An integrated modelling approach to investigate the dynamics of Planktothrix rubescens blooming in a medium-sized pre-alpine lake (North Italy)

Abstract

In a medium-sized pre-alpine lake (Lake Pusiano, North Italy) the cyanobacterium Planktothrix rubescens has strongly dominated the phytoplankton assemblage since 2000 despite improvements in water quality, similarly to what happened in many pre-alpine lakes. The ecological success of the ubiquitous harmful species has been ascribed to largely depend on its eco-physiological traits, lakes re-oligotrophication (and increasing N:P ratios) as well as climate oscillations. Whatever the viewpoint, it has been the dominating algal species over the last two decades but the scientific community is debating about the crucial factors determining the dynamics. A great difficulty is certainly the comprehension of the effects due to human pressures at different scales. Also the natural changes and the interactions within the ecosystem may cause a high uncertainty. The present research focused on the necessity to solve some of the most paradoxical features about P.rubescens large success. An intensive field campaign was conducted to evaluate distributions of phytoplankton taxa, as well as P. rubescens, using spectrally-resolved fluorescence measurements and cell enumeration. These provided a high spatially and temporally resolved database, suitable to calibrate and validate a coupled three-dimensional hydrodynamic and ecological model for lakes ecosystem. The simulations revealed the fundamental role of physiological features. They led to characteristic vertical patterns of distribution, notably a deep chlorophyll maximum, and a visible influence of lake hydrodynamic processes, particularly during high-discharge inflows in summer stratification. The simulations were used to examine growth-limiting factors that help to explain its increased prevalence during a re-oligotrophication phase. A long-term series (1960-2010), assessed over measured data, was reconstructed for some ecological indicators. A natural external phosphorus load was simulated by a hydrological and nutrients transport model (SWAT), after it was calibrated on a natural sub-basin. Data by a paleolimnological survey were used to initialize the lake ecological model to reproduce the past conditions. A specific statistical technique (Spectral Singular Analysis) was used to isolate the trend of air temperature daily series, avoiding the periodic climatic fluctuations. Four different scenarios were simulated to characterize different levels of local and global pressure on lake ecology, combining each alternative driver into the lake model. The integrated lake-basin tool was also proposed as a dynamic tool to simulate the biogeochemical cycle in an alternative pristine ecological state. The output for phosphorus reference conditions was compared to the results by the most traditional methods (previously assessed for subalpine lakes). After decades of lake eutrophication, the simulated temperatures warming did not enhance P. rubescens blooming. Conversely, a positive relation was found when the pressure from the catchment (e.g. phosphorus pollution) was switched off by the simulation, as emerged by the Mann-Kendall statistics on daily model output. In other words, the global warming may have different effects on P. rubescens dynamics, depending on the trophic evolution of a lake. The simulation of a pristine condition projected the lake into an oligo-mesotrophy, as the results of equilibrium between the external phosphorus loading and the trasformations across the internal exchanging pools. The simulation of lake transparency and productivity depicted a good ecological state, but the hypolimnetic waters remained anoxic during the thermal stratification, as confirmed by the paleolimnological survey (to pre-industrial age). P. rubescens persisted in that conditions, but its growth resulted strongly limited by low phosphorus concentration, resulting in a low productivity.