In the summer of 2021, a massive oil spill spread across the Mediterranean towards Turkey and Cyprus from a power plant in Syria. The damaged tanks contained approximately 12,000 m3 of fuel. It was therefore essential to know the position and extent of the spill in order to act.
Among the specialists mobilized by the European intervention center (REMPEC) was the Valencian company Orbital EOS, which provided dozens of images and analyzes of the area. “It is not possible to understand an oil spill coming from an airplane, as we did in the past,” explained its director, Juan Peña, in several international media, including the BBC. “You have to observe it from space.”
Syria is one of the many scenarios in which this small Valencian technology company has played an important role. Using its satellite oil slick detection tool, the company has worked in some of the world’s oil transportation “hot zones” such as India, Singapore, Angola and Nigeria. In April 2024, for example, an explosion at the Campeche well, owned by Pemex, produced a slick of more than 500 km2 in the Gulf of Mexico that its teams followed in detail. And the same thing happened with some oil tankers recently attacked by Yemen in the Red Sea.
On the oil “highways”
For several years, its main client has been the Saudi public company SIRC (Saudi Investment Recycling Company), for those monitoring spills in the Persian Gulf and Red Sea. “They saw our work and, after many conversations, we came to an agreement to offer a service of 30 to 40 images of the area per day, or around 600 images per month,” explains Pablo Benjumeda, operations director at Orbital EOS, at elDiario.es.
Working in this area is not easy, because the concentration of oil tankers passing through it is the largest in the world. “Almost all hydrocarbon traffic between Europe and Asia passes through the Persian Gulf and the Red Sea,” emphasizes Peña. “There is on the one hand the intentional contamination of ships, which empty their tanks to clean them, and on the other the geopolitical situation, since every week Yemen attacks a British or American tanker, full of tons of fuel.”
Immediate reaction
Its main tool, EOS Viewer, is an online platform that combines Earth observation and geospatial analysis. The system is powered by images obtained by NASA and ESA satellites, which it processes to distinguish oil or oil slicks and predict their evolution, a technology that many countries do not have. In addition to regularly taking photos of the areas most traveled by tankers or where platforms and pipelines are located, the client sometimes notifies him to monitor a specific crisis situation, which he can monitor in a few minutes.
“Over the summer, we received internal notification that the Houthis had An oil tanker carrying 150,000 tonnes of crude oil was attacked off the coast of Yemen,” explains Juan Peña. “It was like four times the Exxon Valdez and two times the Prestige.” When the Houthis detonated the charges on the surface of the ship, they caused a series of fires in the tank holes in which the gases emanating from the crude oil burned, as if it were a huge lamp . “On our satellite images, you could see 14 flames at the top of the ship,” says Benjumeda. “That’s why this time there was no stain.”
Machine learning
The story of the two founders of the company is a bit like a movie. Juan and Pablo worked in maritime rescue between 2008 and 2020, participating in boat search missions and commanding remote sensing sensors. “It was a combined operation,” recalls the second. “When there was no emergency for human lives, we carried out flights to monitor pollution, that is, 80% were to search for oil stains.” One day, they decided to leave everything and use this knowledge in the field to establish themselves on their own.
The successes were not long in coming and in 2020 its digital solution was rewarded by the European Space Agency (ESA) for its innovative nature. Today, after having trained a machine learning system with more than 200,000 fragments of images extracted from the public Sentinel-1 satellite, they aspire to achieve, thanks to artificial intelligence, that the system detects spots autonomously with images from other satellites around the world. . “We have a model that works very well and obtains results similar to those of a human expert, which has not been achieved until today,” explains Peña.
Feeding the “flowers” in the ocean
The main problem is that from space, oil and oil stains do not appear as obvious as they might seem and there are many false positives. “What you can distinguish is a spot on a gray background, but it’s not easy, because very often they are not spots,” explains Benjumeda.
Experience has taught them the patterns that distinguish the spilling of a ship or the flowers that leaking oil pipelines attract the ocean. “If crude oil escapes from a pipeline, platform or natural reservoir, it generally comes out of the same point,” explains the expert. “Oil sinks from the sea floor and rises to the surface, drifting with the current, which changes every day, and leaving a succession of spots projected around a fixed point.”
Oil sinks from the sea floor and rises to the surface. It drifts with the current and leaves a flower shape.
Pablo Benjumeda
— Director of Operations of Orbital EOS
When they were tasked with making a historical map of the Gulf of Mexico using AI, several of these floral structures appeared, indicating multiple sources of pollution. “The origin is a set of platforms in the Campeche region that clearly have many associated leaks, both on the surface and in the wells they operate,” Benjumeda explains. “In Singapore, based on 500 images from the last 5 to 7 years, we have an extremely detailed map of chronic pollution, we can see pipe breakdowns, which are numerous, natural leaks, traffic areas , even stains from a year ago,” says Pena.
Often, it is not the countries that resort to the monitoring tool, but the oil companies themselves, because they are more interested than anyone else in knowing whether a leak occurs. “If a company loses crude oil, it loses what it sells,” Peña points out. “We help them at any point in the life cycle of an oil project, from the search for the deposit to the moment when the installation becomes obsolete and must be dismantled,” he explains. “In the North Sea, there are 1,500 platforms and by 2030, more than 600 will be dismantled,” adds Benujumeda. “And this process is going to require vigilance.”
This is seen everywhere on the planet, in places where there is surveillance, ships stain much less than where there is no surveillance.
Juan Peña, CEO of Orbital EOS
What they have found over these years is that knowing someone is watching reduces the number of spills, especially from ships wanting to clean their tanks. “This is seen all over the planet, in places where there is surveillance, ships stain much less than where there is none,” explains Peña. “In the Mediterranean, for example, we see many more illegal discharges below the borders of European countries, which are the ones that prohibit them. » Perhaps expanding this technology to more countries and larger ocean areas, combined with disincentives for those who fail to meet safety standards, can bring us a future with cleaner seas. And may the eyes watching from space make crude oil tides a thing of the past.
