The statistical discrimination between coversan areas in the Netherlands

Abstract

In 1968 Maarleveld published the results of a research of the coversand area in The Netherlands, based on geomorphological evidence and on the evaluation of macroscopical characteristics of the coversands mainly in the 105-75 micron fraction, resulting in a subdivision into 29 subareas. Each subarea could be described in terms of specified amounts (per mil values) of the three characteristics: white, black and green grains. In this study the analytical results of these principal characteristics are consider ed in a mote objective way, in order to arrive at a numerical comparison for all possible pairs of subareas. For any pair this comparison may be expressed as the sample-size N required to detect a difference as large as the one stated between the respective sample means. The N-numbers - which may be considered as difference- or similarity-coefficients have a practical significance in that they provide a measure for the discrimination between any pair of areas as to the characteristic in question: large N-numbers stand for a high degree of similarity, whereas small N-numbers mean that the areas under consideration are very different and consequently may be distinguished with relative ease. The relation between the N-number and the statistical parameters of the characteristics is discussed. A way is indicated how to arrive at defining areas such that they fit classes with a maximum contrast between the analytical data.