HANOI: Bringing early warnings for landslides, rockslides and debris flows directly to communities at risk is the right approach to reducing casualties and property losses in Vietnam's mountainous regions, scientists have said.
Mountain landslides are often triggered by extreme rainfall and can generally be divided into three main types: soil landslides, rockslides and debris flows.
Drawing on field surveys, monitoring data and scientific modelling, experts from the Institute of Earth Sciences under the Vietnam Academy of Science and Technology (VAST) have classified these hazards and proposed early warning mechanisms for nearby communities.
Landslides have become increasingly complex in recent years, affecting larger areas and causing more serious damage.
Besides their direct impact, landslides can trigger cascading hazards such as debris flows, flash floods and river blockages, significantly increasing disaster risks.
Scientists from the Institute of Earth Sciences have conducted numerous surveys and research projects, including "Research on Landslide Types under Extreme Rainfall Conditions in the Northern Mountains and Central Vietnam" and "Developing a Community-Based Village-Level Early Warning System Integrating Landslides, Debris Flows and Flash Floods in Key Areas of the Northern Mountains".
The first project, led by Dr Pham Van Tien and a research team from the institute from January 2024 to June 2026, focused on identifying different types of landslides and clarifying the mechanisms and forces that trigger them during extreme rainfall.
The team conducted more than 20 field surveys in landslide-affected areas. Many took place immediately after disasters, when sites remained unstable and soil and rocks could continue to move, exposing researchers to dangerous conditions.
Among the most notable were a survey following the Lang Nu rockslide disaster in Lao Cai and, most recently, fieldwork in Muong Than Commune, Lai Chau Province.
According to the researchers, direct observation and documentation of physical evidence at disaster sites provide invaluable data for understanding landslide mechanisms, as many important signs quickly disappear after an event.
Their findings indicate that landslides have not only become more frequent in recent years but have also expanded in geographical reach and caused greater damage.
The research has provided an important scientific basis for identifying the characteristics of extreme rainfall, clarifying landslide mechanisms and supporting hazard zoning, residential planning and the introduction of early warning systems in mountainous areas.
The close integration of field surveys, data analysis and simulation models was identified as an effective approach, allowing researchers to move from observing individual events to assessing risks across wider areas. This can improve the practical application of research in disaster management and warning decisions.
The proposed community-level warning model combines scientific data with local observations. Residents are taught to recognise unusual signs such as cracks in the ground, changes in water colour and movements of soil and rocks, enabling them to evacuate before a landslide occurs.
For the first time in Vietnam, the research team has also proposed a landslide classification system based on international standards, using field data, remote-sensing analysis and digital terrain models.
The system provides an important scientific foundation for hazard assessment, forecasting and early warnings, helping reduce disaster risks.
The second project, led by Dr Dao Minh Duc with the participation of scientists and specialists from 2021 to 2025, focused on establishing a scientific and practical basis for community-based warnings of landslides, debris flows and flash floods at village level in northern mountainous provinces.
The project drew on years of field surveys and findings from previous landslide studies.
Residents in three pilot areas became important links in the warning system, working alongside monitoring stations and a data-processing centre.
The pilot locations included Ban Ria Commune in the former Ha Giang Province (now Yen Thanh Commune in Tuyen Quang Province), Cat Thinh Commune in the former Yen Bai Province, and Chieng Dong Commune (now part of Yen Chau Commune in Son La Province).
Using information collected from the monitoring system together with satellite data, the project tested warning messages sent to members of commune-level disaster response teams.
Assessments showed that most of the warning messages corresponded closely with the disasters that subsequently occurred in the pilot communes.
Members of village disaster response teams gradually learned the warning procedures and how to develop their own response plans.
When scientists sent alerts based on data from the automated monitoring system, community response teams knew how to establish warning checkpoints, assign members to monitor developments and prepare evacuation plans.
The early warnings also helped residents take greater precautions in their daily activities and production.
Dr Tien, who participated in the project, said the village-level disaster response teams had operated effectively during the research.
"Based on these research results, we will seek approval to expand the model," he said.
Associate Professor and Doctor of Science Vu Cao Minh, former deputy director of the Institute of Geology, said the projects were moving in the right direction by recognising the central role of residents in disaster prevention.
Minh, who has spent nearly half a century working in Vietnam's geological sector, said disaster prevention in mountainous regions should now be fundamentally community-based.
Identifying different types of landslides under extreme rainfall conditions would allow communities and authorities to recognise threats and issue warnings earlier, he said.
This approach was also consistent with the principle that science should not remain confined to laboratories or academic papers, but must be applied to solve practical problems, support socio-economic development and deliver tangible benefits to communities.
According to Minh, the flash flood in Muong Than, Lai Chau Province, on July 17 offered a costly lesson in assessing and forecasting landslide and flash-flood risks for mountain communities.
The residential area where fatalities occurred was located on flat ground, about 1.5km from the surrounding high hills and far from areas previously identified as being at high risk of landslides. Nevertheless, the disaster still struck.
Landslides and flash floods in northern and central Vietnam have tended to become more frequent and severe. Serious events occurred in the southern Central Highlands in 2023, the northern mountainous provinces in 2024 and the central provinces in 2025, followed most recently by the disaster in Muong Than.
Despite considerable efforts, shortcomings remain, Minh said.
Because landslides and flash floods are highly localised and can develop rapidly, on-site early warnings offer the most effective way to reduce loss of life.
Alongside greater use of forecasting technology, he called for a shift from passive disaster response to risk management and for the "four-on-the-spot" approach to be fully promoted at the grassroots level. - VNS/ANN
