Process-based modelling of turbulent mixing in stratified natural waters

这项研究的长期目标是开发一套新的工具和方法来模拟自然水域中的湍流混合，例如湖泊、河口和海洋，其中水的密度通常随着深度的增加而增加（即流体稳定分层）。理解分层湍流混合对于一系列问题至关重要，从模拟五大湖的生态健康到预测气候变化中的海洋热吸收。

The long-term objective of this research is to develop a novel set of tools and approaches for modelling turbulent mixing in natural waters, such as lakes, estuaries and oceans, in which the density of water typically increases with depth (i.e. the fluid is stably stratified). Understanding stratified turbulent mixing is crucial for a range of problems, from modelling the ecological health of the Great Lakes to predicting oceanic heat uptake in a changing climate. Current methods to predict mixing rely heavily on empirical relations, and a major shortcoming of these relations is that they are largely unaware' of the processes that are responsible for mixing, thus introducing great uncertainties in the predictions. My research will address this issue by adopting a process-based approach, where the critical links between the overall mixing properties of the physical systems and the underlying flow processes will be established. These generic flow processes can manifest themselves in environmental flows as distinct structures known as coherent structures (CSs), the latter serving as the key to understanding the mixing properties of natural flow processes. Through this research, CSs that are relevant to mixing heat, salt and pollutant in lakes and oceans will be identified and examined. The understandings of these CSs will then be utilized to develop processed-based models for mixing. The research will be driven by large-scale numerical simulations of stratified turbulent flows, combined with state-of-the-art mathematical tools for analyzing simulation data and developing new models. The program can be divided into three aims: A. Perform simulations in two flow configurations, i.e. stratified momentum wake and stratified shear box; B. Extract coherent structures (CSs) from the numerical database of stratified turbulence obtained in aim A, using two novel mathematical tools, i.e. dynamic mode decomposition and flow recurrence analysis; and C. Diagnose the mixing properties of the CSs identified and synthesize these properties into predictive mixing models appropriate for natural flow systems. This program will (i) develop a new generation of quantitative tools to infer turbulent mixing rates for numerical