Sustainable solutions to meet the growing food, water and energy needs of the human population are urgently required, especially in developing countries. Demand for hydropower and irrigation water threatens large river systems with potentially irreparable environmental effects (Best, 2019). More than 3,700 dams are proposed in tropical and subtropical countries (Winemiller et al., 2016;Zarfl et al., 2015). The African continent still has a largely untapped hydropower potential: Only 10% of this potential has been exploited, the lowest of any of the world's regions (The World Bank, 2015). The imminent growth of hydropower projects at low latitudes calls for an examination of the environmental effects of tropical dams (Anderson & Veilleux, 2016).Dams disrupt the continuum of rivers, altering natural hydrological regimes . By storing water, dams inevitably increase water residence time. As a consequence, thermal stratification and, subsequently, chemical stratification can develop in a reservoir (Friedl & Wüest, 2002). Hence, downstream water temperature and chemistry depend on the withdrawal depth, and dams thus modify not only downstream water quantity, but also water quality (e.g., Moran et al., 2018;Winton et al., 2019).The alteration of river water quality might have cascading effects and implications for the entire river ecosystem. Water temperature and oxygen concentration, for instance, are key parameters for aquatic species (Caissie, 2006). Water temperature affects growth, metabolism, reproduction, emergence, and the distribution of aquatic organisms, including insects and fish (e.g., Schulte, 2015;Vannote & Sweeney, 1980). Thus, water temperature changes may lead to changes in community composition or even to extinction of some species (Best, 2019). Dissolved oxygen (DO) concentration is also a key parameter for aquatic life (e.g., Kramer, 1987). Low oxygen concentrations alter lifecycle performance, growth capacity, reproductive success and disease vulnerability of fish, whilst hypoxia leads to higher fish mortality (Winemiller et al., 2008).