Richard E. Brandt

Effect of surface roughness on bidirectional reflectance of Antarctic snow

The angular pattern of sunlight reflected by snow is altered by surface roughness, which in the interior of Antarctica is usually in the form of meter-scale longitudinal erosional features (sastrugi), whose axes align with the direction of strong winds. The bidirectional reflectance distribution function (BRDF) changes over the course of a day as the solar azimuth changes relative to the sastrugi axis. The normalized BRDF, or “anisotropic reflectance factor” R, was measured at South Pole Station from a 22-m tower at 600, 660, and 900 nm wavelengths. The R pattern was similar at the three wavelengths; it probably varies little from 300 to 900 nm. Measurements were made at solar zenith angles θ0 from 67° to 90°, over the full range of viewing zenith angle (θr), azimuth angle between Sun and view (ϕ), and azimuth angle between Sun and sastrugi (ϕsas). Variation of R with ϕsas was notable; sastrugi oriented perpendicular to the solar beam cause a reduction of the forward peak, and sastrugi at an oblique angle cause R to lose its symmetry about the solar azimuth. However, the effects of sastrugi are mostly restricted to large viewing zenith angles, so remote sensing of albedo and atmospheric properties can be carried out accurately without knowledge of sastrugi height and orientation if only near-nadir views are used. This recommendation is opposite that for observations of broken clouds over dark surfaces, for which large θr is preferred. A parameterization of R is developed, valid for viewing angles θr ≤ 50°. Sastrugi can cause a reduction of the snow albedo by altering the angle of incidence and by trapping of photons. For the small sastrugi of the Antarctic Plateau, the albedo is unaffected at visible wavelengths but can be reduced by a few percent at near-infrared wavelengths when the Sun is low.

East Antarctic sea ice: Albedo, thickness distribution, and snow cover

Characteristics of springtime sea ice off East Antarctica were investigated during a cruise of the Australian National Antarctic Research Expedition in October through December 1988. The fractional coverage of the ocean surface, the ice thickness, and the snow cover thickness for each of several ice types were estimated hourly for the region near the ship. These observations were carried out continuously during the 4 weeks the ship was in the ice. Thin and young ice types were prevalent throughout the region, and the observations show a systematic increase in the total area-weighted pack ice thickness (including open water area) from only 0.2 m within 50 km of the ice edge to 0.45 m close to the coast. Ice thickness averaged over the ice-covered region only is also relatively thin, ranging from 0.35 m near the ice edge to 0.65 m in the interior. These values are probably typical of average winter thickness for the area. The average snow cover thickness on the ice increased from 0.05 m near the ice edge to 0.15 m in the interior. Average ice concentration increased from less than 6/10 near the ice edge to 8/10 in the interior. The ship-observed concentrations were in good agreement with concentrations derived from passive microwave satellite imagery except in some regions of high concentration. In these regions the satellite-derived concentrations were consistently lower than those estimated from the ship, possibly because of the inability of the satellite sensors to discriminate the appreciable percentage of very thin ice observed within the total area. Spectral albedo was measured for nilas, young grey ice, grey-white ice, snow-covered ice, and open water at wavelengths from 420 to 1000 nm. Allwave albedo was computed by using the spectral measurements together with estimates of near-infrared albedo and modeled spectral solar flux. Area-averaged albedos for the East Antarctic sea ice zone in spring were derived from representative allwave albedos together with the hourly observations of ice types. These area-averaged surface albedos increased from about 0.35 at the ice edge to about 0.5 at 350 km from the edge, remaining at 0.5 to the coast of Antarctica. The low average albedo is in part due to the large fraction of open water within the pack, but extensive fractions of almost snow-free thin ice also play an important role.

Surface Albedo of the Antarctic Sea Ice Zone

In three ship-based field experiments, spectral albedos were measured at ultraviolet, visible, and nearinfrared wavelengths for open water, grease ice, nilas, young “grey” ice, young grey-white ice, and first-year ice, both with and without snow cover. From the spectral measurements, broadband albedos are computed for clear and cloudy sky, for the total solar spectrum as well as for visible and near-infrared bands used in climate models, and for Advanced Very High Resolution Radiometer (AVHRR) solar channels. The all-wave albedos vary from 0.07 for open water to 0.87 for thick snow-covered ice under cloud. The frequency distribution of ice types and snow coverage in all seasons is available from the project on Antarctic Sea Ice Processes and Climate (ASPeCt). The ASPeCt dataset contains routine hourly visual observations of sea ice from research and supply ships of several nations using a standard protocol. Ten thousand of these observations, separated by a minimum of 6 nautical miles along voyage tracks, are used together with the measured albedos for each ice type to assign an albedo to each visual observation, resulting in “ice-only” albedos as a function of latitude for each of five longitudinal sectors around Antarctica, for each of the four seasons. These ice albedos are combined with 13 yr of ice concentration estimates from satellite passive microwave measurements to obtain the geographical and seasonal variation of average surface albedo. Most of the Antarctic sea ice is snow covered, even in summer, so the main determinant of area-averaged albedo is the fraction of open water within the pack.

