DUBAI // UAE-based scientists are helping to predict rainfall during the monsoon season in India as part of a three-year collaboration between the Centre for Prototype Climate Modelling (CPCM) of New York University Abu Dhabi and the Indian Institute of Science in Bangalore.
Called the Monsoon Mission, the project will use historical data in India to inform predictions about future rainfall.
“India has more than a hundred years of data,” said Dr Ajaya Ravindran, senior scientist at NYUAD and one of the project leaders.
The team also includes Dimitris Ginnaakis, assistant professor at the Courant Institute at New York University, and Andrew Majda, professor at Courant and lead principal investigator of the CPCM. The Abu Dhabi and New York teams share a US$550,000 grant from the Indian Institute of Tropical Meteorology.
The scientists will use complex mathematics to identify patterns within the rainfall data with the effort eventually allowing them to build models that help make predictions for the future. They will use a cutting-edge statistical technique called non-linear Laplacian spectral analysis which will be used for the first time to analyse rainfall data.
One challenge for the scientists will be the intermittency of rainfall which makes it harder to find patterns within the data.
Characterised by large amounts of rainfall, the monsoon period lasts from June until September. It not only affects India but the rest of Asia, stretching from Afghanistan to the Philippines.
In India monsoon rainfall has important significance for the power sector and agriculture, said Dr Ravindran. Irregularities in the seasonal rains can disrupt life for hundreds of millions of people.
“The economy in India crucially depends on agriculture,” he said, explaining that a better ability to predict rainfall could assist farmers in decisions such as when to fertilise their fields.
Information about less seasonal rain, on the other hand, can alert decision-makers about possible crop shortages and price hikes.
The project aims to improve the accuracy of predictions in the short term - within the range of ten days or less - as well as within ten days and four months.