Abstract
Coral microatolls track relative sea level (RSL) as they grow and can inform reconstructions of past RSL change, beyond the duration of instrumental observations. However, direct observation of coral growth history requires slabbing cross sections, which is destructive and logistically challenging. To address this complication, recent studies have explored analysis of coral surface morphologies using non-invasive methods. However, these methods are limited; they cannot directly access key features inside the corals that are indicative of past RSL change events. To overcome this limitation, we devised a method to formulate the growth pattern of massive corals through iteration to simulate coral growth in response to RSL changes. Additionally, we developed an alternative model that analytically approximates the same coral growth, suitable for stochastic analyses. To test the effectiveness of these models, we utilized observations of coral microatolls from Sumatra that emerged during the 2004 MW 9.1 earthquake. We first compared actual coral cross sections with coral geometries modeled using the respective RSL histories interpreted for those corals in previous studies. Next, we attempted to reconstruct past RSL history using only coral surface profiles with a stochastic inversion. The forward simulation presented an approximate reproduction of the actual corals, although there were discrepancies due to localized variations in growth rate. The inversion study approximately reproduced the previous interpretation of RSL history; however, the effect of nonuniform growth speed was not negligible. While these results highlight the necessity of a comprehensive understanding of coral growth for a more robust interpretation, the simulators developed in this study are powerful tools for improving this understanding.
Keywords
Bayesian inversion, Coral Microatolls, Holocene, modeling, Relative sea-level change