Эксперты призывают Малайзию внедрить AI-прогнозы смога и переходить на часовые показатели качества воздуха

They said the current 24-hour average was fairly accurate, but showed an average over time and might not reflect real-time situations. Helena Varkkey, an associate professor at Universiti Malaya’s department of strategic and international studies, said Singapore reports both hourly and 24-hour average indices. “It would be great if Malaysia provided this data too, so that people could make more accurate real-time decisions for their health, like to know if it is safe right now to go for a walk outside,” she told FMT. Asked if the current haze was worse than previous years, she said it was hard to say but reminded everyone that the predicted Super El Nino season meant that the impact from any fire would likely be more severe, due to hotter and drier conditions. On the transboundary haze in Asean nations, Helena said countries in the region should move away from a diplomacy-based approach, citing shortcomings in such styles. “There are many frameworks and road maps at the Asean level for haze mitigation, but they lack the proper execution. “Anything that needs to happen cannot happen unless the host country explicitly invites assistance. This has slowed down and blocked effective collective efforts,” she said. Helena also said that Asean should make science-based decisions so that responses and assistance take place when it is scientifically necessary, not when it is diplomatically convenient Need for AI-based prediction M Jayaprakash, a senior principal assistant secretary at GovTech Malaysia, said the current practice of reporting air quality based on a 24-hour average should be reviewed. “Haze conditions can change rapidly with wind direction and weather patterns, making the average less representative of the prevailing air quality at any given time,” he said. The Cambridge-trained atmospheric scientist said haze management should move from a reactive “hotspot-counting” approach towards predicting impacts. “Telling people there are 30 or 40 hotspots in Indonesia does not mean much to them. What matters is knowing when and where the smoke and particles are likely to reach them, so they can be better prepared to face the situation. “The real question is whether the air where I live will become unhealthy tomorrow and when that is likely to happen. This is where environmental intelligence and what I call ‘environmental proofing’ can help us make decisions before the pollution reaches us, rather than after,” he said. In this context, Jayaprakash said Malaysia should make greater use of the AI-based predictive capabilities already locally available, adding that he had been advocating this approach since last year. He said Malaysia already had the capability to combine satellite observations, hotspot data, meteorological conditions, wind patterns and atmospheric modelling to predict where and when smoke is likely to travel. “The priority now is to move from capability to operational use, so that the public and businesses can be given timely information to prepare,” said Jayaprakash.