Openshaw S.J.

Palaeosecular variation observed in speleothems from western China and northern Spain. PhD thesis, University of Liverpool, UK.

This study has produced records of the palaeosecular variation (PSV) of the earth's magnetic field from Speleothems from China and Spain. The ultimate aim of this project was to produce contemporaneous PSV records which would show that Speleothems accurately record ambient geomagnetic field behaviour. From Sichuan Province, China, five Speleothems were collected of which four were studied for their records of PSV. Eight Spanish Speleothems from the Cantabrian coast were collected but their weak magnetisation allowed only one record of PSV to be produced.
All speleothem sub-samples were weakly magnetised and had, on average, initial intensities of <100 x 10-8 Am2kg-1. Despite this, the majority of sub-samples were stable during stepwise alternating-field and thermal demagnetisation and each displayed a single component of magnetisation after removal of any secondary overprints. Rock magnetic experiments were hampered by low mineral concentrations but suggested that the remanences of each speleothem were carried by a mixture of multi and single-domain (titano-) magnetite and also by haematite present in significant quantities. The primary method of remanence acquisition appeared to be a depositional remanence sourced from flooding. This was corroborated by a linear relationship between sub-sample intensities and weight % acid insoluble detritus.
A selection of sub-samples from each speleothem were dated using uranium-thorium disequilibrium and alpha spectrometry. For the majority of sub-samples the low concentrations of uranium, high levels of detrital contamination and initially low chemical yields raised the associated dating inaccuracies above the quoted level for alpha spectrometry of 5-10%. Two Spanish Speleothems had high uranium concentrations and little, or no, detrital contamination. Percent age errors of these Speleothems ranged from 1 to 6%. Comprehensive experiments on the efficiencies of three electrodeposition methods were also undertaken. The most efficient method was found to be a modified version of the Hallstadius method (Hallstadius, 1984), which consistently achieved chemical yields between 40 and 90% for uranium and thorium.
In order to correct more analytically for the presence of detrital contamination, the leachate/leachate method of Schwarcz and Latham (1989) was tested. The maximum likelihood estimation data treatment technique (Ludwig and Titterington, 1994) was used to calculate dates from these analyses. Tests on Mexican speleothem SSJ2 gave excellent results allowing a revised dating scheme to be adopted. Tests on some sub-samples from Chinese Speleothems were generally unsuccessful due to analytical errors.
The isotope 210Pb was used to date the top surface of one speleothem. A constant growth rate was inferred which was significantly less than that calculated from the 230Th - 234U dating method. This was thought to be due to the former techniques inability to resolve growth rates of periods of less than 200 years.
Despite the dating errors associated with each speleothem the records of PSV compare well with each other and with contemporaneous records from China, Japan and also the UK (for the Spanish record). In addition. agreement with PSV data modelled from observatory records suggested that westward drift of the non-dipole geomagnetic field was predominant during the past 10ka.