Please use this identifier to cite or link to this item:
Title: Geographic position of sample grid and removal of uncommon species affect multivariate analyses of diverse assemblages: The case of oribatid mites (Acari: Oribatida)
Authors: Franklin, E.
Moraes, Jamile de
Landeiro, Victor Lemes
Souza, Jorge Luiz Pereira
Pequeno, Pedro Aurélio Costa Lima
Magnusson, William Ernest
Morais, José Wellington
Keywords: Biodiversity Surrogates
Distribution Patterns
Grid Position
Representative Species
Species Discarding
Cost Effectiveness
Matrix Algebra
Community Composition
Cost-benefit Analysis
Data Set
Environmental Change
Habitat Conservation
Habitat Loss
Multivariate Analysis
Population Distribution
Cost Effectiveness
Species Identification
Issue Date: 2013
metadata.dc.publisher.journal: Ecological Indicators
metadata.dc.relation.ispartof: Volume 34, Pags. 172-180
Abstract: In view of the rapid loss of biodiversity, large-scale environmental monitoring programs are urgently needed, over a range of local, regional and global scales. These programs can be made more efficient and cost-effective through shortcuts such as reduction of sampling effort and the use of low-cost surrogates. We revisited a large-scale dataset composed by 161 species recorded in 72 plots of 250 m, distributed over an 8 m × 8 m sampling grid in the tropical rainforest. Samples of litter and soil were collected and oribatid mites were extracted with a Berlese-Tullgren apparatus. Using a "moving window" procedure, we delimited smaller 5 km × 5 km grids in 16 possible positions within the larger grid. We first evaluated which fraction was more important to explain environmental and spatial patterns in the species composition: known environmental or spatial filters representing unknown causes of aggregation, and the confounded variance that might be associated with either or both. We used soil clay content, litter quantity, soil pH, number of trees, and distance to the nearest stream as environmental predictors. The spatial filters were generated using Moran Eigenvector mapping through the Principal Coordinates of Neighbor Matrices technique. To evaluate the influence of these fractions on the species composition, we used partial Redundancy Analysis. Using Principal Coordinates Analysis for abundance and presence/absence data, we evaluated if reduced matrices, discarding sets of less-frequent species, could identify the relationships captured with the complete dataset. All smaller grids contained more than 100 species. The effect of environmental variables on oribatid-mite community composition was always low, and each smaller grid position produced different results. Soil clay content and pH were the main factors associated with oribatid-mite distributions. The effects of unknown spatial patterns were greater than the environmental ones. Independently of the grid position, similar results were obtained for analyses with all oribatid-mite species, to the results obtained from analyses of only the most frequent species. Sets of more frequent and easily identifiable species proved to be a reliable surrogate for the complete assemblage. Omitting identifications of most species will improve the cost-effectiveness of monitoring programs. More emphasis should be placed on investigating the role of spatial heterogeneity and the effects of grid position in relation to patterns in species associations. Efficient biomonitoring could target surrogate species, to enable rapid tracking of environmental change while enlarging the sampling area to provide data for conservation strategies. © 2013 Elsevier Ltd. All rights reserved.
metadata.dc.identifier.doi: 10.1016/j.ecolind.2013.04.024
Appears in Collections:Artigos

Files in This Item:
There are no files associated with this item.

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.