The isotope composition of seawater sulfate is an important tracer of sulfur, carbon, and oxygen cycles in Earth’s deep past. Carbonate-associated sulfate (CAS) extracted by acid digestion is widely used as a proxy for sulfate in paleo-seawater from which the carbonate minerals precipitated. Early and late diagenesis, weathering, and laboratory processing can in some cases compromise original seawater sulfate signals. Here, we report that extracted CAS can also be severely contaminated by recent atmospheric sulfate, especially when the sampled carbonates are from outcrops in arid to semi-arid climates or in heavily polluted regions. Our evidence comes from triple oxygen isotope compositions of sequentially extracted water-leachable sulfate and acid-leachable sulfate from carbonates of diverse ages from northwestern and north-central China and southwestern North America. Independent of the age of the rocks, almost all the water-leachable sulfates and half of the acid-leachable sulfates bear positive 17O anomalies, clearly distinguishable from those of typical seawater sulfate. Because secondary atmospheric sulfate (SAS) is the only source of sulfate known to bear positive 17O anomalies, we conclude that sulfate extracted from carbonate outcrops in these regions has a significant component of SAS. Because SAS generally has a much lower δ34S value than paleo-seawater sulfate, it could shift the δ34S of the extracted CAS to lower values and in some cases even lower than that of the co-occurring pyrite, i.e., the “super-heavy pyrite” enigma reported in geological records. Our findings call for a re-evaluation of many published, outcrop-based CAS data and conclusions.