Vietnam's National Centre for Hydro-Meteorological Forecasting Pairs AI with Google and Weathernews to Sharpen Typhoon Warnings

After Typhoon Yagi's destructive 2024 landfall exposed gaps in track-prediction accuracy, Vietnam's National Centre for Hydro-Meteorological Forecasting (NCHMF), under the Vietnam Meteorological and Hydrological Administration, began integrating artificial intelligence into its operational forecasting for the 2025 typhoon and flood season.
The centre applies machine-learning models trained on radar data, satellite imagery and automated observation stations to generate short-term rainfall nowcasts, and uses AI-powered image analysis to identify cyclone centres and track storm intensity. On 31 July 2025 the Vietnam Meteorological and Hydrological Administration signed a memorandum of understanding with Japan's Weathernews Inc. to supply AI-based typhoon-track predictions and, separately, adopted Google's flood-forecasting and cyclone-prediction models; the Hanoi University of Science and Technology's AI institute was tasked with helping integrate the tools into NCHMF's tropical-cyclone forecasting ahead of the season.
Director Mai Van Khiem reported that 24-hour cyclone-centre position error now runs 90-110 km, in line with regional norms, and that storm and lightning alerts are issued 30 minutes to 3 hours ahead of events; he also said 24-hour advance warnings are associated with roughly 30% lower economic losses, though the centre acknowledged it still lags Japan, China and South Korea technologically, and cites sparse observation networks, budget constraints and a shortage of AI specialists as ongoing limits. No independently audited comparison of AI-assisted versus prior forecasts has been published alongside these figures.

Read the full analysis: https://en.vietnamplus.vn/vietnam-harnesses-ai-to-improve-weather-forecasting-post325894.vnp

Read the full analysis: https://en.vietnamplus.vn/vietnam-harnesses-ai-to-improve-weather-forecasting-post325894.vnp