Central Region faces continued high flood risk as Typhoon Koto approaches

Extreme disasters, 'overlapped storms and floods'

Speaking at the Forum on Scientific and Technological Applications in Disaster Forecasting and Early Warning on the afternoon of November 25, Nguyen Ton Quan, Deputy Head of the Community Disaster Management and Communication Division under the VDDMA, reported grim statistics. Between 2006 and 2025, natural disasters claimed an average of 314 lives annually-including missing persons, and caused economic losses equivalent to 1-1.5% of GDP.

In 2025 alone, the East Sea recorded 19 storms and tropical depressions, a figure second only to the record-breaking year of 2017. Many of these storms moved rapidly with unpredictable path changes. Three consecutive, high-intensity storms made landfall in the Central region, necessitating the Prime Minister to establish a Forward Steering Committee for on-site response.

Extreme rainfall caused historic flooding on more than 13 river routes, accompanied by severe inundation in urban areas ranging from the North to the Central Highlands. A pattern of "successive storms and compounding floods" emerged in many areas. Nineteen hydrological stations across 11 provinces recorded exceptionally high flood levels, with Bac Ninh, Thai Nguyen, Lang Son, and Cao Bang seeing multiple locations surpass all historical benchmarks.

According to Quan, the risk of heavy rain and flooding in the Central region remains high for the remainder of 2025, as Storm No. 15 is projected to enter the East Sea and make landfall shortly.

He emphasized the urgent need to implement Directive 42-CT/TW and the Law on Natural Disaster Prevention and Control. Disaster prevention content must be integrated into sectoral and regional planning, and infrastructure standards must be reviewed and adjusted to enhance resilience. Furthermore, the rain gauge network requires expansion, and the quality of heavy rain forecasting must be improved to support early warnings.

Alongside finalizing risk zoning maps down to the village level, authorities must decisively relocate residents from areas prone to landslides, flash floods, and deep inundation, while transitioning crops and livestock to nature-based adaptations. Dike systems, reservoirs, and disaster prevention infrastructure require priority maintenance and upgrades, with reservoir operations adjusted to ensure sufficient flood control capacity for downstream areas.

Quan also called for strengthening community communication, expanding international cooperation, applying digital technology for real-time disaster monitoring, and mobilizing resources for sustainable disaster recovery.

Early landslide warnings to protect residential areas

Prof. Dr. Do Minh Duc of the VNU University of Science announced that his research team has proposed an early landslide warning model linked to the development of safe residential areas, utilizing data, technology, and community participation.

Prof. Duc argued that in a developing nation, engineering solutions alone cannot block all disasters. Even countries with robust infrastructure like Japan and the United States accept this limitation. Consequently, the team’s priority approach is prevention and avoidance, grounded in science and socio-economic feasibility. A key principle he emphasized is that scientific products must serve the community rather than profit motives. Partnering with the Community Disaster Prevention Fund allows research results to be quickly applied in practice and communicated to residents, particularly regarding the testing of new warning technologies.

Surveys in Mu Cang Chai and various mountainous provinces revealed that Viet Nam is highly susceptible to landslides. Nearly every slope with a cut exceeding six meters in height and a gradient over 45 degrees is prone to sliding. Many reinforced structures have still been dislodged by rock and soil, and massive landslides blocking streams to cause debris flows are clear evidence of these large-scale, sudden risks.

According to Duc, there are two main risk categories. The first involves small to medium landslides caused by rain, which are now better forecasted thanks to a dense rain monitoring network and new techniques. The second category, large-scale rock and soil slides, depends less on immediate rainfall and more on geological structure and accumulated deformation. Although accounting for only 15% of incidents, this group causes over 85% of total damage. The area stretching from Da Nang and Quang Nam to Lam Dong, where project density and traffic infrastructure are high, faces particularly acute risks.

On the ground, warning signs are not difficult to spot. Long cracks, subsiding terraces, tilting houses, and streams that bend abnormally multiple times within a short distance are all traces of old slide masses. However, interpreting these signs requires geological expertise often lacking among local officials and residents.

20-point criteria, "Viet Nam Landslide" App, and Zalo warnings

To bridge this gap, the research team developed a simple checklist of nearly 20 criteria to help authorities and communities self-assess dangerous locations, accompanied by risk zoning maps and lists of specific households in danger zones. Based on this data, the team proposed a "safe residential area" model, referencing Japan’s BOKOMI community model and similar initiatives currently deployed in 1,800 rural areas in China. The model emphasizes five pillars: infrastructure, policy mechanisms, human resources, information technology, and funding, with the ultimate goal of zero human casualties.

The deployment process involves five steps, starting with detailed risk zoning down to the village level, followed by daily forecasting using the Random Forest model and multi-year rainfall data. The subsequent steps include risk identification via field checklists, real-time hourly warnings, and finally, the development of specific response scenarios for each household and evacuation route.

The team utilizes multiple communication channels. When rainfall reaches dangerous thresholds, the system automatically emails Disaster Prevention and Control Steering Committees at all levels and sends Zalo messages to commune leaders, village heads, and residents. The frequency can increase to every 20 minutes as risks escalate. Input data is sourced from over 250 rain gauges managed by the Community Disaster Prevention Fund and technology enterprises, combined with local station networks.

Additionally, the "Viet Nam Landslide" mobile app, available on iOS and Android, provides real-time rain and risk maps for 60 provinces, with details down to the hamlet and house cluster level in many locations. Residents can proactively check their risk status to adjust production and daily activities.

Some areas in Yen Bai, Mu Cang Chai, and Quang Nam have installed siren warning systems that function even during power and signal outages. In recent heavy rain events, the combination of risk maps, Zalo or app notifications, and sirens enabled many households to evacuate early, significantly reducing damage compared to 2017.

Drawing from practical experience, Prof. Duc outlined three groups of lessons: the "three earlies" of detection, warning, and action; the "three rights" ensuring the right people do the right job at the right time; and the "three levels" involving acquiring information, understanding it correctly, and translating it into action.

"Despite ongoing difficulties, if we value planning, prevention, and living in harmony with nature, we can absolutely build residential areas that are safer from landslides," he emphasized.