Double-edged effects of anthropogenic activities on lake ecological dynamics in northern China: Evidence from palaeolimnology and ecosystem modelling

Climate change and anthropogenic activities mutually degraded global lake ecosystems during the Anthropocene. Nutrient mitigation via catchment management has been a constant priority for lake ecological restoration; however, the outcomes after mitigation remain uncertain. In particular, there is limited knowledge on the effects of large-scale forest restoration and soil erosion reduction in catchments on the ecological dynamics of downstream aquatic systems. Here, we study a strategically chosen alpine lake from the Loess Plateau in northern China, which has been extensively afforested in the past decades. We present high-resolution, multi-proxy records to reconstruct environmental and ecological dynamics since 1960. To quantify the potential mechanisms, we combined evidence from palaeolimnological records, contemporary field observations and process-based ecosystem modelling. We developed palaeorecord-based proxies for water transparency, nutrient concentrations and algal composition, and applied these proxies to ecosystem modelling. Our results show that despite the concurrent decline in nutrient levels, the lake ecosystem has experienced large-magnitude rapid shifts since 2000, characterised by unambiguously increased algae proliferation and community changes. The modelling results revealed that the reduction in soil erosion acts as a "double-edged sword"; that is, lower external nutrient loading via soil erosion is ideal for in-lake algae inhibition, whereas increasing light availability counteracts the reduced yet still sufficient nutrient concentrations and, ultimately, promotes temporary algal proliferation via extended habitats and intensive photosynthesis. Intriguingly, lake warming plays a secondary role in driving these ecological dynamics. Our results emphasised the unexpected algal proliferation when the nutrient level declined in lake ecosystems; thus, other compound factors, including water transparency and climate warming, need to be incorporated for reliable prediction and effective restoration. A multi-disciplinary approach integrating palaeorecords as ecological indicators and process-based ecosystem modelling holds great potential for investigating long-term lake ecological dynamics in the context of rapid global changes.