The North American Midcontinent Sea (NAMS) covered a vast area during Late Pennsylvanian glacio-eustatic highstands, providing a laboratory for analysis of the internal watermass dynamics of large cratonic seas (of which few exist in the Recent). In this study, a novel proxy, copper (Cu) isotopes, was used to gain a better understanding of NAMS watermass dynamics. We analyzed the 63-cm-thick Stark Shale (Dennis Formation, Missourian Stage, Upper Pennsylvanian) in the Iowa Riverton core (IRC) at a centimeter scale in order to reconstruct secular variations in Cu-isotopic compositions (δ65Cu) and other geochemical proxies. The gray shale facies yielded δ65Cu of +0.02 ± 0.06‰ (2σ, n = 14), similar to detrital Cu in modern marine sediments (δ65Cu = +0.08 ± 0.20‰; 2σ, n = 42). In contrast, the black shale facies, in which the proportion of authigenic Cu hosted by the organic fraction is 50–100% of total Cu content, exhibits heavier δ65Cuauth values, mostly between +0.09 and +0.43‰. In modern marine systems, one of the main processes leading to 65Cu-enriched sediment compositions is adsorption of aqueous Cu onto Fe-Mn particulates. The black shale facies of the Stark Shale exhibits four peaks in Mo/U ratios that are indicative of an active particulate shuttle because of the tendency of Mo to adsorb more strongly than U to Fe-Mn particulates. The Sr/Ba ratio, a paleosalinity proxy, shows correlated variations, with high values (indicative of more fully marine conditions) linked to low Mo/U and light δ65Cu, and low values (indicative of brackish conditions) linked to high Mo/U and heavy δ65Cu. These considerations suggest that the flux of isotopically heavy Cu to the sediment–water interface via Fe-Mn cycling was enhanced during brackish intervals, with subsequent reductive dissolution of Fe-Mn particulates allowing transfer of hydrogenous Cu to the organic fraction of the sediment. Control of δ65Cu by Fe-Mn cycling is supported by the results of a sequential extraction experiment showing that organic δ65Cu is positively correlated with Mo/U. These findings provide evidence of large fluctuations in watermass salinity and sea-level elevation within the NAMS at timescales of ∼104 yr during core shale deposition, as previously inferred but not conclusively demonstrated until now. The dynamics of particulate shuttle operation suggest that sea-level fluctuations must have had amplitudes of at least a few tens of meters. These fluctuations ended during the final stages of black shale deposition, probably owing to overall sea-level fall and permanent shallowing of the study site into the oxic, low-salinity surface layer of the NAMS.