Material and methods

Study sites and hydropower facilities
The study was conducted in three HP impacted river-reaches at the rivers Sitter (hydropower plant Kubel, KUB; canton of St. Gallen), Hasliaare (hydropower plant Innertkirchen, KWO; canton of Bern) and Linth (hydropower plant Linth-Limmern, KLL; canton of Glarus; Figure 1). Additionally, RF reaches located closely upstream of the hydropower plants were used as hydrological comparisons with relatively constant base-flow conditions.
All HP and RF reaches are of comparable size (Strahler order 5 – 6; natural mean annual discharge 11 – 32 m3/s) and located at similar elevations (590 – 680 m a.s.l.). Dominant grain size (cover > 50%) in all HP reaches is micro- to mesolithal (grain size 2 – 20 cm), whereas it is coarser (> 20 cm) in the RF reaches of the Sitter and Linth.
The KUB hydropower plant has the lowest maximum turbinable peak flow (18 m3/s), flow amplitude (16 m3/s) and ramping rates (0.1 and 0.05 m3/s min). The investigated HP reach was located < 0.5 km downstream of the plant release and is characterized by the highest bankfull width (45 m), a near-natural morphology with a right-sided gravel bar and comparably fine substrate (Dm = 28.1 mm). Since 2016, the hydropower facilities KWO and KLL both include a HP retention volume of 80,000 m3 and 220,000 m3, respectively. The maximum turbinable peak flow, amplitude and ramping rates of the corresponding hydropower plant are higher than at KUB. The investigated HP and RF reaches of the Hasliaare and Linth are characterized by a strongly constrained morphology. The HP reach of the Linth has the lowest bankfull width (15 m), no extended gravel bars, and the coarser sediment size (Dm = 59.9 mm). Moreover, four small run-off power plants are situated between the water release and the HP reach. The riverbed of all HP reaches showed no or insignificant clogging, whereas it was moderate in the RF reaches of the Sitter and Linth. The HP and RF reaches of these two rivers were also characterized by higher cover of algae and mosses than the Hasliaare. Possible relevance of these aspects on drift, stranding and the benthic community composition, will be picked up in the discussion. Key characteristics of the hydropower plants, as well as associated catchments, HP and RF reaches are available in Data S1 (Appendix A).
Experimental setup
To investigate the short- and long-term effects of HP on MIV, we defined three HP scenarios for each HP reach and requested respective experimental flow releases from the hydropower plant operators. We sampled benthic, drifted and stranded MIV and recorded several environmental variables (see below) over a four-week period between March 15th and April 8th 2018. In each HP and RF reach we defined three sampling sites spaced by 10 – 50 m (Figure 2a).