Use of hyperspectral derivative ratios in the red-edge region to identify plant stress responses to gas leaks

Abstract

Hyperspectral features in the red-edge region were tested as an index of plant stress responses to soil–oxygen depletion. The aim was to provide the basis for a warning system to identify natural gas leakage by the spectral responses of plants growing in the affected soil.

Elevated concentrations of natural gas in the soil atmosphere were used to deplete oxygen concentrations around the roots of grass, wheat (Hordeum vulgare cv Claire) and bean (Vicia faba cv Clipper) growing in a field facility. Visible symptoms due to the natural gas included reduced growth of the plants and chlorosis of the leaves.

Spectral responses included increased reflectance in the visible wavelengths and decreased reflectance in the near infra-red. Derivative analysis identified features within the red-edge at 720–730 and 702 nm. Ratios of the magnitude of the derivative at 725 to that at 702 nm were less in areas where gas was present. This ratio enabled identification of stress due to gas leakage up to 7 days before visible symptoms were observed and also at the edges of gassed plots where visible symptoms were not expressed. The technique was able to identify stress responses to long-term leaks in all the crops tested but to short-term leaks only in grass. This study therefore suggests that under appropriate conditions remote sensing could be used to detect pipeline gas leaks from decreases in the ratio of peaks within the red-edge.

Introduction

Gas pipeline engineers have reported vegetation changes around the area of gas leaks in underground pipelines. Leaking pipelines may be detected by the use of remote sensing of the surrounding vegetation to identify early signs of plant stress such as chlorosis of the leaves and poor development of the plants. Remote sensing of stressed vegetation by satellite may offer the potential for early detection of symptoms that are below the subjective detection level, while at the same time reducing the risks associated with aerial inspection of the network.

The stress symptoms in response to gas leaks are believed to be a generic response to the displacement of soil–oxygen from the soil, thus inhibiting root respiration that provides energy for root growth and uptake of water and nutrients from the soil Arthur et al., 1985, Gilman et al., 1982, Hoeks, 1972, Smith, 2002. This effect may be compounded by the oxidation of leaking methane by methanotrophic bacteria that utilise the methane present in natural gas as a carbon-based energy source (Hanson & Hanson, 1996).

Remote sensing has previously been used to detect stress in plants before visible symptoms have been observed. Carter (1993), Carter and Miller (1994) and Carter et al. (1996) used various stresses and plant species to induce changes in reflectance and found that visible reflectance increased consistently in response to stress, and that herbicide-induced stress was detectable 16 days prior to the first visible signs of stress.

The region of the red-edge has also been used as a means of identifying stress. This is the area where there is a sharp change in reflectance between wavelengths 690 and 750 nm which characterises the boundary between dominance by the strong absorption of red light by chlorophyll and the high multiple scattering of radiation in the leaf mesophyll. Derivative analysis of this region shows a peak that can be used to describe changes due to stress. This peak which makes the inflection point of the reflectance spectrum in the red-edge region is defined as the absolute maximum of the first derivative in the range 690–750 nm (λ_p) Filella & Peñuelas, 1994, Horler et al., 1983, Miller et al., 1990. Rock et al. (1988) detected a shift in λ_p, towards the blue, of approximately 5 nm when measuring severe foliage stress on spruce trees due to air pollution. This shift, which was due to a decline in chlorophyll in the pine needles, was detected before visual symptoms became apparent and it was proposed that airborne monitoring could be used to provide an early indicator of vegetation stress.

The wavelength of the red-edge inflection point (λ_p) can be found by plotting the first derivative of the reflectance spectrum, and then manually identifying the highest peak Boochs et al., 1990, Filella & Peñuelas, 1994, Horler et al., 1983, or by fitting a Gaussian curve to the red-edge and extracting the maximum slope wavelength from the coefficient of the fitted curve Miller et al., 1990, Pinar & Curran, 1996. A limitation of both techniques is the assumption that there is only a single maximum in the gradient of the red edge (Lamb et al., 2002). Smoothing of hyperspectral reflectance spectra to remove noise is usually carried out prior to derivative analysis and this often has the effect of smoothing out small spectral features that may contain important stress-related information.

Several researchers have shown that the peak revealed by derivative analysis of the red edge reflectance is actually composed of two or more features. Horler et al. (1983) identified two peaks in derivative spectra; the first at around 700 nm was attributed to the chlorophyll content in the plant leaves and the second at around 725 nm was attributed to cellular scattering in the leaf. Boochs et al. (1990) identified peaks in winter wheat at 703 and 735 nm and Railyan and Korobov (1993) noticed underlying components in the red-edge situated at 700, 715 and 745 nm. Similar effects were seen in the red-edge of a grass canopy by Jago and Curran (1996) and Llewellyn and Curran (1999). While studying grassland canopies at a site contaminated with oil, Jago and Curran (1996) found maxima within the red-edge with peaks at approximately 709 and 693 nm. The position of the major peak changed depending on the amount of contamination within the plot. Llewellyn and Curran (1999) also found multiple first derivative features with peaks at 700 and 729 nm. They found that the shorter wavelength feature indicated grassland with high levels of soil contamination whereas the longer wavelength feature indicated lower levels of contamination; the transition from one red-edge position to another was not a gradual change but instead a switch in dominance of the features. Lamb et al. (2002) found that in leaves with low chlorophyll content the peak at ∼705 nm was dominant but with higher chlorophyll levels the peak at ∼725 nm was dominant. Zarco-Tejada et al. (2002) found that a double peak in the derivative reflectance spectrum of the red-edge at 705 and 722 nm was related to increased chlorophyll fluorescence and chlorophyll concentration. Fluorescence occurs when red and far-red light is emitted from green plants in response to excess stimulation by photosynthetically active radiation. In general, fluorescence is low when photosynthesis is high but increases under stress conditions when the photosynthetic system is unable to respond effectively to light. Chlorophyll fluorescence is at a maximum near 690 and 730 nm Lichtenthaler, 1996, Zarco-Tejada et al., 2002.

Changes in chlorophyll function often precede changes in chlorophyll content so that fluorescence changes may be observed before leaves become chlorotic. D'Ambrosio et al. (1992) measured fluorescence directly and found that the ratio of the fluorescence at 690 nm to that at 735 nm increased as chlorophyll content decreased and suggested that fluorescence could be used as a non-destructive measure of stress.

Although these features have been identified previously, they have not been applied as a method of detecting stress effects that could be used in a remote sensing monitoring system. In this study, three contrasting plant species were exposed to elevated concentrations of natural gas in soil to determine whether gas leak effects on canopy reflectance could be detected by remote sensing. Hyperspectral reflectance spectra were used to identify changes in the spectral response of plants. The shape and position of the red-edge was studied and ratios of derivative peaks used to detect differences between control and gassed plants.