Risk management has reduced vulnerability to floods in many regions, yet their impacts continue to rise. Understanding the drivers of these changing impacts is urgent for effective action, but empirical evidence remains limited, particularly from long-term historical perspectives. Drawing on extensive Chinese historical documents, this study develops a composite index to quantify the overall societal impacts of floods, as manifested across six interrelated subsystems: environment, production, infrastructure, population, economy, and social order. An annual series of the flood impact index for Sichuan, southwestern China, is reconstructed for the period 1644–1911 (the Qing dynasty). Flood impacts exhibit a fluctuating yet overall increasing trend, with three turning points (1727, 1779, and 1856) defining four phases. These phases are characterized respectively by low flood frequency with limited impacts, increasing mortality, recurrent famine, and widespread disruptions to socioeconomic order. Notably, from the nineteenth century, cascading effects became increasingly pronounced, complicating impact chains and amplifying flood impacts across multiple interconnected subsystems. Drawing on the IPCC risk framework and integrating natural and socio-economic indicators, this study identifies the dominant drivers of the stepwise escalation of flood impacts. The increase in impacts from Phase 1 to Phase 2 was driven by rising exposure associated with rapid population growth and cropland expansion. The shift from Phase 2 to Phase 3 was dominated by increasing vulnerability linked to declining per capita cropland availability and frequent warfare. The transition to Phase 4 resulted from the combined effects of rising hazard, exposure, and vulnerability.

Historical experience suggests the need for a holistic, systems-based approach to flood risk management. The Sichuan case illustrates how reductions in vulnerability can be outweighed by rising exposure, a dynamic that remains evident in contemporary climate adaptation. Rather than prioritizing vulnerability alone, hazard, exposure, and vulnerability need to be considered jointly. Moreover, early identification and intervention targeting impact events with cascading potential are critical for limiting damage in increasingly interconnected systems.

How to cite: Chen, S. and Yang, L. E.: Long-term dynamics of flood impacts in Sichuan, China (1644–1911) underscore a holistic approach to flood risk management, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-10896, https://doi.org/10.5194/egusphere-egu26-10896, 2026.

In past decades, urbanisation has risen around the world 1, increasing risk and exposure to shocks 2. Resilience theory offers valuable perspectives for understanding complex socio-ecological systems and their sustainable management 3,4,5, and for improving adaptation to climate change 6. Urban resilience refers to the ability of social, ecological and technical components to withstand, adapt to, and recover from disturbances across spatial and temporal scales 7. Studies have investigated sets of indicators that measure system dimensions separately to assess resilience against hazards (see 8,9). This method allows for the assessment of multiple system components at a given point in time. However, these components interact across spatial and temporal scales, creating temporal trade-offs and path-dependencies. Investigating these dynamics can significantly enhance the understanding of how urban resilience evolves and how its drivers operate over time 4,10,11,12. To advance urban resilience assessment, research should integrate multiple system components and examine their dynamics across different locations, enabling a more contextual understanding of resilience trajectories. In this study, I propose a methodological framework that uses openly published historical information by municipalities to track changes in urban systems over the past 30 years in European cities. The results can inform researchers, urban planners, and policymakers about how changes in the built environment have influenced social and environmental conditions over time, and how these changes are linked to increasing vulnerabilities and risks across urban systems. 

References

[1] Liu, X. et al. "High-spatiotemporal-resolution mapping of global urban change from 1985 to 2015." Nat Sustain. 2020;3(7):564–70. 
[2] Elmqvist, T. et al. "Urbanization in and for the Anthropocene." NPJ Urban Sustain. 2021;1(1):6.
[3] Folke, C. et al. "Resilience thinking: integrating resilience, adaptability and transformability." Ecol Soc. 2010;15(4).
[4] Chelleri, L. et al. "Resilience trade-offs: addressing multiple scales and temporal aspects of urban resilience." Environ Urban. 2015;27(1):181–98.
[5] Elmqvist, T. et al. "Sustainability and resilience for transformation in the urban century." Nat Sustain. 2019;2(4):267–73.
[6] Leichenko, R. "Climate change and urban resilience." Curr Opin Environ Sustain. 2011;3(3):164–8. 
[7] Meerow, S., Newell, J.P. and Stults, M. "Defining urban resilience: A review." Landsc Urban Plan. 2016;147:38–49.  
[8]  Osei-Kyei, R. et al. "Critical analysis of the emerging flood disaster resilience assessment indicators." Int J Disaster Resil Built Environ. 2025;16(3):417–36.
[9]  Zhu, S. et al. "Enhancing urban flood resilience: A holistic framework incorporating historic worst flood to Yangtze River Delta, China." Int J Disaster Risk Reduct. 2021;61:102355.
[10] Meerow, S, and Newell, J.P. "Urban resilience for whom, what, when, where, and why?" Urban Geogr. 2019;40(3):309–29.  
[11]  Sharifi, A. "Resilience of urban social-ecological-technological systems (SETS): A review." Sustain Cities Soc. 2023;99:104910.  
[12]  Casali, Y., Aydin, N.Y., and Comes, T. "A data-driven approach to analyse the co-evolution of urban systems through a resilience lens: A Helsinki case study." Environ Plan B Urban Anal City Sci. 2024;51(9):2074–91.  

