Abteilung Oberflächengewässer

One-dimensional hydrodynamic models are nowadays widely recognized as key tools for lake studies. They offer the possibility to analyze processes at high frequency, here referring to hourly timescales, to investigate scenarios and test hypotheses. Yet, simulation outputs are mainly used by the modellers themselves and often not easily reachable for the outside community. We have developed an open-access web-based platform for visualization and promotion of easy access to lake model output data updated in near-real time (http://simstrat.eawag.ch, last access: 29 August 2019). This platform was developed for 54 lakes in Switzerland with potential for adaptation to other regions or at global scale using appropriate forcing input data. The benefit of this data platform is practically illustrated with two examples. First, we show that the output data allows for assessing the long-term effects of past climate change on the thermal structure of a lake. The study confirms the need to not only evaluate changes in all atmospheric forcing but also changes in the watershed or throughflow heat energy and changes in light penetration to assess the lake thermal structure. Then, we show how the data platform can be used to study and compare the role of episodic strong wind events for different lakes on a regional scale and especially how their thermal structure is temporarily destabilized. With this open-access data platform, we demonstrate a new path forward for scientists and practitioners promoting a cross exchange of expertise through openly sharing in situ and model data.

We analyzed the processes affecting the methane (CH₄) budget in Lake Kinneret, a deep subtropical lake, using a suite of three models: (1) a bubble model to determine the fate of CH₄ bubbles released from the sediment; (2) the one-dimensional physical lake model Simstrat to calculate the mixing dynamics; and (3) a biogeochemical model implemented in Aquasim to quantify the CH₄ sources and sinks. The key pathways modeled include diffusive and bubble release of CH₄ from the sediment, aerobic CH₄ oxidation, and atmospheric gas exchange. The temporal and spatial dynamics of dissolved CH₄ concentrations observed in the lake during 3 years could be well represented by the combined models. Remarkably, the relative contributions of ebullition and diffusive transport to the accumulation of CH₄ in the hypolimnion during the stratified period could not be accurately constrained based only on the observed evolution of CH₄ concentrations in the water column. Importantly, however, our analysis showed that most (∼99%) of the CH₄ supplied to the water column by bubble dissolution and diffusive transport from the sediment is aerobically oxidized, whereas a substantial fraction (∼60%) of the sediment-released bubble CH₄ is directly transported to the atmosphere. Ebullition is thus responsible for the bulk of the emissions from Lake Kinneret to the atmosphere. Therefore, as in all freshwaters, ebullition quantification is crucial for accurately assessing CH₄ emissions to the atmosphere. This task remains challenging due to high spatio-temporal variability, but combining in situ measurements with a process-based modeling can help to better constrain flux estimates.

This paper presents an improvement of a one-dimensional lake hydrodynamic model (Simstrat) to characterize the vertical thermal structure of deep lakes. Using physically based arguments, we refine the transfer of wind energy to basin-scale internal waves (BSIWs). We consider the properties of the basin, the characteristics of the wind time series and the stability of the water column to filter and thereby optimize the magnitude of wind energy transferred to BSIWs. We show that this filtering procedure can significantly improve the accuracy of modelled temperatures, especially in the deep water of lakes such as Lake Geneva, for which the root mean square error between observed and simulated temperatures was reduced by up to 40 %. The modification, tested on four different lakes, increases model accuracy and contributes to a significantly better reproduction of seasonal deep convective mixing, a fundamental parameter for biogeochemical processes such as oxygen depletion. It also improves modelling over long time series for the purpose of climate change studies.

Anthropogenic heat emissions into inland waters influence water temperature and affect stratification, heat and nutrient fluxes, deep water renewal, and biota. Given the increased thermal stress on these systems by growing cooling demands of riparian/coastal infrastructures in combination with climate warming, the question arises on how to best monitor and manage these systems. In this study, we investigate local and system-wide physical effects on the medium-sized perialpine Lake Biel (Switzerland), influenced by point-source cooling water emission from an upstream nuclear power plant (heat emission ∼700 MW, ∼18 W m⁻² lake wide). We use one-dimensional (SIMSTRAT) and three-dimensional (Delft3D-Flow) hydrodynamic numerical simulations and provide model resolution guidelines for future studies of thermal pollution. The effects on Lake Biel by the emitted excess heat are summarized as: (i) clear seasonal trend in temperature increase, locally up to 3.4°C and system-wide volume mean ∼0.3°C, which corresponds to one decade of regional surface water climate warming; (ii) the majority of supplied thermal pollution (∼60%) leaves this short residence time (∼58 days) system via the main outlet, whereas the remaining heat exits to the atmosphere; (iii) increased length of stratified period due to the stabilizing effects of additional heat; (iv) system-wide effects such as warmer temperature, prolonged stratified period, and river-caused epilimnion flushing are resolved by both models whereas local raised temperature and river short circuiting was only identifiable with the three-dimensional model approach. This model-based method provides an ideal tool to assess man-made impacts on lakes and their downstream outflows.

The goal of reducing carbon fuel and thereby saving energy will increase the use of lake water for heating and cooling of riparian infrastructures. This raises the question of which heat use designs meet the ecological and technical requirements for lakes, particularly in regard to climate warming. Thus, this study explores heat use effects on the temperature and stratification of a large, deep, temperate lake by applying the one-dimensional k-epsilon model SIMSTRAT to various forcing scenarios. Several design parameters, such as extraction and discharge depth, and their effects were assessed. Additionally, 21st century climate projections were used to evaluate the effects of climate change relative to those of heat use. Generally, the study showed only minor effects for a realistic heat demand of ±2 W m⁻² quite independent of the heat extraction/discharge modes. Mean water temperature changed less than ±0.2°C as long as there was no discharge into the deepest layers. Water extraction and discharge at the surface had the least thermal influence. To relate to climate change, heat use was scaled up to +85 W m⁻². Resultant simulations showed that such (unrealistic) anthropogenic, lake-based “thermal pollution” would have a comparable influence to that of climate change. Conversely, heat extraction could damp or even compensate climate-induced warming. The present study concludes that (i) there are minor effects on water temperatures, stratification, and seasonal mixing due to heat use of up to ±2 W m⁻² and (ii) those influences are insignificant relative to the expected climate change.

A numerical model was developed for the prediction of the density stratification of lakes and reservoirs. It combines a buoyancy-extended k-ϵ model with a seiche excitation and damping model to predict the diffusivity below the surface mixed layer. The model was applied to predict the seasonal development of temperature stratification and turbulent diffusivity in two medium-sized lakes over time periods ranging from 3 weeks to 2 years. Depending on the type of boundary condition for temperature, two or three model parameters were optimized to calibrate the model. The agreement between the simulated and the observed temperature distributions is excellent, in particular, if lake surface temperatures were prescribed as surface boundary condition instead of temperature gradients derived from heat fluxes. Comparison of different model variants revealed that inclusion of horizontal pressure gradients and/or stability functions is not required to provide good agreement between model results and data. With the aid of uncertainty analysis it is shown that the depth of the mixed surface layer during the stratified period could be predicted accurately within ±1 m. The sensitivity of the model to several parameters is discussed.