Home India Ministry of Earth Sciences Parliament Question: AI in Weather Forecasting...
Date: 2025-08-20 Category: Not Applicable State: Union Government Country: India

Parliament Question: AI in Weather Forecasting

Issued by Ministry of Earth Sciences · Not Applicable

Research with AI Agent Chat with Document Generate Summary Translate Helpful Share Add to Project Create Task

Executive Summary & Key Takeaways

Executive Summary: The Ministry of Earth Sciences (MoES) utilizes AI technology for weather and climate forecasting through the India Meteorological Department (IMD) and other institutions. New technologies are being used to provide weather and rain forecasts to farmers and fishermen, including Gram Panchayat Level Weather Forecasting (GPLWF) and the Bharat Forecasting System (BharatFS). MoES is actively upgrading AI-related technologies for weather forecasting through enhanced computing systems, research collaborations, and capacity building initiatives. Key Points / Main Content: AI in Weather Forecasting: * IMD and other MoES institutions use AI-based tools for experimental weather and climate forecasting, including Advanced Dvorak Technique (AiDT) for cyclone intensity estimation and AI/ML-based hybrid dynamical models for weather prediction. * Research is being conducted in AI/ML for short-range global forecasting, precipitation data downscaling, fire location forecasting, fog forecasting, and lightning/thunderstorm forecasts. * MausamGPT, an AI-based chatbot, is being developed as a climate service advisor for farmers and stakeholders. Forecasting for Farmers and Fishermen: * IMD provides early warning services using the latest technologies. * Gram Panchayat Level Weather Forecasting (GPLWF) launched in collaboration with the Ministry of Panchayati Raj (MoPR), covers nearly all Gram Panchayats across India using a state-of-the-art multi-model ensemble forecast. * GPLWF forecasts are accessible on digital platforms like e-Gramswaraj, Meri Panchayat app, e-Manchitra, and Mausamgram. * Weather forecast information reaches people through Pashu Sakhis, Krishi Sakhis, and Self Help Groups (SHGs). * GPLWF provides localized weather information hourly (up to 36 hours), 3-hourly (36 hours to 5 days), and 6-hourly (5 to 10 days). * Bharat Forecasting System (BharatFS), launched on May 27, 2025, provides high-resolution forecasts down to the panchayat/cluster of panchayats level with a 6km spatial resolution. * Climate Forecast System version 2 (CFSv2) is used for extended range weather forecasts up to 4 weeks. * Agromet Field Units (AMFUs) prepare Agromet Advisories twice a week in English and regional languages. * Impact-based forecasts (IBFs) and advisories are prepared based on severe weather warnings. * Location-specific forecasts and advisories are available through mobile apps like Meghdoot and Mausam, and social media platforms. * IMD has integrated its services with the IT platforms of 18 State Governments. Upgradation of AI Technologies: * MoES has augmented the High Power Computing System (HPCS) with a total computing capacity of 22 PetaFLOPS, including GPU A100. * MoES has dedicated GPUs NVIDIA H100 for AI/ML research in weather forecasting. * IMD is working on AI/ML-based data-driven models along with NWP models. * A Virtual Centre at IITM, Pune, has been established to develop AI/ML-based application tools. * A dedicated functional group has been established in IMD to strengthen R&D activities in AI/ML. * IMD has established specialized GPU and CPU-based infrastructure for AI computing. * IMD has signed MoUs with various Academic Institutions and organizations for collaborations and R&D activities. * Capacity building in the AI/ML domain is being done through training sessions and workshops. * IMD organizes a short-term refresher course on AI/ML every year in May. Impact Analysis: Farmers: * Impact: Access to localized weather forecasts and advisories to make informed decisions on farming operations, leading to improved yields and reduced losses. * Action Required: Utilize available platforms (e-Gramswaraj, Meri Panchayat app, Mausamgram, Meghdoot, Mausam, etc.) to access weather information and advisories. Fishermen: * Impact: Access to timely weather warnings and forecasts to ensure safety at sea and optimize fishing activities. * Action Required: Monitor weather forecasts and warnings provided by IMD and other sources before and during fishing expeditions. Disaster Managers: * Impact: Access to high-resolution weather forecasts for effective disaster preparedness and response. * Action Required: Integrate BharatFS forecasts and other relevant weather information into disaster management plans. General Public: * Impact: Access to improved weather forecasts and warnings for daily planning and safety. * Action Required: Stay informed about weather conditions through available channels (mobile apps, social media, news outlets) and heed warnings issued by authorities. Researchers/Academic Institutions: * Impact: Opportunity for collaboration with IMD and MoES on AI/ML-based weather forecasting research. * Action Required: Participate in collaborative projects with IMD, leveraging expertise in AI/ML to improve weather forecasting models and tools. State Governments: * Impact: Integration of IMD services into state IT platforms to provide localized weather information to farmers and other stakeholders in regional languages. * Action Required: Ensure seamless integration of IMD data into state platforms and promote the use of these platforms among farmers and other relevant users.

