Science1 publisher3 min readPublished
HZDR's SHABA fits each site's seasonal cycle before it flags a vegetation change
A team at Helmholtz-Zentrum Dresden-Rossendorf fitted 26 years of satellite vegetation data around six mines so that seasonal leaf fall is not mistaken for clearing, and released the tool for anyone to run.
The Scientist · Science desk
What happened
- Researchers at Helmholtz-Zentrum Dresden-Rossendorf built SHABA, a satellite tool for detecting and quantifying vegetation change around mining sites, and published the method in the ISPRS Journal of Photogrammetry and Remote Sensing.
- The losses and gains the tool flagged lined up with events already known at those sites, including clearing for site expansion or an access road and gradual regrowth after reclamation work.
- SHABA is already free to use, and the group plans to run it over every mine worldwide and publish the resulting monitoring data at no cost.
Compiled by The ScientistSomething wrong?How this is made
Why it matters
- capability Because the yearly cycle is modelled at each location, a flagged change comes with a date and coordinates, so a reclamation timeline can be tested against an independent series by someone who never visits the site.
- constraint Green cover is what a vegetation index measures, so the tool can establish when cover went and when it returned, but it cannot say whether the ecosystem recovered.
- decision A regulator or metals buyer who wants to lean on this output in a dispute has to characterise its error rate first, because the published check is agreement at six sites.
- precedent A free global layer would hand anyone auditing a supplier a disturbance record the operator did not produce, in places where Gloaguen says the rules differ from Europe's.
Harmonic decomposition, the phrase in the paper's title [3], means fitting periodic terms to each location's annual vegetation cycle and then examining what the fit fails to explain. Departures from the fitted curve are anomalies. Step changes are breaks, and the acronym carries both [2]. Vincent Nwazelibe, the study's first author, said: "A crucial feature is SHABA's ability to factor in typical seasonal changes in vegetation, such as leaf fall or blossoming periods, and also to compare data with previous years' trends" [7]. That, he said, allows the tool to "reliably detect changes in vegetation and pinpoint them precisely in space and time" [8]. Samuel Thiele heads the digital data processing group in the same department. He said SHABA "converts the vegetation data from satellite images into a clear wave diagram, which makes it easy to see and quantify any significant changes in vegetation at a specific location" [9][16].
From 2000 to 2025 inclusive is 26 years [1]. At more than 100 images a year, that is more than 2,600 images per site [2], and more than 15,600 across the six [3].
A circle of 100 kilometre radius covers about 31,000 square kilometres [4], most of it not mine. Cropping cycles and logging inside the buffer move the index for their own reasons. Separating them from the mine's own footprint meant checking the breaks against events already known at each site [10].
Richard Gloaguen heads the exploration department at HZDR's Helmholtz Institute Freiberg for Resource Technology. He described the result as a demonstration of "SHABA's potential as a remote sensing tool for vegetation analysis near mining facilities across a wide variety of climatic regions around the world" [13][16]. The phys.org account of the work gives no detection rate or false-alarm rate, and no comparison against another change-detection method [17]. The published check covers six sites spread across climate zones [4].
The motivation Gloaguen gives is a gap in the record. "In Europe, operators are obligated to ensure that their operations comply with the law. In other regions of the world, regulations may differ from ours," he said [11]. He also said that "up to now, the land footprint of global mining has been inadequately quantified because we had no remote sensing tool to reliably monitor and quantify the environmental impact of the thousands of mines worldwide that underpin metal supply chains" [12]. He said he hopes the tool's "unbiased remote sensing data will facilitate better dialogue between stakeholders on both sides: mining and environmental protection" [14].
What a vegetation index measures is how much green cover a pixel has [3]. A gain after reclamation measures cover coming back, and leaves which species returned, and what the soil and water underneath carry, unanswered.
What to watch
- Whether the promised worldwide run appears, and under what licence the monitoring data is released.
- Whether the ISPRS paper itself reports detection and false-alarm rates that the press summary leaves out.
- Whether any permitting authority or metals buyer cites SHABA output in an audit or enforcement file.