Spectral and spatial indicators of botanical changes caused by long-term hydrocarbon seepage

Recent studies have shown that the presence of anomalous concentrations of gases in the root zone can cause changes in leaf or canopy reflectance. Changes in vegetation patterns due to anomalous hydrocarbon concentrations in the soil were not taken into account in these studies. Since all changes in reflectance found thus far were general stress indicators, the detection of hydrocarbon leaks using remote sensing must be supported by additional indicators of hydrocarbon leakage, such as changes in vegetation patterns. In this study four reflectance indices selected from previous studies were used to detect long term effects of hydrocarbon seepage on vegetation patterns in a meadow near Ojai in the United States. Up to a distance of 20 m from the source of the gas seep, a circle of bare soil prevailed, which was surrounded by a ring of dense vegetation. The patterns observed in the field were translated into a 3 × 3 pixel wide convolution filter, which was applied to a Probe-1 image of the area. Filtering resulted in fewer false anomalies for all tested indices, but the only index that both reduced the number of false anomalies and improved the detection accuracy was the Lichtenthaler index (R440/R740). This index resulted in the best distinction between bare soil and vegetation and could map all four seeps present in the area. Although the three other indices have shown to be successful in mapping chlorophyll differences in gas-affected vegetation, the Lichtenthaler index proves to be a better indicator when the seep is characterised by bare soil contrasting with dense vegetation. Although a scene will never be completely free of false anomalies when vegetation indices as in this study are used to map hydrocarbon seepage, combining spectral information with a filter that searches for specific spatial patterns related to seepage will reduce the number of false anomalies.

► Vegetation diversity changes occur in circular patterns around hydrocarbon seeps.
► Hyperspectral imagery was used to detect these botanical anomalies around seeps.
► Lichtenthaler reflectance index reveals vegetation diversity patterns best.
► Spatial filtering increases the detection rate of seeps in grassland.