A satellite-based study of ocean-atmosphere heat fluxes in the vicinity of John Brewer Reef, Queensland

This study investigates the use of satellite observations for determining the exchange of radiation and turbulent heat fluxes between the ocean and atmosphere. Meteorological observations and radiation fluxes were collected for 45 weeks at John Brewer Reef (18° 38' S, 147° 4 'E), approximately 70 km north-northeast of Townsville, Australia. Estimates of the mean weekly radiation and turbulent heat fluxes through the ocean surface were derived from concurrent satellite observations, and compared with surface-based estimates. The net radiation flux was separated into four components. The downwelling shortwave radiation flux was derived from GMS-3 visible image data, initially on a daily basis, using a model of atmospheric shortwave transmittance. The weekly average downwelling shortwave radiation flux was estimated with a root mean square accuracy of 14.3 Wm-2. The upwelling shortwave radiation flux was parameterised as a simple function of season. Downwelling longwave radiation flux was estimated with a complex scheme. A radiative transfer model (LOWTRAN 6) was employed to calculate downwelling longwave radiation fluxes under clear and totally overcast skies with atmospheric profiles derived from NOAA-9 TOYS sounding data. The cloud fraction, as determined from GMS-3 visible images, was used to produce a weighted average of the two downwelling longwave radiation fluxes. The weekly average downwelling longwave radiation flux was produced with a rms accuracy of 9.1 Wm-2. The upwelling longwave radiation flux was provided primarily by the sea surface temperature, which was obtained on a weekly basis from a product which combined multi-channel satellite-derived temperatures with surface measurements from ships-of-opportunity. The rms accuracy of the resulting weekly average upwelling longwave radiation flux was 3.0 Wm-2. The weekly average net radiation flux was reproduced with arms accuracy of 18.9 Wm-2. The turbulent heat flux was dominated by the latent heat flux, which was modelled according to the bulk aerodynamic approach. The weekly-averaged specific humidity difference between the sea surface and the ten metre level was derived as a simple function of the weekly-averaged precipitable water vapour in the atmosphere, as estimated by NOAA-9 TOVS data. Wind speeds were obtained from a standard meteorological wind analysis which included the region of John Brewer Reef. The weekly-averaged values of the latent heat flux were reproduced with arms accuracy of 39.6 W m-2. The sensible heat flux, which was in general a small term, was estimated as a fraction of the latent heat flux. The total turbulent heat flux was modelled with a rms accuracy of 45.4 Wm-2. The heat balance of a hypothetical column of water near John Brewer Reef was considered. A comparison was drawn between the radiation and heat fluxes through the upper surface of the column and the variations in the heat content of the column as evidenced by temporal changes in its sub~surface temperature profile. A limited quantity of information on the current fields and the spatial distribution of sub-surface water temperatures allowed the discussion of oceanic advection of heat on an order-of magnitude basis only. It was determined that advection made a significant contribution to the heat balance of the column in winter only. The satellite-derived radiation and turbulent heat flux estimates were extended to form spatial averages over a mesoscale region and temporal averages for five weekly periods. The fluxes were calculated for the corresponding area and period with data extracted from marine weather reports, representing the traditional source of mesoscale, mean monthly ocean-atmosphere heat fluxes. Differences between the fluxes derived from the marine observations and those estimated by satellite were larger than, expected, considering the accuracy displayed by the satellite estimates in the neighbourhood of John Brewer Reef. The implication is that the superior temporal and spatial resolution of satellite observations allows better reproduction of the surface heat and radiation fluxes than the traditional approaches.