Supercomputer modeling helps keep responders several steps ahead of a spreading spill.

When a stricken Iranian oil tanker, Sabiti, started leaking its cargo into the Red Sea on October 11, 2019, Ibrahim Hoteit and his team at KAUST soon received a call to assist.

At the time, the Sabiti — laden with 1 million barrels of crude oil — was only 100 km from the Jeddah shore. “I remember the afternoon we got the call,” says Samah El Mohtar, a research scientist working on oil spill predictive modeling. “We worked day and night to forecast the oil’s spread,” says El Mohtar, who was a Ph.D. student in Hoteit’s group at the time.

Predicting when and where the spill would reach the Saudi coastline was critical for minimizing the impact on marine and coastal ecosystems, particularly its unique coral systems, El Mohtar explains.

“In Saudi Arabia, oil spills are also a major threat to water security,” he says. The Kingdom produces 90 percent of its potable water from seawater desalination along the Red Sea coast. “If oil or other pollutants get close to the intake of these desalination plants, they must be shut down.”

Forecasting an oil spill’s spread is a multistep task requiring weather and ocean circulation modeling, as well as data on the spill itself. “Rather than rely on coarse-resolution global models of weather and ocean circulation, we produce our own high resolution, more reliable forecasts for the Red Sea and the Arabian Gulf,” El Mohtar says. “Only after generating our own weather and then ocean circulation and wave forecasts on the KAUST supercomputer, Shaheen, can we run the oil spill forecast model.”

Equipped with the Sabiti spill forecast — which the team processed continuously for 16 hours to generate — Saudi authorities were able to coordinate their containment and clean-up resource allocation, targeting affected areas. “Our forecasts aided in proactive decision making to mitigate the spill’s environmental impact and protect sensitive areas,” Hoteit says.

Read more at KAUST Insight.