Effect of field sampling design on variation partitioning in a dendritic stream network

Abstract

16Variation partitioning is one of the most frequently used method to infer the importance of 17 environmental (niche based) and spatial (dispersal) processes in metacommunity structuring. 18However, the reliability of the method in predicting the role of the major structuring forces is 19 less known. We studied the effect of field sampling design on the result of variation 20 partitioning of fish assemblages in a stream network. Along with four different sample sizes, a 21 simple random sampling from a total of 115 stream segments (sampling objects) was applied 22 in 400 iterations, and community variation of each random sample was partitioned into four 23 another property of sampling design, is the total number of sampling units in the sample, and 50 it is a simple measure of the sampling effort. An additional property is the topology of the 51 sampling units. Topology describes the geometry by which the sampling units are ecologically 52 connected to each other. When sampling units considered being connected, researchers 53 assume that material and individuals can move from one sampling unit to the other one (e.g., 54 Peterson et al. 2013). 55 56 Variation partitioning 57Ecologists try to reveal the mechanisms controlling the distribution of organisms by 58 investigating their spatial distributional patterns. One of the most frequently used statistical 59 methods for quantifying different sources of variation of communities is variation partitioning 60 (or variance partitioning), which was introduced into the ecological methodology by Borcard 61 et al. (1992). In a classical approach, this method uses a sites-by-species community matrix as 62 response data, and a sites-by-environmental variables matrix and a sites-by-spatial variables 63 4 matrix as explanatory data to decompose additively the total variation of the response data 64 into four variance fractions/proportions by fitting canonical ordination models (canonical 65 correspondence analysis [CCA] or redundancy analysis [RDA]) on the data.

Effect of landscape context on fish metacommunity structuring in stream networks

“…HLS, LLR). Nevertheless, it is possible that the EMS approach is just as sensitive to the spatial design of the study as to the method of variance partitioning of environmental and spatial fractions in the metacommunity context (see Sály & Erős, ). Consequently, while empirical studies on the structure of metacommunities are accumulating rapidly, we suggest that simulation‐based studies explore the sensitivity of the EMS analysis to the properties of the sampling design.…”

“…Note, that we did not want to use variance partitioning to quantify the effect of spatial variables and hereby getting inferences on the effects of dispersal and/or drift (neutral stochasticity) mechanisms, since several studies (see Tuomisto & Ruokolainen, ; Gilbert & Bennett, ; Smith & Lundholm, ; Vellend et al ., for details), and our own simulations, specifically developed for stream networks (Sály & Erős, ), show that this method is not really reliably for this purpose. We believe that with the examination of DDRs (see below), we could more reliable quantify the importance of spatial distance on community composition and abundance and hereby getting some inferences on the role of dispersal limitation (Tuomisto & Ruokolainen, ).…”

A comprehensive examination of the network position hypothesis across multiple river metacommunities

“…To account for differences in sample sizes (Sály and Erős 2016), we included the total number of sites (total_SITES) as well as the total number of headwaters (total_HEADW) and downstream sites (total_RIVER) from each catchment. We also included catchment surface (SA) area as the relative importance of assembly processes may vary with spatial extent (Heino et al 2015) and we computed their latitude (LAT) and longitude (LONG) as the average latitude and longitude of their constituent sites, respectively.…”

Ecological drivers of fish metacommunities: Environmental and spatial factors surpass predation in structuring metacommunities of intermittent rivers

“…Sály and Erös (2016) demonstrated that variation partitioning analysis is influenced by the number of sampling sites. Sály and Erös (2016) demonstrated that variation partitioning analysis is influenced by the number of sampling sites.…”