Experiments on the reproducibility and reliability of lichenometric dates, Storbreen gletschervorfeld, Jotunheimen, Norway

Publisher Website
Google Scholar
Abstract
Matthews, J. A. Experiments on the reproducibility and reliability of lichenometric dates, Storbreen gletschervorfeld, Jotunheimen, Norway. Norsk geogr. Tidsskr. 29, 97–109. Experiments are made on the reproducibility of lichenometric dates and a method outlined for obtaining reliable lichenometric dates and a measure of their accuracy. The paper is based on field measurements of Rhizocarpon geographicum growing on 10 former ice margins, Storbreen gletschervorfeld, Jotunheimen, southern Norway. The whole population of largest lichens was measured on each margin and 144 lichenometry curves constructed from randomly sampled sub-sets of the data. The effect on predicted dates of varying the number of fixed points and varying the number of largest lichens per margin is analysed. Reproducibility of dates is measured by the dispersion of individual predictions about the mean prediction. Each mean prediction is the best estimate available of the true age of each margin. It is argued that ‘preferred predictions’ are, in this case, the mean prediction based on five largest lichens per margin and an overall mean of predictions based on one, five and ten largest lichens per margin. Use of families of lichenometry curves is reassessed by comparing mean predicted dates from 40 non-independent lichenometry curves with the ‘preferred predictions’. The ‘preferred predictions’ for the four margins of unknown age are: 1867–71 (M5), 1852–53 (M4), 1824–27 (M3), 1810 (M2). Previously published predicted dates, based on families of lichenometry curves and necessary for the establishment of an areal chronology for the Storbreen gletschervorfeld, are substantiated. Previously published results of lichenometric dating, based on single lichenometry curves, are criticised and widespread use of the principle of reproducibility is advocated.