The Arctic Ocean is an important sink for atmospheric CO2, and there is ongoing debate on whether seafloor seeps in the Arctic are a large source of CH4 to the atmosphere. The impact of warming waters, decreasing sea-ice extent and expanding marginal ice zones on Arctic air-sea gas exchange depends on the rate of gas transfer in the presence of sea ice. Sea ice acts as a near-impermeable lid to air-sea gas exchange, but is also hypothesised to enhance gas transfer rates through physical processes such as increased surface-ocean turbulence from ice-water shear and ice-edge form drag. The dependence of the gas transfer rate on sea-ice concentration remains uncertain due to a lack of in situ measurements. Here we present the first direct estimates of gas transfer rate in a wide range of Arctic sea-ice conditions. The estimates were derived from eddy covariance CO2 and CH4 fluxes, measured from the Swedish Icebreaker Oden during two expeditions: the 3-month duration Arctic Clouds in Summer Experiment (ACSE) in 2014, a component of the Swedish-Russian-US Arctic Ocean Investigation on Climate-Cryosphere-Carbon Interactions (SWERUS-C3) in the eastern Arctic Ocean shelf region; and the Arctic Ocean 2016 expedition to the high latitude Arctic Ocean. Initial CO2 results from ACSE showed that the gas transfer rate has a near-linear dependence on sea-ice concentration, and that some previous indirect measurements and modelling estimates overestimate gas transfer rates in sea-ice regions. This supports a linear sea-ice scaling approach for assessments of polar ocean carbon fluxes. Air-sea gas transfer model assumptions (e.g. Schmidt number dependence) will be examined using simultaneous CO2 and CH4 measurements, and observations in different ice conditions (e.g. summer melt, autumn freeze up, central Arctic and marginal ice zones) will be compared.