Haicheng Initiative
Continental earthquakes:
Roadmap towards earthquake forecasting
The Haicheng Initiative
Discussed and endorsed at the
International Symposium on Earthquake Forecasting, 8-10 July 2025
It has been over three quarters of a century since earthquake prediction was put on the agenda of national and global endeavor of earthquake science. Yet earthquake prediction remains a grand scientific challenge. However, in the long march to earthquake prediction, there were some remarkable achievements. The successful prediction of the 4 February 1975, Haicheng, Liaoning Province, China, Ms7.3 earthquake was one of them. This was the first prediction of a major earthquake which played an important role in the mitigation of earthquake disasters. The prediction itself had shown that, facing the threat of earthquake disaster, science and technology can at least do something constructive and contribute to the reduction of earthquake disaster.
Half century has passed, there have been significant advances in both the study on the predictability of earthquakes and its application for disaster risk reduction (DRR), and new technologies have developed which are potentially useful for the monitoring and modeling of earthquakes and the assessment of seismic hazards at different spatio-temporal scales. Earthquake forecasting itself has shown its interdisciplinary, inter-institutional, and international characteristics. Based on the experiences and lessons of the past half century, and aiming at the goals of
1. More physics in long-and-intermediate-term earthquake forecast and seismic hazard assessment;
2. More insights in short-and-imminent-term earthquake forecast;
3. More and better application of earthquake forecast and seismic hazard assessment for disaster risk reduction
abbreviated as M3, it is time to discuss the vision of earthquake forecast and priorities of science and technology, towards the next half century.
Continental earthquakes pose the most direct and serious threat to people's lives and property. In recent years, interdisciplinary research on continental earthquakes has been conducted, providing scientific basis for assessing the predictability of earthquakes.
Commemorating the 50th anniversary of the successful prediction of the Haicheng earthquake, and promoting to address the scientific challenges of earthquake forecast, the International Symposium on Earthquake Forecasting was held from 8 to 10 July 2025, with about 300 participants from 38 countries/regions, which reached the following consensus on the agenda of developments based on the exchange, discussion, and debate, advocating the joint actions in science and technology.
1. Promoting the study on the predictability of earthquakes and its application for disaster risk reduction (DRR)
1.1 Improving the understandings of seismotectonics
An increasing number of continental earthquakes indicate that major earthquakes primarily occur along tectonic boundary zones. Therefore, understanding the deformation of these tectonic zones and fault movements is an important foundation for recognizing the patterns of earthquake generation. With the application of modern technologies, the precision and resolution of data from topographic surveys, geophysical investigations, and dating methods have significantly improved, providing more reliable information and new perspectives for studying tectonic zone deformation and fault movements.
1.2 Advancing the understandings of deep processes and seismogenesis
Study of seismogenic structure of large earthquakes needs to promote the sharing of new interdisciplinary data and results to help elucidate the deep structure, mechanisms and processes of destructive earthquakes within plate interior, including, but not limited to, geology, geophysics, geodesy and geodynamics.
1.3 Developing earthquake dynamics modeling
With the progress of new observation techniques and theories, such as high-density network, deep drilling observation and optical fiber sensing, scientists have more data support and new perspectives on the understanding of the occurrence process of large earthquakes. The research revealed the diversity of earthquake nucleation and rupture modes, such as pre-slip versus cascade phenomena, and subshear versus supershear rupture. Meanwhile, with big data and lessons learnt from recent earthquakes, there is an ongoing progress of physics-based earthquake source modeling, providing reasonable estimates of the magnitude and likelihood of earthquake rupture.
2. Applying new technologies which are potentially useful for the monitoring and modeling of earthquakes and the assessment of seismic hazards at different spatio-temporal scales
2.1 Applying artificial intelligence to earthquake forecast
The development and application of artificial intelligence (AI) have brought about revolutionary changes to earthquake research, providing more scientific basis and more efficient technical means for earthquake forecasting. On the other hand, without extensive testing against observed seismicity, AI may provide broad avenues for deceptive interpretations/conclusions/ recommendations.
2.2 Applying space technology to earthquake forecast
Earthquakes have a strong impact on the lithosphere, atmosphere, and ionosphere during their preparation, occurrence, and aftermath stage. The development of space technology provides opportunities for studying the multi-layer response to earthquakes and improving monitoring system and tracking capabilities.