How to cite: Casali, Y.: A framework to analyze the evolution of urban systems for resilience assessment, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-1869, https://doi.org/10.5194/egusphere-egu26-1869, 2026.

The Tea-Horse Road area (茶马古道地区) spans the Hengduan Mountains and the eastern edge of the Tibetan Plateau, an area characterized by complex geography and frequent human activity. Over the past two millennia, the region has repeatedly faced floods of varying scales but has demonstrated significant flood resilience. As climate change intensifies, learning from past flood management strategies is crucial to enhancing current resilience. However, due to fragmented literature, discontinuous records, and limited regional attention, no long-term dataset has been available for flood resilience analysis. To fill this gap, this study developed a framework for quantifying long-term flood resilience and constructed the “Tea-Horse Road Flood Resilience Dataset (THR-FRD)”, compiling flood records from AD 0 to 2025. The dataset has a temporal resolution of 50 years, with spatial resolution based on county-level administrative divisions from historical periods. Data sources include local chronicles, archival documents, ethnographic surveys, archaeological evidence, and observational data. The dataset is structured into three core sub-databases: Exposure, Vulnerability, and Risk, to quantitatively assess flood resilience. The Exposure sub-dataset records the frequency, intensity, and affected areas of floods; the Vulnerability sub-dataset analyzes social, economic, and environmental vulnerabilities; and the Risk sub-dataset evaluates the actual damage caused by floods, including casualties, property loss, and infrastructure damage. Flood resilience is assessed through a comprehensive evaluation of exposure, vulnerability, and risk, and can be calculated using a weighted model and normalization method. To maximize the utility of this dataset, the THR-FRD will be open-source and scalable, available in both Chinese and English, and retain original records. It will serve scholars from fields such as history and geography, providing decision support and facilitating interdisciplinary research.

How to cite: Ai, M.: Construction and Application of the Tea-Horse Road Area Flood Resilience Dataset (THR-FRD), EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-14125, https://doi.org/10.5194/egusphere-egu26-14125, 2026.

The Australian tropical savanna is among Earth’s most fire-prone regions. For millennia, Aboriginal Australians used prescribed burning to improve habitats for food plants and herbivores and to mitigate high intensity fires ignited by lightning in the late dry season. However, these practices were rapidly and profoundly interrupted beginning in the late 19^th and early 20^th centuries with the arrival of European pastoralists. Some studies have suggested that as a result of this reduction in early dry season, low intensity burning, late dry season, high temperature fire activity increased, with deleterious effects on ecosystems. However, as Aboriginal burning was curtailed, the introduction of cattle (as well as sheep and donkeys) reduced the grassy fuel layer. Developing a clear picture of baseline fire activity prior to the pastoralist era is important because bushfire intensity modulates greenhouse gas emissions from tropical savannas and modulates savanna and rainforest ecosystem dynamics. Reconstructing bushfire frequency and intensity is complicated by limited historical records of burning prior to the late 20^th century, and few naturally-occurring, high-resolution, fire-sensitive archives.

In order to place 20^th century bushfire into a long-term context, we reconstructed fire activity at sub-decadal resolution for the majority of the last millennium using polycyclic aromatic hydrocarbons (PAH) in three precisely-dated and fast-growing stalagmites from cave KNI-51, located in the tropical savanna of northeastern Western Australia. The molecular weights of PAH are tied to combustion temperature (i.e., higher molecular weights (HMW) form at higher temperature fires), and thus our record preserves evidence of both the timing and temperature of bushfire. In order to integrate the multiple stalagmites used to construct this composite record, we normalized each PAH class (low and high molecular weight) to the total PAH abundance in each sample.

