NOC Surface Flux and Meteorological Dataset


The NOC Surface Flux Dataset contains estimates of the following parameters (including uncertainty) on a 1° monthly mean grid for the global ice-free ocean:

  • Air temperature (at 10 m)
  • Specific Humidity (at 10 m)
  • Wind Speed (at 10 m)
  • Sea level pressure
  • Sea surface temperature (bulk)
  • Latent heat flux
  • Sensible heat flux
  • Net longwave radiation
  • Net shortwave radiation

The dataset covers the period 1973 – end of last full calendar year.

This web page describes the need for these parameters, the way we have estimated, the observations used and how to access the data.

Schematic showing the different components of the ocean heat fluxes.

Maps showing the winter and summer net heat fluxes. During the northern winter (Dec - Feb) strong heat loss can be seen in the northern oceans, especially over the western boundary currents, and strong gain in the southern oceans. The situation is reversed in the northern summer months (Jun - Aug) with a heat gain in the northern oceans and heat loss in the southern oceans.

Net heat flux (W m-2) over the oceans during the northern winter (Dec - Feb)(top) and northern summer (Jun - Aug)(bottom) months. Negative values indicate a cooling of the ocean surface.


The oceans and atmosphere are strongly linked by the transfer of energy (momentum and heat), water, and gases (e.g. carbon dioxide) between them. For example, the wind blows over the ocean transferring energy to the surface. This is turn slows down the wind and creates waves.

Similarly, as the air blows over the oceans it is warmed or cooled by the underlying sea surface, with a transfer of energy between them. Similar processes drive the transfer of water and gases between the oceans and atmosphere.

At the National Oceanography Centre (NOC) we try to understand and estimate these transfers (called fluxes) and the net energy budget on a global scale. This helps us to understand the link between the oceans, the Earth's energy budget and the climate system. The net energy budget of the oceans surface includes contributions from:

  • Solar (shortwave) heating of the sea surface.
  • Thermal and evaporative cooling of the surface (sensible and latent heat fluxes).
  • Radiative (longwave) cooling of the surface.

Measuring and estimating the fluxes

The fluxes between the oceans and atmosphere can be measured directly. This requires the small-scale atmospheric motions to be measured at a high frequency. High frequency measurements of the temperature and humidity are also required. These measurements are costly. As a result they are only made on a few research vessels. This in turn limits the extent of the observations. In order to estimate the fluxes globally we need to turn to other sources.

Using the high-frequency measurements made on board research vessels the fluxes can be parameterised in terms of the bulk atmospheric properties. I.e. simple equations can be defined to estimate the fluxes in terms of a transfer coefficient and the mean temperature, humidity and wind speed at a two known levels. Over the oceans these are usually chosen to be the sea surface and at 10 m above the sea surface.

Estimates of the mean atmospheric parameters can come from a variety of sources including:

  • in situ observations from ships and buoys;
  • estimates from satellite data; and
  • values derived from numerical models.

In the NOC Surface Flux Dataset observations from in situ ship observations are currently used.

In situ ship observations

Merchant, naval and research ships have made weather observations for over 200 years. The first activities to coordinate marine observations date from an International Meteorological Conference in 1853. At this conference a standard format for the collection of the observations was proposed. Brief instructions for making the observations were also issued. The observations are now coordinated and made as part of the World Meteorological Organisation (WMO) Voluntary Observing Ship (VOS) Programme.

The majority of the observations collected as part of the VOS Programme include the parameters required to calculate the fluxes.

  • Air temperature and humidity.
  • Wind speed.
  • Sea level pressure.
  • Sea surface temperature.
  • Cloud cover.
  • Weather conditions.

These measurements are normally made every 3 or 6 hours using basic equipment and methods. Examples include mercury or alcohol thermometers in meteorological screens. Other examples include visual observations of the wind, waves and present weather. Increasingly, automatic weather systems are used to make the observations.

Due to their basic nature, the observations contain errors and biases – some of which can be large. The observations are also sparsely distributed (but much more frequent that research vessel measurements). Before the data can be used biases need to be detected, estimated and corrected. Where this is not possible the observations need to be discarded. Once bias adjusted and quality controlled the data can be used to create global fields and estimate the fluxes.

Calculating the fluxes and creating gridded fields

Map showin average random and sampling uncertainty in the monthly mean net heat flux. Minimum values can be seen over the major shipping lanes (< 20Wm-2) and higher values (> 80 Wm-2) in the data sparse regions)

Mean uncertainty (Wm-2) in the monthly mean net heat fluxes. The major shipping routes can be clearly seen, with values < 20 Wm-2.

The equations used to calculate the fluxes are non-linear. As a result, random errors in the input data may combine to give biased estimates of the fluxes when averaged over a large number of points. To minimise these biases we therefore need to minimise the errors in the input data. This can be achieved by averaging prior to calculating the fluxes. However, the atmospheric parameters can be correlated. For example, the air – sea temperature difference and wind speed is often correlated and this link may be lost on averaging. This will introduce other errors.

To minimise the errors in the input data and to maintain the link between the different parameters the data have been averaged to give daily estimates of the basic meteorological parameters. This timescale is shorter than that typically observed in the atmospheric parameters, allowing the links to be maintained. This also reduces the errors in the input data to the flux calculation.

The data used to create the daily average fields are sparse and unevenly distributed. To take this uneven sampling into account an optimal interpolation (OI, or simple kriging) scheme has been used to grid and average the data. In this scheme, the data are weighted according to their clustering and how confident we are in them. This scheme also has the advantage of filling gaps and providing uncertainty estimates for the gridded fields. Where gaps have been filled or we have limited data the uncertainty is high.

The daily fields have then been used to produce daily estimates of the fluxes and their uncertainty. These have then been averaged to give monthly mean values. The parameterisations of Smith (1980) and Smith (1988) have been used to calculate the fluxes. Further details on the optimal interpolation and calculation of the fluxes can be found in the following papers:

Berry, D. I., and E. C. Kent, 2009: A new air - sea interaction gridded dataset from ICOADS with uncertainty estimates. Bulletin of the American Meteorological Society, 90, 645 - 656. DOI: 10.1175/2008BAMS2639.1.

Berry, D. I., and E. C. Kent, 2011: Air - sea fluxes from ICOADS: the construction of a new gridded dataset with uncertainty estimates. International Journal of Climatology, 31, 987 - 1001: DOI: 10.1002/joc.2059.

Access to the dataset

The monthly mean fields forming the NOC dataset can be downloaded from the US National Center for Atmospheric Research (NCAR) and the British Atmospheric Data Centre (BADC). The daily fields are available on request.

The observations used to construct the NOC Surface Flux Dataset come from the International Comprehensive Ocean – Atmosphere Data Set (ICOADS). Further details and be found on the ICOADS webpage.