New remote sensing methods are required to detect tropical forest degradation. Forest degradation caused by hydrocarbon pollution remains poorly understood. A fieldwork campaign was conducted in the Amazon region of Ecuador where numerous oil spills have occurred during the last decades. More than 1000 samples of leaves were collected at three levels of the canopy profile (upper, medium and understory) and analysed for biophysical, biochemical parameters and spectral signatures. Tropical vegetation in polluted site showed lower leaf chlorophyll content across the vertical profile. In the next step, 28 vegetation indices were applied to atmospherically corrected EOl-Hyperion images and a discriminant function analysis demonstrated that a combination of Sum Green (SG) and NDVI indices explains 74.6% of the separation between polluted and non-polluted sites. Chlorophyll content maps were computed based on the strong correlation (R~2=0.67) between ground truth chlorophyll content at leaf level from top canopy and MERIS Terrestrial Chlorophyll Index (MTCI). The computed chlorophyll content maps revealed lower chlorophyll levels (43.7 μg cm~(-2)) in areas near to petroleum facilities which suggest that vegetation experience vegetation stress symptoms probably caused by pollution. Secondary non-polluted forest reported 52.5 μg cm~(-2) of chlorophyll content and pristine forest 82.5 μg cm~(-2). The results of this study open a possibility to assess forest degradation in petroleum/gas productive areas across tropical forest environments by using hyper-spectral remote sensing methods.
展开▼
