From a global perspective, the ubiquity, frequency, and severity of natural disasters are becoming increasingly prominent. According to reports from the United Nations Office for Disaster Risk Reduction (UNDRR), direct economic losses from natural disasters over the past two decades have amounted to trillions of dollars, with millions of lives lost and billions of people affected. In particular, since the beginning of the 21st century, extreme events such as the 2004 Indian Ocean tsunami, the 2008 Wenchuan earthquake, the 2011 Great East Japan earthquake and subsequent Fukushima nuclear accident, and the 2022 Pakistan floods have served as stark warnings that the traditional research paradigm—characterized by a “siloed” approach focused on individual hazards—is becoming increasingly inadequate for addressing current and future systemic risks. These “black swan” or “gray rhino” events clearly reveal the systemic, complex, and chain-forming nature of natural disasters, whose objects and processes embody a high degree of integration and complexity.
Against this backdrop, the frontier of international natural disaster research is undergoing a profound paradigm shift: moving from mechanistic analyses of single hazards toward systematic simulations and predictions of multi-hazard coupling and disaster chain evolution; and shifting from static risk assessment toward dynamic systemic resilience evaluation that incorporates tipping points and cascading effects. The principles of Earth system science have been incorporated into disaster research, emphasizing the integrated treatment of the lithosphere, hydrosphere, atmosphere, and biosphere as an interacting whole, thereby enabling a deeper understanding of the gestation, initiation, and evolution of catastrophic disasters. The central goal of this transformation is to advance “disaster system science,” achieve early warning and intelligent risk prevention for compound mega-disasters, and this has become a consensus and competitive frontier among the global scientific community.
To achieve this goal, where lies the innovative source of foundational research on natural disaster “mechanism–prediction–prevention and control”? What major scientific questions demand urgent attention? And how can the emerging artificial intelligence research paradigm be integrated to carry out systematic, quantitative, and intelligent studies of extreme complex geological disaster chains?
To address these questions, Chinese scientists (Peng et al., 2026) systematically delineated nine typical disaster chain types driven by endogenic dynamics within the lithosphere, exogenic dynamics within the cryosphere, exogenic dynamics across the atmosphere–hydrosphere interface, and anthropogenic activities within the biosphere. They synthesized the common evolutionary patterns of these disaster chains. Furthermore, from the three dimensions of time, space, and human activities, they distilled three core characteristics: “coupling across spatiotemporal scales,” “exchange across sphere boundaries,” and “significant driving and amplification by human engineering activities.” On this basis, they identified five major fundamental scientific questions that demand urgent attention in future theoretical research: “disaster gestation through multi-sphere interactions, disaster regulation by multi-interface control, disaster propulsion by multi-dynamic coupling, disaster initiation by extreme events, and disaster amplification through multi-disaster chain interactions.”
The authors point out that research on these scientific questions must rely on deep interdisciplinary integration and convergence among Earth system science, physics, mechanics, mathematics, information science, and other disciplines. It requires innovation in research thinking—from geological phenomena to dynamic mechanisms, and from dynamic mechanisms to statistical physics. In particular, they propose the concept of “multi-critical phase transitions” and elaborate on its specific implications in “multi-critical phase transitions during disaster chain dynamics—from single-hazard triggering to chain-scale cascading, and from a single disaster chain to multiple disaster chains.” They innovatively assert that “multi-critical phase transitions” serve as the key to understanding the dynamic processes of disaster chains.
See the article:
Peng J, Cheng Q, Zhang Y, Huang Q, Zhang F. 2026. Mechanism of natural disaster formation: A systematic analysis based on multi-sphere interactions, multiinterface control, multi-dynamic coupling, multi-critical transitions, and multi-disaster chain amplification. Science China Earth Sciences, 69(7): 2529–2560, https://doi.org/10.1007/s11430-025-1963-1
Science China Earth Sciences
Systematic review