28/06/2023

IMPA develops algorithm to predict rainfall in Rio.

Paulo Orenstein (IMPA) e João Carabetta (Escritório de Dados da Prefeitura do Rio de Janeiro)

IMPA is developing an artificial intelligence code to improve rainfall forecasting in the city of Rio de Janeiro, in a partnership with the municipal government and with assistance from Google. The project was announced this Tuesday (27) at the “Google for Brazil” event in São Paulo, which was attended by IMPA researcher Paulo Orenstein.

The city government provides access to the city's rainfall data, and members of the IMPA Center for Projects and Innovation (Centro Pi) create the algorithm by cross-referencing information from various sources such as rain gauges, radar, and satellites. The goal is to know, a few hours in advance, where it is expected to rain and with what intensity, in order to assist the city government's planning and prevent environmental disasters.

“This is a very challenging problem from a predictive point of view because it involves complex data, at multiple scales, and from various distinct sources. It is an excellent opportunity to develop new research directions and have a concrete impact on the city,” Orenstein highlighted.

João Carabetta, head of data at the Rio de Janeiro City Hall's Data Office, states that the partnership demonstrates the government's commitment to using technology to address climate challenges. "By combining the expertise of Google and IMPA with our skills and the city hall's ability to prevent and respond to climate events, we will be able to create a more effective and predictive approach to dealing with extreme weather phenomena."

This project is yet another partnership between IMPA and the Rio de Janeiro City Hall. In 2024, the institute's first undergraduate course, IMPA Tech, will be inaugurated, operating in Porto Maravalley, a technology hub developed by the city hall to also house companies and startups.