From the beginning of 2025 to now, Vietnam has continuously recorded particularly serious natural disasters. Heavy rain, flood overlapping, unusual storms have repeatedly broken historical records, causing heavy damage to people and placing a forecasting and warning system with great challenges.
Professor Tran Thuc, Chairman of the Vietnam Hydrometeorological Association with more than 40 years of experience in hydrometeorology and climate change research, has given recommendations to improve early warning capacity.
Extreme natural disasters increase, climate laws have been broken
Professor, how do you evaluate the unusual natural disaster situation in recent times?
- In recent years, Vietnam has continuously faced many serious natural disasters such as heavy rain, storms and widespread flooding. Some floods have been recorded as "historic disasters" due to water levels on many rivers exceeding records, causing widespread flooding and heavy damage to people, property and infrastructure.
Notably, in 2011, 2016, 2017, 2020, 2024 and 2025, there were particularly major flooding events. Specifically, historical floods have occurred in the Central region (in 2011, 2016, 2020, 2025) and the Northern region (in 2017, 2024, 2025).
In particular, 2025 will be marked by severe and widespread flooding, with water levels exceeding many historical milestones in both the North and the Central regions. This big difference compared to previous years (usually only flooding in single areas or multiple areas but on a smaller scale) has led to hundreds of casualties, lost homes, flooded hundreds of thousands of hectares of rice, causing huge economic losses.
Please tell us more, compared to previous floods, what is the difference in the frequency and characteristics of floods in 2025, is there any station that has exceeded the historical value in the past decades?
- The historic flood in November in the South Central region once again shows the vulnerability of our country to extreme weather phenomena, especially rain and floods. Since the beginning of the year, after the North, North Central and Central Central regions, the South Central regions have become the "flood center", showing the severity of natural disasters in this region.
Traditional flood rules in the South Central region are changing. According to statistics over the past 30 years, major floods in the South Central region often occurred before November 15. However, the major flood in 2025 will appear later than this rule.
Major floods, especially major ones and historic floods have appeared on many rivers from Quang Tri to Khanh Hoa and Lam Dong. At some stations on the Ky Lo River, downstream of the Ba River (in Dak Lak), Dinh Ninh Hoa River (Khanh Hoa), and Da Nhim River (Lam Dong), floods exceeding historical levels have been recorded in the past 30 - 40 years. The flood peak on the Ba River at Phu Lam station during this flood has a repeat cycle estimated at about 50 years (rarely occurring frequency).
This unusual major flood event is a clear demonstration that the traditional flood rules are changing, reflecting an increasingly extreme and unpredictable weather trend due to the impact of climate change leading to changes in circulation, storms and tropical depressions. More frequent and severe natural disasters have created an urgent need to improve forecasting capacity, risk management and natural disaster prevention for this area.
Difficulties in forecasting extreme rainfall
In your opinion, how has the forecasting work of the hydrometeorological sector been implemented in the recent floods? How do you evaluate the strengths and challenges in forecasting and warning work?
- Forecasting work is implemented early and closely following reality. Specifically, since April 2025, weather forecasts have warned of a peak period of heavy rain in the Central and South Central regions in October - November.
The Hydrometeorological Agency has issued bulletins warning of heavy rain, emergency floods and the risk of landslides early (2-3 days in advance). The bulletins are updated with high frequency (according to the time for emergency floods).
Accurately identifying the focus of natural disasters, forecasting that the most severely affected key areas have been accurately zoning off; for example, the recent floods were Khanh Hoa, east of Dak Lak and Lam Dong.
Applying new technology, there has been the participation of high-resolution value forecasting models and AI applications in recognizing rainfall patterns, helping to improve the reliability of short-term newsletters.
However, regarding the ability to forecast extreme rainfall, although the rainfall is forecast, determining the total rainfall at some locally located locations is still limited. This is also a general limitation for tropical and monsoon areas with strong terrain, making it difficult to forecast localized heavy rain within a narrow range.
Forecasts for flash floods and landslides are not detailed enough. Landslide warnings are still mainly at a "high risk in communes and wards" and cannot be localized to each specific landslide point (hillside, small road) in the ward and commune. This is a common limitation in terms of current science and technology in the world.
With delays and data gaps, the network of automatic monitoring stations in upstream areas and deep mountainous areas (where flash floods) in the South Central region is still quite sparse, leading to a lack of real-time measurement data to warn of rapid floods.
3 groups of solutions to improve disaster early warning capacity
To better cope with increasingly extreme weather conditions caused by climate change, what solutions should be implemented, according to the professor?
- In my opinion, to better cope with extreme weather conditions, it is necessary to implement 3 groups of solutions.
Group of technology and infrastructure solutions:
Increase the density of the hydrometeorological monitoring network. Install additional automatic rain gauge stations and specialized hydrological stations in upstream areas and mobile wave "pulling" areas in the South Central region; Strengthen the X-ring weather Radar system in urban areas and mountainous areas to scan clouds and detailed rain, serving extremely short warnings.
Applying AI and in-depth big data. Build machine learning AI models based on local historical flood data to warn of urban flooding (such as Nha Trang, Phan Rang) and landslides in real time; Complete maps of large-resolution natural disaster risk zoning (1:2000 or 1:500) for key residential areas.
Group of mechanism and process solutions:
Change to "Impact-based Forecast". Instead of only announcing "200mm rain", the bulletin needs to specifically describe the consequences: "200mm rain will cause flooding of A road 0.5m deep, risk of landslides at Pass B, water will overflow the C dike". This helps the government and people easily visualize the level of danger.
Mechanism for coordinating the operation of inter-reservoir reservoirs, tightening the operation of inter-reservoir reservoir reservoir processes in river basins in the South Central region (short rivers and slopes). Flood discharge needs to be optimally calculated with rain forecast data to avoid "flood overlapping" for downstream.
Social and communication solution group:
Communicate to the "end-users". Ensure that warning information is reached directly to people in remote areas through multiple platforms even when there is a power outage or local network congestion.
Improve community capacity, instruct people on how to read and understand forecasts and skills to recognize signs of natural disasters on the spot.
Sincerely thank you, Professor!
