MOD-AGE: An age-depth model construction algorithm

We present MOD-AGE, a new system for chronology construction. MOD-AGE can be used for profiles that have been dated using other techniques. The system uses three basic measurements for input data: activities, the atomic ratio or age, and the depth measurement. Based on the measurement results (probability distributions), MOD-AGE estimates the confidence bands of the age-depth model and uses nonparametric methods to avoid the consequences of failure to meet assumptions. Using the normal distribution of the measured data (activities, atomic ratios, depths, etc.), we applied a Monte Carlo simulation to the model age and depth values. Several fitting models could be applied to estimate the relationships. Following multiple tests, we chose the locally weighted scatterplot smoothing (LOESS) method. The stratigraphic correction procedure applied in the MOD-AGE program uses a probability calculus, which assumes that all of the sample ages are correctly estimated. The probability age distribution from the samples was used to estimate the most probable sequence in accordance with the superposition rule. This study presents MOD-AGE as a tool for the chronology construction of speleothems that have been analyzed with the U-series method and compares MOD-AGE to the StalAge algorithm presented by Scholz and Hoffmann (2011). The MOD-AGE software was written in, made by Borland, RAD STUDIO 2007 and is available from the authors as freeware. (C) 2012 Elsevier B.V. All rights reserved.