Designing probabilistic AI monsoon forecasts to inform agricultural decision-making

2026-03-09

AI Summary

This study addresses the challenge of high-risk agricultural decision-making under weather uncertainty, where conventional forecasts fail to meet farmers’ diverse needs. The authors propose a decision-oriented, customized probabilistic monsoon forecasting framework that integrates an AI-driven systematic baseline weather model with a Bayesian-inference-based statistical model of “evolving farmer expectations.” Innovatively employing multi-model ensemble techniques, the approach dynamically predicts the probability of onset events. The method substantially improves forecast skill for the Indian monsoon at extended lead times and was deployed in 2025 through a government initiative to deliver early warnings of an anomalous pre-summer drought to 38 million farmers.

Technology Category

• Reasoning under Uncertainty: Relational Probabilistic Models
• Machine Learning: Mixture of Experts (MoE)
• Humans and AI: Planning and Decision Support for Human-Machine Teams

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Hundreds of millions of farmers make high-stakes decisions under uncertainty about future weather. Forecasts can inform these decisions, but available choices and their risks and benefits vary between farmers. We introduce a decision-theory framework for designing useful forecasts in settings where the forecaster cannot prescribe optimal actions because farmers'circumstances are heterogeneous. We apply this framework to the case of seasonal onset of monsoon rains, a key date for planting decisions and agricultural investments in many tropical countries. We develop a system for tailoring forecasts to the requirements of this framework by blending systematically benchmarked artificial intelligence (AI) weather prediction models with a new"evolving farmer expectations"statistical model. This statistical model applies Bayesian inference to historical observations to predict time-varying probabilities of first-occurrence events throughout a season. The blended system yields more skillful Indian monsoon forecasts at longer lead times than its components or any multi-model average. In 2025, this system was deployed operationally in a government-led program that delivered subseasonal monsoon onset forecasts to 38 million Indian farmers, skillfully predicting that year's early-summer anomalous dry period. This decision-theory framework and blending system offer a pathway for developing climate adaptation tools for large vulnerable populations around the world.

Problem

• probabilistic forecasting
• monsoon onset
• agricultural decision-making
• decision theory
• climate adaptation

Innovation

• decision-theory framework
• blended AI forecasting
• Bayesian inference
• time-varying probability
• subseasonal monsoon prediction