Lambertian Surface Albedo Climatology at 360 nm from TOMS Data Using Moving
Satellite retrieval algorithms, including estimation of the surface UV
irradiance, require accurate knowledge of the surface albedo. At UV wavelengths
the surface albedo is often low, some 3-8%, but snow or ice cover can increase
the surface albedo up to 95%. Therefore, the surface albedo of polar regions
has high annual variation which complicates accurate estimation of the
surface UV irradiance from satellite data. Previously, the surface albedo
climatologies have been constructed assuming that the seasonal surface
albedo cycle is identical from year to year [Herman et al. 1997, Koelemeijer
et al. 2003]. At high latitudes this assumption leads to substantial underestimation
of the surface albedo during snow cover transition periods.The aim of this
work was to construct a global UV range albedo climatology with fine spatial
and temporal resolution and realistic high-latitude values.
Climatology Construction Method
The moving time-window (MTW) method [Tanskanen et al. 2003] is based on
the assumption that the reflectivity values within a certain time-window
around the day of interest form a sample of the reflectivity distribution
whose lower tail corresponds to the clear sky case. An estimate of the
surface albedo is obtained by fitting a linear function to the lower tail
of the cumulative distribution of the sample. The wider the time-window
is, the more information is available about the reflectivity distribution.
However, use of a very wide time-window leads to underestimation of the
surface albedo in case the surface albedo experiences a transient within
the time-window. In order to account for the surface albedo transients
the MTW algorithm aims at choosing an optimum time-window by narrowing
and shifting the window during transients. The MTW algorithm was applied
to the TOMS 360 nm Lambertian Equivalent Reflectivity (LER) time-series
to construct daily surface albedo estimates for the Nimbus-7 period. Prior
to the use of the MTW method the LER data were gridded to a 1 by 1 degree
regular grid defining the grid cell value as the daily minimum LER value.
Finally, the surface albedo climatology was constructed by averaging the
resulting surface albedo estimates for the years 1979-1992.
The surface albedo climatology can be downloaded as a zipped tar package
(7.2 MB). The climatology is stored as 360x180x365 elements of 1 byte binary
data with no additional delimiters nor any metadata. The data set can be
considered as a 3-dimensional array varying in longitude, latitude and
day of the year. The structure of the file is such that longitude is the
fastest varying dimension, latitude is the second fastest, while day of
the year is the slowest varying dimension. The longitude dimension runs
from West to East with 1 degree step starting from -179.5E. Accrodingly,
the latitude dimension runs from South to North with one degree step starting
from -89.5N. The download package contains some documentation and a Fortran
example of how to use the climatology.
When the climatological UV albedo data is used in a publication or presentation,
the source of data should be acknowledged by referring to the appropriate
publication [Tanskanen 2004].
Herman, J. R., E. A. Celarier, Earth surface reflectivity climatology at
340--380 nm from TOMS data, J. Geophys. Res., 102, 28003--28011, 1997.
Koelemeijer, R. B. A., J. F., de Haan, P. Stammes, A database of spectral
surface reflectivity in the range 335--772 nm derived from 5.5 years of
GOME observations, J. Geophys. Res., 105, 5059--5067, 1997.
Tanskanen, A., A. Arola, J. Kujanpää, Use of the moving time-window
technique to determine surface albedo from the TOMS reflectivity data,
In: Proc. SPIE Vol. 4896, p. 239--250, 2003.
Tanskanen, A. Lambertian Surface Albedo Climatology at 360 nm from TOMS
Data Using Moving Time-Window Technique. In: Proceedings of the XX Quadrennial
Ozone Symposium, 1-8 June 2004, Kos, Greece.
Aapo Tanskanen, November 5th, 2004