Sampling effort and information quality provided by rare and common species in estimating assemblage structure

Abstract

Reliable biological assessments are essential to answer ecological and management questions but require well-designed studies and representative sample sizes. However, large sampling effort is rarely possible, because it demands large financial resources and time, restricting the number of sites sampled, the duration of the study and the sampling effort at each site. In this context, we need methods and protocols allowing cost-effective surveys that would, consequently, increase the knowledge about how biodiversity is distributed in space and time. Here, we assessed the minimal sampling effort required to correctly estimate the assemblage structure of stream insects sampled in near-pristine boreal and subtropical regions. We used five methods grouped into two different approaches. The first approach consisted of the removal of individuals 1) randomly or 2) based on a count threshold. The second approach consisted of simplification in terms of 1) sequential removal from rare to common species; 2) sequential removal from common to rare species; and 3) random species removal. The reliability of the methods was assessed using Procrustes analysis, which indicated the correlation between a reduced matrix (after removal of individuals or species) and the complete matrix. In many cases, we found a strong relationship between ordination patterns derived from presence/absence data (the extreme count threshold of a single individual) and those patterns derived from abundance data. Also, major multivariate patterns derived from the complete data matrices were retained even after the random removal of more than half of the individuals. Procrustes correlation was generally high (>0.8), even with the removal of 50% of the species. Removal of common species produced lower correlation than removal of rare species, indicating higher importance of the former to estimate resemblance between assemblages. Thus, we conclude that sampling designs can be optimized by reducing the sampling effort at a site. We recommend that such efforts saved should be redirected to increase the number of sites studied and the duration of the studies, which is essential to encompass larger spatial, temporal and environmental extents, and increase our knowledge of biodiversity.