Meridional heat flux variability at 26.5 degrees N in the North Atlantic Ocean

Data from 4.8 years of current meter mooring arrays located across the western boundary currents at 26.5$\sp\circ$N off the Bahamas are used to investigate meridional heat transport variability. These data are compared to Community Modelling Effort model results and historic hydrographic sections. Important currents here are a northward Antilles Current in the upper ocean and a deep southward DWBC. A large fraction of the entire transatlantic heat flux is observed here. Moored estimates of local heat transport exhibit large variability over a range of $\pm$2 PW on roughly 100 day time scales. Interannual variability is observed in deep water mass transport. An annual cycle of heat transport with a range of 1.4 PW with a summer maximum and fall minimum exists, qualitatively similar to CME model results. This cycle is primarily due to interior wind forcing. It is primarily contained above the thermocline and would strongly influence the annual heat flux cycle through the entire transatlantic section if conditions offshore remained constant. Large northward baroclinic heat flux is observed, causing estimates of net transatlantic heat flux made by combining mooring data with Levitus climatology data in the Atlantic interior to have a mean of 1.55 PW, 0.35 PW greater than the previously accepted 1.2 PW.Local eddy heat flux is small. Since eddy energy is concentrated in the western boundary region, basinwide eddy heat flux is unlikely to be important. Diagnostics of heat flux estimate errors are performed, primarily from CME. This indicates that the geostrophic method yields an underestimate of transatlantic heat flux. An improvement in the most probable error is seen if monthly rather than annual values are used for mean flows. The mean error is also improved if direct rather than geostrophic results are used in the western boundary current region. The barotropic correction term is important, and a major factor in making geostrophic calculations of heat transport underestimates. From the model, deep shear in the ocean seems weak enough to permit monitoring by upper-ocean XBT sampling. With proper instrumentation, widely spaced dynamic height moorings spaced could provide reasonable heat transport estimates