Soil macrofauna as an indicator of soil quality in an undisturbed riparian forest and recovering sites of different ages

Large expanses of riparian vegetation have been destroyed in Brazil, and few studies have focused on biological indicators of soil quality in Brazilian riparian ecosystems. The aim of this study was to assess relationships between the soil fauna and physical, chemical, and biological soil attributes in summer and winter, and to identify the indicators that most effectively distinguish between four study sites in Brazilian Seasonal Semideciduous Forest with different ages: one undisturbed site (NT), and three that have been recovering from disturbance for 5 (R05), 10 (R10), and 20 (R20) years. At each site we overlaid a grid with 30 intersections and selected 15 of them at random for soil sampling (n = 15 points per site). Physical, chemical, and biological (microbiological and soil macrofauna) analyses were performed on these samples. The soil fauna was sampled by two methods (pitfall traps and soil monoliths measuring 25 × 25 × 20 cm) and macrofaunal diversity quantified as richness, cumulated richness, and with Shannon's and Pielou's indices. Patterns in macrofaunal distribution were explored with the parallel discrimination rate and partial redundancy analysis (pRDA) to determine how physical, chemical, and biological attributes affect these patterns. We observed lower Shannon's and Pielou's indices in the winter, for both sampling methods, especially at the most recently recovered site. For monoliths, Hymenoptera were more abundant in the winter and for pitfall traps Mollusca and Chilopoda were only found during summer. Richness, Shannon's and Pielou's indices were only higher at the oldest sites (native and 20 years old) with the monolith data collected in winter. The parallel discrimination rate revealed significant separation between sites, sampling methods, and seasons and indicated Hymenoptera, Diplopoda, Coleoptera, and predators such as Araneae and Opiliones as the principal discriminatory groups. Soil chemistry explained the largest proportion of the pRDA inertia, followed by microbiological and physical attributes. Litter and soil nitrogen content, macroporosity, litter humidity, and microbiological activity and biomass were related most strongly with saprophytic groups and with the oldest sites. These results suggest that organic matter quality and physical soil attributes, in addition to management practices, can accelerate the forest recovery process.