Key Entities Referenced

India Meteorological Department (IMD): The primary agency responsible for weather forecasting and providing weather-related information in India. Ministry of Earth Sciences (MoES): The Indian government ministry responsible for matters relating to Earth sciences, including weather forecasting. Gram Panchayat Level Weather Forecasting (GPLWF): A program launched by IMD in collaboration with the Ministry of Panchayati Raj to provide weather forecasts at the Gram Panchayat level across India. Bharat Forecasting System (BharatFS): An indigenously built numerical weather prediction model for generating high-resolution forecasts, launched on 27 May 2025. Climate Forecast System version 2 (CFSv2): A coupled model used for generating extended range weather forecasts up to 4 weeks at the meteorological subdivision level. Agromet Field Units (AMFUs): Units located at various SAUs, IITs, institutes of ICAR, etc., that prepare Agromet Advisories for their respective districts. MausamGPT Mausam Generative Pre-trained Transformer: An AI-based chatbot trained explicitly as a climate service advisor to farmers and stakeholders. New Delhi, Delhi: Location of the National Weather Forecasting Centre (NWFC).
Official Source Record View Original Source →
See Full Document Text
GOVERNMENT OF INDIA MINISTRY OF EARTH SCIENCES LOK SABHA UNSTARRED QUESTION NO. 4465 TO BE ANSWERED ON WEDNESDAY, 20THAUGUST, 2025 AI IN WEATHER FORECASTING 4465. MS. S JOTHIMANI: Will the Minister of EARTH SCIENCES be pleased to state: (a) whether the Government utilises AI technology to determine changes in season or weather conditions and if so, the details thereof; (b) the details of efforts taken to forecast the weather and rain to farmers and fishermen using new technology; (c) whether there is any upgradation taking place in technologies relevant to AI for weather forecasting; and (d) if so, the details thereof and if not, the reasons therefor? ANSWER THE MINISTER OF STATE (INDEPENDENT CHARGE) FOR MINISTRY OF SCIENCE AND TECHNOLOGY AND EARTH SCIENCES (DR. JITENDRA SINGH) (a) Yes.The India Meteorological Department (IMD) and other MoES institutions have been using AI-based tools for experimental weather and climate forecasting. These include the Advanced Dvorak Technique (AiDT) used for estimate cyclone intensity, AI/ML-based foundation, hybrid (AI+Dynamical) models for weather prediction,etc. The following are the research works done in AI/ML related to weather forecasts.  Short-range global forecasting.  Downscaling of precipitation data.  Fire Location Forecasting.  Fog Forecasting.  Lightening/Thunderstorm Forecasts.  Deep learning for improved global precipitation in a numerical weather prediction system. In addition to the above research work, the MausamGPT (Mausam Generative Pre- trained Transformer), which is an AI-based chatbot trained explicitly as a climate service advisor to farmers and stakeholders also being developed.(b) IMD has been using the latest technologies to provide early warning services to its stakeholders, like farmers, fishermen, etc. Last year, IMD, in collaboration with the Ministry of Panchayati Raj (MoPR), had launched Gram Panchayat Level Weather Forecasting (GPLWF) covering nearly all Gram Panchayats across India using a state- of-the-art multi-model ensemble forecast based on a number of numerical weather prediction models. These forecasts are accessible on digital platforms such as e- Gramswaraj (https://egramswaraj.gov.in/), the Meri panchayat app, e-Manchitra of MoPR, and Mausamgram of IMD (https://mausamgram.imd.gov.in/). The main aims and objectives of GPLWF are to provide weather forecasts up to Gram Panchayat Levels, covering critical parameters such as temperature, rainfall, humidity, wind, and cloud conditions-essential data that farmers need for informed decision-making regarding sowing, harvesting, and irrigation. The platform makes weather forecast information accessible anytime and anywhere at the panchayat level across the country. This weather information reaches a larger number of people through Pashu Sakhis and Krishi Sakhis under the Ministry of Agriculture and Farmers Welfare and the Ministry of Rural Development, as well as other Self Help Groups (SHGs). The GPLWF helps farmers to have access to localized weather information available hourly for up to a 36-hour lead