AI-Based Weather Forecast Tool Helps Farmers Adapt To Climate Change, Unusual Monsoon

Hyderabad: As climate change brings unpredictable weather patterns, farmers in India are increasingly relying on technology to navigate the challenges of a rapidly changing environment. The Indian Ministry of Agriculture and Farmers' Welfare is using artificial intelligence to provide accurate monsoon forecasts. This initiative allows 38 million farmers to receive AI-based predictions about the rainy season up to 30 days in advance, helping them to make better decisions on planting.

This program is a groundbreaking step in using AI to help farmers deal with climate change. Unlike just a weather prediction service, it offers a customised solution for smallholder farmers who depend on monsoon rains for their crops. Researchers believe that the use of AI could greatly benefit millions of farmers globally, especially in tropical areas.

The AI-based forecast for monsoon is the result of the Indian Ministry of Agriculture working with a global research team. Led by the University of Chicago's Human-Centred Weather Forecasts Initiative, the project included researchers from IIT Bombay, IISc Bangalore, and UC Berkeley. After testing different models, they found that Google's Neural GCM model and ECMWF's AI Forecasting System were the most reliable for predicting the monsoon. To boost accuracy, they combined these AI models with historical rainfall data from the India Meteorological Department. The project received partial funding from AIM for Scale, a global initiative supported by the Gates Foundation and the UAE, focused on agricultural innovations for farmers in developing countries.

Accurate AI forecasts empower farmers to adapt

Economist Michael Kremer says that disseminating AI weather forecasts provides significant benefits, generating over $100 for each dollar invested by the government. India is at the forefront of using AI to enhance various sectors, especially agriculture. This year was crucial for launching the program, unlike a typical monsoon season—where rains arrive over southern India in June and steadily advance north—the 2025 monsoon started early and stalled for nearly three weeks. AI forecasts accurately predicted this delay, helping farmers plan their activities effectively.

Parasnath Tiwari, a farmer from Madhya Pradesh, received the forecast on his phone and used the information to prepare early and switch to more lucrative crops, assured by the forecast’s prediction of a long monsoon season. "Before this, I mostly relied on my own experience and local knowledge to know when the monsoon would arrive," he said. "The forecast about the arrival of the monsoon was accurate…I have increased trust in the forecast, and I will rely on the information shared by scientists in the future."

Crafting a smarter, farmer-centred forecast

A collaboration led by the University of Chicago’s Human-Centred Weather Forecasts Initiative, along with researchers from IIT Bombay, IISc Bangalore, and UC Berkeley, has improved monsoon predictions using seven different weather models. They assessed these models over nearly 60 years of data to better meet farmers' needs for accurate forecasts.

In a typical monsoon season, rains arrive over southern India in June and steadily advance north (Getty Images)

With AI advancements since 2022, the initiative found that AI models like Google’s NeuralGCM and ECMWF’s AIFS significantly enhanced prediction accuracy. By combining these models with extensive rainfall data, they created a new forecasting approach that can predict monsoon rains 30 days in advance, providing farmers with essential information to make timely decisions.

Reaching the last mile

The Ministry of Agriculture and Farmers' Welfare created a 30-day weather forecast to assist farmers, reaching 38 million directly through SMS and nearly 1 million more in Odisha via voice messages. Precision Development (PxD), a nonprofit aiding smallholder farmers, helped design and test these messages.

Real-time forecasts were generated from May to July, and these were transformed into messages sent to farmers before the monsoon season. Researchers plan to conduct surveys to see how many farmers read the messages and change their behaviour based on the information. Initial findings suggest that farmers, like Parasnath Tiwari from Madhya Pradesh, shared this information with their peers, leading to benefits for others in their community.

Professor Ramesh Chand from Niti Aayog emphasised the importance of tailoring weather information to meet farmers' needs, stating that this initiative is highly valuable for helping them make informed decisions about farming based on clear and relevant forecasts.

The future of AI and agriculture

ECMWF Director-General Florence Rabier emphasises the goal of making weather forecast data more accessible. The AIFS model is helping to provide accurate and timely weather forecasts, benefiting millions of farmers and users. With 35 nations collaborating, ECMWF aims to push forward AI in weather science for a broader impact.

Researcher Mayank Gupta highlights that AI models are not only more accurate but also eliminate the need for supercomputers, as they can run on desktops. They can be customised for specific weather needs at a lower cost and time, making forecasting more efficient and widely accessible.

Researchers are collaborating with AIM for Scale to replicate similar programs in low- and middle-income nations, inspired by the Ministry of Agriculture and Farmers’ Welfare as a model for innovation. They are also training government meteorologists in these countries to effectively use AI models.