Solar Radiation
Documentation written by F. Fetterer; edited by M. Serreze
Global (or total downwelling shortwave) radiation fields are constructed using measurements from the Russian North Pole drifting stations, western drifting stations, land stations, and empirically derived values from earlier Russian studies. The radiation fields are derived in the same way as those presented in Serreze et al. [1998], but are improved with additional data.
Ocean data sources
Radiation data from the Russian North Pole (NP) drifting stations were compiled by AARI scientists as monthly averages for the mean position of stations NP-17 through NP-31 [Marshunova and Mishin, 1994], and obtained from Daily Arctic Ocean Rawinsonde Data from Soviet Drifting Ice Stations, a CD-ROM product of AARI, University of Washington Polar Science Center, and NSIDC.
Radiation data from U.S. drifting station "T-3," primarily for the spring through autumn months of 1953, 1957-1959, and 1971-1973 were digitized from tables in Marshunova and Chernigovskii [1971] and Weller and Holmgren [1974].
Monthly radiation means and positions from U.S. drifting station "ARLIS II" for January 1964 through May 1965 were digitized from tables in Roulet [1969].
Measurements collected during the Surface Heat Budget of the Arctic (SHEBA) field program provided data for 1997 through 1998.
The total number of station months from these combined data sources is over 70 from June through September, and over 40 from March through May, as well as October. Very few data points are available from November through February, when polar darkness dominates most of the Arctic.
Land data sources
The Global Energy Balance Archive (GEBA, [Gilgen and Ohmura, 1999; Ohmura and Gilgen, 1991] http://bsrn.ethz.ch/gebastatus/) provided monthly means for global radiation for 57 land stations north of 60 degrees North. Station records range in length from one to 45 years. Stations in the Russian Technological Handbook of the Climate Of Russia (Arctic Region) Solar Radiation (provided in English on this Atlas), are for the most part represented in GEBA, but because the tables of mean monthly values in the Handbook are generally the result of more years of data than the GEBA averages, with few exceptions we used the Handbook values for a station rather than the GEBA values when both were available. Monthly values for nine Alaskan stations (1961 through 1990) were obtained from the NOAA Solar and Meteorological Surface Observations (SAMSON) archive (available from NSIDC). Monthly records of at least 20 years duration for 14 sites in the Canadian Arctic, obtained from the Atmospheric Environment Service, Ottawa, and climatological monthly means for two Eurasian sites from Gavrilova [1963] were also included.
Several years of monthly data from the early 1970s were obtained for the Baffin Bay region as a result of field experiments [Muller and Coauthors, 1976; Jacobs et al., 1974]. Records of up to five years in length (1995 through 1999) for 13 stations over the Greenland ice sheet came from the National Atmospheric and Space Administration (NASA) -sponsored Program for Arctic Regional Climate Assessment (PARCA).
Ancillary data sources
The assembled archives provide poor coverage over the North Atlantic Ocean and for coastal Greenland. To obtain coverage for these areas, it was necessary to resort to calculated values provided by Marshunova and Chernigovskii [1971] and Smetannikovoi [1983]. These estimates are based on consideration of clear sky radiation and attenuation by aerosols, cloud amount, and cloud type.
Monthly mean global radiation values from the data sets described above were interpolated to the EASE-Grid using Cressman interpolation [Cressman, 1959]. The Cressman interpolation routine is given by:
where xi is the monthly mean value of radiation at location i within a radius R of the grid location, and n is the number of stations found within the search radius. The weights wi are given by
where d is the distance from the grid cell to the observation location.
Using the monthly mean values directly may introduce biases related to latitudinal dependencies of the flux. Therefore, prior to the application of the Cressman interpolation, an estimate of clear sky downwelling shortwave flux was subtracted from the station value. This step essentially normalizes the data with respect to latitudinal variations in solar zenith angle and day length, and with respect to associated path-length dependencies of non-cloud atmospheric absorption and scattering. The clear sky downwelling shortwave flux is added back to the gridded values after interpolation.
The clear sky downwelling shortwave flux is based on a radiative transfer model [Schweiger and Key, 1994]. The clear sky flux calculations use calculated or derived values for surface albedos, optical depth, month-specific total precipitable water, and total column ozone. Uncertainties in these quantities have little effect on the resulting clear sky radiance. For instance, for a 20 percent error in assumed albedo of 0.85, the resulting error in the calculated flux of 400 Wm-2 is only 5 Wm-2 [Serreze et al., 1998].
The record length for the land station data varies widely, and some land and ocean data are given as climatological monthly means only. Data from the North Pole drifting stations are from a different position for each month. To reduce spatial biases that would be introduced by interpolating drifting station and other data without regard to position and record length, a two-step interpolation procedure was used. First, individual monthly means at the average position for that month from the North Pole drifting stations were interpolated to the ocean EASE-Grid cell centers. The initial search radius R was set to 500 km. If data from fewer than five years were available (that is, fewer than five monthly mean values) within that radius, the search was enlarged to 750 km. The grid cell was coded as missing if fewer than five monthly mean values resulted from the enlarged search.
In the second step, climatological monthly means for land stations were calculated from the station record of monthly means. Climatological means based on less than three years were discarded. These values, plus those calculated values described in the "Ancillary data sources" section above, were combined with the ocean grid cell values from the first step of the interpolation. Land and ocean points in the combined data set were then interpolated to EASE-Grid cell centers, using a search radius of 500 km, enlarged if necessary to 750 km, 1000 km, and 1250 km, with a minimum of two values for each radius. The larger radii were necessary because of the relatively great distance between land stations.
In order to fill grid cells that would otherwise be coded as missing for months from November through February (when polar darkness dominates), grid cells where the clear sky downwelling shortwave flux is less than 15 Wm-2 were simply ascribed the clear sky value multiplied by an assumed effective cloud transmittance of 0.60. Calculations of effective cloud transmittance are described in Serreze et al. [1998].
The gridded radiation products are climatological monthly means. The browse versions of these files are shown as .gif format images with a color bar and contours. These browse files are for the purpose of quickly visualizing the content of the corresponding ASCII data files. The IDL routine used to color-map the images gives a smooth and visually pleasing result, but keep in mind that the gridded ASCII files have a value only for every grid cell. The grid cell centers are shown as red dots. For information on the structure of the gridded files, see the section on "EASE-Grid".
The large search radius used for the Cressman interpolation tends to smooth over inhomogeneities in the fields that might result from the data sampling problems related to varying station density and caused by using records of different length. At the same time, real local variations are smoothed over. The results for the North Atlantic and coastal Greenland are based on empirically derived values and should be viewed with caution.
The data sets described above were checked for obvious outliers but were otherwise used as provided. Radiometers generally have an accuracy of ±5 to ±10 Wm-2. However, the accuracy may be less if sensors were calibrated in non-Arctic conditions. The extreme solar zenith angles of the Arctic can lead to large measurement errors, due to limitations in the cosine response of the instruments, and to multiple reflections within the radiometer domes.
The section of the Atlas containing gridded field data also contains a monthly climatology of direct, total (or global) and net radiation by M. S. Marshunova.