Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2023 Oct 2;195(11):1261.
doi: 10.1007/s10661-023-11837-y.

Trace metal content prediction along an AMD (acid mine drainage)-contaminated stream draining a coal mine using VNIR-SWIR spectroscopy

Affiliations

Trace metal content prediction along an AMD (acid mine drainage)-contaminated stream draining a coal mine using VNIR-SWIR spectroscopy

Jamie-Leigh Robin Abrahams et al. Environ Monit Assess. .

Abstract

The current study investigated the use of VNIR-SWIR (visible/near infrared to short-wavelength infrared: 400-2500 nm) spectroscopy for predicting trace metals in overbank sediments collected in the study site. Here, we (i) derived spectral absorption feature parameters (SAFPs) from measured ground spectra for correlation with trace metal (Pb, Cd, As, and Cu) contents in overbank sediments, (ii) built univariate regression models to predict trace metal concentrations using the SAFPs, and (iii) evaluated the predictive capacities of the regression models. The derived SAFPs associated with goethite in overbank sediments were Depth433b, Asym433b, and Width433b, and those associated with kaolinite in overbank sediments were Depth1366b, Asym1366b, Width1366b, Depth2208b, Asym2208b, and Width2208b. Cadmium in the overbank sediments showed the strongest correlations with the goethite-related SAFPs, whereas Pb, As, and Cu showed strong correlations with goethite- and kaolinite-related SAFPs. The best predictive models were obtained for Cu (R2 = 0.73, SEE = 0.15) and Pb (R2 = 0.73, SEE = 0.21), while weaker models were obtained for As (R2 = 0.46, SEE = 0.31) and Cd (R2 = 0.17, SEE = 0.81). The results suggest that trace metals can be predicted indirectly using the SAFPs associated with goethite and kaolinite. This is an important benefit of VNIR-SWIR spectroscopy considering the difficulty in analyzing "trace" metal concentrations, on large scales, using conventional geochemical methods.

Keywords: Floodplain sediments; Heavy metal; Predictive modeling; Reflectance spectroscopy; Remote sensing.

PubMed Disclaimer

Conflict of interest statement

The authors declare no competing interests.

Figures

Fig. 1
Fig. 1
Localities (red dots) for overbank sediment sampling and ground hyperspectral data collection along the Blesbokspruit River, Mpumalanga, South Africa. Overbank sediment samples were collected at two sites roughly 5 m apart at each of the six different localities. Flow direction is indicated by the black, dashed arrow. Also shown is a wetland (green dash lines) and acid ponds (yellow rectangle)
Fig. 2
Fig. 2
Median raw ground-derived spectra of overbank sediments at each of the 12 sample sites (Fig. 1) along the Blesbokspruit River, highlighting the wavelengths associated with atmospheric water
Fig. 3
Fig. 3
Continuum-removed ground spectra of overbank sediments at each of the 12 sample sites (Fig. 1) along the Blesbokspruit River, with wavelengths related to atmospheric water removed and the spectral subsets used to derive the SAFPs enclosed in black rectangles
Fig. 4
Fig. 4
Definition of absorption-band position, depth, and asymmetry (modified after Van der Meer, 1999)
Fig. 5
Fig. 5
Measured vs. predicted concentrations of a Pb, b Cd, c As, and d Cu in overbank sediments in the study area (red dots with sample identification). The 1:1 control lines are shown in solid black and the regression lines are shown in dotted grey line

References

    1. ASD Inc. (2009). FieldSpec®3 user manual. ASD Document 600540 Rev. H.
    1. Austin PC, Steyerberg EW. The number of subjects per variable required in linear regression analyses. Journal of Clinical Epidemiology. 2015;68(6):627–636. doi: 10.1016/j.jclinepi.2014.12.014. - DOI - PubMed
    1. Balsam WL, Wolhart R. Sediment dispersal in the Argentine Basin: Evidence from visible light spectra. Deep Sea Research Part II: Topical Studies in Oceanography. 1993;40:1001–1031. doi: 10.1016/0967-0645(93)90046-P. - DOI
    1. Baran A, Tarnawski M. Assessment of heavy metals mobility and toxicity in contaminated sediments by sequential extraction and a battery of bioassays. Ecotoxicology. 2015;24(6):1279–1293. doi: 10.1007/s10646-015-1499-4. - DOI - PMC - PubMed
    1. Bell FG, Bullock SET, Halbich TFJ, Lindsay P. Environmental impacts associated with an abandoned mine in the Witbank Coalfield, South Africa. International Journal of Coal Geology. 2001;45(2-3):195–216. doi: 10.1016/S0166-5162(00)00033-1. - DOI