The KNI-51 stalagmite record reveals that high temperature fire was a regular component of the Australian tropical savanna throughout the last millennium, suggesting late dry season fires were commonplace. However, soon after the arrival of European pastoralists in the 1880s, the frequency of high temperature fires decreased markedly and remained low until the record end of the KNI-51 record in 2009 CE. This shift in bushfire regime, which is apparent based on decadal averages of normalized HMW PAH and through breakpoint analysis, occurred despite severe reductions of early dry season burning by Aboriginal Australians. It also occurred during a monsoon rainfall regime, determined using oxygen isotope ratios from the same stalagmites, that was close to the last millennium average. Thus, after discounting prescribed burning and hydroclimate, we ascribe this decrease in high temperature bushfire to reductions by cattle of grassy fuel loads. The anomalous nature of the 20^th century Australian tropical savanna pyroscape in the area of KNI-51 highlights the complexities associated with re-establishing the pre-pastoralist era bushfire regime in this region.

How to cite: Denniston, R., Ondei, S., Argiriadis, E., and Bowman, D.: Declining High Temperature Bushfire in Australian Tropical Savanna Following Arrival of European Pastoralists, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-2157, https://doi.org/10.5194/egusphere-egu26-2157, 2026.

In oceanic temperate forests, as in more fire-prone ecosystems, fire contributes to shape the environment, define relationships between nature and society, orient forest uses, and influence biogeochemical cycles in ways that still need to be better understood. Fire weather is expected to increase also in these ecosystems over the coming decades, raising major concerns, and highlighting the need to better understand their past dynamics and biogeochemical implications. The Fontainebleau forest, located in France’s Ile-de-France region, has been documented since the 11^th century, when it first became a royal forest, and is now a famous and highly frequented forest. At the crossroads of multiple uses, its management has evolved in response to numerous and sometimes antagonistic activities, such as hunting, timber harvesting, sand and sandstone quarrying, and, in recent centuries, the development of tourism, outdoor activities, and a significant artistic movement, the Barbizon school. The ecological and biogeochemical role of fire in such a socio-ecosystem is to be clarified. Our hypothesis is that the large amount of documentation available on this forest can help better understand past fire dynamics and their biogeochemical implications. What does available documentation reveal? A selection work in the existing iconographic archives led to the creation of a corpus representing fire in the Fontainebleau forest comprising 10 postcards, 9 engravings, 1 painting and 15 photographs dating from 1860 to 1911, as well as contemporary images of the ecological succession after a fire. Combining these images with the data from the 3FD database ⁽¹⁾, we extract different types of information. In particular, we give visual evidence of type of fire (understory or peat), most exposed vegetation, and evolution of the management practices of fire (organisation of the reaction to fire). We also show how the geographical information and the images themselves  can help  to set up an experimental design and conduct field work, which will then enable us to carry out and interpret biogeochemical analyses.  We also discuss the originality of this material.

Reference :

(1) Chevalier, M., Abiven, S., & Lebrun Thauront, J. (2024). Fontainebleau Forest Fires Database (3FD), version 1.0 [Data set]. In Fire Ecology. Zenodo. https://doi.org/10.5281/zenodo.13305154

How to cite: Rabotin, T., Abiven, S., Cointe, B., Tenu, C., Lebrun-Thauront, J., and Mertens, K.: Contributions of a century-old iconographic corpus to improve the understanding of past fire dynamics in the Fontainebleau forest, France, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-17907, https://doi.org/10.5194/egusphere-egu26-17907, 2026.

Climate change has intensified extreme rainfall events, while rapid urban expansion has reduced rainwater infiltration. Together, these processes have disrupted urban hydrological systems and increased the frequency and severity of urban flooding, posing growing threats to lives and property. Conventional flood mitigation strategies largely depend on extensive grey infrastructure, such as pipes and tunnels, designed to rapidly evacuate stormwater. However, many of these systems were developed in the last century and are increasingly economically and ecologically unsustainable under intensifying rainfall extremes. In response, Nature-based Solutions (NbS) have gained prominence as sustainable approaches that work with natural processes to enhance flood resilience. Although NbS are often framed as a modern response to climate change, similar principles have long existed in traditional ecological and planning practices. However, Traditional Ecological Knowledge (TEK) is frequently regarded as fragmented or highly context-specific, which limits its systematic integration into contemporary flood resilience frameworks. As a result, it remains unclear whether such historically grounded practices can be translated into generalisable, scientifically testable principles applicable to modern NbS design and flood risk assessment.

