Many European lakes have suffered from reed die-back since the 1950s. Previous studies have concluded that this is due to a combination of several interacting factors, but possibly also a single threat with high impact might be responsible for the phenomenon. In this study, we investigated 14 lakes in the Berlin area differing in reed development and sulphate concentration from 2000 to 2020. To unravel the decline of reed beds in some of the lakes with coal mining activities in the upper watershed, we compiled a comprehensive data set. Thus, the littoral zone of the lakes was divided into 1302 segments considering the reed ratio relative to segment area, water quality parameters, littoral characteristics and bank usage of the lakes which all have been monitored for 20 years. We ran two-way panel regressions with a within estimator to consider the spatial variation between and within the segments over time. The regression results revealed a strong negative relationship between reed ratio and sulphate concentrations (p<0.001) as well as tree shading (p<0.001) and a strong positive relationship with brushwood fascines (p<0.001). Taking only sulphate into account, reeds would have covered an additional area of 5.5 ha or 22.6% in 2020 (total reed area: 24.3 ha) in the absence of increased sulphate concentrations. In conclusion, changes in water quality upstream the catchment cannot be ignored in the development of management plans for downstream lakes.