C12B-01 10:25h
Variability of Decimetre and Centimetre Scale Ice Surface Roughness and the Potential Consequences on the CryoSat Radar Altimeter Signal
Snow and ice surface roughness affect the backscatter of the pulse emitted by a radar altimeter, and hence the accuracy of the surface elevation calculated from the waveform echo, but the influence of surface roughness has not been quantified. As part of the CryoSat calibration/validation field campaigns on the Devon Ice Cap in 2004, surface roughness measurements were made at 0.1-7km intervals along a 48km transect from near the summit to the southern margin. Measurements were made at the decimetre scale by surveying and at the centimetre scale using digital photography. The data collected were subjected to wavelet analysis to define characteristic roughness wavelengths, and the fractal dimension associated with each of these was calculated using the semi-variogram method. Vario functions were calculated for the photographic data. The survey results show that wavelength scales depend on orientation and distance from the ice cap summit, the fractal dimension depends on the wavelength scale and the orientation, and both are significantly affected by storm events. Profiles aligned with the easterly prevailing wind direction, and thus perpendicular to the predicted satellite track, proved to be more sensitive to meteorological events than those normal to the dominant winds. Wavelet and fractal analysis of the photographic data was less conclusive, potentially due to the `noisier' nature of the data at this scale, where `noise' is actually the superimposition of small scale wavelengths onto larger ones. Vario analysis showed the characteristic wavelengths at the centimetre scale to increase with distance from the summit, although the abrading effect of storm events caused a decrease in wavelength. The amplitude of the roughness also increases with distance from the summit, although following a period of calm this value is significantly decreased along the transect. Orientation with respect to the prevailing wind direction is also a significant factor. Analysis of the return waveforms acquired by an airborne radar altimeter concurrently with ground data will allow the impact of the different roughness scales and orientations to be assessed.
C12B-02 10:40h
Observations of Surface Roughness at Several Scales in Jakobshavn Isbrae Drainage Basin, Greenland
The ice surface in the region of Jakobshavn Isbrae and its drainage basin exhibits a fascinating wealth and complexity of surface forms, which are the manifestation of glaciologic and environmental morphogenetic processes operating at different scales. The fast-moving Jakobshavns Isbrae stands out in any satellite image by its heavily crevassed surface, with surface provinces that extend kilometers to tens of kilometers and crevasses of 10-100 meters, whereas the ice surface in the surrounding slow-moving ice is characterized by much smaller features such as sastrugi, melt ponds, wind- and ablation ridges and smooth snow, which are equally important in understanding the complexity of a changing environment. The observation of such complexity necessitates different scale-dependent observation methods including field, aerial and satellite data and a definition of surface roughness that allows to capture the spatial complexity of the interface between snow/ice and atmosphere. Analysis methods range from structural geology and continuum mechanics to geostatistical surface characterization and classification. Application to Jakobshavns Isbrae is of particular interest as the ice-stream system has long been considered the prototype of a time-independent system but has lately exhibited dramatic changes.
C12B-03 10:55h
Analysis of Surface Roughness Derived from Airborne Laser Altimetry on the Greenland Ice Sheet and Comparison with Stratigraphic Records
While repeat satellite altimetry and volume balance studies indicate that the central parts of the Greenland Ice Sheet is in balance, repeat airborne laser altimetry have revealed significant thinning of the ice sheet at lower elevations, enough to raise sea level by about 0.15 mm/yr. A major impediment for assessing the long-term significance of observed changes is the relatively short record of observations currently available. Over time intervals of a few years, the elevation of the ice surface responds to interannual fluctuations in snowfall which may lead to apparent thickening or thinning trends that are not sustained over longer time periods. Since 1995, a suite of shallow to intermediate depth cores has been collected in Greenland by NASA's Program for Arctic Regional Climate Assessment (PARCA) to support the efforts to establish the mass balance of the ice sheet. One major difficulty in extracting inter-annual variability from these cores is to estimate the contribution of small-scale spatial variability influenced by surface roughness such as sastrugi. Previously we demonstrated the use of airborne laser profiling for estimating surface roughness in Greenland's Summit region. Now we extend this analysis to other parts of the Greenland ice sheet. High resolution surface elevation data were collected by NASA's Airborne Topographic Mapper around several ice core sites in 1995. To compute the surface roughness, first the large-scale slope is removed by piecewise fitting of linear segments using least-squares method. Then the surface roughness is computed as the standard deviation of the elevation fluctuations for flight segments of predetermined length. To obtain the estimated variance of interannual variability, the variance of the surface effects (roughness) was subtracted from the total variance of the ice core records. We also analyzed the spatial coherence of surface irregularities by computing correlation functions and variograms. Where differently-oriented flight lines were available, spatial anisotropy was also investigated. As we demonstrate, the spatial coherence obtained from the high resolution topography provides important information on the optimum spacing of multiple cores.
C12B-04 11:10h
Determination of Snow and Ice Surface Roughness and its Importance for Ablation
Study of the surface roughness of snow fields, glaciers, and ice sheets requires measurement and analysis of the surface's three-dimensional features, anisotropies, and complex microtopography. Observing that the notions of relief and surface roughness differ only with respect to scale, we consider surface roughness a spatial variable defined as the derivative of (micro)topography. Spatial snow and ice surface roughness can be measured with the Glacier Roughness Sensor, a multichannel instrument that collects data at 0.2~m across-track, 0.1~m along-track resolution and subcentimeter vertical accuracy, with differential kinematic GPS data for positioning. Roughness data are analysed using the geostatistical classification method. Results of the classification provide (1) information on the morphogenesis of snow or ice surface types of a given environment, (2) subscale information for the interpretation of satellite data and facilitate (3) segmentation of a study area into characteristic surface classes. One application is monitoring the extent of ablation in the Greenland Ice Sheet. By deriving a mathematical relationship between aerodynamic roughness length and spatial surface roughness, calculating roughness from actual observations and driving energy balance models with the range of resultant values and micro-meteorological data, it could be shown that melt energy varies with a factor of two dependent on surface roughness. Hence surface roughness is an important geophysical variable that needs to be considered in any scenario of a warming ice sheet.
