How IMD intends to leverage AI for improved weather predictions
As the India Meteorological Department (IMD) is gearing up to celebrate its 150th anniversary on Monday, it is turning to artificial intelligence (AI) and faster supercomputers to enhance weather forecasting models. Speaking to PTI, IMD Director General Mrutyunjay Mohapatra shared test beds are being set up in Odisha and Madhya Pradesh to better comprehend the physical processes behind thunderstorms and heavy monsoon rains. Additionally, the IMD will team up with academic institutions to jointly develop tools that boost forecast accuracy.
Panchayat Mausam Seva, National Framework for Climate Services
The IMD is set to commence its 150th-anniversary festivities on Monday with the launch of "Panchayat Mausam Seva," a service providing weather forecasts to farmers in every village. The National Framework for Climate Services will also be introduced to incorporate climate information into every sector and activity. Moreover, Mohapatra highlighted the IMD's progress, from issuing cyclone warnings at Kolkata port in the late 19th century to providing nowcasts directly to users' mobile phones today.
Collaboration with academic institutions, expert group on AI-ML
The IMD has established an expert group on AI and machine learning (ML) and is partnering with NITs, IITs, IIITs, research and development, and academic institutions for joint tool development. Mohapatra stated AI-ML can significantly enhance decision-making and weather forecasting in nowcast and short to medium-range forecast models. He highlighted they have weather data since 1901 in digitized form. "AI is data science where we can utilize all this historical data to develop tools and methods for forecasting," Mohapatra added.
Upgrading high-performance computing systems for better numerical modeling
To improve numerical modeling capabilities, the IMD is upgrading its high-performance computing systems. Mohapatra revealed, "The Ministry of Earth Sciences (MoES) is in the process of procuring high-performance computing systems. Our capacity will increase from 10 petaflops to 30 petaflops." "With that, the numerical model runtime, resolution, and post-process data generation will improve," he elaborated. Faster supercomputers will enable the IMD to generate more sector-specific and location-specific information, aiming to enhance the weather modeling system's resolution from 12km to 6km.
Predicting severe weather events challenging sometimes
Mohapatra noted that early warnings and forecasts provided by the IMD have significantly improved over the years, resulting in minimal loss of human life from cyclones and heavy rains. He mentions, "Forecast accuracy has increased by 40-50% in the last five years Despite these improvements, the IMD chief acknowledged there are occasional misses, like the cloudburst phenomenon, lightning, and small-level severe weather events. He emphasized the ongoing need to enhance the observational system to detect every weather accurately.
