Several theories have been proposed to explain the functioning of river systems, predicting changes in communities and the architecture of food webs along the longitudinal gradients (e.g., Vannote et al., 1980; Junk et al., 1989; Thorp et al., 2008). One of the earliest theories, known as the River Continuum Concept (RCC) (Vannote et al., 1980), predicts longitudinal gradients in community composition, food webs, and ecosystem-scale processes from headwaters to mouth. Evidence from different climatic regions shows increasing ecological diversity (e.g., increase in species richness and number of links in trophic webs) from headwaters to mouth (e.g., Pease et al., 2012). However, most large river basins in different parts of the world have been significantly altered in their hydrology and geomorphology due to dam construction (Dudgeon et al., 2005) and South America is no exception (Agostinho et al., 2007; Winemiller et al., 2016). The Serial Discontinuity Concept (SDC) (Ward and Stanford, 1983; 1995) arose as a modification of the RCC to explain the functioning of flow-regulated ecosystems. This theory postulates that fluvial ecosystems have a predictable longitudinal gradient in biophysical conditions, which depend on environmental conditions and the position, functioning, and distance to dams (Ward and Stanford, 1983; 1995). Often, river stretches are interrupted by a series of consecutive dams. The spatial distribution of barriers along the longitudinal gradient of the river (i.e., dam cascades) can strongly affect downstream ecological patterns (e.g., Barbosa et al., 1999; Santos et al., 2018). The cumulative effects of such series of dams on changes in hydrology, sediment and nutrient retention, and physico-chemical processes represent a stronger impact than the combined effects of independent dams (Santos et al., 2018). These environmental changes can affect local communities, including shifts in fish composition (Barletta et al., 2010), and the structure of food webs (Ruhí et al., 2016). Some studies have found pauperization of some functional groups upstream and downstream of dams (e.g., de Bem et al., 2021), and an increase in the proportion of generalist species of fishes after dams (Turgeon et al., 2019). Declines in downstream diversity, population sizes, and genetic diversity in upstream reaches of the dam cascade due to isolation caused by habitat fragmentation have also been reported (Ribolli et al., 2021). Moreover, habitat fragmentation due to dams could result in species loss but also the addition of other species downstream of the dams, thus altering beta diversity (i.e., changes in species diversity between communities within a region) by modifying longitudinal patterns of turnover (degree of species replacement) and nestedness (subset of species present in the most diverse site) (Ganassin et al., 2021).

The presence of dams also affects the availability of basal food sources, leading to changes in the ecological niche utilized by the species (Guo et al., 2023). Stable isotope analyses (SIA) facilitate the study of global changes in ecological niches using, for example, isotopic niches areas as proxies of trophic diversity in food webs (Newsome et al., 2007) over a relatively long-time span (as stated originally by Layman et al., 2007a). In particular, SIA has enabled the identification of changes in community functioning indirectly induced by the presence of dams (primarily changes in hydrological flows affecting basal resources) (Guo et al., 2023).

The aim of this work was to determine the changes in ecological fish diversity, as a combination of different metrics of diversity (i.e., species richness, beta-diversity, trophic and isotopic diversity) along the fluvial gradient of a major South American river, the Uruguay River, within the framework of existing theories for regulated and unregulated rivers. We hypothesized that, in contrast to expectations from unregulated systems, dams would promote a decline in ecological niche, species diversity and a higher species turnover along this long river longitudinal gradient.

92 fish species were collected in total from all sampled sites, accounting for 1364 samples for stable isotopes and 2309 for dietary analysis (Table S3 in SM).

In general, fish species richness (alfa diversity) increased along the longitudinal gradient (R2 = 0.49; p = 0.01) but showed sharp declines in the sites located downstream of the dams. The greatest loss of species occurred downstream of the dam between sites 2 and 3, with a loss of 44% of the species, and downstream of the Salto Grande Dam, between sites 8 and 9, with a loss of 43% of the species (Table 1; Fig. 2a). The migratory species Megaleporinus obtusidens and Prochilodus lineatus occurred from site 5 and downstream, while Salminus brasiliensis was found at site 7 and distributed downstream along the rest of the longitudinal gradient (Table S3 in SM).

