Phytoplankton usually responds directly and fast to environmental fluctuations, making them useful indicators of lake ecosystem changes caused by various stressors. Here, we examined the phytoplankton community composition before, during, and after a simulated 1-month heat wave in a mesocosm facility in Silkeborg, Denmark. The experiment was conducted over three contrasting temperature scenarios (ambient (A0), Intergovernmental Panel on Climate Change A2 scenario (circa +3 °C, A2) and A2+ %50 (circa +4.5 °C, A2+)) crossed with two nutrient levels (low (LN) and high (HN)) with four replicates. The facility includes 24 mesocosms mimicking shallow lakes, which at the time of our experiment had run without interruption for 11 years. The 1-month heat wave effect was simulated by increasing the temperature by 5 °C (1 July to 1 August) in A2 and A2+, while A0 was not additionally heated. Throughout the study, HN treatments were mostly dominated by Cyanobacteria, whereas LN treatments were richer in genera and mostly dominated by Chlorophyta. Linear mixed model analyses revealed that high nutrient conditions were the most important structuring factor, which, regardless of temperature treatments and heat waves, increased total phytoplankton, Chlorophyta, Bacillariophyta, and Cyanobacteria biomasses and decreased genus richness and the grazing pressure of zooplankton. The effect of temperature was, however, modest. The effect of warming on the phytoplankton community was not significant before the heat wave, yet during the heat wave it became significant, especially in LN-A2+, and negative interaction effects between nutrient and A2+ warming were recorded. These warming effects continued after the heat wave, as also evidenced by Co-inertia analyses. In contrast to the prevailing theory stating that more diverse ecosystems would be more stable, HN were less affected by the heat wave disturbance, most likely because the dominant phytoplankton group cyanobacteria is adapted to high nutrient conditions and also benefits from increased temperature. We did not find any significant change in phytoplankton size diversity, but size evenness decreased in HN as a result of an increase in the smallest and largest size classes simultaneously. We conclude that the phytoplankton community was most strongly affected by the nutrient level, but less sensitive to changes in both temperature treatments and the heat wave simulation in these systems, which have been adapted for a long time to different temperatures. Moreover, the temperature and heat wave effects were observed mostly in LN systems, indicating that the sensitivity of phytoplankton community structure to high temperatures is dependent on nutrient availability.

Shallow lakes are globally the most numerous water bodies and are sensitive to external perturbations, including eutrophication and climate change, which threaten their functioning. Extreme events, such as heat waves (HWs), are expected to become more frequent with global warming. To elucidate the effects of nutrients, warming, and HWs on zooplankton community structure, we conducted an experiment in 24 flow-through mesocosms (1.9 m in diameter, 1.0 m deep) imitating shallow lakes. The mesocosms have two nutrient levels (high (HN) and low (LN)) crossed with three temperature scenarios based on the Intergovernmental Panel on Climate Change (IPCC) projections of likely warming scenarios (unheated, A2, and A2 + 50%). The mesocosms had been running continuously with these treatments for 11 years prior to the HW simulation, which consisted of an additional 5 °C increase in temperature applied from 1 July to 1 August 2014. The results showed that nutrient effects on the zooplankton community composition and abundance were greater than temperature effects for the period before, during, and after the HW. Before the HW, taxon richness was higher, and functional group diversity and evenness were lower in HN than in LN. We also found a lower biomass of large Cladocera and a lower zooplankton: phytoplankton ratio, indicating higher fish predation in HN than in LN. Concerning the temperature treatment, we found some indication of higher fish predation with warming in LN, but no clear effects in HN. There was a positive nutrient and warming interaction for the biomass of total zooplankton, large and small Copepoda, and the zooplankton: phytoplankton ratio during the HW, which was attributed to recorded HW-induced fish kill. The pattern after the HW largely followed the HW response. Our results suggest a strong nutrient effect on zooplankton, while the effect of temperature treatment and the 5 °C HW was comparatively modest, and the changes likely largely reflected changes in predation.