Regional Analysis and modelling of river temperature

水温被认为是一个关键的河流变量。河流的极端水温与重要渔场的关闭和鱼类死亡有关。在接下来的五年里，我打算专注于一个新的研究方向，即河流水温的区域分析和建模。将追求以下具体目标：1)开发识别热均质区域（THR）的方法，即具有相似热状态的河流群；2)开发创新的方法来表征和模拟每组河流的季节性热状态；3)在非平稳环境下，开发模型来评估每一组的水温异常和极端情况；4)开发加拿大首个区域确定性水温模型框架。***定义热均匀区域的新方法将利用小波分析，并通过样本阶数统计量的比较来检验均匀性。水温时间序列的季节成分将使用灵活的Beta分布进行建模。本文将建立一个带有外生变量的非平稳区域自回归模型（ARX）来模拟水温异常（减去季节趋势后的日温度残差）。***极值将使用全区域非平稳峰值超过阈值（POT）方法、分位数回归、copulas和递进追踪回归进行建模。***最后，将利用CEQUEAU水文和水温模型建立区域确定性模型框架。拟议的研究不仅是实施有用管理工具的必要步骤，而且还将有助于发展中领域的多学科研究和培训。拟议研究的影响不仅限于渔业管理，还将对市政用水的进出管理产生影响，因为水质高度依赖于水温

Water Temperature is recognized as a key river variable. Extreme water temperatures in rivers have been linked to the closing of important fisheries and fish mortality. Over the next five years, I intend to focus on a new research direction related to regional analysis and modelling of river water temperature. The following specific objectives will be pursued: 1) develop methods to identify thermally homogenous regions (THR), i.e. groups of rivers with similar thermal regimes; 2) develop innovative approaches to characterize and model the seasonal thermal regime of rivers within each group; 3) develop models to assess water temperature anomalies and extremes within each group in a non-stationary context; and 4) develop the first Canadian framework for regional deterministic water temperature modelling. ***New methods for the definition of thermally homogenous regions will make use of wavelet analysis and homogeneity will be tested with the comparison of sample order statistics. The seasonal component of the water temperature time series will be modelled using the flexible Beta distribution. A non-stationary regional AutoRegressive model with eXogenous variables (ARX) will be developed to model the water temperature anomalies (daily temperature residuals once the seasonal trend has been subtracted). ***Extremes will be modelled using a fully regional non-stationary Peaks-Over-Threshold (POT) method, quantile regression, copulas and progression pursuit regression. ***Finally, a regional deterministic model framework will be developed with the CEQUEAU hydrological and water temperature model. ***The proposed research is not only a necessary step in the implementation of useful management tools, but it will also contribute to multidisciplinary research and training in a developing field. The impact of the proposed research is not limited to fisheries management and will have repercussions in the management of municipal water intakes and outflows because water quality is highly dependent on water temperature.**