Amazonian rainforests once thought to hold an innate pristine wilderness, are increasingly known to have been densely inhabited by populations showing a diverse and complex cultural background prior to European arrival. To what extent these societies impacted their landscape is unclear. Amazonian Dark Earths (ADEs) are fertile soils found throughout the Amazon Basin, created by pre-Columbian societies as a result of more sedentary habits. Much is known of the chemistry of these soils, yet their zoology has been neglected. Hence, we characterised soil macroinvertebrate communities and activity in these soils at nine archaeological sites and adjacent reference soils in three Amazonian regions, totaling eighteen sampling sites. Furthermore, we characterized various soil chemical and physical properties associated with soil fertility.

The current dataset contains data on soil macroinvertebrate biodiversity (26 taxa), with a special focus on termites, earthworms and ants. It also contains data on soil macromorphology, bulk density and porosity, soil carbon, nitrogen, macro and micronutrients, magnetic susceptibility and apparent electrical conductivity. The results show similar overall macrofauna morphospecies richness in ADE and adjacent soils, but distinct communities in each soil type. They also show higher soil fauna activity in ADEs when compared to adjacent reference soils, associated with greater earthworm populations. Finally, they also confirm the well-known high soil fertility in ADEs compared with adjacent soils. Land use was an important determinant of both macrofauna biodiversity and soil fertility. These findings support the idea that humans have built and sustained a contrasting high fertility ecosystem that persisted until our days, altering biodiversity distribution patterns in Amazonia.

Soils were sampled in Brazilian Amazonia in the municipalities of Iranduba-AM, Belterra-PA and Porto Velho-RO. In each region, paired sites with anthropogenic dark earths (ADE) and nearby reference (REF) non-anthropogenic soils were sampled under three different land-use systems: native secondary vegetation (dense ombrophilous forest) classified as old secondary forest when >20 years old, or young regeneration forest when <20 years old, and agricultural systems (maize in Iranduba, soybean in Belterra, and introduced pasture in Porto Velho).

At each site, soil and litter macrofauna were collected using the Tropical Soil Biology and Fertility (TSBF) method (Anderson and Ingram, 1993) at five sampling points (soil monoliths 25x25 cm up to 30 cm depth) within a 1 ha plot, four at the corners and one on the center of a 60 x 60 m square, resulting in an “X” shaped sampling design. Each soil monolith was divided into surface litter and three 10 cm-thick soil layers (0-10, 10-20, 20-30 cm). Macroinvertebrates (animals with >2 mm body width) were manually hand-sorted and fixed in 92% ethanol. Earthworms, ants and termites were identified to species or genus level and other macroinvertebrates were sorted into morphospecies with higher taxonomic level assignations. Density (number of individuals) and biomass of the soil macrofauna surveyed using the TSBF method were extrapolated per square meter.

For the earthworms, ants and termites (ecosystem engineers) additional samples were performed, especially in forest sites to better estimate species richness of these taxa. Earthworms were collected at four additional cardinal points of the grid at all sites, and hand-sorted from holes of similar dimensions as the TSBF monoliths. Termites were sampled in the forest sites only forests (except one of the REF young forests at Porto Velho), in five 20 m² (2 x 10 m) plots (close to the five main soil monoliths) by manually digging the soil and looking for termitaria in the soil, as well as in the litter and on trees using a modification of the transect method (Jones and Eggleton, 2000). Ants were sampled in 10 pitfall traps (300 ml plastic cups) set up as two 5-trap transects on the sides of each 1 ha plot, as well as in two traps to the side of each TSBF monolith (distant ~5 m) only in the forest systems of Iranduba and Belterra (not Porto Velho). Each cup was filled to a third of its volume with water, salt and detergent solution. Termites and ants were preserved in 80% ethanol and earthworms in 96% ethanol and the alcohol changed after cleaning the samples within 24 h. All the animals (earthworms, ants, termites) were identified to species level or morphospecies level (with genus assignations) by co-authors SWJ/MLCB (earthworms), AA (termites) and ACF/RMF (ants).

Soil samples for chemical and particle size analysis were collected from each TSBF monolith after the soil fauna hand-sorting. Around 2 to 3 kg soil from each depth (0-10, 10-20, 20-30 cm) and the following soil properties were evaluated according to standard methodologies (Teixeira et al., 2017): pH (CaCl₂); Ca²⁺, Mg²⁺, Al³⁺ (KCl 1 mol L^-1); K⁺, P, Fe, Zn, Mn, Cu and Ni (Mehlich-1); Pseudo-total contents of trace elements (Ba, Cd, Co, Cu, Ni, Pb, Se and ZN) were determined by acid digestion (HNO₃ + HCl); Fe (sulfuric extract); total nitrogen (TN) and carbon (TC) using an element analyzer (CNHS). Base saturation and cation exchange capacity (CEC) were calculated using standard formulae (Teixeira et al., 2017) and particle size fractions (% sand, silt, clay) were obtained following standard methodologies (Teixeira et al., 2017). Soil magnetic susceptibility (MS) and apparent electrical conductivity (EC_a) (Siemens per meter – S m^-1) were obtained using a KT-10 S/C magnetic susceptibility/conductivity meter (Terraplus) with 10 Hz of operating frequency.

            Soil macromorphology samples were taken close to the TSBF monolith (~2 m) using a 10 x 10 x 10 cm metal frame. The collected material was separated into different fractions including: living invertebrates, litter, roots, pebbles, pottery sherds, charcoal (biochar), non-aggregated/loose soil (NA), physical aggregates (PA), root-associated aggregates (RA), and fauna-produced aggregates (FA) using the methodology proposed by Velasquez et al. (2007). Soil bulk density and total porosity were determined using undisturbed core samples (0.05 m diameter, 0.05 m depth) collected at ~2 m from the TSBF samples following the method proposed by Teixeira et al. (2017).

All data is provided in excel format, and includes 11 tabs in the data file: Metadata, Coordinates, Soil chem, BD+POR, Macromorph, Macro_den, Macro_bio, Morpho_TSBF, Add_worm, Add_ants, Add_termites. The Legend tab provides a detailed explanation for each variable included in each table, including the units used for each. The Metadata tab describes the main types/classes of variables included in each table. Soil chem, BD+POR and Macromorph tables contain the data about chemical and physical soil variables. The Macro_den and Macro_bio contain the data about density and biomass on all the soil invertebrate taxa found, respectively. The Morphosp_TSBF, Add_worm, Add_ants and Add_termites tables contain the invertebrate species/morphospecies occurrence in TBSF and extra samples for earthworms, ants and termites, respectively.

References used in methods section

Anderson, J.M., Ingram, J.S.I., 1993. Tropical Soil Biology and Fertility: A handbook of methods, 2 edition. ed. Oxford University Press, Oxford.

Jones, D.T., Eggleton, P., 2000. Sampling termite assemblages in tropical forests: testing a rapid biodiversity assessment protocol. Journal of Animal Ecology 37, 191–203.

Teixeira, P.C., Donagemma, G.K., Fontana, A., Teixeira, W.G., 2017. Manual de métodos de análise de solo, 3^o. ed. Embrapa, Brasília.

Velasquez, E., Pelosi, C., Brunet, D., Grimaldi, M., Martins, M., Rendeiro, A.C., Barrios, E., Lavelle, P., 2007. This ped is my ped: Visual separation and near infrared spectra allow determination of the origins of soil macroaggregates. Pedobiologia 51, 75–87. doi:10.1016/j.pedobi.2007.01.002