Fig. 2.

Changes in fish ecological diversity at community level along the longitudinal gradient of Uruguay River: a. Species richness; b. Trophic diversity (number of trophic links); and c. Mean standard ellipse area (SEAc- refering to the isotopic niche breadth), at each sampling sites. Boxes indicate the SEAc with Bayesian models and 50, 75, and 95% credibility intervals. Dashed lines indicate the location of dams.

The parameters of the Bayesian model were: 95% confidence intervals, number of chains = 2; iterations = 200000; burning = 10000, and thins = 10, and the ellipse built considered 40% of data points, as implied by SEAc.

(0.16MB).

Remarkably, despite the sampling effort being greater in the upper section (by performing three different methods), we found lower species richness there than in the middle and lower sections. Also, despite the relatively low number of sampling points in space and time, the patterns found were strong and aligned with expectations.

Of the three different beta-diversity estimation, two of them, multiple beta-diversity and turnover, provided support to the hypothesis that longitudinal trends in diversity existed and were interrupted by dams. Multiple beta-diversity (βSOR, estimated as the multiple comparison of all possible pairs of sites) was high (βSOR = 0.82), with turnover being the component with the highest contribution (βSIM = 0.72), while nestedness had a low contribution to the difference in species composition between local communities along the gradient (βNES = 0.10). In the analysis comparing each site with the reference site (i.e., site 1), pairwise beta diversity also increased along the longitudinal gradient, showing a marginally significant relationship with the distance to the headwaters (R2 = 0.259; p = 0.06). Turnover also significantly increased along the longitudinal gradient (R2 = 0.305; p = 0.04), but nestedness did not show a clear pattern (R2 = 0.110; p > 0.05) (Fig. 3a; Table S5 in SM). The highest pairwise beta diversity (i.e., the highest differences between communities) and species turnover were found between sites 1 and 9 (downstream of Salto Grande Dam). In contrast, beta diversity and species turnover from the step-wise comparison between consecutive sites decreased along the longitudinal gradient, with a significant relationship with the distance to the headwaters (R2 = 0.864; p < 0.001 and R2 = 0.571; p < 0.004, for total beta diversity and species turnover, respectively), while nestedness again did not show a significant pattern with the distance to the headwaters (R2 = 0.044; p > 0.05) (Fig. 3b; Table S5, SM). In this case, the highest beta-diversity and turnover were observed between sites 2 and 3 (i.e., the sites separated by two dams), whereas the lowest beta diversity was observed in the stretch between sites 6 and 8.

Fig. 3.

Changes in pairwise beta-diversity and its components (turnover and nestedness) along the longitudinal gradient of the Uruguay River: a. Pairwise comparison between each site and the first site (site 1) at the headwaters; b. Step-wise: pairwise comparison between each site and the neighbor site, located downstream. Dashed lines indicate the presence of dams between those sites.

(0.27MB).

Trophic diversity (expressed as the number of trophic links) followed a similar pattern to taxonomic richness, i.e., a general trend of increase along the longitudinal gradient showing interruptions in the areas with dams. The number of trophic links recorded in the network ranged from 31 to 289, with the minimum reported at sites 2 and 3 and the maximum at site 12, showing an increase along the longitudinal gradient (R2 = 0.55, p = 0.006), and decreased after the dams (i.e., from 36 to 31 between site 1 and 2, with the strongest decline observed after the Salto Grande dam, from 238 to 105 trophic links) (Fig. 2b).