Visible and near-ultraviolet absorption spectrum of ice from transmission of solar radiation into snow

Snow is a scattering-dominated medium whose scattering is independent of wavelength at 350-600 nm. The attenuation of solar radiation in snow can be used to infer the spectral absorption coefficient of pure ice, by reference to a known value at 600 nm. The method is applied to clean Antarctic snow; the absorption minimum is at 390 nm, and the inferred absorption coefficient is lower than even the lowest values of the Antarctic Muon and Neutrino Detector Array (AMANDA) experiment on glacier ice: The absorption length is at least 700 m, by comparison with 240 m for AMANDA and 10 m from laboratory attenuation measurements.

The ablation zone in northeast Greenland: ice types, albedos and impurities

Ice types, albedos and impurity content are characterized for the ablation zone of the Greenland ice sheet in Kronprinz Christians Land (80° N, 24° W). Along this ice margin the width of the ablation zone is only about 8 km. The emergence and melting of old ice in the ablation zone creates a surface layer of dust that was originally deposited with snowfall high on the ice sheet. This debris cover is augmented by locally derived wind-blown sediment. Subsequently, the surface dust particles often aggregate together to form centimetre-scale clumps that melt into the ice, creating cryoconite holes. The debris in the cryoconite holes becomes hidden from sunlight, raising the area-averaged albedo relative to surfaces with uniform debris cover. Spectral and broadband albedos were obtained for snow, ice hummocks, debris-covered ice, cryoconite-studded ice and barren tundra surfaces. Broadband ice albedos varied from 0.2 (for ice with heavy loading of uniform debris) to 0.6 (for ice hummocks with cryoconite holes). The cryoconite material itself has albedo 0.1 when wet. Areal distribution of the major surface types was estimated visually from a transect video as a function of distance from the ice edge (330 m a.s.l.). Ablation rates were measured along a transect from the ice margin to the slush zone 8 km from the margin (550 m a.s.l.), traversing both Pleistocene and Holocene ice. Ablation rates in early August averaged 2 cm d -1 . Impurity concentrations were typically 4.3 mg L -1 in the subsurface ice. Surface concentrations were about 16 g m -2 on surfaces with low impurity loading, and heavily loaded surfaces had concentrations as high as 1.4 kg m -2 . The mineralogical composition of the cryoconite material is comparable with that of the surrounding soils and with dust on a snowdrift in front of the ice margin, implying that much of the material is derived from local sources. A fine mode (clay) is present in the oldest ice but not in the nearby soil, suggesting that its origin is from wind deposition during Pleistocene glaciation.

Spectral Bidirectional Reflectance of Antarctic Snow: Measurements and Parameterization

The bidirectional reflectance distribution function (BRDF) of snow was measured from a 32-m tower at Dome C, at latitude 75°S on the East Antarctic Plateau. These measurements were made at 96 solar zenith angles between 51° and 87° and cover wavelengths 350–2400 nm, with 3- to 30-nm resolution, over the full range of viewing geometry. The BRDF at 900 nm had previously been measured at the South Pole; the Dome C measurement at that wavelength is similar. At both locations the natural roughness of the snow surface causes the anisotropy of the BRDF to be less than that of flat snow. The inherent BRDF of the snow is nearly constant in the high-albedo part of the spectrum (350–900 nm), but the angular distribution of reflected radiance becomes more isotropic at the shorter wavelengths because of atmospheric Rayleigh scattering. Parameterizations were developed for the anisotropic reflectance factor using a small number of empirical orthogonal functions. Because the reflectance is more anisotropic at wavelengths at which ice is more absorptive, albedo rather than wavelength is used as a predictor in the near infrared. The parameterizations cover nearly all viewing angles and are applicable to the high parts of the Antarctic Plateau that have small surface roughness and, at viewing zenith angles less than 55°, elsewhere on the plateau, where larger surface roughness affects the BRDF at larger viewing angles. The root-mean-squared error of the parameterized reflectances is between 2% and 4% at wavelengths less than 1400 nm and between 5% and 8% at longer wavelengths.

