Frequency and risk analysis of extreme hydrological phenomena

对洪水、河流流量减少或强降雨等极端水文事件的准确估计对于管理水电设施、建立城市下水道网络的流量和规划水利工程等活动至关重要。这一估计依赖于加拿大多年来收集的水文数据。根据水文记录，加拿大科学家和工程师对未来发生的罕见水文事件进行了估计。估计中涉及的一个重要因素是统计频率模型与观测数据的拟合。有许多模型可供选择，也有许多方法可以将模型与数据进行拟合。这项研究的重点是寻找有助于选择最佳频率模型来拟合观测数据的方法，以及最合适的拟合这些模型的方法。一个重要的目标是为未来罕见的水文事件制定稳定的、物理上合理的、不高估或低估这些事件的真实价值的估计程序。错误的估算可能会导致水利设计和水资源系统管理的错误。我们预计，这项研究的结果将为加拿大水文学家和工程师在水文风险分析和水管理的频率分布和拟合方法之间的选择提供改进的指导方针。我们还预计，这项研究将有助于更好地量化在估计罕见水文事件时所固有的统计误差，因为到目前为止，在加拿大这样一个地理辽阔的国家的不同地区收集的水文测量量有限。对收集到的数据进行适当的分析，有助于正确估计未来必然发生的罕见但严重的水文事件。科学家、工程师、规划者、社会工作者和决策者需要做好充分准备，以预测和管理这类未来事件。

An accurate estimation of extreme hydrological events such as floods, low river flows or severe rainfall is essential for activities such as managing hydro-electrical facilities, establishing flow capacities of urban sewer networks and planning hydraulic works on water streams. This estimation relies on hydrological data gathered in Canada over many years. Based on hydrologic records, Canadian scientists and engineers make estimates of rare hydrologic events that would occur in the future. An important element involved in the estimation is the fitting of statistical frequency models to the observed data. There are many models to choose from, and many methods for fitting the models to the data. This research focuses on finding means that help in selecting the best possible frequency models to fit the observed data and the most appropriate methods of fitting these models. An important goal is to arrive at estimation procedures of future rare hydrological events that are stable, physically justifiable, and that do not overestimate or underestimate the true value of the events. An incorrect estimation may lead to errors in the hydraulic design and the management of water resources systems. We expect that the results of this research will provide Canadian hydrologists and engineers with improved guidelines for choosing between frequency distributions and fitting methods for hydrological risk analysis and water management. We also expect the research to help better quantify the statistical errors inherent in the estimation of rare hydrological events due to the limited amounts of hydrological measurements gathered up until now in various parts of a geographically vast country such as Canada. The proper analysis of this gathered data is instrumental to the proper estimation of rare but severe hydrological events that are bound to occur in the future. Scientists, engineers, planners, social workers and decision makers need to be well prepared to forecast and manage such future events.