Harmonic decomposition of vegetation indices time series for assessing mining impacts

Remotely sensed time-series data provide a powerful tool for environmental monitoring, particularly for assessing heterogeneous spatio-temporal vegetation dynamics in mining environments. Here, we present a new approach, SHABA (Seasonal Harmonic Anomaly Break Analysis), for remotely monitoring the effects of mines on vegetation. SHABA combines Seasonal and Trend decomposition (LOESS), Fast Fourier Transform-based seasonality (e.g., HANTS) and heuristic-based breakpoint detection to identify rapid and long-term vegetation changes. This allows us to quantify browning and greening intensity as deviations from local year-specific periodic and trend behaviour, and identify abrupt but potentially subtle changes (breakpoints). We apply this approach to MODIS EVI data from six mining sites (Aitik, Roșia Poieni, Trident, Lumwana, Carajás, and Vametco). Our results show spatially explicit, significant negative change magnitudes within primary mine footprints, reflecting vegetation loss driven by distinct phases of clearing for infrastructure expansion. Beyond operational boundaries (secondary footprints), change magnitudes are more subtle and exhibit heterogeneous greening–browning patterns, arising from either or a combination of direct mining effects and indirect land-use pressures associated with mine site establishment. SHABA workflow is transferable and can be applied globally to different mines to detect vegetation changes and, when interpreted, supports environmental reporting, impact assessment, and post-mining remediation.