As climate change increasingly disrupts agricultural patterns, farmers face growing challenges in managing their crops and resources. Reliable and timely weather information is crucial in making informed decisions that can optimize productivity and reduce losses. In response to these challenges, PxD, in partnership with the Ministry of Agriculture and Farmers’ Welfare (MoA&FW), Government of India, and the Development Innovation Lab India (DIL-India) at the University of Chicago Trust, has launched a transformative initiative aimed at delivering high-quality weather forecasts to inform farmer decision-making. This initiative is designed to equip farmers with the tools they need to adapt to changes in weather patterns and strengthen their resilience.

Climate Challenges and the Role of Weather Forecasts

Climate change is disrupting weather patterns, making rainfall less predictable, and putting agricultural livelihoods at risk. In the absence of reliable forecasts, farmers face challenges in planning their growing seasons, leaving them more vulnerable to losses.

Studies highlight how farmers adjust their practices in response to receiving forecasts about the timing of the monsoon. In Telangana, farmers for whom the forecast was “good news” (i.e., the growing season would be longer than expected) leveraged the extended growing season, prompting them to increase cultivated land, introduce new crops, and boost pre-harvest investments. Conversely, farmers for whom the forecast was “bad news” (i.e., the growing season would be shorter than expected) adjusted their strategies to minimize potential losses by reducing the amount of land they cultivated. Another study found that accurate forecast information enables farmers to optimize their inputs, increase yields, and reduce costs.

With the low cost of digital dissemination, delivering forecasts at scale is effective and affordable, offering significant potential to strengthen agricultural resilience.

Pilot Study Results: Monsoon 2024

During the Monsoon 2024 season, a pilot program disseminated weather forecasts across five Indian states — Uttar Pradesh, Madhya Pradesh, Rajasthan, Telangana, and Chhattisgarh. The Indian Meteorological Department (IMD) provided total seasonal rainfall forecasts, which were sent via SMS to 8.6 million farmers in May and June. Additionally, monsoon onset forecasts, developed by a researcher at the Potsdam Institute for Climate Impact Research (PIK), were shared with 0.85 million farmers in Telangana. These long-range forecasts, generated 40 days in advance, predict the likely start date of the monsoon season within a 9-day range, offering farmers a valuable early window to plan their agricultural activities effectively.

Dissemination was conducted using the Ministry’s existing m-Kisan system, which facilitated the distribution of two rounds of SMS to the 9.45 million farmers mentioned above.

Phone surveys with a sub-sample of farmers showed that 89% of farmers expressed interest in continuing to receive forecasts, and 90% found the information useful for making planting decisions.

Scaling Impact: India’s First AI-Based Weather Forecasting Program (2025)

Building on the success of our 2024 pilot, PxD led the message design and dissemination strategy for a historic expansion: India’s first-of-its-kind government-led AI weather forecasting program reaching 38 million farmers across 13 states during Monsoon 2025.

This groundbreaking initiative represents the largest targeted deployment of AI weather forecasts to farmers anywhere in the world. Working alongside Development Innovation Lab – India and Nobel Laureate Professor Michael Kremer, we transformed cutting-edge AI models-Google’s Neural GCM and ECMWF’s Artificial Intelligence Forecasting Systems (AIFS)-into actionable, farmer-friendly guidance delivered through the Ministry’s m-Kisan platform.

The AI Revolution in Weather Forecasting

Since 2022, an AI-driven revolution has transformed weather science, delivering more accurate predictions across many settings. These models can be tailored to meet user needs and enhance our ability to predict complex phenomena, such as the onset of the Indian monsoon, weeks in advance. The Ministry of Agriculture took a major step forward in applying this revolution to benefit crores of farmers.

The forecasts used in this program were a blend of two open-access AI models that, in rigorous evaluations, clearly outperformed other available forecasts at predicting monsoon onset locally for farmers.

Real-World Impact

The program’s effectiveness was demonstrated when this year’s monsoon arrived early but then stalled for 20 days mid-season. The AI-based forecasts correctly identified this pause in monsoon progression, and the government sent updated information to farmers every week until continuous rains arrived in their area.

As Additional Secretary Dr. Pramod Kumar Meherda noted: “This program harnesses the revolution in AI-based weather forecasting to predict the arrival of continuous rains, empowering farmers to plan agricultural activities with greater confidence and manage risks.”

Farmer-Centered Design

PxD’s role was critical in ensuring the complex AI-generated forecasts became actionable guidance that farmers could understand and trust. We worked directly with farmers to design messages that were clear, timely, and relevant to their specific agricultural decisions, demonstrating how advanced technology can be made accessible to those who need it most.

This achievement demonstrates PxD’s unique ability to translate cutting-edge technology into farmer-centered solutions at unprecedented scale.