Geometric morphometrics and geological shape-classification systems

N. MacLeod
DOI: https://doi.org/10.1016/S0012-8252(02)00068-5
IF: 12.038
2002-01-01
Earth-Science Reviews
Abstract:Many areas of geological inquiry involve the description and/or comparison of shapes. While various morphometric tools have long been available to facilitate these types of comparisons, by far the most common approach to such form-classification has been via the creation of a semiquantitative scale of morphological exemplars, type specimens, etc. to which unknown structures, objects, or specimens can be compared. Such form-scales are ubiquitous—either in terms of text-based descriptions or illustration sets—throughout the geological literature. However, students, and even experienced geologists, often have difficulty using such scales and achieving consistent results. Investigations of three such scales drawn from the fields of sedimentology, paleontology, and geomorphology using the analytical tools of geometric morphometrics suggests that one reason for this difficulty is that the exemplars drawn from sets of real objects (1) often exhibit shape differences other than those under nominal consideration or (2) are used to represents object classes whose boundaries are insufficiently documented or described. Herein, strategies are developed that employ the ordination and modeling capabilities of eigenshape analysis to correct these deficiencies and devise sets of new, more representative, and easier to use shape-classification systems. By employing these approaches, augmented where necessary with formal statistical analyses, geologists can improve the sophistication, accuracy, and reproducibility of their morphological inferences. In doing this, they will also improve the reliability of their hypotheses tests.
What problem does this paper attempt to address?