A Look at the Surface-Based Temperature Inversion on the Antarctic Plateau

Data from radiosondes, towers, and a thermistor string are used to characterize the temperature inversion at two stations: the Amundsen-Scott Station at the South Pole, and the somewhat higher and colder Dome C Station at a lower latitude. Ten years of temperature data from a 22-m tower at the South Pole are analyzed. The data include 2- and 22-m temperatures for the entire period and 13-m temperatures for the last 2 yr. Statistics of the individual temperatures and the differences among the three levels are presented for summer (December and January) and winter (April–September). The relationships of temperature and inversion strength in the lowest 22 m with wind speed and downward longwave flux are examined for the winter months. Two preferred regimes, one warming and one cooling, are found in the temperature versus longwave flux data, but the physical causes of these regimes have not been determined. The minimum temperatures and the maximum inversions tend to occur not with calm winds, but with winds of 3–5 m s 1 , likely due to the inversion wind. This inversion wind also explains why the near-surface winds at South Pole blow almost exclusively from the northeast quadrant. Temperature data from the surface to 2 m above the surface from South Pole in the winter of 2001 are presented, showing that the steepest temperature gradient in the entire atmosphere is in the lowest 20 cm. The median difference between the temperatures at 2 m and the surface is over 1.0 K in winter; under clear skies this difference increases to about 1.3 K. Monthly mean temperature profiles of the lowest 30 km of the atmosphere over South Pole are presented, based on 10 yr of radiosonde data. These profiles show large variations in lower-stratospheric temperatures, and in the strength and depth of the surface-based inversion. The near-destruction of a strong inversion at South Pole durin g7ho n 8September 1992 is examined using a thermal-conductivity model of the snowpack, driven by the measured downward longwave flux. The downward flux increased when a cloud moved over the station, and it seems that this increase in radiation alone can explain the magnitude and timing of the warming near the surface. Temperature data from the 2003/04 and 2004/05 summers at Dome C Station are presented to show the effects of the diurnal cycle of solar elevation over the Antarctic Plateau. These data include surface temperature and 2- and 30-m air temperatures, as well as radiosonde air temperatures. They show that strong inversions, averaging 10 K between the surface and 30 m, develop quickly at night when the sun is low in the sky, but are destroyed during the middle of the day. The diurnal temperature range at the surface was 13 K, but only 3 K at 30 m.

Comment on ''Snowball Earth: A thin-ice solution with flowing sea glaciers'' by David Pollard and James F. Kasting

[1] Pollard and Kasting [2005] (hereinafter referred to as PK) have coupled an energy-balance climate model to an ice-shelf flow model, to investigate the Snowball Earth episodes of the Neoproterozoic, 600–800 million years ago, when the ocean apparently froze all the way to the equator [Hoffman and Schrag, 2002]. PK’s particular concern was to investigate the possibility that over a wide equatorial band where sublimation exceeded snowfall, the bare ice may have been thin enough to permit transmission of sunlight to the water below, providing an extensive refugium for the photosynthetic eukaryotes that survived the Snowball events. This possibility was first proposed by McKay [2000], whose model of radiative transfer and heat conduction predicted tropical ice only a few meters thick for ice albedo up to 0.7, under assumed conditions of sunlight and temperature at the Snowball equator. However, when spectral resolution was incorporated into the radiative transfer model [Warren et al., 2002], the ice albedo had to be reduced below 0.4 to obtain the thin-ice solution under otherwise identical conditions. It seemed unlikely that such dark sea ice could avoid melting under the equatorial Sun. But even if it could avoid melting, the thin ice would risk being crushed by the inflow of kilometer-thick ‘‘sea glaciers’’ from higher latitudes [Goodman and Pierrehumbert, 2003]. To further investigate the feasibility of tropical thin ice, it was therefore necessary to couple a climate model to models of sea-ice thermodynamics and sea-glacier flow. This is what PK have done. Surprisingly, they conclude that thin ice (<3 m thick) ‘‘may have prevailed’’ in a 20-degree latitude band centered on the equator. We argue here that this conclusion is too optimistic, because PK’s thin-ice solution apparently required that several controlling variables be set outside their measured ranges: the albedo of cold glacier ice, the depth of transition from snow to ice, and the thermal conductivity of ice. We then raise the more general question of how wide is the thin-ice domain in the parameter-space of model variables. 2. Choices of Model Variables That Favor Thin Ice 2.1. Albedo of Cold Glacier Ice [2] As sea glaciers flowed equatorward into the tropical region of net sublimation, their surface snow and subsurface firn would sublimate away, exposing bare glacier ice to the atmosphere and solar radiation. This ice would be freshwater (meteoric) ice, which originated from compression of snow, so it would contain numerous bubbles, giving a high albedo. The albedo of cold (nonmelting) glacier ice exposed by sublimation (Antarctic ‘‘blue ice’’) has been measured as 0.55–0.65 in four experiments in the Atlantic sector of Antarctica [Bintanja and van den Broeke, 1995; Bintanja et al., 1997; Liston et al., 1999; Reijmer et al., 2001], 0.63 in the Transantarctic Mountains [Warren et al., 2002], and 0.66 near the coast of East Antarctica [Weller, 1968]. (Weller’s blue-ice albedo is often quoted as 0.69 [e.g., Weller, 1980; King and Turner, 1997], but that is an average value, which included a few times with patchy snow cover on the ice.) Not all glacier ice is bubbly: under several hundred meters of ice thickness, the pressure becomes so great that each air bubble dissolves in the ice to form a clathrate crystal, and the ice becomes relatively clear [Price, 1995]. However, when this ice becomes depressurized as its cover sublimates away (as in the Antarctic blue-ice areas where the albedos were measured), the bubbles reform [Lipenkov, 2000], so that the albedo of 0.55–0.66 is observed at the surface. [3] In contrast to glacier ice, which forms by compression of snow, sea ice forms by freezing of liquid water. In their model, PK tried two values of scattering coefficient, each of which they applied uniformly to both sea ice and glacier ice. With albedo 0.47 (a reasonable value for meter-thick snowfree sea ice [Brandt et al., 2005, Figure 1]), the sea glaciers terminated at 9 degrees latitude, leaving an equatorial strip of thin ice. But with albedo 0.64 (i.e., within the measured range for glacier ice), the sea glaciers advanced to the equator, with a thickness of 1 km. To add realism to the model, it would be good to assign different values of scattering coefficient to these two very different types of ice.

