Cloudy with a chance of monsoons
In India, where 42 percent of workers depend on agriculture for their livelihoods, supercomputers are giving farmers new hope for a better harvest. More than half of arable land in the country is dependent on rainfall, which means the monsoon season, lasting from June to September, has an outsized impact on the agriculture sector and other industries linked to it.
As it stands, extreme weather events cost India US$9–10 billion annually. With the growing threat of climate change, productivity decreases in major crops could rise to as much as 40 percent by 2100 unless farmers adapt.
Enter Pratyush and Mihir—‘rising sun’ and ‘sun’ in Hindi—a supercomputing duo with a combined 6.8 petaFLOPS of power allowing the government to monitor monsoons and predict extreme weather events.
The fourth most powerful weather research computer in the world, Pratyush can simulate world weather conditions at a high resolution of 12 kilometers, in 64 different layers of the atmosphere. For India, it can achieve a resolution as fine as three kilometers. As a result of this added computing power, the number of farmers who receive weather information through text messages is expected to increase from 24 million to 45 million.
In climate models, researchers use equations to represent the physical principles of our world. Weather forecasting is done by solving these equations. Pratyush has to crunch the numbers at every single imaginary point on the globe placed 12 square kilometers apart, then repeat this again for all the different layers in the earth’s atmosphere.
The calculations are then run 20 to 30 times while changing variables like temperature and humidity to ensure predictions are as accurate as possible. Clearly, monitoring monsoons is no mean feat.
