Revisiting the silicon isotopic signal of sponge skeletons and its implications

This study investigates the relationship between the silicon (Si) isotopic composition of sponge skeletal silica (δ30SibSi) and seawater characteristics in sponge habitats, specifically the concentration of dissolved silicic acid and its Si isotopic signature (δ30SidSi). Initially, these correlations were considered a promising calibration proxies for paleoceanographic reconstructions, but the incorporation of subsequent data points into the dataset over the past decade has highlighted complexities in how sponges fractionate silicon isotopes during silicification processes. We revisit the historical dataset, including a detailed examination of each datapoint to identify biases related to environmental, biological, and taxonomic factors. We also contribute new isotopic data obtained by multi-collector inductively coupled plasma mass spectrometer analysis, specifically targeting underrepresented low-silicic-acid environments. This revised dataset highlights that anomalies in the calibration, in particular species with fused skeletal frameworks, remain incongruous. We found that part of the problem is that the relationship between silicic acid concentration and δ30SibSi in the revised dataset of only Demospongiae follows a distinct, statistically robust, non-linear trend different from the weak, linear fit in Hexactinellida. Consequently, isotopic data from these two sponge classes should not be combined for calibration analysis, if possible. Yet, while the robust non-linear regression for only Demospongiae revitalizes the proxy, the relationship becomes asymptotic at silicic acid values above 200 μM, limiting its applicability to Cenozoic and Mesozoic conditions and excluding early Paleozoic scenarios with high concentrations of silicic acid. Practical recommendations for using and improving the proxy are discussed.