Documentation

Algorithm for Surface Air-Sea Fluxes:

We used the COARE 3.0b algorithm to compute bulk air-sea fluxes, including latent and sensible heat flux, net heat flux, and other associated fluxes including evaporation, evaporation minus precipitation, sensible heat flux due to rain, buoyancy flux, and wind stresses.

These flux data are available for display and delivery at hourly, daily, 5-day, monthly, and quarterly resolution from www.pmel.noaa.gov/tao/drupal/flux/.

The COARE 3.0b algorithm is detailed in

Fairall, C.W., E. F. Bradley, J. E. Hare, A. A. Grachev, and J. B. Edson, 2003: Bulk Parameterization of Air-Sea Fluxes: Updates and Verification for the COARE Algorithm. J. Climate, 16, pp 571-591. The fortran version of the software is available from
ftp://ftp1.esrl.noaa.gov/BLO/Air-Sea/bulkalg/cor3_0/fortran3_0/.

Implementation of this algorithm by the TAO Project of NOAA/PMEL was guided by

Cronin, M.F., C. Fairall, and M.J. McPhaden, 2006: An assessment of buoy derived and numerical weather prediction surface heat fluxes in the tropical Pacific. J. Geophys. Res., 111, C06038, doi:10.1029/2005JC003324.

Temporal Resolution

Hourly:

To compute high resolution fluxes, we use hourly data for all inputs and the warm-layer and cool-skin corrections which are built into the COARE 3.0b algorithm. The hourly input data are either delayed mode data or realtime data, with realtime allowing very recent fluxes to be available. All of the high resolution data sets described below are computed from both delayed mode and realtime hourly input data. These hourly input data are mainly averages of higher resolution data, including two minute radiation data and ten minute surface meteorological data. There are four high resolution data sets, each corresponding to an entry in the averaging menu on the flux page.

The first two, called "High Resolution" and "High Res (Rel Wind Speed)", both require a complete set of input hourly data at each time, i.e., wind speed, sea-surface temperature, air temperature, relative humidity, shortwave radiation and longwave radiation. In the case of "High Res (Rel Wind Speed)" we also require that 10 meter or shallower high resolution ocean current data are available so we can compute the wind speed relative to the sea surface (which we assume the currents at 10 m or shallower are respresentive of). "High Resolution" uses absolute wind speed. These first two high resolution flux data sets are therefore our best possible estimates of the surface fluxes, which include both the cool skin and warm layer effects enabled by the COARE algorithm. The availability of these data is limited primarily by the availability of longwave radiation, since longwave is available at relatively few buoy sites compared to the other input data.

To greatly increase the availability of high resolution surface fluxes at more buoy sites, we relax the requirement that we have longwave radiation data available for the two data sets identified by the menu entries "High Resolution no LW" and "High Res no LW (Rel Wind Speed)". These data sets use a longwave radiation climatology computed from daily OAFlux net longwave radiation and sea-surface temperature data. This climatology, interpolated to the nominal location of each buoy, is used at all times, rather than any observed longwave radiation. These two high resolution flux data sets, which also include the cool skin and warm layer effects, are slightly lower quality estimates of the surface fluxes because we use the longwave climatology rather than longwave data. However, the fluxes are not very sensitive to the input longwave data compared to other inputs. A detailed analysis of the sensitivity of the fluxes computed using longwave input data versus using the longwave climatology shows very slight differences in the fluxes, at most 0.4% (sensible heat flux at only one buoy site) and more typically much smaller, on the order of 0.01% or less.

Fluxes are also insensitive to input sea level pressure, so for all four high resolution flux data sets we use the ICOADS climatology of sea level pressure at the nominal location of each buoy to replace any missing data. Similarly, when rain-rate data are not available we simply set the rain-rate to zero and at such times we do not compute evaporation minus precipitation, or the sensible heat flux due to rain.

There are fewer moorings from which relative vis-a-vis absolute wind speed can be computed, but at these sites, flux estimates are slightly more accurate (see below). Note that net shortwave radiation is independent of wind speed.

Hourly Q_net is computed using the net shortwave radiation and net longwave radiation output from the COARE v3.0b algorithm. This is done so that the net heat flux includes any effects of the warm layer and cool skin corrections. For those high resolution fluxes computed using a longwave climatology, Q_net is therefore not computed.

The delayed mode and realtime hourly input data come from one of the following buoy types: TRITON buoys in the Pacific and Indian Oceans maintained by JAMSTEC, ATLAS Next Generation buoys in the Pacific, Atlantic, and Indian Oceans maintained by NOAA/PMEL, T-FLEX buoys in the Alantic and Indian Oceans maintained by NOAA/PMEL, and ATLAS Refresh buoys in the Pacific Ocean maintained by NOAA/NDBC. TRITON delayed mode and realtime hourly data are hourly averages centered at 30 minutes after the top of the hour. ATLAS Next Generation hourly delayed mode data are computed from higher resolution data, i.e., two minute and ten minute data and centered at the top of the hour. T-FLEX realtime hourly data are spot samples at the top of the hour, rather than hourly averages, per se. Future hourly T-FLEX delayed mode input data will be computed in the same way as the ATLAS Next Generation input data, from higher resolution data. ATLAS Refresh hourly data are mostly averages of higher resolution realtime data, i.e., two minute radiation data, ten minute surface met data. Hourly realtime Refresh barometric pressure and ocean currents are provided directly by NDBC.

For detailed information on the history of the various buoy systems at specific sites, refer to the following links:

Pacific Ocean: TAO ATLAS Next Generation / TRITON

Pacific Ocean: TAO ATLAS Next Generation / Refresh

Atlantic Ocean: PIRATA ATLAS Next Generation / T-FLEX

Indian Ocean: RAMA ATLAS Next Generation / T-FLEX

As more delayed mode data from recent deployments become available, fluxes computed from these new data will replace those computed previously from realtime data. In each data file delivered at every data time there is a source code to indicate if the data were computed from entirely delayed mode data (source code 6) or if any realtime data data were used (source code 2).

Daily:

As for hourly fluxes, we provide two daily flux data sets, one using absolute wind speeds and the other using relative wind speeds. The cool skin, warm layer calculation is turned off for calculations using daily averages. Therefore, daily averaged net shortwave radiation, net longwave radiation, and sensible heat due to rain are all independent of wind speed.

To assess the sensitivity of the latent and sensible heat fluxes and wind stress to the use of relative wind speeds compared to absolute wind speeds, we computed the differences between these two daily data sets at mooring sites in the Pacific, Atlantic and Indian Oceans where 10 m current velocity data were available. The results are summarized in the following tables:

To minimize biases in flux estimates from using daily rather than high resolution wind speeds, all daily wind speeds are corrected prior to input into the bulk algorithm using an estimate of the mesoscale gustiness computed from high resolution data (Cronin et al. 2006).

5-Day, Monthly, and Quarterly Averages:

5-Day, monthly, and quarterly averages of surface fluxes are computed from daily fluxes which use absolute wind speed as input.

Input Data Quality:

For all of the flux calculations, we use only data with the highest two levels of data quality from the TAO, TRITON, PIRATA, and RAMA arrays.

Climatologies for Surface Fluxes:

The surface heat flux climatologies, which are shown in time series displays and used to compute anomalies, are based on the monthly Objectively Analyzed Air-Sea Fluxes (OAFlux) data from Woods Hole Oceanographic Institution. These include latent and sensible heat fluxes, evaporation, net shortwave and longwave radiation, and net heat flux. The climatology for E-P was computed by subtracing the Xie and Arkin rain climatology (1997) from the OAFlux evaporation