HP scenarios
The HP scenarios consisted of the combined increase of flow amplitude and up-ramping rate (Figure 2c and Table 1). To reduce impacts on hydropower flexibility and production, in each river all field experiments were carried out over two consecutive days. The first day, we targeted a combination of three flow amplitudes and three up-ramping rates. The same three scenarios were repeated the next day but the drift nets were placed further towards the river center to achieve higher peak flow velocities as at day one (Figure 2a and Table 1). The selection of the targeted up-ramping rates (1, 2 and 3 cm/min) and peak flow velocities (0.4 and 0.8 m/s) was based on the results of previous flume experiments, which showed distinct drift responses in these ranges (Schülting et al., 2018b). In terms to achieve the targeted HP scenarios and to define appropriate sampling sites, for each HP reach we set up and calibrated 2D unsteady hydrodynamic-numerical models to pre-define the operation mode of each hydropower plant (KUB) or retention basin (KWO, KLL).
In contrast to KUB and KWO, at KLL the maximum turbinable peak flow had to be kept constant to avoid flow velocities > 1 m/s and thus ensure wadable conditions. With this limited discharge setting, for the Linth, it was also necessary to define an alternative HP scenario with a reduced up-ramping rate (0.5 cm/min) as for the Sitter and Hasliaare. Moreover, due to different hydrological conditions (e.g., melting snow, rain) and operational characteristics of the hydropower plants and/or retention basins (e.g., turbine type, concessional limitations), base flow and down-ramping rates varied between the HP reaches. These constraints resulted in some differences between the targeted (modeled) and the measured flow amplitudes, peak flow velocities and up-ramping rates (Appendix B in Data S1). Nevertheless, measured flow amplitudes in the Sitter and Hasliaare increased approximately twofold respectively threefold from HP scenario 1 (SC1) to SC2 and ca. fourfold respectively five-sixfold from SC1 to SC3, whereas it remained almost constant in the Linth (Figure 2c and Table 1). An increase in mean and maximum up-ramping rate from SC1 to SC2 and to SC3 in addition to a higher peak flow velocity at day two was also achieved but varied between HP reaches. Mean up-ramping rate between SC1 and SC3 increased 2.5 times in the Sitter, 1.5 to 4 times in the Hasliaare and 1.5 to 2 times in the Linth. From day one to day two, peak flow velocity over all scenarios increased in average 4 times in the Sitter, 0.25 times in the Hasliaare and 0.15 times in the Linth (Table 1).