KUALA LUMPUR: The increasingly accurate weather forecasting capabilities enabled by the use of artificial intelligence (AI), satellite data and weather radar by the Malaysian Meteorological Department (MetMalaysia) now play a crucial role in translating early warnings into rapid action on the ground.
MetMalaysia director-general Dr Mohd Hisham Mohd Anip said these capabilities have strengthened operational planning among rescue agencies and enhanced community preparedness, particularly in areas prone to flooding and landslides during the North-East monsoon, which began last November and is expected to continue until March.
He said that during this period, between five and seven episodes of continuous heavy rainfall are forecast.
According to him, the integration of radar and satellite data allows for near real-time detection of weather systems such as storms and heavy rainfall, which are then analysed using AI to generate short-term forecasts, or nowcasting, of up to three hours ahead.
“This forecasting method achieves an accuracy rate of more than 90%, surpassing conventional methods.
“Further improvements in accuracy for longer forecast periods will be implemented under the weather system enhancement project, which is currently in the procurement phase,” he told Bernama recently.
He added that MetMalaysia has also introduced strategic innovations through the development of a new smart weather forecasting system, including more accurate warning lead times and an extension of the operational forecast period from seven to 14 days to provide alerts on severe weather events.
In addition, he said, MetMalaysia operates a high-resolution forecasting model of up to 333m in major urban areas such as the Klang Valley, Johor Baru and George Town, enabling more localised rainfall predictions.
Mohd Hisham said a multi-model ensemble approach is also employed to identify forecast uncertainties and enhance accuracy and consistency, while AI is used to generate extreme weather probability forecasts to support risk assessment and more effective decision-making.
“This innovation is supported by a comprehensive early warning system, which includes a nationwide observation network comprising weather radars, satellites, upper-air stations and surface stations, as well as data processing using high performance computing through numerical weather forecast modelling,” he said.
