Development of an improved hourly ensemble streamflow forecasting system for flood forecasting on small watersheds

加拿大拥有丰富的水资源，使其成为水电和洪水管理等与水有关的领域的领导者。尽管如此，近年来重大洪水造成了数百万美元的损失，已经提出了多项倡议，以评估更好地预测其时间和规模的方法，以帮助公共安全官员规划应急响应干预措施。在过去的25年里，水文气象咨询公司Hydro Météo帮助市政当局、公共安全机构和流域管理组织预测洪水，防止冰塞，并为魁北克省的河流提供仪器。他们在水文预报方面的专业知识使他们能够为客户提供定制分析，从而缩短响应时间，并尽可能地防止对基础设施的破坏和对人类生命的风险。目前，Hydro Météo使用的预报系统基于一个简单而巧妙的水文模型，该模型以每日时间步长运行，他们利用不同来源的天气预报运行他们的模型，将降水和温度转化为河流流量。Hydro Météo目前正在过渡到使用小时时间步长的多模型方法，以提高预报分辨率和准确性，特别是在较小的流域（20至2000平方公里）。然而，有效地实施一个预报系统，使用多种水文模型和多种天气预报产品的每小时的时间尺度仍然是一个关键问题，仍然是在科学界争论。这个拟议的项目将评价和制定方法，以便根据模型在不同水文制度中的表现进行次每日水文预报。这将有助于改善洪水风险管理的短期预测，也有助于水电运营公司更准确地估计流入水电水库的水量。

Canada is blessed with plentiful water resources which have contributed to making it a leader in water-related fields such as hydropower and flood management. Nonetheless, major floods have caused millions of dollars of damages in the recent years and multiple initiatives have been put forth to evaluate methods to better predict their timing and magnitude to help public safety officials plan emergency response interventions. For the last 25 years, the hydrometeorological consulting firm Hydro Météo has helped municipalities, public safety agencies and watershed management organizations to forecast floods, prevent ice jams and instrument rivers across the province of Québec. Their expertise in hydrological forecasting allows them to provide custom analyses for their clients, improving response times and preventing damages to infrastructures and risks to human life as much as possible.Currently, Hydro Météo uses a forecasting system based on a simple, yet skillful, hydrological model running on a daily time step, and they run their model using weather forecasts from different sources to transform precipitation and temperatures into river streamflow. Hydro Météo is currently transitioning to a multi-model approach using an hourly time step to improve forecast resolution and accuracy, especially on smaller (between 20 and 2000 square kilometers) watersheds. However, efficiently implementing a forecasting system using multiple hydrological models and multiple weather forecast products on hourly timescales remains a key issue which and is still debated in the scientific community. This proposed project will evaluate and develop methods to allow sub-daily hydrological forecasting based on model performance in different hydrological regimes. This will make it possible to improve short-term forecasting for flood risk management, but also for more precise estimates of inflows to hydropower reservoirs for hydropower operating utilities.