Key points are not available for this paper at this time.
• Spectral characteristics of vegetation across browning stages are systematically analyzed. • A Browning Vegetation Index (BVI) is proposed, integrating red-edge, NIR, and SWIR bands. • BVI outperforms 11 standard indices in distinguishing browning vegetation stages. • Validated BVI demonstrates robustness for large-scale vegetation browning monitoring. Browning vegetation—an early-warning indicator of ecosystem degradation and environmental stress—presents a significant challenge for large-scale monitoring. Conventional indices like the Normalized Difference Vegetation Index (NDVI) are susceptible to saturation effects and have difficulty in detecting subtle or early-stage browning, while most stress-sensitive indices target single stressors and provide inadequate results across broad areas. To overcome these limitations, field spectra were first collected and linear spectral unmixing was applied in order to characterize vegetation at multiple browning levels. Building on these insights, a novel Browning Vegetation Index (BVI) was developed specifically for Sentinel-2 imagery. This index strategically leverages the synergistic relationship between chlorophyll degradation (captured by the red-edge band) and moisture loss (reflected by the NIR and SWIR bands) to maximize spectral contrast across browning stages. We demonstrated the capability of the BVI for large-scale mapping in the Tianshan Mountains, where the annual spatial transition from healthy green to browning vegetation was traced using Canny edge detection and Otsu adaptive thresholding. Validation with high-resolution Gaofen-2 data across four Tianshan subregions yielded precision, recall, and F1 scores exceeding 0.85, along with an overall accuracy of 0.883. Against 11 established indices, the BVI consistently delivered superior sensitivity to varying browning intensities. Finally, cross-continental tests in the United States, Afghanistan, India, and China confirmed BVI’s robustness and transferability for the fine-scale extraction of browning vegetation across vast and varied landscapes.
Liu et al. (Tue,) studied this question.