Identifying plants under natural gas micro-leakage stress using hyperspectral remote sensing
Introduction
Natural gas is an important source of clean energy. In recent years, the demand and consumption of natural gas have been increasing. Pipeline transportation is important for transporting natural gas, and the safety of pipelines cannot be ignored. Once natural gas leaks during transportation, it might have harmful impacts on the environment, economy, and safety (Lukonge and Cao, 2020; Oh et al., 2018; Ren et al., 2018). Therefore, it is important to develop efficient methods for monitoring pipelines and the surrounding environment.
Although the leakage of large gas fluxes can be easily detected by existing methods, the direct detection of micro-leaks is relatively complicated (Lukonge and Cao, 2020). However, natural gas leakage can be detected by plants, as natural gas replaces oxygen in the soil, causing a lack of oxygen in the roots of vegetation and indirectly affecting plant growth (Smith et al., 2004). In this case, the plant responds with spectral abnormalities, reduced leaf greenness, and sparsity (Sanches et al., 2013b). The indirect detection of natural gas can be achieved by monitoring abnormalities during plant growth on the surface of the leak. Hyperspectral remote sensing technology was found to be useful for detecting vegetation anomalies caused by liquid and gaseous hydrocarbons (HCs) (Noomen et al., 2012; Sanches et al., 2014). In this study, we designed a small-scale natural gas leakage simulation field experiment under controlled conditions to evaluate the potential use of hyperspectral remote sensing to detect natural gas leakage.
In recent decades, many studies have been conducted on the use of hyperspectral remote sensing to detect HC stress in plants. Smith et al. (2004) found that the canopy spectral reflectance of vegetation exposed to natural gas increased in the visible (VIS) region and decreased in the near-infrared (NIR) region, and that stress caused by natural gas in plants could be detected using the red edge characteristic of the first derivative spectrum. Noomen et al. (2008) performed a covariance analysis between the spectral indices of two plant species (maize and wheat) stressed by natural gas leakage and oxygen concentration and found that there was a high correlation between hyperspectral indices and oxygen concentration, which could be used to detect the anomalies caused by natural gas in plants. Sanches et al. (2013a) found that for plants under diesel and gasoline stress, the stressed vegetation could be distinguished based on the analysis of leaf and canopy spectral red edge, thereby indirectly detecting gasoline and diesel pollution. Gürtler et al. (2018) used hyperspectral methods to analyze the spectral changes of plants contaminated by HCs at the leaf and canopy scales and found that the reflectance of contaminated plants increased in the VIS region and decreased in the NIR and shortwave infrared (SWIR) regions. Analyzing canopy spectra could effectively distinguish contaminated plants from uncontaminated ones. Lassalle et al. (2018) conducted a greenhouse experiment on three plant species that were exposed to oil pollution for a long duration and performed a linear discriminant analysis on the canopy spectra. They found that the original and converted spectra could distinguish between oil-stressed and healthy plants. Ran et al. (2020) performed an analysis of variance on the spectra of bean, corn, and grass canopies under natural gas leakage stress, and constructed a natural gas stress index (NGSI), which could successfully identify stressed plants. In recent years, numerous studies have focused on delineating vegetation abnormalities caused by liquid HCs using remote sensing (Huang et al., 2019; Lassalle et al., 2019a; Lassalle et al., 2019b; Lassalle et al., 2019c), but relatively little attention has been paid to gaseous HCs (Ran et al., 2020). Therefore, research in this area is necessary. In this study, a field experiment was designed to simulate natural gas leakage, with the aim of constructing an index model to monitor the vegetation health of diverse plant species and help indirectly monitor underground natural gas pipeline micro-leakage.
The common processing approach in this field is to eliminate or reduce the effects that are not related to the features of interest in the targets and to enhance the spectral features, such as spectral transformations and vegetation indices (Sanches et al., 2013b; Sanches et al., 2014). Smith et al. (2004) conducted a derivative analysis of the canopy spectra and constructed a spectral index (FD725/FD702) based on the first derivative peaks in the red edge. This achieved good results at detecting natural gas leakage stress for three plant species, namely, grass, beans, and winter wheat. Noomen et al. (2006) performed continuum removal (CR) processing on a maize canopy spectrum under natural gas, methane, and ethane stress, and proposed band depth and normalized band depth to effectively extract the changes in spectral absorption characteristics induced by stress. Sanches et al. (2014) used CR to detect the stress caused by liquid HCs on two plant species (perennial soybean and brachiaria) and constructed the plant stress detection index. It was found that the index could detect stress in brachiaria and perennial soybean earlier than the red edge position. Lassalle et al. (2019c) tested 33 vegetation indices in a bramble study under mixed petroleum HCs and heavy metals stress and found that 14 indices could effectively distinguish each mixture from the control treatment.
A novel spectral analysis method was used in a study to extract spectral features using variational mode decomposition (VMD), which adaptively decomposes the signal and has a strong local decomposition ability (Dragomiretskiy and Zosso, 2014). This method was initially used to analyze extracted noise and signals (Feng et al., 2017; Ma et al., 2020). Yang et al. (2018) used VMD for the first time to extract information on corn stressed by heavy metals (Cu, Pb) and diagnosed the degree of contamination. This indicated the potential of VMD for hyperspectral data processing. This method was adopted in this study to extract the spectral features of vegetation under gaseous HC leakage stress.
