Stable phytoplankton community compositions in Lake Mead (Nevada-Arizona, USA) during two decades of severe drought

Lake Mead, a large reservoir on the Colorado River and a critical drinking water source for the southwestern United States, typically exhibits high water quality, characterized by low nutrient and chlorophyll- a concentrations. This stability persists despite the inflow of highly treated wastewater since the 1960s and significant water level declines since 2000, driven by the ongoing Megadrought and basin-wide consumptive use. Such environmental changes may alter phytoplankton communities, potentially leading to increased cyanobacteria abundance, which could negatively impact water quality and the aquatic ecosystem through harmful algal blooms and toxin production. Here we analyzed 17 years of phytoplankton community structure and chlorophyll- a concentrations in Lake Mead, alongside quantitative water quality data, including nutrients, temperature, and water clarity, to assess the effects of environmental changes on phytoplankton communities. Contrary to the hypothesis that cyanobacteria abundance would have increased throughout the reservoir, our results indicate that phytoplankton community structures have remained largely stable, except for shallow areas where increases in temperature or phosphorus levels were observed. Additionally, we evaluated machine learning models for predicting changes in phytoplankton community structures. While the models confidently predicted changes in total phytoplankton biovolume and chlorophyll- a concentrations within the input parameter boundaries, predictions of peak biovolume showed considerable uncertainty, emphasizing the importance of incorporating uncertainty analysis in forecasting and communicating results. This study underscores the current buffering capacity of large, oligotrophic reservoirs like Lake Mead to maintain stable phytoplankton communities despite environmental changes. However, it also highlights the potential for significant community shifts if this buffering capacity is exceeded. • 17 years of phytoplankton data shows stable assemblages in most of Lake Mead. • Large, oligotrophic reservoirs are resilient to environmental change. • Temperature and nutrient increases elevated phytoplankton biovolume locally. • Machine learning can predict large-scale phytoplankton structures. • Uncertainty quantification is essential when making biovolume forecasts.