Parameterizations for narrowband and broadband albedo of pure snow and snow containing mineral dust and black carbon

The reduction of snow spectral albedo by black carbon (BC) and mineral dust, both alone and in combination, is computed using radiative transfer modeling. Broadband albedo is shown for mass fractions covering the full range from pure snow to pure BC and pure dust, and for snow grain radii from 5 µm to 2500 µm, to cover the range of possible grain sizes on planetary surfaces. Parameterizations are developed for opaque homogeneous snowpacks for three broad bands used in general circulation models and several narrower bands. They are functions of snow grain radius and the mass fraction of BC and/or dust and are valid up to BC content of 10 ppm, needed for highly polluted snow. A change of solar zenith angle can be mimicked by changing grain radius. A given mass fraction of BC causes greater albedo reduction in coarse-grained snow; BC and grain radius can be combined into a single variable to compute the reduction of albedo relative to pure snow. The albedo reduction by BC is less if the snow contains dust, a common situation on mountain glaciers and in agricultural and grazing lands. Measured absorption spectra of mineral dust are critically reviewed as a basis for specifying dust properties for modeling. The effect of dust on snow albedo at visible wavelengths can be represented by an “equivalent BC” amount, scaled down by a factor of about 200. Dust has little effect on the near-IR albedo because the near-IR albedo of pure dust is similar to that of pure snow.

Transmission of Solar Radiation by Clouds over Snow and Ice Surfaces: A Parameterization in Terms of Optical Depth, Solar Zenith Angle, and Surface Albedo

A multilevel spectral radiative transfer model is used to develop simple but accurate parameterizations for cloud transmittance as a function of cloud optical depth, solar zenith angle, and surface albedo, for use over snow, ice, and water surfaces. The same functional form is used for broadband and spectral transmittances, but with different coefficients for each spectral interval. When the parameterization is applied to measurements of ‘‘raw’’ cloud transmittance (the ratio of downward irradiance under cloud to downward irradiance measured under clear sky at the same zenith angle), an ‘‘effective’’ optical depth t is inferred for the cloud field, which may be inhomogeneous and even patchy. This effective optical depth is only a convenient intermediate quantity, not an end in itself. It can then be used to compute what the transmittance of this same cloud field would be under different conditions of solar illumination and surface albedo, to obtain diurnal and seasonal cycles of cloud radiative forcing. The parameterization faithfully mimics the radiative transfer model, with rms errors of 1%‐2%. Lack of knowledge of cloud droplet sizes causes little error in the inference of cloud radiative properties. The parameterization is applied to pyranometer measurements from a ship in the Antarctic sea ice zone; the largest source of error in inference of inherent cloud properties is uncertainty in surface albedo.