Natural disasters in 2025 converge many extreme and dangerous factors

This judgment was shared by Dr. Truong Ba Kien, Deputy Director of the Climate Change Research Center (CCRC) under the Viet Nam Institute of Meteorology, Hydrology and Climate Change (IMHEN). In reality, despite early forecasts, many localities still suffered severe damage due to heavy rain and historic floods that far exceeded conventional response scenarios.

"Matching pieces" cause continuous rain

Sir, the weather patterns causing rain in the Central region over recent months are assessed as familiar, appearing almost every year. How did extreme factors manifest in the recent heavy rains, and what were the causes?

Dr. Truong Ba Kien: For forecasters, the weather patterns behind the recent rainfall in the Central region are not new: tropical storms and depressions in the East Sea, reinforced cold air, strong northeast and east winds, the equatorial trough, and disturbances within the easterly wind belt. In fact, the Central region faces these conditions every year.

Looking at the recent heavy rain in Khanh Hoa, Gia Lai, Dak Lak, and Kon Tum, the picture becomes quite clear. Strong reinforced cold air surged down to the South Central region and collided with the southern Truong Son range, creating wind bands blowing toward the Dak Lak–Khanh Hoa coastline.

At the same time, in the lower atmosphere (around 925–700 hPa), a very strong easterly wind band appeared. This wind brought large amounts of moisture from the sea directly inland. When it met the coastal mountains, the moist air mass was blocked and forced to rise, generating prolonged orographic rainfall that lasted for hours, even days.

Meanwhile, a strong and persistent subtropical high-pressure system maintained the continuous flow of moist easterly winds over an extended period. As a result, rain did not fall in a single episode but repeatedly came down on areas already flooded, worsening the inundation situation.

On top of these factors, the equatorial trough and atmospheric oscillations in the equatorial region were also strongly active, causing further disturbances in the atmosphere. This made the rainfall difficult to end.

This rain episode reflects special features of natural disasters in 2025, encapsulated in three words: compound, prolonged, and amplified.

Compound means multiple hazardous weather patterns occur simultaneously and overlap. For example, strong cold air came down at the time when the easterly wind band was active, along with the equatorial trough shifting northward. These "matching pieces" formed a "rain-producing machine" with nearly maximum efficiency.

Prolongation occurred because heavy rainfall did not happen in a single spell but continued from one episode to the next. The soil was already saturated, and rivers and streams could not drain floodwaters in time. Therefore, just a few more days of rain were enough for flooding to surpass historical levels.

Amplification stems from sea surface temperatures in the East Sea being warmer than the multi-year average, allowing the atmosphere to hold more moisture than before. As a result, under the same weather patterns, the rainfall this time was significantly higher than 10–20 years ago.

The recent historic floods on many rivers in the Central region were the “product” of this convergence: the weather patterns were familiar, but the climatic background and surface conditions had changed, resulting in a much more extreme flooding event.

Forecasts must specify when and where the impacts will occur

From the recent natural disasters in the Central region, in your view, to what extent should Vietnam continue investing in order to forecast abnormal heavy rain and related hazards such as landslides and major floods?

Dr. Truong Ba Kien: It can be affirmed that Vietnam’s hydrometeorological forecasting capacity has made very significant progress in recent years under the attention and decisive direction of the Party and the Government. The Secretariat issued Directive No. 10 on strengthening the Party’s leadership over hydrometeorological work, along with the Government’s decisions on modernizing the sector (including Decision No. 1970/QD-TTg dated January 23, 2021, approving the Hydrometeorology Sector Development Strategy to 2030, with a vision to 2045). These provide an important foundation for synchronous investments in observation networks, radar systems, satellites, lightning detection, and the national forecasting center.

A clear demonstration is that during the 2025 storm season, storm forecasts were issued earlier, improving from 3 days to about 5 days, thereby allowing governments and communities a wider “action window.” Storm orbit forecasts have also achieved high reliability, with significantly reduced errors, especially within 48–72 hours before landfall. Forecasts of storm intensity and storm-induced rainfall have become increasingly accurate, helping localities be more proactive in evacuations, reservoir operations, and protecting critical infrastructure.

