Response of dung beetle diversity to remediation of soil ecosystems in the Ecuadorian Amazon

Ethics statement

This study was authorized under research permit 016-2018-IC-DPAO/AVS and authorization for the mobilization of specimens 005-2019-MOV-FAU-DPAO/AVS, both issued by the Ministerio del Ambiente del Ecuador.

Study area

The study was performed in the Sucumbíos (0°5′S, 76°53′W) and Orellana (0°56′S, 75°40′W) provinces in the Ecuadorian Amazon (Fig. 1). The land use of the localities is distributed in 70% for Natural Forest, 22% for crops and pastures, and 8% for urban and industrialized areas (Ministerio de Agricultura y Ganadería, 2015). The predominant climate is Humid Tropical Rainforest according to the Holdridge climate classification (Holdridge, 1967), with a mean annual temperature of 24.5 °C and heavy rainfall throughout the year (4,132 mm). Meteorological data from an automatic weather station (0°20′S, 76°52′W) located ~22 km from the most distant sampling site were used to compute monthly values of precipitation, temperature, and relative humidity (Table 1).

The paired-site sampling per ecosystem type were established 1 year after remediation process has been completed in ecosystems previously affected by hydrocarbon activities (Sensitive Ecosystems and Agricultural Soils). In addition, two types of non-contaminated soil ecosystems were included as controls (Natural Forests and Palm Plantations). As described earlier, Sensitive Ecosystems had lower pollutant levels than the Agricultural Soils (Table 2). Further, our Sensitive Ecosystems were beside fragmented Natural Forests and were revegetated with native trees as described by Villacís et al. (2016a) (Fig. 2A), whereas one of the Agricultural Soils was beside a Palm Plantation and three were in deforested areas (probably to be used later for cattle raising or cropland areas) (Fig. 2B). We delimited four sites (1 ha) into fragmented Natural Forests of around 400 ha (Fig. 2C) and delimited four sites (1 ha) into Palm Plantations (8 years) of the Instituto Nacional de Investigaciones Agropecuarias (Fig. 2D), both without a history of oil spills or remediation activities.

Sampling design and dung beetle trapping

Sixteen sample sites (four per ecosystem type) were established within a ~30 km radius (Fig. 1). In the Natural Forests and the Palm Plantations, the sampling sites (1 ha, 100 m × 100 m) were established 200 m from the border, whereas the remediated ecosystems had irregular geometry but the sampling area was also 1 ha. Before dung beetle sampling, a soil sample from the layer 0.0–0.2 m was collected next to each pitfall trap and later mixed to form a composite soil sample per site (16 in total). To determine the total petroleum hydrocarbon and polycyclic aromatic hydrocarbon concentrations (Table 2) GC2 014 gas chromatographs were used (Shimadzu Scientific Instruments, Inc., Kyoto, Japan). In addition, the concentration of cadmium (Cd), nickel (Ni), and lead (Pb) in soils (Table 2) was determined by using atomic absorption spectrometry (AA-6800; Shimadzu Scientific Instruments, Inc., Columbia, MD, USA) as indicated by the EPA SW-846 method (Le Blanc & Majors, 2001). The analyses were performed at the Soil Laboratory of the Universidad de las Fuerzas Armadas, Ecuador.

In each sampling site, six pitfall traps were baited with pig dung and installed in the center of each sampling site separated 10 m from each other. Pitfall traps consisted of plastic containers of 0.8 L (15 cm depth × 10 cm diameter) buried up to their rims in the soil and containing solution 50:50 of water with alcohol. The traps remained active at the sites for 120 consecutive h per month during 1 year (February 2018 to January 2019). The amount of dung per trap was ~50 g and was replaced every 24–36 h. In April 2019, the traps were not evaluated due to conflict with the landowners, who prevented the entrance to the sampling sites in La Joya de los Sachas locality.

Dung beetles were preserved in 70% ethanol and identified to species-level using dichotomous keys (Chamorro et al., 2018; Vaz-De-Mello et al., 2011) and voucher specimens from the Museo de Historia Natural “Gustavo Orcés” (Escuela Politécnica Nacional, Quito, Ecuador). Some specimens were pinned and deposited in the Museum of Zoological Researches (Universidad de las Fuerzas Armadas, Sangolquí, Ecuador).

Data analysis

All analyses were performed using the software INFOSTAT (Di Rienzo et al., 2008) in interface with R (R Core Team, 2013). To assess the sampling effort, species accumulation curves were created using the Clench model, which estimates the probability of finding new species as field sampling effort increases (Clench, 1979; Soberon & Llorente, 1991). In addition, the richness observed in each type of ecosystem was evaluated using the non-parametric estimator Chao 1, which is an estimator of the number of species in a community based on the number of rare species in the sample (Chao, 1984; Colwell & Elsensohn, 2014).

