Back to Search Start Over

Chemometric models for quantifying soil properties from its alkane patterns

Authors :
Jiménez-González, Marco A.
Almendros Martín, Gonzalo
Álvarez, Ana María
Carral, Pilar
Rosa Arranz, José M. de la
Sanz Perucha, Jesús
Jiménez-González, Marco A.
Almendros Martín, Gonzalo
Álvarez, Ana María
Carral, Pilar
Rosa Arranz, José M. de la
Sanz Perucha, Jesús
Publication Year :
2017

Abstract

The soil lipid fraction consists of a complex mixture of low-polarity compounds of microbial or plant origin [1]. Homologues series of alkanes are present in all soil types: long-chain homologues are frequently considered to derive from epicuticular waxes of vascular plants, whereas short-chain homologues are thought to derive mainly from microorganisms. In general the free lipids (i.e., directly extracted with organic solvents) amount to a little proportion of the soil organic matter (2¿ 50 g/kg), but may have a direct influence in soil processes viz, enhancing soil water repellency or promoting soil aggregation. On the other hand, alkanes as many other soil lipids, are also a valuable source of environmental information: In fact, homologous series of alkanes may vary in terms of the vegetation and microbial activity, but also reflect different soil abiotic and ecophysiological properties with a bearing on the selective accumulation of the individual alkane molecules. In this exploratory research we focus exclusively on the n-alkanes found in the free soil lipid fraction, which are studied by gas chromatography¿mass spectrometry (GC¿MS), using the characteristic ion trace at m/z 85 to identify and quantify the alkane series. For this study we collected 35 topsoil samples from Spanish soils (0¿10 cm depth) in contrasted environments. The lipids were Soxhletextracted for 12 h with petroleum ether (40¿60 ºC). Thehomologous series were considered in the range between C15 (pentadecane) to C35 (pentatriacontane). The proportions of the different homologues were used as descriptors to explore possible correlations with a large series of soil properties. Complementary chemometric methods were used to examine the information latent in the alkane patterns viz, partial least squares regression (PLS), multidimensional scaling (MDS) and principal component analysis (PCA). We found a large forecasting potential of the n-alkanes mainly as regards soil physical properties (i.e., a

Details

Database :
OAIster
Notes :
English
Publication Type :
Electronic Resource
Accession number :
edsoai.on1105214042
Document Type :
Electronic Resource