Thanks to these advances, during most major rain spells in 2025, meteorological agencies were able to issue early warnings several days in advance, followed by short-term and very short-term reports. Many localities proactively suspended school activities, operated reservoirs, closed coastal areas, and evacuated residents based on these forecasts. This is clear evidence of the modernization of the current forecasting system.

However, to accurately forecast abnormal heavy rain down to the level of individual communes and wards and to provide specific warnings of flash floods and landslides, much work still needs to be completed. The hydrometeorology sector requires methodical investment aligned with the orientations set out in existing resolutions and strategies, focusing on four key priorities.

First, increase the density of observation stations and local sensors, especially automatic rain gauges, while investing in high-performance computing (HPC, GPU) infrastructure to rapidly process massive data volume.

Second, develop high-resolution numerical prediction models, operated in the form of super-ensembles on high-performance computing systems, in order to provide more detailed simulations of heavy rain, floods, and other extreme phenomena.

Third, accelerate the application of artificial intelligence and big data, enabling computers to automatically detect early signals of strong convection, hailstorms, and extreme rainfall—phenomena that often form rapidly and are difficult to forecast using traditional methods.

More importantly, it is essential to build an impact forecasting system. This system goes beyond providing figures on rainfall or water levels; it transforms these data into flood maps, landslide-prone locations, and disrupted transport routes. Based on that, local authorities can proactively evacuate residents, deploy response forces, close roads, or protect infrastructure in a timely manner, rather than merely receiving general information such as "heavy rain" or "major flood" as before.

In other words, while we already have a solid foundation thanks to the drastic direction of the Party and the Government, to keep pace with the growing risks of natural disasters under climate change, investment in forecasting and early warning should be regarded as essential infrastructure, on par with roads, electricity, water, healthcare, and education.

Response scenarios detailed to commune and village levels

With experiences in developing disaster-impact scenarios for Viet Nam, what recommendations does the Institute offer for creating disaster scenarios detailed to each locality along with concrete response actions?

Dr. Truong Ba Kien: In practice, we need a fundamental shift in thinking: from disaster prevention to proactive disaster risk management based on science, data, and technological advances. Risk management is not about waiting for storms or floods to respond; it means proactively identifying and assessing hazards and controlling risk right from the stage of planning  and socio-economic development.

From practical experience, the Institute recommends localities build disaster scenarios according to the following guidelines.

First, scenarios must be based on science and modern climate data, not solely on historical records. It is essential to use updated climate change scenarios, new extreme rainfall thresholds, and detailed flooding, flash flood, and landslide maps down to the commune and village levels, ensuring they accurately reflect the actual conditions of each area.

Next, it is necessary to shift from forecasts of how many mm of rain to forecasts of “what and where impacts will occur.” For example, instead of merely reporting very heavy rain, specify which roads are likely to be cut off, which residential areas face deep inundation, which sites are prone to landslides, and approximately how many households will need evacuation.

In addition, specific action thresholds must be established. When rainfall exceeds dangerous levels or water levels reach alarm thresholds, the system should automatically activate response measures, such as evacuating residents, closing roads, securing dams and reservoirs, and relocating people out of high-risk areas—rather than relying on step-by-step manual directives.

Another very important aspect is integrating risk scenarios into development planning. Urban, transportation, irrigation, and new residential planning projects must be based on climate risk maps, not solely on past rainfall and flood statistics as before.

Finally, it is essential to enhance local capacity in risk governance through staff training, developing response procedures tailored to each scenario, conducting periodic drills, and regularly updating scenarios after every disaster event.

In the context of climate change, which is making natural disasters increasingly extreme and unpredictable, the key is not only to reduce losses once a disaster occurs but to proactively manage risks early based on science and detailed impact analysis for each locality. This is the sustainable direction to improve resilience and protect communities.

Sincerely thank you, sir