Dam removal has become an increasingly popular river restoration strategy with the aim of reestablishing more natural hydrologic and sediment regimes and increasing habitat availability for aquatic organisms. However, dam-removal research is in its early stages and the ecological impacts are not fully known. For fish, the removal of lowhead dams has been shown to restore hydrologic connectivity and reestablish lotic-type species; yet the magnitudes and rates of recovery are variable across studies. Food webs represent complex trophic networks and are fundamental in understanding the organization of natural assemblages and community dynamics and in describing ecosystem processes. Therefore, incorporating a food-web approach in dam removal studies might be expected to be an important step in quantifying ecosystem responses to dam removal. Here, I investigated the consequences of two lowhead dam removals in the Olentangy and Scioto Rivers (Columbus, Ohio) for (1) fish assemblage structure and (2) fish-centered food webs 2-3 years following removal.
The disturbance of dam removal had consequences for both fish assemblages and food-web structure. Upstream fish assemblages experienced a significant decline in species richness in the year following dam removal, which was accompanied by a shift in assemblage composition (ANOSIM: R = 0.714, p = 0.001). In the Olentangy River, assemblage structure shifted significantly over time (year 1-2) and included the re-colonization of benthic insectivores (ANOSIM: R = 0.136, p = 0.019). Species richness increased over time in both the Olentangy and Scioto Rivers (linear mixed models: OR – F2,16= 9.70, p = 0.002; F2,12 = 26.50, SR – p < 0.0001). Species richness returned to pre-dam removal levels in the Scioto River, but was still 58% lower than pre-dam removal levels in the Olentangy River upstream actively restored reach (i.e., where extensive channel modifications occurred) and in other experimental reaches.
Shifts in assemblage composition and species richness were not accompanied by changes in the relative abundance of top predators, intermediate, or basal species; however, these metrics might not capture important shifts in trophic guilds that could have implications for other food-web properties. Similar to species richness, network connectivity (i.e., link density, connectance) decreased significantly at the Olentangy River upstream actively restored reach following dam removal (paired t-test: connectance – t = -6.46, p = 0.023; link density – t = -13.54, p = 0.005) and did not return to pre-dam removal levels. Whereas link properties decreased on the Scioto River experimental reaches in the year following dam removal, they increased to pre-dam removal levels by second year (linear mixed models: connectance – F2,12 = 8.38, p = 0.005; link density – F2,12 = 42.26, p < 0.0001). Food-chain length (FCL) – a key measure of food-web architecture – declined from 4.37 to 2.75 in the Olentangy River actively restored reach, and declined significantly in the upstream control and passively restored reaches over time (linear mixed model: F2,14.7 = 7.43, p = 0.003). FCL also decreased in the upstream reach of the Scioto River by the second year after dam removal (linear mixed model: F1,6.4 = 28.93, p = 0.001; Tukey’s HSD: p = 0.0004). Significant reductions in channel width (i.e., ecosystem size) and alterations to hydrological connectivity could be linked to observed shifts in food-web architecture. Changes in ecosystem size could have consequences for individual traits of fish species (e.g., body size) or other food-web properties (e.g., connectance) that serve as proximate mechanisms regulating FCL. For example, FCL increased linearly with body length (linear regression: R2 = 0.221, p = 0.043) meaning that larger-bodied taxa held higher trophic positions. Larger habitats tend to support larger organisms and, thus, longer food chains (Power, 2002). Therefore, individual traits could be important mechanisms that help to explain the relationship between environmental drivers of food-web structure.
The results of this investigation have both applied and theoretical implications for understanding the role of disturbance in structuring fish assemblages and food webs in river systems. More specifically, dam-removal disturbance appeared to strongly influence network stability via a loss of diversity and trophic connections on the Olentangy River. Yet, the rebound of food-web properties on the Scioto River by the second year following dam removal suggests that food-web networks in larger river systems could be more resilient than in smaller systems. In the first few years following dam removal, active channel restoration did not enhance fish community diversity or FCL and exhibited a less complex fish-based food web than in a passively restored reach, although the benefits of this restoration may take additional time to be realized. The number of small dam removal projects is expected to grow considerably in the coming years, and increasing our understanding of the consequences of dam removal for the ecological integrity of river ecosystems is critical for effective river conservation and management.