In the case of isotopic niche area, SEAc presented a strong decrease in the upper reaches coinciding with the presence of the dam cascade (99% probability of having a higher SEAc in site 2 than in site 3, site located downstream the dam cascade), followed downstream by increases and a more even distribution of the isotopic niche in the middle and lower stretches of the river (Table 1; Fig. 2c). Similarly to species richness and trophic diversity, the SEAc showed a reduction after Salto Grande, with a 96% probability of SEAc being higher in the pre-Salto Grande site than in the post-Salto Grande location.

The general longitudinal patterns found (i.e., increased species richness, beta-diversity and turnover of species, trophic diversity, and ecological niche diversity) support the general framework proposed by the River Continuum Concept (Vannote et al., 1980). However, we also found clear evidence of the negative impacts of dams on fish diversity as predicted by the SDC (Ward and Stanford, 1983; 1995), shown by the sharp declines in species richness, trophic diversity, and ecological niche diversity in the sections immediately downstream of the dams, as well as the changes in turnover in paired sites around dams.

Similar evidence partially supporting the RCC was found in the Tennessee River (USA) and major tributaries (23 reservoirs), where fish species richness and the number of functional attributes increased longitudinally (ca. 1500 km) despite being highly dammed (Besson et al., 2023). However, fish species richness and abundance often decrease at sites located further downstream of a cascade of dams, as found in three Brazilian basins (Iguaçu, Paranapanema, and São Francisco River basins) (Ganassin et al., 2021). The decrease in species richness after the dams may be a consequence of the changes in environmental conditions (Pelicice et al., 2018). In the dam cascades located in the upper Uruguay River, we observed an abrupt reduction in species richness from site 2 to site 3, with the latter having the lowest species richness of the entire river. This likely reflects the limitation of fish migration upstream of the dams, as suggested elsewhere (e.g., de Bem et al., 2021; Ribolli et al., 2021). In this sense, we observed the presence of migratory species in the middle reach of the Uruguay River, from site 5 to the mouth, which are sites located downstream of the dam cascades. This had been observed in previous studies where richness of migratory species increased towards downstream reaches of dammed rivers (Pelicice et al., 2018).

The negative impacts on fish ecological diversity indicators, along with their recovery as the distance to the dams increased, align with the predictions of the Serial Discontinuity Concept (SDC) and its modifications (Ward and Stanford, 1983; 1995), thus supporting our hypothesis.

Furthermore, the predominant influence of species turnover on beta diversity among all sites support the notion that dams act as environmental filters, promoting species sorting (López-Delgado et al., 2020) and the introduction of new ones. The step-wise beta-diversity decreasing along the longitudinal gradient suggests that the variation in species composition among neighboring sites decreased as the river potentially receives new species from tributaries in the longitudinal gradient. However, we observed that the highest beta-diversity and turnover occurred downstream of the dam cascade in the upper part of the river. There, we observed the effects of dams preventing the accumulation of species, altering the downstream composition as found in other dammed rivers (Agostinho et al., 2008).

These findings align with our findings at the species richness level, underscoring the impact of dam-induced habitat fragmentation on beta diversity (Edge et al., 2017). Similar patterns have been observed in studies on fish (e.g., Lansac-Tôha et al., 2019; Ganassin et al., 2021) and macroinvertebrates (Wang et al., 2021a).

Empirical evidence shows that taxonomic diversity is usually high in sites located in a highly connected longitudinal gradient (i.e., more central locations), but beta diversity is often low as species compositions are similar due to the absence of dispersal restrictions (Gianluca et al., 2017; Henriques-Silva et al., 2019). These central communities seemed to be present in our study system in the dam-free areas, primarily between sites 6 and 8, where we found high alpha diversity and low differences in community composition. These stretches may also act as species hubs that facilitate biodiversity recovery (Pachla et al., 2022).

