AI Pilot To Improve Local Monsoon Onset Forecasting

The government has implemented an AI-enabled pilot to improve local monsoon onset forecasting for farmers ahead of the Kharif 2025 season. The initiative, conducted with the Development Innovation Lab–India, demonstrates how advanced modelling can support agricultural planning at scale.

Blended AI Model for Local Forecasts

The pilot used an open-source blended model combining Google’s NeuralGCM, ECMWF’s Artificial Intelligence Forecasting System and 125 years of IMD rainfall data. The system generated probabilistic forecasts for local monsoon onset—information vital for determining optimal sowing dates.

Dissemination Through M-Kisan Portal

Localised forecasts were delivered via SMS through the M-Kisan platform to 3.88 crore farmers across 13 states. Messages were issued in Hindi, Odia, Marathi, Bangla and Punjabi to ensure accessibility. The initiative focused exclusively on forecast delivery, with no financial assistance attached to the pilot.

Farmer Response and Behavioural Impact

Post-forecast feedback surveys were conducted through Kisan Call Centres in Madhya Pradesh and Bihar. Findings indicated that 31–52% of surveyed farmers adjusted their agricultural decisions, particularly by modifying land preparation practices, sowing timelines, crop choice and input use.

Exam Oriented Facts

• AI model blended NeuralGCM, ECMWF-AIFS and 125 years of IMD rainfall data.
• Forecasts sent to over 3.88 crore farmers in 13 states via M-Kisan.
• Five languages used: Hindi, Odia, Marathi, Bangla and Punjabi.
• 31–52% farmers altered planting decisions based on the forecasts.

Government’s Push for Tech-Driven Agriculture

Minister of State for Agriculture Ramnath Thakur informed the Lok Sabha that the pilot shows strong potential for AI-driven advisory systems to support climate-smart agriculture and improve decision-making for millions of farmers.