In the heart of Southeast France, an ancient agricultural practice is being revived with the help of cutting-edge technology. Terraces, those stepped structures carved into hillsides, have long been a testament to human ingenuity, transforming steep slopes into arable land. Now, a team of researchers, led by Michael Vincent Tubog from the Physics and Geology Department at Negros Oriental State University in the Philippines, is using machine learning to map and revitalize these historic terraces, with potential implications for the energy sector’s land management strategies.
The Roya Valley, nestled in the French Alps, is home to countless dry stone terraces that have lain dormant for decades. But when Storm Alex struck in October 2020, these terraces proved their mettle, showcasing remarkable resilience against erosion. This event sparked a renewed interest in these structures, not just for their historical significance, but also for their potential to support modern agricultural initiatives and agri-tourism.
Tubog and his team set out to develop a semi-automatic method for detecting and mapping these terraced areas using LiDAR (Light Detection and Ranging) and orthophoto data. They turned to French national repositories for their data, processing it with GIS (Geographic Information System) software and analyzing it through a Support Vector Machine (SVM) classification algorithm. The results were impressive: the model identified 18 terraces larger than 1 hectare in Saorge and 35 in La Brigue.
“The terraces we found are not just historical artifacts,” Tubog explained. “They represent a sustainable land management practice that can help mitigate risks in mountainous regions. By mapping these terraces, we’re providing a tool for future landscape restoration and food resilience planning.”
The accuracy of the model varied between the two sites, with a user accuracy of 97% in Saorge and 72% in La Brigue. This disparity, the researchers noted, reflects site-specific differences such as terrain steepness, vegetation density, and data resolution. These findings underscore the importance of accounting for local geomorphological and data-quality factors to improve model performance.
So, how might this research shape future developments in the field? For one, it highlights the value of machine learning in terrace mapping, a task that has traditionally been labor-intensive and time-consuming. By automating this process, researchers and land managers can cover larger areas more efficiently, identifying potential sites for agricultural revitalization and risk mitigation.
Moreover, the study’s findings could have significant implications for the energy sector. As renewable energy projects increasingly encroach on rural and mountainous landscapes, understanding and preserving these terraces could help balance energy development with land conservation. For instance, solar farms could be designed to coexist with terraces, using them to minimize environmental impact and enhance biodiversity.
The research, published in the journal ‘Land’ (translated from French), is a testament to the power of interdisciplinary collaboration. By combining geology, agriculture, and machine learning, Tubog and his team have opened up new avenues for sustainable land management. As we look to the future, their work serves as a reminder that sometimes, the key to progress lies in looking back at traditional practices and giving them a modern twist.
The energy sector, with its growing focus on sustainability and land stewardship, would do well to take note. After all, the terraces of the Roya Valley are more than just steps carved into a hillside; they are a blueprint for a more resilient and sustainable future.