Local flood forecasting capability for fluvial and estuarine floods: Use of GridStix for constraining uncertainty in predictive models

该项目旨在利用大量联网的GridStix深度传感器，提高对重要地点洪水泛滥和水位高度的预测，以提高洪水预警能力。该项目包括改进连接传感器和分布式计算资源的软件。这将允许分布式液压路径模型运行，并有可能通过实时使用传感器信息来减少预测中的不确定性。由于GridStix也有车载计算能力，因此也有可能为面临洪水风险的特定地点建立一个廉价的本地预报系统。所涉及的科学问题包括如何最好地使网络健壮，如何最好地约束洪水路径模型中的不确定性并改进其预测，以及如何最好地实现局部洪水预测模型。这项研究将在经常发生河流洪水的里布尔河和迪河口的潮汐系统上进行。这项研究是兰开斯特大学和布里斯托尔大学、普罗德曼海洋学实验室和环境局合作进行的。

This project aims to make use of lots of networked GridStix depth sensors to improve predictions of flood inundation and water level elevation at important locations with a view to improving flood warning capabilities. The project involves improving the software that links the sensors and distributed computing resources. This will allow distributed hydraulic routing models to be run, with the possibility of reducing the uncertainty in their predictions by using the sensor information in real-time. Since the GridStix also have on-board computing capabilities there is also a possibility of building a cheap local forecasting system for specific points at risk of flooding. The science questions involved include how best to make the netwroking robust, how best to constrain the uncertainty in flood routing models and improve their predictions, and how best to implement the local flood forecasting models. The research will be implemented on the River Ribble, subject to regular fluvial flooding, and the tidal system of the River Dee Estuary. The research represents a collaboration between Lancaster and Bristol Universities, the Proudman Oceanographic Laboratory and the Environment Agency.