Here, we present a systematic interpretation of flood management strategies in ancient Chinese civilisation through the lens of Feng Shui. Feng Shui is an indigenous planning philosophy centred on the concept of harmony between humans and nature and has been widely applied in traditional village site selection and layout. This study focuses specifically on the local water management principles embedded within Feng Shui. We synthesise ancient texts and classical literature to reconstruct traditional water-planning concepts and relate them to contemporary hydrological and geomorphological theory. Using spatial statistical and mathematical fitting analyses across more than 300 historical villages, we demonstrate the consistency and non-site-specificity of these principles. Furthermore, hydrodynamic simulations of a representative village show that Feng Shui–inspired water systems can effectively reduce flood depths and peak flows under present-day extreme rainfall scenarios, through mechanisms such as distributed storage, controlled diversion, and flow-path reorganisation.

Together, these results indicate that traditional village planning embodied core principles analogous to those underpinning modern NbS. Our findings provide quantitative evidence for the scientific basis, adaptability, and flood mitigation effectiveness of traditional ecological knowledge. More broadly, this study demonstrates a methodological pathway for translating TEK into scientifically grounded frameworks by integrating historical analysis, spatial statistics, and numerical modelling, highlighting its potential relevance for contemporary flood resilience assessment and NbS design.

How to cite: Li, X., Liang, Q., and Chen, H.: Decoding Chinese Ancient Culture-related Nature-based Solutions for Flood-Resilience Using Modern Informatics, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-6015, https://doi.org/10.5194/egusphere-egu26-6015, 2026.

By the 18th century, China had established a relatively systematic and stable framework for relief institutions and bureaucratic operations. This study introduces the agent-based analytical framework FRAMα to reproduce the local governance network embedded in this bureaucratic structure. FRAMα is a reduced version of the empirically informed flood resilience agent-based modelling framework FRAMe, in which only the most essential mechanisms are retained.

Using a county affected by the 1870 Yangtze River flood as a case, the study describes local flood response conditions during the event. Scenario analysis shows that, although the bureaucratic system was relatively well developed, local governance outcomes varied substantially under different network configurations. A centralized governance structure relied heavily on the stability of key nodes, particularly on whether the local chief official (county magistrate) continued to fulfill their responsibilities. When this node remained functional, local governance exhibited a high level of operational resilience. Once the node ceased to function, system resilience declined rapidly and flood losses increased accordingly.

By transforming the recurrent historical issue of “officials’ dereliction of duty” into an analytical object of governance network structure, this study extends existing research on Qing dynasty relief and bureaucratic governance. It offers a new perspective for understanding the resilience of institutional operation in historical disaster governance and highlights the importance of shared responsibility and substitution mechanisms for contemporary flood resilience building.

How to cite: Feng, W., Chen, S., and Yang, E. L.: Modelling Local Governance Structure and Flood Resilience in the 1870 Yangtze River Flood, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-5457, https://doi.org/10.5194/egusphere-egu26-5457, 2026.

How to cite: Zhang, Y. and Yang, Y.: From Adaptation to Breakdown: Traditional Knowledge and the 1929 Famine in Shaanxi, China, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-680, https://doi.org/10.5194/egusphere-egu26-680, 2026.

While forensic methodologies for disaster analysis have been proposed and applied for more than a decade, a structured meta-analysis of multi-hazard events—revealing patterns and existing gaps across the disaster risk management cycle—remains significantly underexplored. This study presents a comparative analysis of five major multi-hazards events using the PARATUS approach, which integrates disaster analysis (Forensic analysis) with risk analysis (Impact Chains), specifically we compared: the 2017 Hurricane Irma (Sint Maarten), the 2018 Vaia Windstorm (Italian Alps), Gloria Storm 2020 (Catalonia), the 2021 La Soufrière volcanic eruption (Saint Vincent), and the 2023 Kahramanmaras earthquakes (Türkiye).

Beyond variations in compound and cascaded hazard combinations, these events encompass a range of geophysical and environmental conditions across areas with distinct socio-economic patterns. The PARATUS framework was selected for its structured, temporal alignment with disaster phases: pre-disaster conditions, hazard and impact analysis, recovery, and resilience building and the use of innovative conceptualisation tools such as impact chains.

