This work investigates how internal wave-driven turbulence varies in time, from hourly to yearly timescales, and in space, across two distinct regions of the Arctic Ocean. We apply a shear-based fine-scale parameterization to mooring records in Nares Strait and on the Beaufort Sea shelf-slope that sampled the upper stratified water column every 30-45 min and span 2003-2006 and 2003-2004, respectively. In doing so, we generate over 600,000 estimates of the internal wave-driven dissipation rate. These estimates exhibit large temporal variability in both regions, spanning over 3 orders of magnitude. Despite these wide ranges, we find distinct distributions at each site. In Nares Strait, the time series of dissipation shows systematic variation at tidal frequencies, and tidal forcing appears to influence dissipation more strongly than winds, sea ice, and stratification on daily timescales. On longer timescales, dissipation exhibits a weak seasonal cycle, being elevated when the stratification is high and during the ice melt season. In the Beaufort Sea, we detect no dominant timescales or significant relationships with forcing metrics, but note that the dissipation rate is typically 2 orders of magnitude lower than that in Nares Strait. This region is characterized as being in a turbulent mixing regime for only 2% of the record, compared to 73% of the Nares Strait record, implying that turbulence here is rarely energetic enough relative to the stratification to drive a turbulent heat flux. Inferred Beaufort Sea heat fluxes are an order of magnitude lower than the O(1) W m −2 average value found in Nares Strait. Plain Language Summary Arctic Ocean mixing rates and how they vary in space and time have important consequences for the transport of heat, salt, and nutrients. However, the tasks of understanding how mixing changes water properties and of predicting future changes to Arctic Ocean mixing rates remain challenging, particularly given that most observations used to study mixing in the Arctic Ocean to date are geographically and temporally limited. In this study, we infer mixing rates by using an estimation technique that relates ocean turbulence to properties of internal waves in the ocean interior. We apply this technique to highly resolved, multiyear measurements obtained from two distinct regions of the Arctic Ocean. Using these estimates, we characterize the statistical distributions of mixing metrics and find strong regional differences between the two study sites. We find that one region is highly energetic and displays systematic patterns in mixing intensity at tidal and seasonal timescales. Turbulent energy at the other site is 2 orders of magnitude weaker, and active turbulent mixing occurs extremely infrequently. These results help to put previous mixing observations into a more informed context and provide insight into the underlying causes of Arctic Ocean mixing patterns.