Influence of hydrodynamics on phytoplankton behaviour in lakes in urban areas through high-frequency measurements and three-dimensional numerical modeling : application to Lake Champs-surMarne (France) and Lake Pampulha (Brazil)

Abstract

The main objective of this thesis is to characterize the hydrodynamics of shallow urban lakes to better understand its influence on spatial and temporal variability of phytoplankton biomass through high-frequency measurements and three-dimensional numerical modelling. To achieve this objective, two experimental study sites were investigated: Lake Champs-sur-Marne, a small and shallow urban lake in France, and Lake Pampulha, a medium-sized and shallow urban reservoir, in Brazil. The research shows that a calibrated three-dimensional hydrodynamic model and an ecological model using high-frequency monitoring are essential for understanding shallow lakes hydrodynamic and the temporal and spatial behaviour of phytoplankton and cyanobacteria. It was highlighted strategies to assess spatial and temporal variability of cyanobacteria biomass and how relevant is to understand lake dynamics for management purpose and the importance to well describe the mixing and stratification and to measure the external forces for appropriate temporal and spatial dynamic based in ecosystem modelling approach. In this research it was studied two different lakes. The first study site was Lake Champs-sur-Marne, a small and shallow urban lake that was formed during sand extraction quarry. The lake does not have water inflow that renews the water accumulation. It is filled mainly by groundwater, flowing from the water table of the nearby Marne River. The second study site was Lake Pampulha, a medium-sized and shallow urban reservoir. Lake Pampulha is an artificial reservoir that is filled by 8 tributaries, with a residence time of 89 days (considering the annual mean inflow). Both lakes have problems with cyanobacteria blooms. Lake Champs-sur-Marne is frequented by children from a nearby urban area however, water recreational activities in the lake have been repeatedly banned in the summer because of potential health risks caused by toxic cyanobacteria blooms. Lake Pampulha was built for many purposes, including drinking water supply. However, due to an intense urbanisation process in the lake catchment from the 1970s on, the lake water quality and ecological state was severely compromised resulting in a hypereutrophic reservoir, leading to the interruption of drinking water provision from the 1980s, with frequent cyanobacteria blooms. In the French lake, a three-dimensional hydrodynamic and ecological model was calibrated using high-frequency measurements of chlorophyll-a, temperature, and water velocity. The chlorophyll-a was measured at 1.5 m depth. The water temperatures were measured at surface (0.5 m), middle (1.5 m) and bottom (2.5 m) depths. The velocities were measured with a depth range of 0.45 m from the surface to 2.7 m at the bottom. After the model calibration the phytoplankton behaviour was verified through scenarios with different nutrient concentration using the ecological model and a scenario using a tracer concentration to evaluate the wind influence. The Period 1 (lasting 7 days, from 23rd June 2015 to 30th June 2015, corresponding to 159 hourly values for each depth) was used to evaluate hydrodynamic and the phytoplankton considering transport simulation as a conservative tracer. Periods 2 (lasting 14 days, from 13 July to 27 July 2015, corresponding 342 hourly values for each depth) and Period 3 (lasting 19 days, from 14 July 2016 to 01 August 2016, corresponding to 456 hourly values for each depth) were selected to run the coupled hydrodynamic and ecological model. Period 4 (lasting 19 days, from 19 September 2016 to 12 October 2016, corresponding to 567 hourly values for each depth) was used to assess the measured and calculated velocities. The temperature at three measured depths was validated for all of the four periods. The research carried out in Lake Champs-sur-Marne showed that:

- High-frequency monitoring enables an analysis of the high temporal variation of biomass in the lake and the influence of different external forces on hydrodynamic behaviour;

- The three-dimensional model Delft-3D was useful in the analysis of the influence of lake hydrodynamics on phytoplankton concentrations and spatial heterogeneity;

- Alternation and duration of mixing and stratification conditions and the complex hydrodynamic functioning of the lake could be represented by the three-dimensional model;

- The three-dimensional model can help to improve lake monitoring design, highlighting regions with different behaviour;

- Mathematical modelling is a great tool to simulate different scenarios in cases where there are field data gaps;

- The measurements and simulations showed that the hydrodynamics of Lake Champs-sur-Marne are very unstable and dynamic. The lake reacts rapidly and intensely to meteorological forcing. However, the three-dimensional hydrodynamic model was able to represent these complex hydrodynamics behaviour;

- The measurements and simulations showed that the lake does not have velocity stratification in terms of power frequency. It was shown that, the velocity behaviour was very similar and presented a strong correlation with wind intensity, in which wind intensity of 3.0 m/s may impact the entire vertical column, therefore, being important an in loco meteorological station.

In the Brazilian lake, the three-dimensional hydrodynamic model Delft3D-Flow was calibrated and validated using high-frequency temperature measurements. The calibration period lasted 18 days, from 16th May to 03rd June 2016, corresponding to 440 hourly data for each one of the four respectively depths: surface (0.5 m), 2.5 m, 5.5 m and bottom (9.5 m). The validation period lasted 16 days, from 29th May to 14th June 2016, corresponding to 388 hourly values of water temperature for each of the same four depths. A second validation period lasted 88 days, from 15th May to 10th August 2015, corresponding to 2107 hourly values, for surface (0.5 m) depth was also simulated. The research carried out in Lake Pampulha showed that: - The main sensitivity parameters to calibrate the thermal behaviour of the lake are the Wind factor and Dalton coefficient;

- High-frequency measurement with hourly time step could detect sudden changes of water temperature with different amplitudes depending on the depth and using three-dimensional model it was possible to investigate the cause of sudden changes in the water temperature;

- The three-dimensional model, associated with high-frequency measurement, showed that detecting colder freshwater current and knowing the time trajectory throughout the deeper layer of the lake may contribute to implement and leverage restoration techniques and support water management for urban lakes that have hypoxia condition at the bottom of the lake;

- A three-dimensional model could accurately reproduce the alternation of stratification and mixing conditions along all simulated period, allowing a deeper analysis of a shallow tropical lake hydrodynamics

As shallow urban lakes have been the subject of much less study, this research could add important contributions to knowledge about the influence of hydrodynamics on phytoplankton behaviour in shallow and urban lakes. It was shown that shallow lakes are very instable and react strongly and rapidly to meteorological forces. Phytoplankton biomass displays high heterogeneities in space and time that result in high complexity to be able to measure its behaviour. Using a three-dimensional model, it was shown that hydrodynamics play an important role to better understand algal blooms and scum formation. This study shows that a calibrated and validated three-dimensional hydrodynamic and ecological model using high-frequency monitoring is essential for understanding water quality in shallow urban lakes. Otherwise, using the results of modeling, suggestions could be made to improve the current monitoring locations to better elucidate cyanobacteria behaviour. Therefore, these scientific contributions are promising to help in assessing and measure the distribution and resuspension of nutrients, sedimentation, and pollutants in lakes, and for understanding phytoplankton blooms more broadly.