Large variability in planktic foraminiferal Mg/Ca and δ18O revealed by nearby surface sediments

Foraminiferal Mg/Ca and δ¹⁸O are widely used to reconstruct ocean temperatures. Comparisons between proxy-derived temperatures from surface sediments and modern gridded climatologies are commonly used to infer the recording season and habitat depth of planktic foraminifera, while down-core proxy variations are typically interpreted as reflecting past oceanographic changes at the site or region. Using these approaches requires the assumption that proxy-derived temperatures from a single site represent the mean hydrographic conditions of the corresponding spatial grid used in proxy–model or proxy–proxy comparisons. However, the extent to which this assumption holds across spatially distributed surface sediments remains poorly constrained. Sediment heterogeneity, sampling, and foraminiferal ecological processes could introduce additional variability into foraminiferal proxy data. To address these issues, here we estimated upper-ocean temperatures from the Mg/Ca ratio and δ¹⁸O of both surface and subsurface-dwelling foraminifera from multiple surface sediments within seven 1°×1° grids, which correspond to the typical spatial resolution of gridded climate fields, around the Okinawa Islands in the Northwest Pacific. The results suggest that the spread of Mg/Ca- and δ¹⁸O-derived temperatures within individual grid cells reaches up to ~4 °C, which is comparable to the typical glacial–interglacial temperature range in this region, despite the nearshore setting and lack of strong dynamic ocean processes. The Mg/Ca and δ¹⁸O-derived temperature variability differ among species, with subsurface dwellers exhibiting larger variability (~1.1 ºC, 1σ) than surface dwellers (~0.6 ºC, 1σ). To further characterize the contributions of individual processes to observed proxy variability, we used the forward model Sedproxy to simulate the variability induced by seasonal and depth occurrence of foraminifera. This variability is largely attributable to seasonal occurrence in surface-dwelling species, whereas in subsurface-dwelling species it cannot be explained by seasonality alone and likely also reflects variability in calcification depth within the upper thermocline, where temperatures change most rapidly with depth. In summary, our results attempt to quantify the contributions of ecological and sampling-related processes to proxy variability within a grid of nearby surface sediments. We therefore suggest that such variability provides an estimate of proxy uncertainty that should be taken into account in paleoceanographic reconstructions. While the magnitude may depend on regional setting, systematic assessments across regions and species are needed to better constrain proxy uncertainty and avoid over-interpreting proxy-derived temperature differences.