Long-term low-level ingestion of cadmium (Cd) causes human health problems, and in Australia, vegetables supply ~40% of the Cd in the typical diet. Plants take up Cd from the soil; however, the uptake is poorly predicted by simple soil tests, such as the total concentration of Cd (Cdt). Therefore, a greater understanding of Cd behaviour in soils is needed to improve the prediction of Cd uptake by plants and open a new path to minimise the risks for human health. The objectives of the research in this thesis were to: identify key soil properties affecting Cd behaviour, identify/develop selective methods to measure them, and to formulate a conceptual model of Cd partitioning. These objectives were based on the hypothesis that empirical modelling informed by a better understanding of Cd chemistry would accurately describe Cd partitioning in soil. To test the hypothesis, the key properties were measured on soils from the peri-urban fringe of Greater Sydney (n = 41) and a series of models of increasing complexity were fitted to the data. A model with three explanatory variables— log10 Cdt, pH and log10 ECEC (effective cation exchange capacity)—explained 94.6% of variation in log10 CdCa (the concentration of Cd in solution in a suspension of soil in 10 mM CaCl2), which strongly supported the hypothesis. The study also indicated that the explanatory variables, Cdt, pH and ECEC, may describe Cd behaviour in many soils, and that for these general models, partition coefficients, such as log10 (Cdt/CdCa), are unsuitable dependent variables. The preceding model used Cdt as an explanatory variable, notwithstanding that labile Cd (CdE) was mechanistically preferable. However, CdE can only be measured using isotopic techniques: a requirement that has constrained the evaluation of CdE as an index of Cd behaviour and bioavailability. Therefore, a simple proxy measure of CdE was investigated. The literature indicated that solutions of chloride salts might selectively extract CdE, and Cd extracted into 1 M NH4Cl (CdNH4Cl) was compared with CdE measured by stable isotope dilution ICPMS. For 23 soils from the partitioning study, 1 M NH4Cl failed to completely extract CdE, unless the pH was less than 5. The cause(s) of this effect will be investigated with the aim of developing a universally applicable measure of CdE that does not require isotopic measurements. All models of Cd uptake by plants rely on soil properties measured on homogenised samples, although the distribution and bioavailability of Cd vary spatially in the field. Were such variability to increase at the micro-scale, its effects could erode the accuracy with which models could predict Cd behaviour and uptake. Consequently, I tested whether the distribution of Cd could be mapped by using synchrotron micro-x-ray fluorescence spectroscopy (micro- XRFS): the most sensitive method of observation. The soils examined contained 0.3–6.4 mg Cd/kg, i.e. were typical agricultural soils, and one was spiked to ~100 mg Cd/kg. Micro-XRFS mapped the Cd in the spiked soil, and in one particle in the other soils. For typical agricultural soils, the sensitivity realised in this study would have been sufficient to characterise the average Cd binding site, but fell at least 10-fold below that needed to map the Cd distribution in them. The research satisfied the objectives, advanced knowledge of Cd behaviour in soils, and provided new research leads. These leads include the possibility of developing general models of Cd partitioning in soils, derivatives of which may predict Cd uptake by plants. The accuracy of these models may be strengthened by the use of CdE as an explanatory variable, but may be weakened by the effects of in situ variation in the distribution of Cd. The benefits to human health of agricultural practices that decrease dietary Cd justify continuation of research to develop models that accurately predict Cd uptake by plants. / Doctor of Philosophy (PhD)
Identifer | oai:union.ndltd.org:ADTP/235516 |
Date | January 2008 |
Creators | Milham, Paul J., University of Western Sydney, College of Health and Science, Centre for Plant and Food Science |
Source Sets | Australiasian Digital Theses Program |
Language | English |
Detected Language | English |
Page generated in 0.0018 seconds