Discussion
In this study, we demonstrate that iron flocculation resulting from open-cast lignite mining has the ability to restructure the lotic food web, including the abundance of benthic invertebrates and fishes, growth of YOY top predators and the strength of the linkage to the adjacent terrestrial ecosystem. While several studies have shown the consequences of iron concentrations for single species or specific communities, our study is among the first that integrates consequences to stream functioning and structure, i.e., the animal-mediated fluxes occurring across the land-water interface.
Fluxes from land to water occur primarily in the appearance of terrestrial carbon from plants, either in the particulate form as plant litter or in dissolved form, mostly as humic substances. Several studies could demonstrate that terrestrial carbon has the ability to fuel aquatic food webs and can be traced up to the highest trophic levels (Karlsson et al. 2012, Scharnweber et al. 2014, Tanentzap et al. 2014). However, another pathway of terrestrial particulate carbon into aquatic systems, especially important for lotic systems, is the one via terrestrial insects falling on the water surface and being consumed by freshwater fish (Baxter et al. 2005). In the famous greenhouse experiments of Nakano et al. (1999), terrestrial insects were prevented from falling into a forest stream, and the strong impact of these pathways could be demonstrated, being able to shape the feeding ecology (Fausch et al. 1997) and overall fish biomass (Kawaguchi and Nakano 2001).
In the present study, we demonstrate how the human impact on land-use, more precisely via the contamination of iron resulting from lignite mining is able to alter these linkages and shape fish communities. Generally, fish abundance and biomass are known to be low in lotic systems characterized by high iron concentrations (Vuori 1995, Amisah and Cowx 2000), and our results are in line with these previous findings. The low abundance of benthic invertebrates as food, as demonstrated in this study, might play a crucial role in this correlation. Such low availability of aquatic prey at high iron section of the Spree River might force some fish species to integrate terrestrial insects into their diet. Here, we could demonstrate this switch using a combination of techniques illustrating the short-term (metabarcoding of gut content), but also long-term (stable isotopes of hydrogen) resource use. Indicated by the DNA-sequences from metabarcoding, perch and bleak caught upstream of the dam had a higher number of terrestrial species in their gut content, but this result was not significant for roach. While perch and roach are known to usually ingests only minor proportions of terrestrial insects (e.g., Persson 1983, Svanbäck and Bolnick 2007, Marklund et al. 2019), bleak has an upwards facing snout showing the adaptions feeding on insects from the water surface (Biro and Musko 1995). The metabarcoding approach does not allow for a qualitative assessment of the ingested prey, as biomass estimates are not possible. However, the pattern was further confirmed by the stable isotope data, and for perch, bleak and also roach, the incorporation of terrestrial resources to the diet increased significantly from 34% on average at low iron sections to 74 % on average. Terrestrial species identified in the gut content of the fish included a wide range of species and taxa, including earth worms, ants or flying insects and most identified species (except for two species) appeared only once in the entire dataset. Fish that incorporate terrestrial species into their diet may therefore not specialize on specific prey, but might respond rather opportunistically to organisms that end up on the stream surface.
Significant differences between the terrestrial reliance of fish individuals upstream and downstream of the dam was pronounced, as demonstrated from the two methods integrating different time scales. However, some uncertainty in the absolute values of the terrestrial contributions have to be considered. When correction for the proportion of water δ2H entering the consumer, we applied and correction factor of 0.2, as suggested by Wilkinson et al. (2015). However, this correction factor is still under debate and the literature also shows that the value of ω can be assumed as 0.17 (Solomon et al. 2009) or averaging 0.25-0.30 (Brett et al. 2018). The outputs of the mixing models are very dependent on this assumption, thus making it difficult to compare the obtained estimates to previous reported terrestrial contributions from other systems.
