Antarctic Surface Energy Budget Data Set Validation

Data Details FTP Site Problems Credits

Validation Data

In order to assess the validity of the model output, the radiative and turbulent heat fluxes were compared to surface measurements. The radiative fluxes are compared to measurements taken at Amundsen Scott South Pole Station for the year 1987. The measurements at the South Pole were taken by Dutton et al. (1989) from April 1986 to February 1988. The approximate absolute errors in the measurements are 2% for the shortwave and 5% for the longwave. The South Pole measurements were compared to the average of the four model grid points that surround the pole. Estimates of the sensible and latent heat fluxes at the surface over the Ross Ice Shelf by Stearns and Weidner (1993) were used to determine the validity of the turbulent heat fluxes. The sensible and latent heat flux estimates were averaged over the time period 1984-1990 since a significant amount of monthly mean data for a given station and year may be missing. In addition, an area average was computed based on data from six automated weather stations. This alleviates the problem of rapidly varying topography when comparing a point measurement to values within the 200 km ARCSyM grid cell. The NCEP-NCAR Reanalysis Product (Kalnay et al., 1996) is also examined. In this analysis, the model results that do NOT include the APP-x data are referred to as the baseline results and the model runs that include the APP-x data are called the APP-x model runs.

Radiative Flux Validation

Figure 1 shows the downwelling longwave flux comparisons at the South Pole. The APP-x model runs show a positive improvement from the baseline runs every month of the year. The improvements, however, are only large during the summer months when the cloud amount from the APP-x model runs is as much as 40% greater than the cloud amount from the baseline runs. For instance, the downwelling longwave flux from the APP-x model run is about 30 W/m^2 closer to the surface measurements than the baseline model output in January. From March to October, the downwelling longwave radiation is underestimated by as much as 23 W/m^2 in the APP-x and baseline cases. For all other months, the downwelling longwave flux from the APP-x model run is within the measurement error. The APP-x results show an average improvement of about 7 W/m^2 over the baseline results throughout the year. Further, the results from the ARCSyM experiments with the APP-x data are on average 31 W/m^2 closer to the surface measurements than the NCEP-NCAR reanalysis.

The surface absorbed (net) shortwave radiation at the South Pole (Figure 2) does not show an improvement with the incorporation of the APP-x data in January and November, even though the downwelling longwave radiation is significantly better during those months. This may suggest that the earth-atmosphere albedo is too large at that time because the surface absorbed shortwave radiation is underestimated by up to 13 W/m^2. Also, the shortwave cooling effect of clouds is more sensitive to changes in cloud optical depth than the longwave warming effect (Pavolonis and Key, 2003), so it is possible that the monthly average cloud optical thickness near the South Pole may be slightly overestimated during the summer. As one may expect, since the downwellinglongwave radiation from the NCEP-NCAR reanalysis is greatly underestimated, the surface absorbed shortwave flux is greatly overestimated and is characterized by errors much greater than for either ARCSyM run.

The upwelling longwave flux at the South Pole is shown in Figure 3. Even though the downwelling longwave flux and the absorbed shortwave flux from either ARCSyM run compare much more favorably with the surface measurements than the NCEP-NCAR reanalysis, the upwelling longwave flux (or surface temperature) is less accurate than the NCEP-NCAR reanalysis, except in January. This suggests that any errors in the turbulent heat flux from the NCEP-NCAR reanalysis discussed below cancel the errors in the net radiation. Also, the differences between the APP-x case and the baseline case are generally small. One may then conclude that the response of the surface temperature to changes in the downwelling longwave radiation and net shortwave radiation are small for these 48 hour model integrations. If the model integrations were longer, the differences in surface temperature between the APP-x run and the baseline run would likely be greater.

Turbulent Flux Validation

In Figure 4, the comparisons between the sensible heat flux estimates from Stearns and Weidner (1993) and the modeled results are shown. The results that include the APP-x data compare slightly more favorably with the estimates of sensible heat flux than those for the baseline case seven months out of the year. When the baseline case compares more favorably, the differences between the two results are still fairly small. The average difference between the model results and the station estimates are 5.8 W/m^2 for the baseline case and 5.6 W/m^2 for the APP-x case. As mentioned earlier, the response of surface temperature to a change in the downwelling longwave radiation and the absorbed shortwave radiation is fairly small in the 48 hour integrations. This also seems to be reflected in the sensible heat flux, so the two model runs produce similar results. The errors in the NCEP-NCAR reanalysis are much greater than the errors in either ARCSyM result every month except January and February.

Figure 5 shows the latent heat flux comparisons. As for the sensible heat flux, the differences between the APP-x case and the baseline case are small. The average error for both type of simulations is about 3 W/m^2, with the greatest errors occurring during the summer months. In the winter, the magnitude of the latent heat flux over the Ross Ice Shelf is a negligible component of the surface energy budget. The errors in the NCEP-NCAR reanalysis are considerably larger than those associated with either ARCSyM run, especially in the summer.


Dutton, E.G., R.S. Stone, and J.J. DeLuisi, 1989: South Pole surface radiation balance measurements, April 1986 to February 1988. NOAA Data Rep. ERL ARL-17, 49 pp.

Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-yr reanalysis project. Bull. Amer. Meteor. Soc., 77, 437-471.

Pavolonis, M.J. and J.R. Key, 2003: Antarctic cloud radiative forcing at the surface estimated from the AVHRR Polar Pathfinder and ISCCP D1 data sets, 1985-1993. J. Appl. Meteor. (in press). Click here for a PDF version.

Stearns, C.R. and G.A. Weidner, 1993: Sensible and latent heat flux estimates in Antarctica. Antarctic Meteorology and Climatology: Studies Based on Automated Weather Stations. Antarctic Research Series, Vol. 61, Amer. Geophys. Union, 109-138.