The longitudinal patterns in the isotopic niche area were less clear than those observed with alpha and beta-diversity. Several not mutually exclusive factors might be operating on the variability of the isotopic signal. A series of both biotic and abiotic factors directly or indirectly promoted by dams (not analyzed in this work), can promote either high or low isotopic diversity (Box 1). The main abiotic factors potentially affecting ecological diversity include spatial and temporal environmental heterogeneity (Layman et al., 2007b; Turner et al., 2015) and hydrological heterogeneity determined by variability in water flows (Trudeau and Rasmussen, 2003). Biotic factors mainly include a high diversity of basal resources, a local increase in species richness, and the use of a wide diversity of environments and food sources (Delong et al., 2011). Also, the arrival of individuals that move long distances allows for the incorporation of isotopic signals from diverse environments, increasing the isotopic niche area in a given location (Rasmussen et al., 2009; Wang et al., 2021b). This phenomenon could well explain the higher isotopic niche area found for the fish community at the areas near the river mouth and at site 5, where fish could freely move and assimilate resources with potentially contrasting isotopic values along a stretch of 1000 km. Some of these factors could have operated in the long unregulated and thus highly variable stretch in the middle section of the Uruguay River.

The dimensions of our river (ca. 1800 km long, with an average discharge above 4000 m3/s) and the scale of our study allowed us to detect the recovery of the system after the negative impacts of dams.

Our results suggest that the dam-free stretches of the Uruguay River play a crucial role in maintaining the river's ecological diversity, as suggested by the recovery of ecological niche areas, mainly in the middle section. Fish passage constructions often fail as upstream movements are allowed but the return of adults and their offspring is not (Agostinho et al., 2007). Therefore, the maintenance of long dam-free reaches and undammed tributaries in the watershed is key to maintaining fish diversity in impacted rivers (López-Delgado et al., 2020). Tributaries can provide food resources and serve as a refuge and feeding areas (Santos et al., 2020; Pachla et al., 2022) and can be used as spawning areas for migratory fish (Reynalte-Tataje et al., 2012) and nursery (Corrêa et al., 2011).

Our results contribute to enriching theoretical frameworks such as RCC and SDC by evidencing a general pattern of longitudinal increases in fish species richness, beta diversity and ecological functions, but also clear negative effects of dams, mainly of the cascade of dams in the upper section of the river. Our findings highlight the need to consider diverse dimensions of diversity (e.g., species richness, beta-diversity, trophic diversity, ecological niche), as dams’ effects go beyond the loss of species. The size and flow of this river, as well as the focus on the entire longitudinal gradient, enabled us to detect ecosystem recoveries within the dam-free section. In the Neotropics, numerous large dams are projected in the coming decades (Winemiller et al., 2016; Pelicice and Castello, 2021). Notably, in the Uruguay River, two additional consecutive hydroelectric dams, Garabí (planned operational power: 2700 MW/h) and Panambí (1800 MW/h), are scheduled in the middle section (1000 km stretch where the river currently flows freely), coinciding with the highest ecological diversity of fish. This calls for attention from the relevant authorities in the different countries involved. The Uruguay River is situated at a triple multinational border, which also implies that environmental monitoring data may be collected at different times and at a specific number of sites and may imply differences in fishing techniques and effort. Through this work, we identified the need to establish and sustain standardized transnational monitoring programs to generate scientific knowledge that properly informs policy and management.

Conceptualization: AL-R, MM, IG-B; Methodology: AL-R, IG-B; Investigation: AL-R, AD, SdAS, JP, FTM, DR-T, EZ-F, IG-B; Resources: DR-T, EZ-F, IG-B; Data curation: AL-R, SdAS, IS, JP, IG-B; Formal analysis: AL-R; Writing-original draft: AL-R, MM, IG-B; Writing review & editing: all; Visualization: ALR; Supervision: MM, IG-B; Project administration & Funding acquisition: DR-T, EZ-F, IG-B.

This research project was partly funded by the Scientific Research Sectorial Commission (Uruguay) (Project CSIC I+D_2016_577-348) and the National Agency for Innovation and Research (ANII) (Project ANII-FCE_2_2016_1_126780).

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.