A data scientist from Mali is part of a team using machine learning to help predict—and hopefully prevent—destructive locust swarms in West Africa’s Sahel region.
During the last major locust outbreak in West Africa in 2003-2005, authorities spent more than $450 million to end the plague, which caused an estimated $2.5 billion in crop damage. Another disaster occurred in 2012, threatening the food security of more than 50 million people, but successful control operations lessened the impact.
Fatumata Haidaramachine learning engineer at the SERVIR West Africa Small Innovation Grant Program, explains that he saw a need for better locust prediction, which led to the incorporation of machine learning techniques to improve locust prediction accuracy in P_Locust Platform.
“This work is vital because locust swarms can destroy crops, leading to severe food shortages and economic instability,” he says, “By improving the platform, we aim to provide more accurate forecasts, enabling timely interventions that protect agricultural livelihoods ».
Haidara explains that her research focuses on analyzing weather patterns, vegetation status and ecological conditions to predict locust outbreaks: by integrating different environmental indicators and real-time data, researchers can provide earlier and more accurate warnings to those in charge decision making.
“This is critical for early intervention, which helps prevent crop damage, ensuring food security and economic stability for communities in West Africa,” he says, “With the platform’s improved accuracy and extended turnaround times, Policy makers and communities can take proactive decisions to reduce the impact of locust infestations.”
Machine learning in Mali
Haidara grew up in Sikaso, a prosperous farming town in Mali, where her parents owned fields and she often helped water the vegetable plants.
“This experience made me curious about automating agricultural tasks to simplify them and protect crops from climate hazards,” he says, “This early fascination led me to pursue graduate studies in physics and eventually a career in artificial intelligence applied to climate science’.
Haidara explains that her current role as a machine learning engineer allows her to meet the challenges associated with managing and mitigating the effects of locust infestations, which can devastate crops and livelihoods in the region.
“It is vital for scientists from the Global South to explore solutions to global challenges because they possess an inherent understanding of their local context, culture and specific issues facing their communities — this deep local knowledge allows them to develop more relevant and sustainable solutions,” he says, adding that empowering local scientists strengthens scientific capacity and helps communities be better prepared to face future challenges independently.
Drones could reduce farmer-herder conflict in Benin
Elsewhere in West Africa, another team of researchers is using high-tech solutions to help farmers and ranchers: Daniel David Tossou, Director of ATLAS-GIS Sa and Benin Flying Labs, says his team’s drone flights have produced high-resolution land-use maps analysis and then The team did topographical surveys to draw up plot plans for the prospective landowners along a cattle crossing corridor.
For decades, the lack of grazing land for livestock and the expansion of agricultural land due to poor soils has led to conflict between ranchers and farmers, but now, drones have helped find a path that minimizes conflict.
“A 156-kilometer corridor is now demarcated with cement markings, and county and community authorities have issued regulations calling for herders to use the corridor,” Tosou says, adding that the corridor issue is a national concern as conflicts between farmers and herders continue. to cause death and great damage, with loss of human life.
“Awareness campaigns are being carried out in the areas where the corridor is located, using mass media (awareness and information panels) and village meetings,” he says, “however, there has been some resistance over the past year, without dramatic consequences, unlike other areas where the phenomenon is still widespread.