Abundance, richness, and the Shannon diversity index were estimated for the four studied ecosystems using pooled data per month and sampling site as recommended by Magurran (1998). The Shannon diversity index takes into account the proportion of each species in an ecosystem studied and has been postulated as a unifying measure for the partitioning of diversity at multiple levels (Konopiński, 2020).

Abundance, richness, and diversity were also analyzed using repeated measures (by month). Differences between ecosystems were analyzed using analysis of variance with mixed models for a complete randomized design with four replications. In addition, we performed orthogonal contrasts. The first contrast evaluated differences in abundance, richness, and diversity between remediated soil ecosystems and non-contaminated soil ecosystems. The second contrast evaluated the differences between Agricultural Soils and Palm Plantations, and the third contrast evaluated the differences between Sensitive Ecosystems and Natural Forests. Furthermore, we tested differences between ecosystems, months, and interactions by using the post hoc test of Di Rienzo, Guzmán and Casanoves (DGC) (P < 0.05). The normality of the data was verified using the Shapiro-Wilks test, and the homoscedasticity was modeled using independent variances.

The Sørensen index was used to compare the similarity of dung beetle species composition between each type of ecosystem evaluated. Finally, a dendrogram was prepared using this information (Beals, 1984).

Diversity of dung beetles

We collected 7,506 individual beetles of the Scarabaeinae subfamily, belonging to 13 genera and 37 species (Table 3). Abundance varied greatly, ranging from one to 1,502 individuals (an average of 202.9 individuals ± 52.8 SE per species). Canthon aequinoctialis (20% of total abundance), Ontherus sulcator (13%), Dichotomius ohausi (10%), and Deltochilum howdeni (9%), accounted for 52% of all individuals collected. Fifty-one percent of the species were classified as rare, with a relative frequency of less than ten percent. Twenty-two percent of the total species collected were found in all four evaluated ecosystems, while 12 others were exclusive to one of them: 11 in the Natural Forest and one in the Sensitive Ecosystem. Canthidium aurifex, Eurysternus atrosericus, Ontherus sulcatur, Onthophagus osculatii, and O. nyctopus are new provincial records, whereas O. hircus is a new record for Ecuador (Table 3).

Difference in diversity between habitats over sampling time

As the sampling time increased, the species accumulation curve stabilized after the 7th month in the Natural Forests, whereas in the Sensitive Ecosystems, Agricultural Soils and Palm Plantations, the slope did not reach values close to 0 (Fig. 3). The non-parametric estimator of richness (Chao 1) showed that 99% of the species in Natural Forest, 97.63% in Sensitive Ecosystems, 94.34% in Agricultural Soils, and 87.35% in Palm Plantations were recorded. Cluster analysis showed that Agricultural Soils and Palm Plantations presented a species similarity of 32.4%, and both were similar to the Sensitive Ecosystems at 26.8%. The Natural Forests were similar to other ecosystems only at 8.53% (Fig. 4).

The average values of abundance, richness, and the Shannon index differed between ecosystems within each month (there was a significant statistical interaction ecosystem × months; P = 0.0073; Table 4). The orthogonal contrast showed that non-contaminated soil ecosystems contained higher abundance, richness and diversity of beetles in comparison to remediated soil ecosystems (F_{1, 132} = 313.51, P < 0.0001). Natural Forest presented higher abundance, richness and diversity than Sensitive Ecosystems (F_{1, 132} = 313.51, P < 0.0001) and Palm Plantations presented higher abundance, richness and diversity than Agricultural Soils (F_{1, 132} = 51.60, P < 0.0001; Fig. 5).

The highest monthly average values for abundance (January and November), richness (January, February, September, October and November) and Shannon’s index (September and November) were recorded in the Natural Forests (Fig. 6). Meanwhile the average values of abundance, richness and diversity for Sensitive Ecosystems, Palm Plantations and Agricultural Soils were lower than in the Natural Forests, showing little variation during the 11 months of sampling.

Dung beetle diversity

The species presented in this study represent 17% of the 220 species of dung beetles registered in Ecuador (Chamorro et al., 2018) and more than 50% of previously registered species in the Orellana and Sucumbíos provinces. Five species are new for these provinces, whereas Ontophagus hircus was recorded for the first time in Ecuador. This demonstrates that beetle diversity must be studied not only to understand the effects of ecosystem disturbance but also to complete the species inventory of the Ecuadorian Amazon.

The stabilization of the species accumulation curves in the Natural Forest supports the registration of almost all species in that ecosystem. As for the other ecosystems (Sensitive Ecosystems, Agricultural Soils, and Palm Plantations), the slope of the curves suggests that we did not sample the total species richness of dung beetle communities (Gering, Crist & Veech, 2003).

