Accurate reconstruction of ancient regional and global climates using oxygen isotope ratios (δ¹⁸O) from fossil shells and carbonate rocks requires determining the ancient geographic locations of samples. However, different Global Plate Models (GPMs) – which reconstruct past continental positions – often produce conflicting estimates of these locations. These discrepancies in latitude and longitude become larger for older time periods, potentially affecting climate reconstructions such as latitudinal sea surface temperature (SST) gradients. Because global mean surface temperatures (GMSTs) are calculated using SST data weighted by latitude, differences between GPMs could also impact GMST estimates.
In this study, we evaluate how uncertainties in Early Paleozoic climate reconstructions arise from variations in GPMs, using a massive Ordovician δ¹⁸O dataset from carbonate rocks, brachiopods, and conodonts. By applying four publicly available GPMs to locate these samples in their ancient geographic context, we measure differences in both regional and global temperature estimates. Our results reveal temperature differences as large as ~9.5°C regionally and ~10°C globally, directly tied to inconsistencies in GPM-based paleogeographic reconstructions. These variations highlight that the choice of plate model can introduce uncertainties comparable to the climate signals being studied. Therefore, paleoclimate studies should consider the uncertainties in ancient geographic reconstructions to improve the reliability of climate interpretations.

Palaeoclimate, Ordovician, Global Plate Models, Palaeogeography