To effectively detect HC stress in plants, we also proposed a new index based on the abovementioned spectral analysis method, namely the variational mode decomposition index (VMDI). This index is based on two bands (616 and 829 nm) that are sensitive to stress in three different plant species. The primary objective of this study was to design the VMDI, evaluate the stress recognition ability for all plant species during the entire growth period using the proposed index, and compare it with three indices proposed in previous studies. The findings of this study may be significant for the other wide-range pipeline inspection of natural gas leakage using airborne platforms in the future.
Section snippets
Plant species
Wheat (Triticum aestivum L.) is the world's third-most important food crop after corn and rice (Asseng et al., 2011), and is widely cultivated globally. Considering that the height of corn in the late growth stage makes data collecting relatively difficult and that rice requires a demanding growth environment, wheat was assumed to be a suitable and representative crop in this study for evaluating leakage.
Cynodon dactylon L. (Pers.) is one of the most widely used perennial warm-season grass in
Visible symptoms of stress
Because of the stress of natural gas, the physiology and morphology of plants change, which may result in obvious changes that are visible to the naked eye. In this study, the notable stress symptoms caused by wheat, grass, and soybean growing on the gassed plots were the yellowing of leaves, decreased plant height, and canopy cover sparsening. Moreover, bare land appeared in areas with higher concentrations of gas when the vegetation became too sparse. Notably, compared to the control wheat (
Conclusion
Natural gas pipeline leakage increases the concentration of natural gas in the soil, thereby affecting the health of plants and causing notable visual symptoms and abnormal canopy spectral reflectance for plants. VMD was applied to analyze the canopy spectrum, along with an improved spectral feature selection method (SD) that could determine the two bands (616 and 829 nm) that were sensitive to gas stress for all three plant species. The VMDI was then constructed. Subsequently, the recognition
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
This work was supported by the National Natural Science Foundation of China [41871341,41571412], and the Fundamental Research Funds for the Central Universities [2020YJSDC03].
References (41)
- et al.
Measurement of total flavone content in snow lotus (Saussurea involucrate) using near infrared spectroscopy combined with interval PLS and genetic algorithm
Spectrochim. Acta A Mol. Biomol. Spectrosc.
(2010) - et al.
Remote sensing for vegetation monitoring in carbon capture storage regions: a review
Appl. Energy
(2019) - et al.
Investigating potato late blight physiological differences across potato cultivars with spectroscopy and machine learning
Plant Sci.
(2020) - et al.
Determination of changes in leaf and canopy spectra of plants grown in soils contaminated with petroleum hydrocarbons
ISPRS J. Photogramm.
(2018) - et al.
Hydrocarbon micro-seepage detection from airborne hyper-spectral images by plant stress spectra based on the PROSPECT model
Int. J. Appl. Earth Obs. Geoinf.
(2019) - et al.
Comparison of two feature selection methods for the separability analysis of intertidal sediments with spectrometric datasets in the German Wadden Sea
Int. J. Appl. Earth Obs. Geoinf.
(2016) - et al.
Application of PROSPECT for estimating Total petroleum hydrocarbons in contaminated soils from leaf optical properties
J. Hazard. Mater.
(2019) - et al.
Detection and discrimination of various oil-contaminated soils using vegetation reflectance
Sci. Total Environ.
(2019) - et al.
Monitoring oil contamination in vegetated areas with optical remote sensing: a comprehensive review
J. Hazard. Mater.
(2020) - et al.
Impact fault detection of gearbox based on variational mode decomposition and coupled underdamped stochastic resonance
ISA Trans.
(2019)
Continuum removed band depth analysis for detecting the effects of natural gas, methane and ethane on maize reflectance
Remote Sens. Environ.
Spectral and spatial indicators of botanical changes caused by long-term hydrocarbon seepage
Ecol. Inform.
Acoustic data condensation to enhance pipeline leak detection
Nucl. Eng. Des.
Pipeline corrosion and leakage monitoring based on the distributed optical fiber sensing technology
Measurement
Multi-variable selection strategy based on near-infrared spectra for the rapid description of dianhong black tea quality
Spectrochim. Acta A Mol. Biomol. Spectrosc.
Unravelling remote sensing signatures of plants contaminated with gasoline and diesel: an approach using the red edge spectral feature
Environ. Pollut.
Assessing the impact of hydrocarbon leakages on vegetation using reflectance spectroscopy
ISPRS J. Photogramm. Remote Sens.
Spectroscopic remote sensing of plant stress at leaf and canopy levels using the chlorophyll 680-nm absorption feature with continuum removal
ISPRS J. Photogramm. Remote Sens.
Use of hyperspectral derivative ratios in the red-edge region to identify plant stress responses to gas leaks
Remote Sens. Environ.
Decomposition ensemble model based on variational mode decomposition and long short-term memory for streamflow forecasting
J. Hydrol.
Cited by (11)
Predicting TFe content and sorting iron ores from hyperspectral image by variational mode decomposition-based spectral feature
2023, Journal of Cleaner ProductionForecasting of wind speed under wind-fire coupling scenarios by combining HS-VMD and AM-LSTM
2023, Ecological InformaticsProximal hyperspectral sensing of abiotic stresses in plants
2023, Science of the Total EnvironmentCitation Excerpt :HSI measurements can be used to examine the spectral properties of plants located close to leakage points, enabling leakage rates of underground pipelines to be calculated. An experiment designed by Pan et al. (2022) to simulate an underground gas pipeline leak and the effects of gas exposure on different plant cultivars provided an index model for monitoring plant health. This paper presented the variational mode decomposition index (VMDI), which employs two bands at 616 and 829 nm, that have a high correlation with gas stress.