The objective of this study is to quantify the impact of freshwater stratification on the vertical gradients of partial pressure of CO 2 ( p CO 2 ) and estimates of air-sea CO 2 exchange in Hudson Bay during peak sea-ice melt and river runoff. During the spring of 2018, we sampled water in Hudson Bay and Hudson Strait for dissolved inorganic carbon, total alkalinity, salinity, the oxygen stable isotope ratio in the water (δ 18 O), and other ancillary data. The coastal domain and regions close to the ice edge had significant vertical concentration gradients of p CO 2 across the top meters of the ocean due to the presence of a stratified fresh layer at the surface. The p CO 2 and salinity in the central (where sea-ice melt was significant) and the southeast (where river runoff and sea-ice melt were significant) side of the bay generally increased with depth, with average gradients of 4.5 μatm m –1 and 0.5 m –1 , respectively. Ignoring these gradients causes a significant error in calculating air-sea CO 2 fluxes, especially when using shipboard underway systems that measure p CO 2 at several meters below the sea surface. We found that the oceanic CO 2 sink in Hudson Bay is underestimated by approximately 50% if underway p CO 2 system measurements are used without correction. However, we observed that these gradients do not persist for more than 5 weeks following ice melt. We have derived a linear correction for underway p CO 2 measurements to account for freshwater stratification during periods of 1–5 weeks after ice breakup. Given the lack of measurements in stratified Arctic waters, our results provide a road map to better estimates of the important role of these regions in global carbon cycles.