Received: 27 Apr 2018 – Discussion started: 01 Jun 2018
Abstract. In this study, we propose a statistical method to validate sea-level reconstructions using geological records known as sea-level indicators (SLIs). SLIs are often the only available data to retrace late-glacial relative sea level (RSL). Determining the RSL from SLI height is not straight forward, the elevation at which an SLI was found usually does not represent the past RSL. In contrast, it has to be related to past RSL by investigating sample’s type, habitat and deposition conditions. For instance, water distribution at which a specific specimen is found today can be related to the indicator's depositional height range. Furthermore, the precision of dating varies between geological samples, and, in case of radiocarbon dating, the age has to be calibrated using a non-linear calibration curve. To avoid an a-priori assumption like normal-distributed uncertainties, we define likelihood functions which take into account the indicative meaning’s available error information and calibration statistics represented by joint probabilities. For this conceptional study, we restrict ourselves to one type of indicators, shallow-water shells, which are usually considered as low-grade samples giving only a lower limit of former sea level, as the depth range in which they live spreads over several tens of meters, and does not follow a normal distribution. The presented method is aimed to serve as a strategy for glacial isostatic adjustment reconstructions, in this case for the German Paleo-Climate Modelling Initiative PalMod (https://www.palmod.de/en) and by extending it to other SLI types.
How to cite. Latinović, M., Klemann, V., Irrgang, C., Bagge, M., Specht, S., and Thomas, M.: A statistical method to validate reconstructions of late-glacial relative sea level – Application to shallow water shells rated as low-grade sea-level indicators, Clim. Past Discuss. [preprint], https://doi.org/10.5194/cp-2018-50, 2018.
By using geological samples we are trying to validate the models that are reconstructing the sea level in the past 20 000 years. We applied proposed statistical method using 4 types of shells that were found in the area of the Hudson Bay on 140 members of model ensemble. After the comparison of the the results with studies from this area, we concluded that the method is suitable for validation of model ensemble based sea-level change caused by land movement of the Earth due to ice-age burden.
By using geological samples we are trying to validate the models that are reconstructing the sea...