Estimates of the recent mean and time varying water mass transformation rates associated with North Atlantic surface-forced overturning are presented. The estimates are derived from heat and freshwater surface fluxes and sea surface temperature fields from six atmospheric reanalyses-the Japanese 25-yr Reanalysis (JRA), the NCEP-NCAR reanalysis (NCEP1), the NCEP-U.S. Department of Energy (DOE) reanalysis (NCEP2), the European Centre for Medium-RangeWeather Forecasts (ECMWF) Interim Re-Analysis (ERA-I), the Climate Forecast System Reanalysis (CFSR), and the Modern-Era Reanalysis for Research and Applications (MERRA)-together with sea surface salinity fields from two globally gridded datasets (World Ocean Atlas and Met Office EN3 datasets). The resulting 12 estimates of the 1979-2007 mean surface-forced streamfunction all depict a subpolar cell, with maxima north of 45°N, near σ = 27.5kgm-3, and a subtropical cell between 20° and 40°N, near σ = 26.1 kgm-3. The mean magnitude of the subpolar cell varies between 12 and 18 Sv (1 Sv ≈ 106m3 s-1), consistent with estimates of the overturning circulation from subsurface observations. Analysis of the thermal and haline components of the surface density fluxes indicates that large differences in the inferred lowlatitude circulation are largely a result of the biases in reanalysis net heat flux fields, which range in the globalmean from -13 to 19 Wm-2. The different estimates of temporal variability in the subpolar cell are well correlated with each other. This suggests that the uncertainty associated with the choice of reanalysis product does not critically limit the ability of the method to infer the variability in the subpolar overturning. In contrast, the different estimates of subtropical variability are poorly correlated with each other, and only a subset of them captures a significant fraction of the variability in independently estimated North Atlantic Subtropical Mode Water volume.