Online Supplementary Tables

The relationship between icehouse climates and oceanic anoxic events remains controversial, particularly regarding if glacial conditions inhibit or promote widespread ocean anoxic conditions. To address this, we reconstruct redox variations through detailed analysis of pyrite framboids size distributions and morphologies from six sections spanning the uppermost Devonian to Middle Permian in the Baoshan Block, which located in the northern margin of Gondwana during the Late Paleozoic. Our research shows that the redox conditions during the Early Carboniferous can be compared with those in various regions worldwide, while its biotic composition, including biodiversity and abundance, is primarily correlated with those of Eurasia. On the contrary, the redox conditions became frequent anoxic/euxinic when the ice age began during the Visean and Ice Age peak during the Sakmarian, which were likely caused by glacio-eustatic sea-level fluctuations and intensified upwelling. Following the end of the glaciation (late Kungurian-Wordian), the Baoshan Block once again experienced predominantly stable dysoxia, coincided with the relatively abundant fauna. This study suggests that the icehouse climate, to some extent, promoted anoxic events in shallow marine settings.
A total of 210 rock samples were selected from fresh surfaces without veins, roots, or strongly weathered surfaces. The scanning electron microscopy (SEM) are employed to analyze pyrite framboids. The accuracy of this statistical analysis highly depends on the careful identification of crystal types and the number of measured pyrite framboids. A detailed identification of crystal morphology and compositional analysis by EDS are required before measurement since pyrite framboids may undergo post-depositional oxidation or overgrowths, and can be easily confused with euhedral pyrite or other metal crystals under low-magnification SEM observation. The polished surface (~ 1 cm2) of each sample needs to be fully observed and statistically analyzed. To ensure accuracy, each sample is scanned at a 100 μm scale, with measurements, imaging, and EDS analysis conducted within a <10 μm field of view. All framboids exposed on the sample surface were measured, with a minimum of 100 pyrite framboids measured per sample, as conditions allowed. Under the above experimental standards, the average observation time for a sample is approximately 2.5–3 hours.
The supplementary tables are used to comprehensively summarize key data (mean, median, minimum, and maximum diameters, standard deviation, number of framboids, and redox interpretation) derived from over 15,000 individual framboid measurements, providing a quantitative foundation for interpreting redox fluctuations. This dataset can be used to clarify the paleo-oceanic redox history during the Gondwanan glaciation, providing a reference for investigating paleo-oceanic redox environments during the transition between icehouse and greenhouse climates.