A Trend-Surface Analysis Trial for the Planation Surfaces of North Cardiganshire

Abstract

An objective study is made of the landscape of the Aberystwyth hinterland. The results show that, of the earlier studies, E. H. Brown's division of the area into three was the most rational. A CONSIDERABLE amount of attention has been given to the origin and form of the relief of north Cardiganshire during the last hundred years or so. The different workers1 have distinguished a variety of plateau surfaces, usually at dissimilar levels and even the idea of a single concave surface has been put forward.2 There is the same diversity of opinion regarding the origin and age of the landscape.3 By distinguishing three surfaces, E. H. Brown4 has advanced what is perhaps the most rational solution. These are the Coastal Plateau, the High Plateau surrounding the Plynlimon massif and the intervening Middle Plateau (Fig. i). The Middle Plateau is breached by the gorge of the River Ystwyth in the south; but it also appears in the south-west beyond the River Wyre, resulting in the attenuation of the Coastal Plateau. There remains some doubt over the validity of this three-fold division, one reason being that the Middle Plateau has been considered to lack the homogeneity that Brown attributed to it.5 Trendsurface analysis offers a convenient and objective method for examining the landscape of the Aberystwyth hinterland and determining the number of surfaces that comprise it. The technique of trend-surface analysis has been developed very largely by the geologist, but it is being used increasingly by a wider range of earth scientists. There is an extensive literature on its application, yet although its use in geomorphology was suggested by H. Svensson some years ago,6 the method has been employed infrequently for separating planation surfaces.7 It provides a simple quantitative means for testing the validity and relevance of multiple-surface hypotheses and it can, of course, reveal surfaces that have hitherto been unrecognized. Hence, as R. J. Chorley and P. Haggett suggest,8 for areas such as Wales where the planation surfaces that have been recognized are both numerous and disputed, trend-surface analysis presents a judicious method of reaching a final solution. Trend-surface analysis is a form of multiple regression analysis applied in three dimensions. The trend surface demonstrates the systematic changes in the distribution of the independent variable, while deviations between the observed values and the predicted surface indicate the small-scale fluctuations that are superimposed on the broader pattern. The goodness-of-fit of the surfaces can be assessed by several statistical tests, but a study of the distribution of the deviations (or residuals) can also be revealing. Using the I :63 360 Ordnance Survey Sheet 127 (Aberystwyth), all the summit heights whose measurements were shown, and all the estimates that could be made of the highest