Regional geochemical mapping in Great Britain with particular references to sources of error.

This thesis describes the research and development of procedures for the preparation of regional geochemical maps of Great Britain. The maps are produced at the 1:250,000 scale and represent the distribution of approximately twenty elements of economic or environmental significance. They are based on the analysis of drainage samples but the application of alternative sample types to regional geochemical mapping in Great Britain is briefly discussed. The first part of the study describes an orientation study carried out in the northern Highlands of Scotland. The area was selected since it presents the most difficult problems for sampling while the natural variation in heavy metal concentration can be assessed as a basis for evaluating pollution in industrialized areas. Investigation of sampling and sample preparation methods indicates that -100 BSI mesh (150 um) material is the optimum size fraction for the regional survey (it provides adequate geochemical contrast; it includes all the size fractions in which metals are enriched; and it is available in amounts which are adequate for all analyses required). Grinding samples improves precision and reduces the preferential volatilization of metals co-precipitated with hydrous manganese and iron oxides when optical emission spectroscopic analysis is used. A study of the dispersion of metals in the drainage system of an area of Pb-Ba-Mo mineralization and an area of U-Pb mineralization respectively indicates that a sampling interval of 1 per 2 km2 based on second or third order streams is adequate to define anomalous metal concentrations and regional variation is represented on the same interval. Modifications to the sampling procedure to reduce the effects of the secondary environment are complicated by the primary association of Mn and Fe with mineralization. The ratio of analytical and sampling error for regional variance shows that procedural precision is adequate to prepare regional geochemical maps with the exception of zirconium and zinc at low concentrations. The second part of the thesis describes the procedures developed for the systematic determination of accuracy and precision in the data for the geochemical maps. Changes in accuracy are monitored by analysis of standards and the re-determination of sub-sets of samples using a neutron activation method. Samples are assigned random numbers in the field and are reordered sequentially prior to preparation and analysis in order to determine systematic errors within batches of samples. Sampling, sub-sampling and analytical precision are monitored using analysis of variance methods on the results obtained on duplicate samples and analyses. The system for the efficient processing of this information is described and examples of the different types of error presented. Finally the application of an accurate and precise instrumental neutron activation analytical method to the geochemical mapping programme is described. The method provides a reference analytical procedure for orientation, error control and follow-up. The concentration of several metals not readily amenable to standard multi-element methods such as Au, W, Se, Sb, As, Hf, Cd, Cs and several REE can also be determined by the method. Conclusions are made with particular reference to the geochemical maps for Mo and U in Sutherland and Caithness, Scotland.