Metacommunity theory meets restoration: isolation may mediate how ecological communities respond to stream restoration

Abstract

An often-cited benefit of river restoration is an increase in biodiversity or shift in composition to more desirable taxa. Yet, hard manipulations of habitat structure often fail to elicit a significant response in terms of biodiversity patterns. In contrast to conventional wisdom, the dispersal of organisms may have as large an influence on biodiversity patterns as environmental conditions. This influence of dispersal may be particularly influential in river networks that are linear branching, or dendritic, and thus constrain most dispersal to the river corridor. As such, some locations in river networks, such as isolated headwaters, are expected to respond less to environmental factors and less by dispersal than more well-connected downstream reaches. We applied this metacommunity framework to study how restoration drives biodiversity patterns in river networks. By comparing assemblage structure in headwater vs. more well-connected mainstem sites, we learned that headwater restoration efforts supported higher biodiversity and exhibited more stable ecological communities compared with adjacent, unrestored reaches. Such differences were not evident in mainstem reaches. Consistent with theory and mounting empirical evidence, we attribute this finding to a relatively higher influence of dispersal-driven factors on assemblage structure in more well-connected, higher order reaches. An implication of this work is that, if biodiversity is to be a goal of restoration activity, such local manipulations of habitat should elicit a more profound response in small, isolated streams than in larger downstream reaches. These results offer another significant finding supporting the notion that restoration activity cannot proceed in isolation of larger-scale, catchment-level degradation.

Erratum

“…We used a mixed model (R function lme ) to perform a repeated measures ANOVA with season, network position (i.e., order), and restoration state (i.e., reach) as fixed factors and site as a random factor. The results of the repeated measures ANOVA on local diversity were similar to those presented in Swan and Brown (). In the original paper, we report a difference in local diversity between restored and unrestored reaches in HW sites, but no difference between reaches at MS sites (Table ).…”

“…Here, we report results for the local diversity and temporal variability analyses based on the correct data set. Further, we discuss the implications of the new results relative to those reported in Swan and Brown ().…”

“…In Swan and Brown (), we presented benthic invertebrate community responses to restoration activity at different points in a river network. We tested the hypothesis that community response to restoration would be stronger in headwater (HW) streams vs. larger mainstem (MS) streams.…”

“…Post‐publication we discovered a data handling error that affected 2 analyses: local diversity (Swan and Brown : Table 3, Figure 3) and temporal variability (Swan and Brown : Table 3, Figure 4). The error occurred because these two analyses had been performed on a preliminary dataset that had not undergone appropriate quality control (QC).…”

“…Because the hypothesis being tested was that manipulation of local in‐stream environment would have larger effects on benthic communities in HW than in MS, we decided that sites with only riparian modification were not an appropriate test of the hypothesis. The error did not affect the spatial dissimilarity analysis (Swan and Brown : Table 3, Figure 5) because it was conducted at a later date on the final, correct data set, which was uploaded along with the final paper at the time of publication (https://portal.lternet.edu/nis/mapbrowse?packageid=knb-lter-bes.3180.110). Here, we report results for the local diversity and temporal variability analyses based on the correct data set.…”