Catalog Home Page

Quantification and characterisation of carbon in deep kaolinitic regoliths of South-Western Australia

Sangmanee, Podjanee (2016) Quantification and characterisation of carbon in deep kaolinitic regoliths of South-Western Australia. PhD thesis, Murdoch University.

[img]
Preview
PDF - Whole Thesis
Download (2MB) | Preview

Abstract

Greenhouse gas mitigation is a key element of global climate change agreements, and soil carbon is a major component of these strategies. However, the estimation of soil carbon is presently limited to depths of 0.3 m despite some soil profiles containing carbon at greater depths. Low concentrations (0.01 - 0.5% TOC) of soil carbon remain even after deforestation in the last 80 years, in deep kaolinitic soil profiles in south-western Australia but the form, distribution and stability of this has not been reported. A similar situation pertains elsewhere in the world.

Therefore, using kaolin-based mixtures and soil profiles from this region, the main aims of this thesis were to (i) expand the methods used for deep soil carbon quantification (Chapters 3, 4) and (ii) characterise carbon in deep soil (Chapters 4, 5, 6). This laboratory-based study employed standard samples (such as lignin, humic acid, cellulose and chitin mixed with kaolinite) and their combinations with variation in concentrations (0.008 - 11.55% TOC) analysed in parallel with field samples.

Quantitatively, the limit of detection (LOD) and limit of quantification (LOQ) of the Walkley-Black method (0.015 and 0.050% TOC, respectively) were approximately five times lower than the Heanes method (0.085 and 0.281% TOC, respectively) evaluated using pure kaolinite as background. Based on the LOQ, the Walkley-Black method was slightly superior to the Heanes method.

Therefore, the former method was further evaluated against dry combustion (Leco) using 94 calibration and 27 validation samples from deep soils (1 - 35 m depth) with a concentration range of 0.01 - 0.536% TOC. The predictive equation [TOC (%)]actual = 1.66[TOC (%)]WB + 0.018 (R2 = 0.91) obtained from the validation set agreed well with the benchmark dry combustion values (RMSE = 0.017).

A model for quantification of deep soil carbon using near infrared spectroscopy (NIR) was tested for sensitivity with standard samples in the concentration range 0.00 - 1.50% TOC. Positive prediction values were observed when employing square root of %TOC as input data. Then, models were developed from the 121 field soil samples previously used in the wet digestion analysis. The first derivative and standard normal variation (SNV) coupled with exclusion of bands in the range 5600 - 5000 cm-1 were suitable pre-processing approaches which gave an LOD of 0.001% TOC, with a very strong correlation (R2 = 0.98, RMSE = 0.0.32) but less accurate compared to the Walkley-Black method.

For characterisation, the ability of mid infrared spectroscopy (MIR) to characterise carbon in small concentrations was determined. Diffuse Reflectance Infrared Fourier Transform spectroscopy (DRIFT) had superior sensitivity to attenuated total reflectance spectroscopy (ATR). The best LODs were achieved when using kaolinite as a background. The LODs of DRIFT were 1.92% TOC for lignin or humic acid and 1.00% TOC for cellulose or chitin. However, key lignin and humic acid bands were concealed when mixed with cellulose and chitin at the same proportion. The identification of specific carbon structures from the mid infrared region was difficult in a mixture of several carbon types due to peaks overlapping. Therefore, qualification of deep soil samples was hampered by this method.

On the other hand, gas chromatography and mass spectrometry (GC/MS) was a promising approach for deep soil carbon. The concentration of residual carbon, potential carbon derived from exogeneous organic matter, in the form of low molecular weight compounds (LMWCs) was determined to be in the range 3.15 - 14.27 μg C g soil-1. Three compound classes were typically observed from samples across three different field locations: 1) terpenes, 2) fatty acids, amides and alcohols, and 3) plant steroids; indicating the influence of above ground input and roots of the past and present vegetation. In conclusion, (Z)-docos-13-enamide and bis(6- methylheptyl) phthalate were the main components throughout the soil profiles representing 53 - 81% of the LMWC, particularly at depths of 18 - 19 m.

Pyrolysis and off-line thermochemolysis using tetramethylammonium hydroxide (TMAH) were employed to characterise macro organic carbon (MOC) in deep soil to a depth of 29 m. Pyrolysis using the temperature range 250 - 600 ºC was suitable to characterise all laboratory standard samples at 3.85% TOC. For TMAH thermochemolysis, pre-concentration of the sample with 0.1 M NaOH was required before analysis. Unfortunately, the method was limited to characterisation of soil at 0 - 0.1 m depth where biomarkers such as lignin, polysaccharides, proteins and terpenes were present. Distribution of carbon species throughout the profile was revealed by pyrolysis. The coincident evidence from pyrolysis and thermochemolysis implied influences of vegetation, fire events and traces of microorganisms at 0 - 0.1 m depth, while lignin compounds were detected consistently down to 29 m by the pyrolysis method. It is concluded that MOC occurs in multiple chemical forms in surface soil but only occurs in lignin derived forms in deep soil.

A key issue for global mitigation studies is how deep soil carbon will respond to both land-use change (deforestation, reforestation) and climate change itself. The studies presented in this thesis provide a suite of methods suitable for the quantification and characterisation of deep soil carbon in kaolinitic regoliths and these can be extended to deep soils in other parts of the world. Identification of LMWC and MOC shed some light on sources and distributions of carbon throughout a soil profile.

Publication Type: Thesis (PhD)
Murdoch Affiliation: School of Veterinary and Life Sciences
Supervisor: Dell, Bernard
URI: http://researchrepository.murdoch.edu.au/id/eprint/35708
Item Control Page Item Control Page

Downloads

Downloads per month over past year