HORSE: A NEW QUANTITATIVE STRATIGRAPHIC METHOD THAT HARMONIZES REGIONAL AND GLOBAL STRATIGRAPHIC RECORDS FOR GEOLOGICAL TIMESCALE AND EVOLUTIONARY STUDY

A uniform and high-resolution geologic timescale is essential for discovering macroevolutionary patterns and their drivers on regional and global scales. The discipline of quantitative stratigraphy combines stratigraphy, statistics, and computer science to bring stratigraphic records into an optimized timescale model that allows them to be correlated. Graphic Correlation (GC) and Constrained Optimization (CONOP), both of which sequence geologic events, can generate a global composite that includes all events and minimizes inconsistencies among those records. On the other hand, Horizon Annealing (HA) follows to the principles of GC and CONOP, but attempts to sequence stratigraphic "horizons" while preserving local details. Here we report on the HORizon SEquencing (HORSE) method, which generalizes the idea of HA with an implementation of parallel computing and genetic algorithms, enabling its application to massive data. For empirical assessment, we employ three datasets of different scales. Among the four methods, HORSE greatly outperforms HA in computation efficiency and shows remarkable similarities to CONOP in terms of global event sequencing, while preserving the sophisticated local information (e.g., local first appearance and last appearance of taxon) for paleobiogeographic and paleoecological studies, which is beyond the capabilities of CONOP.