C12B-05 11:25h
Surface roughness change associated with fresh snow metamorphosis
Snow surface roughness is an important in understanding surface-atmosphere interactions, and is used in bulk transfer modeling of sublimation. At the plot scale and finer, it is hypothesized that the surface roughness of a snowpack varies rapidly after a fresh snowfall as the snowpack metamorphoses. This change is driven by wind and radiation loading. Four laboratory experiments were performed to estimate mass loss to sublimation under controled conditions. Surface and near surface meteorological variables were measured and used to estimate sublimation from bulk transfer modeling. Comparison of the measured and modeled sublimation indicates that there is a rapid decrease in the roughness of the snow surface as the crystals at and near the surface undergo destructive metamorphosis. Daily measurements of meteorology and sublimation by {\it Williams} (1959) were used with an estimate of time since the last snowfall events and indicated the opposite of the laboratory experiments; for modeled sublimation estimates to equal sublimation measurements, the snow surface roughness must be less for fresh snow than for aged snow. Other datasets [e.g., {\it Sverdrup}, 1936] are examined to illustrate the systematic change in snowpack surface roughness. Sverdrup, H.U., The eddy conductivity of the air over a smooth snow field, {\it Geofysiske Publikasjoner}, 11(7), 69pp, 1936. Williams, G.P., {\it Evaporation from snow covers in Eastern Ontario}, National Research Council of Canada, Division of Building Research, Research Paper No. 73, 1959.
C12B-06 11:40h
Snow HDRF Measurements on Various Snow Surfaces with the new IAC-Gonio-Spectrometer
This work presents a field Gonio-Spectrometer developed at the Institute for Atmospheric and Climate Science, Swiss Federal Institute of Technology, Zurich (IAC-ETH). The main motivation to built this Gonio-Spectrometer was the study of the Hemispherical Distribution Reflectance Factor (HDRF) of dry snow on the Ice Sheet of Greenland and to examine the influence of the HDRF on the surface energy balance. The surface Albedo is of great importance for both, large scale and small scale climate modelling and energy balance studies. Especially for remote regions, satellites provide an extraordinary means to measure reflected sunlight. However, raw satellite data have to undergo several corrections depending on the viewing angles of the sensors relative to the targets and the irradiance source (sun). The function that describes the distribution of reflected radiance with angle is called Bidirectional Reflectance Distribution Function (BRDF). The HDRF is the commonly used dimensionless form of the angular distribution of reflectance. BRDF and HDRF are a functions of four angles: incoming (solar) zenith angle and azimuth, and outgoing (reflected) zenith angle and azimuth. In situ measurements of HDRF data, a combination of multidirectional and hyperspectral data, require complex and demanding experiments. Therefore, existing data sets a rare. However, the advent of new satellite systems that offer hyperspectral resolution and off-nadir tilting capability ask for ground truth data sets. The IAC-Gonio-Spectrometer measures the HDRF with a distance of 1 meter between sensor and target. The sensor, an optic cable, can be placed on an arbitrary place on the hemisphere and always points towards the same surface area. Depending on the viewing geometry, the diameter of the footprint area varies from 5~cm (at nadir) to 20~cm (at 75 degree zenith angle). The pointing accuracy, analyzed in a laboratory experiment with a laser beam, was measured at $\pm$ 2.5~cm. In the summer field season of 2004, ninety complete samplings of snow HDRF were accomplished on the Ice Sheet of Greenland. The measurements were made for various solar illumination geometries and various snow surface properties. The snow surface changed from very smooth, either after a snowfall or during wind drift, to very rough, after rime incidents. New snow showed the highest anisotropy in the HDRF distribution while a rough surface exhibited a more isotropic distribution. Moreover, the HDRF became more anisotropic with increasing solar zenith angle with values over 200% for wavelength around 1000~nm.
C12B-07 11:55h
A 44 kyr Paleo-Roughness Record From Antarctica
Two 788 m high resolution conductivity records from ice cores drilled at Dome C, Antarctica, provide an unprecedented opportunity to examine the past roughness of the ice sheet surface. We measured the distribution of the depth differences between matched synchronous events in the cores. This enabled the mean surface height distribution to be estimated for time intervals in the past. The technique was confirmed by checking the results against the conductivity profile of a third nearby core (120m depth), enabling uncertainties to be established. The mean standard deviation of surface roughness was 0.029 m for 0 to 11.5 kyr BP and 0.031 m for 18 to 45 kyr BP in ice equivalence. A record of this nature allows us to assess the reliability of the fine scale data in an ice core record. Specifically how well the preserved information represents the mean conditions at the time of deposition and the extent to which short term events may be missing from the core. This is of particular interest when matching volcanic horizons with other paleo records and when studying rates of volcanism. We find the probability of a typical volcanic signal being indistinguishable from background noise is ~ 6% during the Holocene but ~ 26% during the last glacial period. The 44 kyr roughness record, originating from dunes and sastrugi at the ice sheet surface, is likely to be a function of past wind speed, accumulation rate and temperature. Future developments constraining this relationship should allow conclusions regarding past local wind speed to be drawn from the record.