Young PhD Breakthrough in Flood Forecasting

Turning Point

Dr. Tran Ngoc Vinh's success stems from his courageous decision to leave his comfort zone and the stressful experiences he experienced during many years of studying and working in Korea and the United States. His journey, from the early days of studying Physics and Mathematics in Vietnam, through Korea, to the United States, is a story of the will to overcome limits and the desire to find solutions to protect the community from natural disasters.

Graduating from the University of Natural Sciences (VNU), Vinh started his career as a researcher at his old school. However, the decision to go to Korea to do research was a real turning point. “The cultural and language differences made me feel self-conscious, and I often doubted whether I had the ability to continue studying. But that was the most important turning point in my career. Being away from my family, I had to stand on my own two feet: set my own goals, make my own plans, and strive to achieve them,” Dr. Vinh shared.

Tran Ngoc Vinh presented on urban flooding phenomenon at the University of Michigan, USA

Now looking back on his nearly 5-year PhD journey at Ulsan University, South Korea, he realizes that the pressure he put on himself was even greater than the requirements of his supervisors. “There was a period when my life was almost unbalanced: During my PhD, I often worked more than 15 hours a day, many nights I worked until 6-7 am to go to sleep. However, it was that harsh period that helped me realize my true passion and the path I wanted to pursue, laying the foundation for breakthrough projects later on,” he recalls.

Dr. Vinh's original major is Earth Science, specifically hydrometeorology. During his undergraduate studies at the University of Natural Sciences, he was fortunate to participate in many research projects of his teachers and go on field trips everywhere, especially in the Central region of Vietnam, focusing on the phenomenon of flooding. Field trips to the Central provinces, witnessing firsthand the fierce floods that swept away many lives, properties, and material possessions of the people, left him with many emotions. "During my work, I always struggled with two questions: Is it possible to predict natural disasters? And what can we do to minimize the damage caused by floods?", he said.

Dr. Tran Ngoc Vinh, researcher at the University of Michigan, USA

Although he loves Physics and is passionate about traditional flood forecasting models, he also sees their limitations: “Human power is limited and cannot forecast on a large scale.” On the contrary, AI can process huge amounts of information quickly and accurately, with discipline. He believes that combining AI with physical models and human experience is the way to overcome the weaknesses of each tool.

From that thought, the research team led by him at the University of Michigan (USA) conducted the research project "Artificial intelligence improves the accuracy, reliability and economic value of continental-scale medium-term flood forecasts", to be completed in 2023. The research developed a new hybrid modeling framework combining AI and the current US flood forecasting model - National Water Model (NWM), which can reduce errors in national flood forecasting programming, providing more accurate predictions of where floods will occur.

Dr. Vinh and the scientists found that when the AI was used in conjunction with the NWM model developed by the US National Oceanic and Atmospheric Administration (NOAA), the resulting hybrid model was four to six times more accurate. The AI was trained on NOAA data for the US, but the system could be customized for any country.

The research results show that the hybrid model not only outperforms NWM, but also outperforms an advanced AI model developed by Google, especially in extreme flood situations. The work was published in the top journal of the American Geophysical Union (AGU), which publishes fewer than 100 studies per year.

Testing across the US

Dr. Tran Ngoc Vinh, Top 20 Golden Globe Science and Technology Awards 2025, owns 8 national patents in Korea, 29 scientific articles published in international scientific journals in the Q1 category...

Dr. Vinh shared: “To announce the results, we had to overcome many storms and a huge amount of work. But that is proof of the journey of constantly searching for better solutions to forecast natural disasters, a goal I have cherished since the first days of entering the profession.”

As the main author of this research project, for many years he has been working tirelessly, from collecting and processing input data, including meteorological data (rain, temperature, wind, etc.), flood flow data, simulation data from NWM to designing the overall research framework, building simulation scenarios to evaluate the effectiveness of the model, proposing to preside over the writing of the manuscript, participating in the review process of the research project...

“In terms of research, I will integrate artificial intelligence into the Earth model system to improve simulation capabilities and calculation speed. The goal is to forecast large-scale natural disasters such as storms in a longer period of time, for example, the next 10 days. When we have accurate and earlier forecast results, we will be better prepared, minimizing human and property damage” Dr. Tran Ngoc Vinh (University of Michigan, USA)

In particular, he was directly responsible for designing, programming, and training the AI model to test across the US with more than 42,000 flood events, with a forecast period of 1-10 days. "It can provide probabilistic forecast scenarios - a key factor for decision-making in high-risk contexts and can run on a regular computer, without the need for a supercomputer," Dr. Vinh stated the model's superior features.

In parallel with flood research, Dr. Tran Ngoc Vinh also published a work on urban flooding in the journal Nature Cities, pointing out the "flood loop - upgrading sewer systems - flooding", warning that the current approach to designing drainage systems is not optimal. He hopes to apply this research into practice in Vietnam, not only focusing on flood forecasting, but also providing prevention solutions, as well as serving the design of disaster prevention works, optimizing flood drainage systems in large urban areas such as Hanoi and Ho Chi Minh City.

Source: https://tienphong.vn/tien-si-tre-dot-pha-trong-du-bao-lu-post1780398.tpo