Fatores ambientais na distribuição espacial de assembleias de colêmbolos (Collembola: Entomobryomorpha) edáficos e redução do esforço amostral na Amazônia

Abstract

The destruction of natural landscapes has intensified ecological and conservation studies of soil fauna. Arthropods are often used as a model for studies on integrity and functioning of terrestrial ecosystems. Using two collection methods, we investigated the spatial distribution pattern and the sample sufficiency of edaphic springtails to create one collection protocol in different vegetation types in the Amazon. Standardized samples were taken by RAPELD system in the years 2011, 2012 and 2013 in three forests in the Amazon. In each area, 30 plots of 250 m, spaced 1 km from each other were used. The depth of the litte, the clay and phosphorus contente were used to assess the ecological relationships with the community of springtails. In Reserva Ducke the inventory was carried out in 2011 using the Berlese-Tullgren apparatus. In total, 4,126 individuals, 28 species of three families springtails were recorded. The reduction in sampling effort by the similarity with the maximum effort composition in reduced effort was efficient (rP> 0.6), but the maintenance of ecological responses was not showed because the the ecological relathionship captured with the maximum effort was not recovered in reduced efforts. Therefore, the reduction of the sampling effort-with the Berlese method was possible only for research taxonomic aspects. In 2012 we conducted a survey in Maracá Ecological Station with two collection methods, Berlese-Tullgren and pitfall and evaluated the cost- benefit of these two techniques. Combining the methods, there were 4,453 individuals and 15 species. The methods, Berlese-Tullgren and Pitfall, detect different species compositions without overlapping information. However, using only Pitfall, the ecological pattern captured with the two combined methods is recovered. We concluded that the use of only one method of collection is viable, enabling the pitfall as an effective method for sampling springtails. In the years 2012 and 2013 we carried out surveys in Ducke, Maraca and Viruá National Park using Pitfall. In total, 10,295 springtails and 20 species were recorded in the three forests. The community of springtails edaphic was influenced by environmental factors in the three areas. The species composition changes according to the litter gradient in Ducke and Maraca with the clay content in the Ducke and Viruá, and the fóforo content of the soil in Viruá. The reduction from 10 to five subsamples per plot was efficient to sample the three areas, since the reduced effort recorded high similarity in species composition and recovered the ecological pattern recorded with the maximum effort. We concluded that the reduction of subsamples was effective with Pitfall in three areas in the Amazon. Reduction of sampling effort provides the protocol simplification that require less effort in the field and laboratory in terms of time and cost, contributing with data to the maintenance of biodiversity.