Such diet shift towards a higher integration of terrestrial insects was also reported from Brook and Brown Trout (Salvelinus fontinalisand Salmo trutta ) that compensated for the low availability of aquatic prey in stream sections of high heavy metal concentrations (Kraus et al. 2016). However, the feeding on terrestrial organisms may come at a cost. First, fish individuals feeding on the water surface might be more vulnerable to attacks from aerial predators. In addition, aquatic resources are generally rich in PUFAs, while terrestrial insects are characterized by low concentrations of these important biomolecules (Twining et al. 2016, Twining et al. 2021). Thus, fish individuals feeding to a high degree on terrestrial insects and therefore facing a diet of lower quality when it comes to PUFA content, might have to internally synthesize PUFAs to meet the physiological need (Scharnweber et al. 2021), which could potentially affect growth and fitness (Scharnweber and Gårdmark 2020, Závorka et al. 2022).
A reduction of energy at the base of the food web was further translated up the food chain, thus it most likely reduced the condition in juvenile and adult pike, which is the most abundant top predator in this system. Juvenile pike in their first year caught in high-iron sections were on average 9.9 cm smaller compared to juveniles caught in sections of clear water (data not shown). Turbidity in the high-iron sections may affect the foraging abilities of pike as a visual predator by altering the reactive distance to the prey fish, with consequences for attack rates and the amount of consumed prey (Jacobsen and Engström-Öst 2018). However, pike might compensate low visibility by other sensory modes, such as the ones from the lateral line (Raat 1988). Instead, it can reasonably be assumed that the differences in total length of juvenile pike is rather caused by the overall low availability of food (i.e., low abundances of prey fish) and likely the limited aquatic vegetation, which is important as habitat for the juvenile pike (Jacobsen and Engström-Öst 2018). Unfortunately, low sample sizes of adult pike individuals did not allow further comparisons of cohorts, but it can be assumed that pike of all cohorts living in high-iron sections of the Spree River have an overall lower fitness. Our results thus highlight the broad impact of iron load on each level of the aquatic food web. The results presented here, however, represent a single timepoint of the year. We would like to emphasize that longer studies including the seasonal and temporal variances of abundances and fish resource use are needed to draw more general conclusions on the post-mining effects on the lotic food webs.
Besides the overall low availability of aquatic prey, other factors may add to the detrimental effects of iron (III) for fish communities. For example, iron could directly impact fishes, as precipitates can clog gills and eggs (Andersson and Nyberg 1984, Gerhardt 1992), but further indirect effects are possible. The growth of submerged macrophytes, for instance, can be hampered under brown water conditions (Blanchet et al. 2022). Macrophytes provide a complex habitat, favoring fish and invertebrate diversity and richness simultaneously (Soukup et al. 2022). Certainly, more studies are needed to understand the underlying and potentially multifactorial causes for the low fish abundance found in this study.
Intensive land-use as from mining activities may not be the only human-induced perturbation acting on the river food web, as multiple stressors often act jointly (Sabater et al. 2019). The modification of the river flow via the building of dams poses one of the biggest impacts on rivers across the globe (e.g., Dudgeon et al. 2006, Schmutz and Moog 2018). The disruption of the river flow by damming and the formation of a reservoir is a drastic process, which will alter many aspects of the ecosystems, e.g. fish migration and the transport of nutrients. An additional factor may occur from the modification of the downstream thermal regime, which depends on the mode of operation and specific mechanism of the water release from the reservoir (Olden and Naiman 2010). While many large dams may release cold water from below the thermocline of the reservoir, more shallow reservoirs may alter the downstream thermal regime by warming with profound impact on metabolic rates, physiology and life-history traits of aquatic species (Lessard and Hayes 2003, Olden and Naiman 2010). The rather shallow Spremberg reservoir releases the water from the bottom, but the reservoir itself is usually not stratified (data not shown). In the study presented here, the river water downstream of the dam was generally warmer during summertime, but colder during wintertime (Appendix 3). This indicates that the Spremberg Reservoir severely alters the thermal regime further downstream. This pattern could potentially affect the abundance of invertebrates and fish, and further the growth of pike. The study set-up does not allow for the differentiation between the effects of iron flocculation and the changes of the water temperature downstream of the dam and further studies are need to investigate the cumulative stressor of human-induced perturbations on our river systems.