 |
| Titel |
Use of near infrared reflectance spectroscopy (NIRS) for determining soil content in chlordecone (organochlorine pesticide) |
| VerfasserIn |
Bernard Barthes, Didier Brunet, Magalie Lesueur-Jannoyer, Raphaël Achard, Luc Rangon, Thierry Woignier |
| Konferenz |
EGU General Assembly 2011
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 13 (2011) |
| Datensatznummer |
250048797
|
|
|
|
|
|
| Zusammenfassung |
| Chlordecone is a toxic organochlorine insecticide that was used in banana plantations until
1993 in the French West Indies. Now previously contaminated soils have become new
sources of contamination for natural water and cultivated roots and tubers, and represent an
important concern of public health, at last in the French West Indies. The conventional
method for analyzing chlordecone in soils involves extraction with two solvents and
quantification by gas chromatography-mass spectroscopy (GC-MS), which has a low
detection limit but is expensive, and the whole process takes several hours. This does not
easily allow analyzing large amounts of samples such as those required for characterizing soil
contamination and its variations on vast areas. Near infrared reflectance spectroscopy
(NIRS) has been reported to allow accurate prediction of diverse soil properties.
The present study aimed at assessing the potential of NIRS for characterizing soil
contamination by chlordecone over a set of 236Â samples collected at 0-30 and
30-60Â cm depths in Andosols (109 samples), Nitisols (72Â samples) and Ferralsols
(55Â samples) from Martinique. Over the set, chlordecone content ranged from 0 to
20 mg kg-1.
Using partial least square regression, chlordecone content determined through GC-MS
could be correctly predicted by NIRS (Q2Â =Â 0.73, Rval2Â =Â 0.82, RPDÂ =Â 2.2 for the total
set), especially for samples with chlordecone content  12 mg kg-1 (which is huge), suggesting under-prediction by NIRS. Nevertheless, when
the set was divided into three or four classes of chlordecone content (Â 10, or
 15 mg kg-1, respectively), contamination class was correctly
predicted for ca. 80% samples.
In addition, looking at B-coefficients (i.e. regression coefficients in the equation that
expresses chlordecone content as a function of absorbance at every wavelength) indicated that
wavelengths around 1830-1840Â nm contributed heavily to the prediction of chlordecone
content by NIRS. The 1860Â nm region has been attributed to the absorption due to the sixth
overtone of C–Cl bond stretching, which could be linked to the fact that the molecule of
chlordecone includes ten C–Cl bonds (its formulae is C10Cl10O). This strongly
suggests that chlordecone content was directly predicted by NIRS, though it was very
low.
Thus NIRS could be considered a time- and cost-effective method for characterizing soil
contamination by chlordecone in the soils from Martinique. |
|
|
|
|
|
|