6. Cooperative Programs
Spectral Albedo Observation on the Snow Field at Barrow, Alaska
Teruo Aoki, Tadao Aoki, Masashi Fukabori, and Yuji Zaizen
Meteorological Research Institute, Tsukuba, 305-0052, Japan
Yoshihiro Tachibana
Research Institute of Civilization, Tokai University, Hiratsuka 259-12, Japan
Fumihiko Nishio
Hokkaido University of Education, Kushiro 085, Japan
Introduction
There are possibilities of remote sensing for snow impurities by observation of the visible albedo and the snow grain size using the near infrared albedo [Warren, 1982]. We will estimate the snow impurities and snow grain size with the satellite data of the next generation. The basis of these studies is a multiple scattering radiative transfer model for the atmosphere-snow system that simulates the spectral albedos at the snow surface and the top of the atmosphere. To validate this model, field observations for the spectral albedo and snow physical parameters have been done on the snow field at Barrow, Alaska, from April 14 to 27, 1997. Under a clear sky, even a gentle slope of the snow surface affects the snow albedo [Grenfell and Warren, 1994]. As a simpler case in which it is unnecessary to consider such an effect, the relationship between the spectral albedo and snow physical parameters under an overcast condition on April 21 after new snowfall, is discussed in this study.
Observation
The observation site was a snow field on tundra 5 km northeast of Barrow town. The snow depth was 60 cm and the surface was covered by new snow of 1 cm depth that fell the day before the observation. The grain shape of new surface snow was dendrites with crystal size (radius) of 1-2 mm and branch size (radius) of 25-50 mm. The second layer (1-24 cm snow depth) was fine-grained old snow with a grain radius of 100-150 mm, and the third layer (24-44 cm) was faceted crystals with a grain radius of 250-400 mm. The bottom layer (44-60 cm depth) was hoar crystals with a grain radius of 1-3 mm. Snow impurities were filtered within a day by Nuclepore filters with the pore size of 0.2 mm after melting the snow samples of the surface and 5-10 cm depth snow layers. The concentrations of impurities were estimated by direct weight measurements of the Nuclepore filters, before and after filtering, with a balance. The observed concentration of snow impurities was 2.46 ppmw for surface snow and 1.19 ppmw for 5-10 cm depth. The sky condition was overcast with altocumulus. Air and snow surface temperatures were 14.2ºC and 11.0ºC, respectively at 1200 local standard time (LST).
The spectral snow albedo was observed by a grating type spectrometer, "FieldSpec FR", made by ASD Inc. (U.S.). It is necessary to observe the downward and upward flux to obtain the albedo. The downward flux was observed by directing the optical fiber of the spectrometer toward the surface of a standard white reflection plate that is set horizontally above the snow surface. The upward flux was observed by directing the optical fiber toward the underside of the standard white reflection plate. The scanning spectral range of the spectrometer is 0.35-2.5 mm with the spectral resolution of 3 nm for the wavelength of 0.35-1.0 mm and 10 nm for 1.0-2.5 mm. The spectral snow albedo is the average of five spectral albedos obtained from five pairs of measurements for the downward and snow reflected solar fluxes. It takes several minutes to obtain these quantities.
Results and Discussion
The observed spectral snow albedo was compared with the theoretical calculations by a multiple scattering model for the atmosphere-snow system in which snow grains are assumed to be mutually independent ice particles and radiative transfer processes are based on the Mie theory for single scattering and the "doubling and adding" method for multiple scattering [Aoki et al., 1997]. We calculated the theoretical spectral albedos for some combinations of snow layer structure and snow impurities and obtained the best fitting of theoretical spectral albedo to the one observed (Figure 1). The theoretical calculations were carried out for two-layer snow models with three geometric depths (d1), 2, 5, and 10 mm, with a density of 0.05 g cm-3 and reff = 25 mm in the first layer and semi-infinite geometric depth with reff = 100 mm in the second layer. The observed geometric depth of the surface layer was 1 cm, but we could not measure the density due to the very low density and thinness of the layer. It is known that the mean density of new snow consisting of spatial dendrites is 0.036-0.059 g cm-3 except in a snowstorm where the minimum is 0.02 g cm-3 [Kajikawa, 1989]. There was no snowstorm when new snow fell on the surface on April 20 at Barrow. We, therefore, assume the density of new snow to be 0.05 g cm-3. For snow impurities in the theoretical calculation, the first snow layer is contaminated with a 0.1 ppmw internal soot mixture and a 2 ppmw external dust mixture. The second snow layer is contaminated by a 0.1 ppmw internal soot mixture and a 1 ppmw external dust mixture.
