Geospatial analysis and modelling of riverine systems: realising the potential for satellite supplied data assimilation

该项目旨在分析创建河流环境数字孪生模型的潜力和实用性，供水资源管理人员和工程师使用。河流在空间和时间上的高变化率使其难以进行现场监测，特别是在较长的一段时间内。因此，卫星遥感已被确定为改进系统管理的一个关键工具。许多河流特征已经从空间测量，但在大多数情况下，这些仅限于概念验证研究，缺乏转化为可操作和适用产品的研究。因此，目前有一个研究利基被发现在生产数字孪生虚拟环境供电地球观测，以解决真实的世界工程管理问题。首先，这将涉及检查哪些河流特性是最重要的不同类型的河流管理者或工程师。这些将被组织成那些目前可以通过地球观测测量，那些理论上可以测量，但目前缺乏算法或方法来实施和那些将需要现场传感器或更高分辨率的图像。前一类将被开发成一个新的仪表板，类似于一个低级别的数字孪生模型，它将整理这些现有的方法，并为管理者提供一个最先进的决策依据。云计算软件，如谷歌地球引擎，将被用来提供框架，以组成这一同化所需的大数据和处理能力，使其可行的使用。使用此仪表板作为基础，有三个工作流程要进行。首先，类似的系统和API对于非专家来说非常难以使用，与计算机游戏所采取的巨大可视化飞跃相比，这些可以被视为笨重和没有吸引力。谷歌已经创建了一些概念性的洪水地图，这些地图使用游戏引擎来提供一个真实的想法，预测洪水在3D山谷中会是什么样子，这个工作包的目标是将其中一些改进纳入仪表板。第二，建议开展一系列研究，对已确定地区的卫星数据、航空遥感和无人驾驶飞机遥感的不同分辨率进行比较。将评估直接用例，以概述对公司支付专有卫星数据或现场方案以增强其数字孪生的业务案例的理解。这还包括评估示例场景中所需的结对率（虚拟环境需要更新的频率），以便更好地理解系统。最后，将整理一个工作包来调查目前无法通过卫星遥感测量的特征的可能方法，可以通过航空遥感或传感器系统进行测量。预计管理人员的一些关键变量将不可用，因此数字孪生模型的使用将依赖于与模型无缝匹配的替代方法。因此，其中一些方法将在实地方案中用传感器或较小规模的遥感方法进行测试，评估其可能的贡献和相互作用。最终产品将是对卫星地球观测能力的全面了解，以充分发挥其革命性河流管理和工程的潜力。这将通过对所有这些可用数据和算法进行整理的数字孪生模型的建设和改进来促进。沿着评估这一虚拟环境不同方面可能的改进和可行性的工作包，包括其视觉交流能力、在特定情况下对更高质量数据的需要以及在卫星数据无法完成任务的情况下增加外部数据集以进行补偿的能力

This project aims to analyse the potential and practicalities of creating a Digital Twin of a riverine environment for use by water managers and engineers. The high rate of spatial and temporal change in rivers make them difficult to monitor in-situ especially over an extended period of time. As a result, satellite remote sensing has been identified as a key tool to improve system management. Many riverine characteristics have been measured from space but for the most part, these are limited to proof-of-concept studies and a lack of research into conversion into actionable and applicable products. Thus, there is currently a research niche to be found in producing a Digital Twin virtual environment powered by Earth Observation to solve real world engineering management problems. First this would involve examining which river characteristics are most important for different types of river manager or engineer. These will be organised into those that can currently be measured through earth observation, those that theoretically can be measured but currently lack algorithms or methodology to be implemented and those that would require in-situ sensors or higher resolution imagery. The former category will be developed into a novel dashboard, analogous to a low-level Digital Twin, which will collate these existing methods and provide a state of the art for managers to rely on for decision making. Cloud computing software, such as Google Earth Engine, will be relied upon to provide the framework to compose the big data required for this assimilation and the processing power to make it feasible to use. Using this dashboard as a foundation there are three workflows to be undertaken. Firstly, similar systems and APIs are very difficult for a non-expert to use and when compared, to the vast visualisation leaps taken by computer games, these can be viewed as clunky and unappealing. Google has created some conceptual flood maps which use gaming engines to give a real idea of what predicted flood water would look like in a 3D valley and this work package will aim to incorporate some of these improvements into the dashboard. Secondly, it is proposed that a set of studies will be undertaken to compare different resolutions of satellite data, aerial and drone based remote sensing in the areas already identified. Direct use cases will be assessed, in order to outline an understanding of what the business cases are for companies to pay for proprietary satellite data or field programmes to enhance their Digital Twins. This would also include an appreciation of the required twinning rate (how often the virtual environment would need to be updated) during the example scenarios in order to better understand the system.Finally, a work package will be put together to investigate the possible ways in which characteristics, not currently measurable through satellite remote sensing, can be measured through aerial remote sensing or sensor systems. It is expected that some of the key variables for managers will not be available and therefore the use of a Digital Twin will be reliant on alternative methods fitting seamlessly into the model. Therefore, some of these methods will be tested in field programmes with sensors or smaller scale remote sensing methods assessing their possible contribution and interactions. The final product will be a comprehensive understanding of the ability for satellite Earth Observation to live up to its potential to revolutionise river management and engineering. This will be facilitated through the construction and improvements made to a Digital Twin collating all these available data and algorithms. Along with work packages assessing the possible improvements and viability of different aspects of this virtual environment including its ability to visually communicate, the need for higher quality data in specific cases and the ability to add external datasets to compensate where satellite data cannot carry out the task