Using big data to design resilient coastal cities

利用大数据设计弹性沿海城市

• 批准号: 2273819
• 金额: --
• 依托单位: University of Liverpool
• 依托单位国家: 英国
• 项目类别: Studentship
• 财政年份: 2019
• 资助国家: 英国
• 起止时间: 2019 至 无数据
• 项目状态: 未结题
• 来源: https://gtr.ukri.org/projects?ref=studentship-2273819
• 关键词: Using; big; data; design; resilient

Coastline monitoring has long played a pivotal role in the management of the physical and economical characteristics of coastal cities. The susceptibility to fast-onset erosion defines an aspect of the vulnerability of these coastal cities, whilst resilience is contrastingly measured by the ability to prepare and plan for, absorb, recover from, and more successfully adapt to adverse events (NRC, 2012). Coastline monitoring is crucial in increasing the resilience of a coastal setting and cities around the world have developed innovative strategies to combat the impacts of coastal erosion and flooding. One such strategy is the implementation of remote monitoring techniques to obtain a comprehensive understanding of the spatio-temporal dynamics of the coastline. The largest use of remote sensing in coastal management is for measuring the change in coastal environments over time, likely due to increasing resolution in remote sensing imagery (Green et al., 1996). The application of remote sensing technology to monitor coastlines can identify intricacies in coastal dynamics that are very complex and likely not picked up by standard beach profiling records (Wallbridge, 2018). The chapters of this thesis represent how performing data analytics on high resolution remote sensing data provides a comprehensive understanding of local coastal dynamics, as well as how humans are interacting with coasts. The analysis presented observes a detailed look into spatio-temporal trends, using a combination of image processing techniques and machine learning. Firstly, the thesis observes how stationary X-band radar data designed for physical coastline monitoring can be repurposed to observe human mobility trends across a beach environment. A particle tracking technique is couples with a machine learning classifier to identify hotspots of human movement, while retaining anonymity. Secondly, the thesis utilises novel radar-derived digital elevation models (DEMs) with a high temporal resolution to identify trends in intertidal morphological dynamics over various temporal scales, from short-term observation over a single tidal cycle, to annual changes. Through understanding how these short-term dynamics contribute to long-term change, we can begin to provide insightful and pivotal information for coastal managers in decision making processes such as intervention and long-term defence strategies. The third chapter combines aspects of human mobility and population growth with the changing coast. Temporal clustering techniques are used on open source population and shoreline change datasets to characterise the interactions between coastal cities and coastal dynamics. Through this technique, areas of high and low risk to coastal hazards such as coastal erosion and flooding can be identified, as well as an understanding of how variations in shoreline change can impact population dynamics.

长期以来，海岸线监测在沿海城市的自然和经济特征管理中发挥着关键作用。对快速侵蚀的敏感性是这些沿海城市脆弱性的一个方面，而复原力则是通过准备和计划、吸收、恢复以及更成功地适应不利事件的能力来衡量的（NRC，2012年）。海岸线监测对于提高沿海环境的复原力至关重要，世界各地的城市都制定了创新战略，以应对海岸侵蚀和洪水的影响。其中一项战略是采用远程监测技术，以全面了解海岸线的时空动态。遥感在沿海管理中的最大用途是测量沿海环境随时间的变化，这可能是由于遥感图像分辨率的提高（绿色等人，1996年）。应用遥感技术监测海岸线可以确定海岸动态的复杂性，这些动态非常复杂，可能无法通过标准的海滩剖面记录获得（Wallbridge，2018年）。本论文的章节介绍了如何对高分辨率遥感数据进行数据分析，以全面了解当地的海岸动态，以及人类如何与海岸互动。该分析使用图像处理技术和机器学习相结合的方法，详细研究了时空趋势。首先，本文观察如何固定的X波段雷达数据设计的物理海岸线监测可以被重新利用，以观察人类的流动趋势在海滩环境。粒子跟踪技术与机器学习分类器相结合，以识别人类运动的热点，同时保持匿名性。其次，论文利用新的雷达衍生的数字高程模型（DEM），具有高的时间分辨率，以确定在不同的时间尺度上，从短期观测在一个单一的潮汐周期，以每年的变化趋势在潮间带形态动态。通过了解这些短期动态如何促进长期变化，我们可以开始为沿海管理人员提供有见地的关键信息，以帮助他们制定干预和长期防御战略等决策过程。第三章将人口流动和人口增长与不断变化的海岸结合起来。时间聚类技术用于开源人口和海岸线变化数据集，以模拟沿海城市和沿海动态之间的相互作用。通过这一技术，可以确定海岸侵蚀和洪水等沿海灾害的高风险和低风险地区，并了解海岸线变化如何影响人口动态。