period, 3-hourly from 36 hours to the next five days, and every 6 hours from the next 5 days to 10 days. On 27 May 2025, the Government launched the indigenously built Bharat Forecasting System (BharatFS), a state-of-the-art numerical weather prediction model for generating high-resolution forecasts. It promises finer and accurate rain forecasts down to the panchayat/cluster of panchayats level. The BharatFS has a spatial resolution of 6km compared to the previous 12 km resolution of the global forecasting system (GFS). It also has a capability to provide predictions of rainfall upto 10 days, covering the short and medium range. Thus, it would help to provide a forecast at the panchayat/cluster of panchayats level for the public, farmers, disaster managers, and other stakeholders. The Climate Forecast System version 2 (CFSv2) coupled model is used for generating extended range weather forecasts (upto 4 weeks) at the meteorological sub-division level. Based on observed and forecasted weather, Agromet Field Units (AMFUs) covering 127 agroclimatic zones located at various SAUs, IITs, institutes of ICAR, etc., prepare Agromet Advisories twice a week (every Tuesday and Friday) in English as well as in Regional languages for their respective districts to help the farming community make appropriate decisions on day-to-day farm operations. Along with the AAS bulletins, daily weather forecast and nowcast information are also issued by Regional Meteorological Centers (RMCs) and Meteorological Centers (MCs) of IMD. Impact-based forecasts (IBFs) and appropriate advisories for agriculture are also being prepared by AMFUs based on the severe weather warnings for different districts of various States and UTs across the country issued by the National Weather Forecasting Centre (NWFC), New Delhi, and RMCs and MCs of IMD. Technological advancements have further enhanced accessibility, enabling farmers to receive location-specific forecasts and advisories through mobile apps such as ‘Meghdoot’, ‘Mausam’ and Social media platforms like WhatsApp, Facebook, etc. Additionally, IMD has integrated its services with the IT platforms of 18 State Governments, allowing farmers to access information in both English and regional languages.(c)-(d) Yes. MoES recently augmented the High Power Computing System (HPCS) with a total computing capacity of ~22 PetaFLOPS, with about 10% of the total capacity of the new HPC systems having Graphics Processing Unit(GPU) (A100). Apart from this, MoES has a separate GPU’s (NVIDIA H100) dedicated for AI/ML research in weather forecasting. IMD is working on AI/ML-based data driven model along with NWP models for further improving the weather forecasting skills over Indian region and the sea area of North Indian Ocean including the Bay of Bengal and the Arabian Sea in various spatio-temporal scales to make weather warnings more accurate, timely and actionable, helping communities to better prepare for and respond to weather hazards. The details are given in Annexure-1.Annexure-1  Virtual Centre at IITM, Pune, has been established by MoES to develop AI/ML/DL based application tools.  A dedicated functional group has been established in IMD under MoES to strengthen the R&D activities in AI/ML.  IMD has established a specialized GPU and CPU-based infrastructure for AI computing.  IMD has signed MoUs with various Academic Institutions like IITs, IIITs, NITs, ISRO, DRDO, Ministry of Electronics and Information Technology (MeitY), etc, for collaborations and R&D activities, utilizing facets of various AI/ML applications to weather and climate.  The capacity-building in AI/ML domain with respect to weather and climate are being done by nominating scientists in training sessions and workshops.  IMD organizes a short-term refresher course on the Fundamentals of Artificial Intelligence and Machine Learning every year in May. The usage of AI-based monitoring tools and forecasting models is as follows:  To Estimate Tropical Cyclone Intensity, satellite-based AI-enhanced Advanced Dvorak Technique (AiDT), as given by Cooperative Institute for Meteorological Satellite Studies, is utilised by IMD apart from other products  IMD also uses AI-based model guidance from the European Centre for Medium-Range Weather Forecasting (ECMWF) for tropical cyclone genesis, track, and intensity prediction. ********

Continue your research