AI-powered weather insights improve Brazil’s maritime planning

By | 2 minute read | September 15, 2020

storm cloud

With a length of 8,000 kilometers, Brazil’s coastline challenges scientists wishing to model patterns of waves, tidal surges, sediments and weather. Besides the difficulty of sampling such a vast expanse, limited marine and weather data is available from Brazilian sources. The result is that coastal enterprises lack reliable forecasts to inform their response to severe sea and weather conditions.

Personnel at oil and gas producers, shipping companies, ports, yacht clubs, and municipalities often react to rough conditions after the fact rather than plan ahead. This can disrupt business operations, destroy property and, most importantly, put lives at risk.

Say an oil tanker is scheduled to arrive in port on a specific day and time. If bad weather and rough seas delay the arrival, port managers can adjust if they expect the delay. But if the ship is late without warning, employees, terminals and warehouses face disruptions. Similarly, accurate forecasts help businesses and individuals prevent losses from coastal floods and storms by giving them adequate time to prepare.

SIAA responds to a pressing need for better forecasts

I founded SIAA in response to this need. With a background in oceanography and a prior position as a port consultant, I saw firsthand the importance of better planning tools.

For the first few years, SIAA developed analytical models of Brazil’s ocean, river, reservoir and weather conditions to generate improved forecasts. We were hindered, however, by the lack of accurate global weather data. And although our team built a high-performance computational cluster to power the models, it was unreliable and costly to manage.

Consultants from BuildIT, an IBM Business Partner, solved these problems by leading us to IBM Watson, IBM Cloud, and The Weather Company.

We populated our forecasting models with data from The Weather Company Data Packages, whose extensive data sets cover the world for insights into current conditions, approaching weather and historical patterns. Then, BuildIT helped us power our analytics from the IBM public cloud. Its virtual platform is far more reliable and cost effective than the hardware cluster it replaced.

BuildIT also helped enhance our platform with IBM Watson technology, which contributes predictive analytics powered by AI, machine learning and neural networks.

Aimar delivers insights that inform advance planning

These moves primed the platform for commercial use. Called Aimar, its web interface provides high-resolution forecasts of marine and weather conditions for businesses, municipalities, the military and individuals in Brazil, and beyond. The insights help anyone needing to plan ahead in response to severe coastal conditions.

Combining weather data with oceanography creates interesting possibilities. Analysis of wind, rain, tides and currents can predict emerging flood conditions. We are also developing a service to improve shipping routes. By analyzing oceanic patterns and weather, the platform will help shipping firms map out the safest, fastest and most fuel-efficient routes from port to port.

Reliable forecasts and the social good

The early response to the Aimar platform has been quite positive. As it catches on, we at SIAA believe these reliable forecasts will lead to better planning—better planning that will help save lives, conserve resources, and avoid damages for the businesses and people of Brazil and all of Latin America.