By categorising multi-hazard events according to Tilloy et al. (2019), the meta-analysis provides evidence that these events amplify impacts and hinder response. Evidence for this amplification is found across the following categories: independent hazards (e.g., concurrent volcanic, pandemic, and disease events), triggering hazards (e.g., an earthquake cascading into landslides and liquefaction), and compound hazards (e.g., consecutive severe storms).

The meta-analysis underscore the relevance of investigating the social dimension of risk to formulate effective long-term risk-reduction and mitigation strategies. This is evident across the sections of Paratus' forensic framework: first, the pre-disaster conditions are shaped by institutional, social, economic, and environmental vulnerabilities, often driven by unplanned development, poverty, and weak governance. Subsequently, during events, response and early warning systems are frequently hindered by poor coordination and inadequate communication with marginalised groups. Furthermore, post-disaster recovery, while focused on restoring infrastructure and finance, often adopts top-down approaches that neglect community engagement and long-term equity.

Despite significant progress in hazard and risk understanding and the identification of necessary risk management measures, this advanced knowledge has not yet been fully translated into consistent application, updated regulations, or comprehensive resilience planning. Consequently, critical resilience gaps persist, including unaddressed infrastructure vulnerabilities, insufficient community preparedness, fragmented emergency coordination, and a lack of financial risk-transfer mechanisms.

Finally, the forensic analysis framework is well-suited to meta-analysis due to its comprehensive, methodical structure, which ensures consistent, multidimensional data synthesis across diverse disaster events.

How to cite: Olaya Calderon, L. J., Cocuccioni, S., Romagnoli, F., Ramadhani Prasetya, S., Kumbikano, M., Pantaleoni, N., Kundak, S., Göksu, Ç., Atun, F., and Pittore, M.: Decoding Multi-Hazard Disasters: A Forensic Meta-Analysis using the PARATUS Forensic Analysis Framework, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-8293, https://doi.org/10.5194/egusphere-egu26-8293, 2026.

Disruption of travel due to extreme weather and other forms of damage, such as network isolation, travel delays, and associated wider economic losses, can far exceed direct damages. The scale of these indirect impacts depends critically on how operators and travellers respond to disruptions, which combines centralised responses with operational actions and behavioural adaptation in shaping the system performance.

This study proposes an innovative framework for multimodal transport systems that integrates decentralised operator response and passenger disruption-aware routing to evaluate indirect disruption impacts. Historical flood events are used to define plausible stress scenarios that locally reduce network capacity and service quality. When disruption occur, operators respond independently by prioritising either disrupted service recovery (e.g., speed up early road clearance and recovery process) or reinforcement of non-disrupted modes and corridors (e.g., adding bus frequencies and train short turns). These decentralised actions modify travel conditions and perceived generalised costs, and passengers subsequently reselect modes and routes through a logit-based choice model, leading to the change of travel demand at origin-destination level.

The framework is applied to road and rail networks in Great Britain using observed demand and future demand scenarios in 2030 and 2050 derived from long-term housing plans. By comparing indirect disruption impacts under a road-only system with those under an integrated road-rail system, the analysis highlights the extent to which multimodal connectivity mitigates indirect damages and reduces network isolation. Additionally, by capturing the interaction between disruption, decentralised response, and passenger behavioural change, the framework produces risk-weighted post-disruption capacity gaps that identify where congestion and service shortfalls persist. The results explicitly identify corridors and modes where capacity investment is most effective under future demand growth and plausible disruption conditions, providing actionable insights for long-term capacity planning and transport resilience. Indirect impacts are not just a property of infrastructure damage, but of how systems adapt.

Keywords: indirect disruption impacts; decentralised response; multimodal transport; integrated capacity planning and resilience

How to cite: Li, Y., Pant, R., Russell, T., Thomas, F., Hall, J., and Parlantzas, N.: Assessing Travel Disruption and Decentralised Responses in Multimodal Transport Systems , EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-21536, https://doi.org/10.5194/egusphere-egu26-21536, 2026.

How to cite: Ho, T. P., Feng, W., Cheng, S., Ai, M., and Yang, L. E.: Flood risk and social resilience evolution in the Vietnamese Mekong Delta in the documented history, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-18044, https://doi.org/10.5194/egusphere-egu26-18044, 2026.