Abstract. Observational evidence of a quasi-quadrennial oscillation (QQO) in the polar mesosphere is presented based on the analysis of 24 years of hydroxyl (OH) nightglow rotational temperatures derived from scanning spectrometer observations above Davis research station, Antarctica (68∘ S, 78∘ E). After removal of the long-term trend and solar cycle response, the residual winter mean temperature variability contains an oscillation over an approximately 3.5–4.5-year cycle with a peak-to-peak amplitude of 3–4 K. Here we investigate this QQO feature in the context of the global temperature, pressure, wind, and surface fields using satellite, meteorological reanalysis, sea surface temperature, and sea ice concentration data sets in order to understand possible drivers of the signal. Specifically, correlation and composite analyses are made with data sets from the Microwave Limb Sounder on the Aura satellite (Aura/MLS v4.2) and the Sounding of the Atmosphere using Broadband Emission Radiometry instrument on the Thermosphere Ionosphere Mesosphere Energetics Dynamics satellite (TIMED/SABER v2.0), ERA5 reanalysis, the Extended Reconstructed Sea Surface Temperature (ERSST v5), and Optimum-Interpolation (OI v2) sea ice concentration. We find a significant anti-correlation between the QQO temperature and the meridional wind at 86 km altitude measured by a medium-frequency spaced antenna radar at Davis (R2∼0.516; poleward flow associated with warmer temperatures at ∼0.83±0.21 K (ms−1)−1). The QQO signal is also marginally correlated with vertical transport as determined from an evaluation of carbon monoxide (CO) concentrations in the mesosphere (sensitivity 0.73±0.45 K ppmv−1 CO, R2∼0.18). Together this relationship suggests that the QQO is plausibly linked to adiabatic heating and cooling driven by the meridional flow. The presence of quasi-stationary or persistent patterns in the ERA5 data geopotential anomaly and the meridional wind anomaly data during warm and cold phases of the QQO is consistent with tidal or planetary waves influencing its formation, which may act on the filtering of gravity waves to drive an adiabatic response in the mesosphere. The QQO signal plausibly arises from an ocean–atmosphere response, and appears to have a signature in Antarctic sea ice extent.