age of these large biogenic structures in the semiarid zone, where termites have a limited range due to water deficit. We aimed to understand the process of mounds building on different types of soils at the ecotonal semiarid region of Minas Gerais state. In addition, we tested the hypothesis that these mounds have been produced by termites during the Holocene, and that the paleoenvironmental conditions during mound formation were different from today, helping to enhance the understanding of vegetation dynamics in the semiarid Caatinga Biome. For this, five mounds with and without living termites (M1, M2, M3, M4 and M5) and five adjacent soil profiles (P1, P2, P3, P4 and P5) were selected for morphological description, followed by physical, chemical, mineralogical and micromorphological analysis, as well as δ13C and 14C dating analysis. The occurrence of these large termite mounds in the semiarid landscape is unrelated to soils or parent material, and represent the products of cumulative termite activity in the edge of the semiarid domain. Mounds and adjacent soils formed after the mid Holocene (younger than 4000 years), based on 14C dating, and may be even older. The calculated rates of termite accretion (based on 14C dating) are consistent with previous reports of tropical soils, in the range of 0.027 to 0.066 cm soil/Ha/y. A comparative estimate based on mound volume and density was 0.188 cm soil/Ha/y, up to 3.5 times greater than the previous approach based on 14C dating. These figures suggest the production of a soil mantle in the range of 270–660 cm in 10,000 years, accounting for erosion. The mounds and adjacent soils have differences in micropedological features, notably related to bioturbation (excremental infillings), and contrasting textural variations with depth, with higher and more uniform clay and silt contents in mounds, and a kaolinitic/illitic mineralogical composition throughout. The chemical fertility of murundus depends on the type of underlying soil from which it was built, and does not increase in all cases. In the murundus studied, the microgranular structure typical of oxic materials is welded and massive, indicating that the natural compaction of soils dominated by kaolinite/illite under the semi-arid climate prevents the stability of loose aggregates formed by termite bioturbation, reorganizing the termitic microaggregate groundmass. This clearly shows the polycyclic (polyphasic) biogenic nature of mounds and adjacent soils, corroborated by carbon isotopes studies. Soils and mounds are younger than mid Holocene in the first 100 cm, and experienced alternations of dry and wet climate spells from the mid-Holocene onwards, resulting in apparent changes in vegetation at the ecotonal, southern limit of semi-arid Caatinga, over the last 5000 years. Since we did not reach the base of mounds, they should be even older; mounds represent fine proxies for Holocene paleoclimate reconstructions, if a more detailed pedostratigraphic study is carried out. A general model of termite mounds formation in relation to the interconnected pedological, geomorphological and vegetational evolution is presented, summarizing all present knowledge on these outstanding biogenic landforms.