The latest published IPCC AR6 report indicated that Asian cities, particularly Chinese ones, have experienced severe flood consequences. These cities are affected by frequent cyclone-enhanced surges and intensive rainstorms. More than 157 Chinese cities have suffered from severe urban floods since 2006.

Uphill challenges would further exacerbate urban flood risk with large populations and socio-economic assets due to rapid urbanisation and climate change. For example, a 24-h accumulated rainfall of 460 mm hit Beijing on July 21, 2012, leading to a severe urban surface water flood and causing 79 deaths. The urban flood occurred in Zhengzhou on July 20, 2021, with a 1-h rainfall of 202 mm, resulting in the death of 339 people. This year also hit Beijing outskirt with over 700mm accumulated rainfall for 4 days during the late July and caused over 20 casualties under the adverse current from Typhoon Doksuri. The intensive low-pressure current further hits the North East China and Harbin area that caused severe damages on August 2-6, 2023. 

Hence, these floods have caused significant impacts on injuries, casualties, and economic losses during the last decade especially damaged the urban area with increasing populations and their property assets. Therefore, protecting our communities and improving urban flood resilience has become an urgent issue for stakeholders to improve the adaptability of the urban systems with increasing urban flood risk.

In 2015, the Sponge City Program (SCP) initiated had been conducted for urban flood management. Nature-Based Solutions (NBS) have been used to absorb stormwater as a “Sponge” via restored urban vegetation and soil surface. The SCP is largely increasing the flood protection standard reaching a 1-in-30-year return period flood event in the Chinese cities where adopted the practice. The SCP has only been initiated for about 7-8 years, which implies the local authorities still need more time to implement and expand the “Sponge” coverage in the urban districts. In recent years, some urban floods have occurred in SCP selected pilot cities (e.g., Shenzhen 2019, Wuhan 2020, etc.).

Indeed, these urban floods still caused severe consequences and impacts on the local communities and impacted the transport system (roads and subways), public services, and businesses. As urban floods become more frequent and severe, only relying on physical infrastructure could be insufficient to foster flood resilience under uncertainties.

There is a growing popularity for using human-centric information to foster urban flood resilience, with the inspiration from various globally recognised disaster reduction frameworks (i.e., the Sendai Framework Disaster Reduction) on the Disaster Risk Reduction (DRR). Under enormous challenges on climate change and urbanisation, the urgency and necessity to reduce flood risk is a "must-have" strategies across the urban future. DRR is aligned with the Sendai Framework, the United Nations 2030 Agenda for Sustainable Development, and the UN Sustainable Development Goals. 

Indeed, we would like to emphasis that the Public responses to urban floods can differ in cities or periods. Public responses are crucial for improving urban flood resilience during preparedness and response.

In this study, we have investigated via public response analytics. Hence, stakeholders and policymakers can make better decisions and co-productions in response to future flood events. Digital media has recently applied to understanding the explicit facts, constraints, and challenges during the past urban flood events.

However, we have found most of the existing studies applied social media data (e.g., Weibo, Twitter, Facebook, etc.) at a small scale (e.g., district level in the city area). Existing studies were conducted for some specific urban flood events. It is hard to compare public responses among different cities. The major reason for these gaps can be summarised as digital media data analysis requiring much time and data resources.

Since 2015, building on advances in digital media mining, an inspiring technology that could be transformative in this space, namely, the “Global Knowledge Graph (GKG) of the Global Database of Events, Language and Tone (GDELT)” project. This tool is powerful that the GKG contains graph data. The GKG is coded from the contents of news media articles from nearly all corners of every country using automated algorithms. It offers various themes to describe the characteristics (e.g., time, emotions, locations) of natural disasters.

Notably, the GKG can overcome a significant challenge in assessing media trends (i.e., access to data). It provides structured data which can be analysed directly. Moreover, news media tends to be more critical and objective than social media because news articles are selected by reporters or editors.

Hence, the GKG is a valuable source via media data analytics and useful in urban flood resilience and planning research. This research aims to answer the research question: how to enhance urban flood risk management and resilience via media data analytics?

There are actually about 27 prefecture-level SCP pilot cities were taken as cases . A systematic investigation of news media responses to urban floods is conducted using the GKG data. 

Firstly, the pattern of media attention and public sentiment is investigated, followed by an investigation of the relationship between news media responses and rainfall. After that, the variation of news media responses during an urban flood event is described. Finally, we discussed the opportunities and challenges in GKG analytics and other new technologies for enhancing urban flood resilience.

Through this study, we will contribute to and implement more resilient urban flood management in China's sponge cities and extensively in other Chinese cities. We hope readers can enjoy our findings and this study recommended the application of new technologies (e.g., GDELT GKG) for improving flood resilience. That will help enhance knowledge transformation among stakeholders, particularly from the previous flood events and experiences (e.g., reflections from communities and media). 

In future studies, incorporating social-economic data will be helpful for a better understanding of preparedness and response disparities of urban flood resilience. 

Please feel free to follow up the paper authored by Xiaohui Lu and us for the full findings here: https://www.nature.com/articles/s41598-022-24370-8#citeas