2.3 Encouraging comprehensive field investigation of earthquakes
Rapid initiation of comprehensive field investigations after large earthquakes plays an important role in understanding earthquake occurrence and earthquake predictability. To systematically explore related scientific issues, the following aspects are to be stressed in comprehensive investigations of significant earthquakes:
a)Multi-disciplinary synergies and international collaborations in earthquake investigations;
b)Application of new technologies;
c)Rigorous testing against registered observables, scientific guidelines, standards;
d)Data-sharing policies of earthquake monitoring agencies.
2.4 Revolutionizing test sites and in-situ experiments for earthquake forecast and prediction
The design, construction, operation, evaluation, and optimization of the test sites for earthquake forecast, especially recent developments of in-situ experiments, monitoring of the variation of Earth media, sustainability of the test sites as ‘big science’ infrastructures, networking of testing areas as a unified system, and virtual test sites based on remote sensing and numerical simulation are the important issues in earthquake forecast and prediction.
3. Improving the practice of earthquake forecast and prediction
3.1 Improving long-to-intermediate term earthquake forecast
Accurate earthquake risk assessment is the foundation for the effectiveness of operational earthquake defensive measures, which demands quantitative and refined forecast of large earthquakes. With the development of earthquake observation and exploration, theory and technology, the physical model of earthquake source has progressively improved, the mechanism of strong earthquake propagation became clearer, and numerical simulation technology of the dynamic process of strong earthquake propagation has gradually matured. The long- and intermediate-term probabilistic earthquake forecast and numerical forecast have good foundations in theoretical, computational and observational resources. The forecast must pass through extensive testing against observed seismicity.
3.2 Keeping exploring short-to-imminent term earthquake forecast
The nearly 50 years of earthquake forecast practice in China as well as other countries/regions has shown that precursors were observed before certain earthquakes. However, identifying earthquake precursors is still difficult; most of so-called well-known “significant earthquake precursors” still require durable testing against seismic reality for reliable operational forecasting. In the detection of earthquake precursors, all of the non-tectonic factors affecting the observation need to be excluded. Meanwhile, the observed anomalies need to be explained by the related seismogenic processes. Breakthrough of short-to-imminent-term earthquake forecast relies on the study of the precursors.
3.3 Facilitating the transition of earthquake forecast from science to application
Based on the new understandings of earthquakes and new technologies of the monitoring and modeling for earthquake forecast, and summarizing the experiences and lessons over the past half century, it has been possible to significantly enhance the technical readiness level (TRL) of the existing methods of operational earthquake forecast (OEF) and seismic hazard assessment (SHA) - the empirical, the statistical, the physics-based, and the numerical, making earthquake forecast and SHA applicable to the society for disaster risk reduction (DRR).
Effective forecasting requires a holistic, step-by-step approach—progressively narrowing magnitude, location, and time within physical and data limits. OEF must be reliable, testable, supported by evidence, but not necessarily perfect. Evaluating OEF requires tracking both successes and failures. Rigorous, prospective testing is needed, properly compared with the results of random guesses to assess error rates and optimized parameters through user-specific cost-benefit analyses. A shared theoretical framework is essential for evaluating forecasting tools consistently.
3.4 Making earthquake forecast useful through disaster scenario simulation
It is essential to know the expected ground shaking before the earthquake occurrence. Such information can be attained by modelling a wide set of possible scenarios for earthquakes that satisfy the alert conditions. An upper bound for the expected ground shaking can also be determined by considering the maximum magnitude for the impending earthquakes plus an additional safety term.
In addition to ground motion, earthquakes may lead to cascading disasters and pose serious threats to social and economic development, people's life and property. Therefore, it is very important to explore the formation mechanism of earthquake related disasters, accurately assess their risks, and formulate the mitigation and emergency plan to ensure the safety of people's life and property.
3.5 Investigating the social aspects of earthquake forecast
Earthquake forecast is a practice-oriented problem crossing the border of natural science, engineering, and social sciences. This character was much reflected during the occurrence of Haicheng earthquake by the disciplines of seismo-sociology and preparedness for earthquakes. Systematic planning and communication of earthquake forecasts at the regional, national, and global scales play important roles for future developments. In this regard different countries/regions have accumulated rich experiences and lessons which need to be addressed comprehensively.
Earthquake sequences are not random, but rather lack comprehensive geneses: the successful prediction of the 1975 Haicheng earthquake is an early proof.
We call for the international wisdom and joint action towards earthquake forecasting, for the sustainability of the society.