The community structure in the non-contaminated soil ecosystems tended toward high abundance, richness, and diversity when compared to the remediated soil ecosystems, indicating that the ecosystems has likely not fully recovered. According to a recent study, dung beetles bioaccumulate more Pb than other insects such as cicada and longicorn (Zhou et al., 2019). In another Coleoptera family, the Carabidae, modifications of the digestive tract were reported in individuals exposed to concentrations of 10 mg kg⁻¹ Pb (Conti et al., 2017). In addition, the presence of heavy metals in soils negatively influences small mammals (Richardson et al., 2015), which decreases food sources for dung beetles. These factors may negatively affect dung beetle diversity in hydrocarbon-degraded ecosystems, but further research should be carried out to elucidate the effects of other oil pollutants found in the studied ecosystems on these organisms.

Similarly, cluster analyses showed an acute division between the Natural Forests and other ecosystems evaluated. This could be because Natural Forest fragments within human modified landscapes constitute wildlife refuges (Blaum et al., 2009). This trend of decreased dung beetle diversity between the Natural Forest and the other ecosystem types follows a general decreasing gradient of diversity and an increase in anthropomorphic disturbances due to contamination and land use (Aninta et al., 2019; McCain, 2005; Stevens & Gavilanez, 2015). For example, Scarabaeinae diversity in Palm Plantations was similar to that found in Agricultural Soils. This is consistent with previous studies of Fitzherbert et al. (2008) and Harada et al. (2020), which reported that agrochemicals could favor the degradation of soil and nutrients and hence diminish dung beetle diversity.

Habitat and temporal variation of dung beetle diversity

Dung beetles are very sensitive to habitat disturbance (Audino, Louzada & Comita, 2014; Campos & Hernández, 2015; Da Silva, Vaz-de-Mello & Di Mare, 2013). For example, changes in dung beetle community structure have been reported under low vegetation cover (Nichols et al., 2007) as a result of intense solar radiation on the soil surface, which accelerates the decomposition rate of food sources (Méndez et al., 2019). Moreover, chemical perturbations affect dung beetle communities in several ways such as changes in composition, diversity and population (Hutton & Giller, 2003; Correa et al., 2022). Therefore, the presence of hydrocarbons and heavy metals in Agricultural Soils and Sensitive Ecosystems as well as landscape modifications at oil exploitation sites could influence abundance, richness, and diversity of dung beetle assemblages.

The diversity of dung beetles is determined by regional rainfall patterns (Novais et al., 2016). Our results indicated that the diversity of dung beetles in the Natural Forests was higher during the month with lower levels of precipitation (October 238 mm month⁻¹) which is consistent with the study of Ibarra-Polesel, Damborsky & Porcel (2015), which investigated dung beetles in subtropical ecosystems. However, in contrast, several other studies have demonstrated that higher beetle diversity is linked to the months with elevated values of precipitation (Escobar et al., 2008; Nunes et al., 2016; Rangel-Acosta & Martínez-Hernández, 2017).

We may have found higher dung beetle diversity during the months with the lowest levels of rainfall due to the interference of other environmental factors. For example, alteration of microclimates and microhabitats (Medina, Escobar & Kattan, 2002; Noriega & Realpe, 2018; Sánchez-Hernández et al., 2022), changes in trophic structure (Novais et al., 2016), as well as altitudinal gradient effects (Noriega & Realpe, 2018) could have affected the mobility, displacement, and genetic flow of organisms between ecosystems (Harvey, Gonzalez & Somarriba, 2006). The Natural Forests in the provinces where the study was conducted are under a higher degree of environmental disturbance (deforestation, forest fragmentation, oil spills, population growth, etc.) (Rivera-Parra et al., 2020) and, in general, the entire Amazon basin is under massive degradation due to deforestation (Marin et al., 2022).

Implications for the conservation

Our study provides the first quantitative data on dung beetle communities in ecosystems affected by hydrocarbon activities in the Ecuadorian Amazon. The diversity of dung beetles provides a useful tool for assessing the temporary status of remediated sites previously affected by hydrocarbon activities. Although differences in beetle diversity were found between remediated ecosystems, similar recommendations for conservation measures can be made for both Agricultural Soils and Sensitive Ecosystems. Therefore, Agricultural Soils and Sensitive Ecosystems should not only be based on the levels of hydrocarbons and heavy metals but also on the diversity of bioindicators.

The presence of dung beetles in remediated ecosystems provides a baseline for implementing strategies to improve the existing diversity. However, the conservation of biodiversity in remediated ecosystems depends not only on remediation activities, but also on other anthropogenic activities in the Amazonian tropical forests (Rivera-Parra et al., 2020).