• search hit 7 of 7
Back to Result List

Identification of potential toxic elements (PTE) in Technosols and in the hyperaccumulator plant Brassica juncea with imaging spectroscopy

  • Vital, fertile soil is a limited resource and must be considered equal to air and water in its status as a protected good. In Europe, there are around 2.5 million potentially contaminated due to natural and anthropogenic activities. One third of these areas have been identified and 15% of them have already been remediated. An efficient approach for remediation is phytoremediation, a cost-effective and ecosystem-friendly method compared to conventional methods. We showed that hyperaccumulator plants, which have a high tolerance to the accumulation of potentially toxic elements (PTE), are suitable for phytoremediation methods. Imaging spectroscopy allowed us to monitor and optimize the process of accumulation. Our study was conducted in two steps: First we analysed the potential of multivariate procedures using Partial Least Squares Regression (PLSR) and Random Forest Regression (RFR) to predict highly contaminated areas with an inhomogeneous distribution of PTE concentrations from image data. Therefore, we used contaminated soilVital, fertile soil is a limited resource and must be considered equal to air and water in its status as a protected good. In Europe, there are around 2.5 million potentially contaminated due to natural and anthropogenic activities. One third of these areas have been identified and 15% of them have already been remediated. An efficient approach for remediation is phytoremediation, a cost-effective and ecosystem-friendly method compared to conventional methods. We showed that hyperaccumulator plants, which have a high tolerance to the accumulation of potentially toxic elements (PTE), are suitable for phytoremediation methods. Imaging spectroscopy allowed us to monitor and optimize the process of accumulation. Our study was conducted in two steps: First we analysed the potential of multivariate procedures using Partial Least Squares Regression (PLSR) and Random Forest Regression (RFR) to predict highly contaminated areas with an inhomogeneous distribution of PTE concentrations from image data. Therefore, we used contaminated soil samples analysed for heavy metal contents as training data. The fully-trained models assisted to monitor PTE contaminated areas during phytoremediation. In a second step, we investigated hyperaccumulator plants and their spectral fingerprints during PTE uptake to detect the highest possible amount of PTE the plant can tolerate. Soil samples were measured after different preparation steps (“oven-dried”, “sieved”, “ground”, “Loss on Ignition (LOI)”) with a HySpex VNIR-1600 and HySpex SWIR 320m-e hyperspectral sensor under laboratory conditions. The spectral range of both sensors covers 450 nm to 2500 nm with a spectral sampling interval of 3.7 nm for the VNIR sensor and 6.25 nm for the SWIR sensor. The resulting spectral library was used for the PLSR and RFR analysis. Considering the optimal coefficient of determination (R2), PLSR showed an improving performance and accuracy with increasing preparation steps: R2_Cr: 0.52–0.78; R2_Cu: 0.36–0.73; R2_Ni: 0.19–0.42 and R2_Zn: 0.41–0.74. In comparison, RFR showed a weaker estimation performance, even when using higher sample preparation levels (R2_Cr:0.36–0.62; R2_Cu: 0.17–0.72; R2_Ni: 0.20–0.35 and R2_Zn: 0.26–0.67). The results indicate that PLSR provides a more robust estimation than the user-friendly RFR method. Additionally, the PTE estimation performance in strong heterogeneous soil samples can be improved by pre-treatment of soil samples in the laboratory. Second, we cultivated Brassica juncea in a greenhouse and applied the plants with different controlled zinc, nickel and copper concentration levels (low, medium, high). Measurements with a HySpex VNIR-SWIR hyperspectral sensor (408-2500 nm) and a point spectrometer PSR+ from Spectral Evolution (350-2500 nm) were conducted in-situ and in the laboratory. A principal component analysis (PCA) was performed on reflectance spectra to identify and visualize spectral changes with increasing PTE uptake. In addition, different indices were calculated, such as Leaf Area Index (LAI) and Red-Edge Inflection Point (REIP). First results revealed changes at the chlorophyll feature between 500-600 nm and at the REIP position with increasing PTE concentration. Further spectral features and indices will be investigated to check hyperspectral responses of PTE concentration. With reflectance imaging we are already able to detect differences of the PTE accumulation within the hyperaccumulator plants.show moreshow less

Export metadata

Additional Services

Search Google Scholar
Metadaten
Author: Friederike Kästner, Magdalena Sut-LohmannORCiD, Mark Grimm, Hannes FeilhauerORCiD, Theres Küster, Thomas RaabORCiDGND
URL:http://is.earsel.org/workshop/12-IS-Potsdam2022/wp-content/uploads/2022/06/programm-booklet_long_complete_v2.pdf
Title of the source (English):12th EARSeL Workshop on Imaging Spectroscopy
Document Type:Conference publication not peer-reviewed
Language:English
Year of publication:2022
First Page:152
Faculty/Chair:Fakultät 2 Umwelt und Naturwissenschaften / FG Geopedologie und Landschaftsentwicklung
Einverstanden ✔
Diese Webseite verwendet technisch erforderliche Session-Cookies. Durch die weitere Nutzung der Webseite stimmen Sie diesem zu. Unsere Datenschutzerklärung finden Sie hier.