Fig. 1. Observed spectral snow albedo on April 21, 1997, at Barrow and theoretical results two-snow layer models with three geometric depths (d1), 2, 5, and 10 mm with a density of 0.05 g cm-3 and reff = 25 mm in the first layer, and semi-infinite geometric depth with reff = 100 mm in the second layer. The first snow layer is contaminated by a 0.1 ppmw internal soot mixture and a 2 ppmw external dust mixture. The second snow layer is contaminated by a 0.1 ppmw internal soot mixture and a 1 ppmw external dust mixture.
If the crystal grain size (reff = 250-400 mm) is assumed for snow grain size in the first layer, the theoretical spectral albedo does not agree with the observed one in the near infrared. It means that the optically effective snow grain size is on the order of branch size for the snow of dendrites but is not of the crystal size.
Good agreement is obtained for the model with a geometric depth of d1 = 2 and 5 mm for the first layer at almost all wavelengths. In the case of d1 = 10 mm, which is the value observed in this study, the theoretical spectral albedo is higher than the observed one by 0.05 to 0.1 at the wavelengths 1.0<l<1.35 mm. If we assume the value of snow density to be 0.02 g cm-3, which is the minimum value of Kajikawa [1989], the geometric depths d1 = 2 and 5 mm in Figure 1, respectively, become d1 = 5 and 12.5 mm. These values are consistent with the observed geometric depth.
On the other hand, soot concentration of 0.1 ppmw with an internal mixture in both snow layers leads to good agreement between theoretical and observed spectral albedos. The arctic background concentration of snow impurities ranges from 0.005 to 0.045 ppmw [Warren and Clarke, 1986]. Our value of 0.1 ppmw is twice as high as the value of Warren and Clarke [1986]. Since our observation site was close to Barrow town, there is a possibility that the snow was polluted by locally emitted soot. Dust concentrations of 2 ppmw in the first layer and 1 ppmw in the second layer are consistent with the measured results obtained by using Nuclepore filters.
Acknowledgments. We are indebted to Y. Nakajima and Y. Tsuruga of RESTEC (Remote Sensing Technology Center), and Y. Saruya for their logistic support. We also thank D. Endres and M. Gaylord of CMDL, and G. W. Sheehan of Ukpeagvik Inupiat Corporation/Naval Arctic Research Laboratory for their help in this field experiment. Discussions with K. Stamnes, M. Jeffries, A. J. Alkezweeny and S.-I. Akasofu of the Geophysical Institute, University of Alaska Fairbanks, were fruitful for this study. This work was done as part of the ADEOS Field Campaign supported by NASDA (National Space Development Agency of Japan).
References
Aoki, Te., Ta. Aoki, and M. Fukabori, Approximations for phase function in calculating the spectral albedo of snow surface with multiple scattering. Pap. Meteorol. Geophys., 47, 141-156, 1997.
Grenfell, T. C., and S. G. Warren, Reflection of solar radiation by the Antarctic snow surface at ultraviolet, visible and near-infrared wavelengths. J. Geophys. Res., 99, 18,669-18,684, 1994.
Kajikawa, M., Relation between new snow density and shape of snow crystals, (in Japanese with English abstract) Seppyo, 51, 178-183, 1989.
Warren, S. G., Optical properties of snow. Rev. Geophys. Space Phys., 20, 67-89, 1982.
Warren, S. G., and A. D. Clarke, Soot from arctic haze: radiative effects on the arctic snowpack, Glaciolog. Data, 18, 73-77, 1986.