South Africa is bracing for a repeat of last season’s locust swarms, the worst in decades, and is banking on a tracking technology backed by the institute of Microsoft Corp’s late co-founder, Paul Allen, to tackle the infestation.

Agri Eastern Cape, which represents farmers in the coastal province, started using the Allen Institute for AI’s Earth Ranger system in April to pinpoint where swarms will erupt in September and October, according to Jason Kümm, manager of rural safety and communications, at the organisation.

“Things are quietening down but these locusts have laid eggs,” he said in an interview. “Earth Ranger has allowed us to understand where this has taken place. When conditions are right and these eggs hatch, we will have the resources in place.”

Swarms that spread across the Eastern Cape and neighboring provinces in the summer growing season were as big as 10 000 ha in some cases and the infestation rivaled those seen in 1935 and 1983.

The locusts razed pastures used to feed livestock across large parts of the region, including in areas that were recovering from eight-year droughts. The heaviest rains since records began in 1921 in many districts allowed the insects to multiply rapidly. While citrus-growing areas were infested, the locusts didn’t cause significant damage to fruit trees.

The area is also home to dairy farms and sheep and angora goats, whose hair is used in mohair garments.

“This is one of the biggest we have experienced in history,” said Gunther Pretorius, manager for economics and natural resources at Agri Eastern Cape.

Earth Ranger allows farmers to call in reports of swarms, and whether the locusts are at a juvenile hopping stage or already airborne. Those are then logged and tracked on the system, allowing for more efficient allocation of pesticides and resources. They can also decide whether ground teams or aircraft are needed to tackle the insects.

Offered as a free service, the tracking system was developed in 2015 and was first used to measure the size of Africa’s savanna elephant population by assessing data from aerial surveys.