Google Research’s NeuralGCM, a physics-informed AI model trained on decades of weather data, helped the University of Chicago deliver SMS alerts predicting the onset of India’s monsoon to 38 million farmers. Combined with ECMWF’s AIFS and historical records, the system forecast timing up to a month ahead and captured an unusual dry spell, informing decisions on when to plant or switch crops. Open-sourced and efficient enough to run on a laptop, the model aims to broaden access to high-quality forecasting. University of Chicago research indicates month-ahead accuracy can markedly lift smallholder incomes, highlighting AI’s growing role in climate resilience and agriculture.
Related articles:
Pangu-Weather: A fast and accurate AI weather forecasting system
Climate-smart agriculture: practices and impacts (World Bank)
Understanding the advance of the Southwest Monsoon (India Meteorological Department)
