C41D-01 INVITED 08:00h
Ultra Wideband Radar Mapping of Near Surface Internal Layers: Systems, Results and Analysis
We developed two radar systems for mapping near-surface internal layers. We developed one of these systems to operate over the frequency range 500 to 2000 MHz for surface-based measurements, and the other to operate over the frequency range from 600 to 900 MHz for airborne measurements. Both systems are designed to operate in frequency modulated continuous wave (FM-CW) mode with less than 200 mW of transmit power. We have used the airborne system to collect data over flight lines flown by a NASA P-3 aircraft as a part of NASA's Program for Arctic Regional Climate Assessment (PARCA) initiative during the 2002 and 2003 field seasons. These data show that we can map layers to a depth of about of 150 m in the dry snow zone, 50 m in the percolation zone, and 20 m in the melt zone. During the 2002 field season, one of the flight lines passed over the NASA-U\_1 ice core site with coordinates of $73.84\deg$N and $49.49\deg$W. The ice core was analyzed to determine density at a mean sampling interval of 1.04 m and a variance of 0.05 m. Using these density data, we generated the dielectric profile, which was input into the radar waveform simulator to generate the radar return as a function of depth at the core site. We compared the simulated waveform with the measured data to identify and date a few layers. We tracked the dated layers over a distance of several kilometers to compute spatial and temporal variations in the accumulation rate. During the 2004 field season, we used the surface-based system to collect data over a 10 km x 10 km area at the Summit Camp in Greenland, in conjunction with several in-situ measurements of snow density and layering. The results from the surface-based experiment show that we can map annual layers to a depth of about 200 m and with about 10 cm resolution. We will provide an overview of the radars developed for mapping of near-surface layers and the waveform simulator. We will show results from airborne and surface-based experiments and compare theoretical and experimental data.
C41D-02 INVITED 08:15h
Characterizing Subglacial Interfaces With Airborne Radar Sounding Techniques
Ice sheets are sensitive indicators of global change including sea-level rise. An ice sheet's subglacial interface is an important factor controlling its dynamic behavior. In particular, the grounding zones of ice streams and subglacial lakes are complex systems involving the interaction of the moving ice mass with underlying materials such as liquid water, saturated lubricating tills, and rough or frozen bedrock sticky spots. Imaging and characterizing the subglacial environment of ice sheets is fundamental to understanding these complex systems. Airborne radar sounding is a powerful and well-known technique for studying ice sheets and glaciers and their contiguous underlying environments. We present results from data acquired in 2001 over the ice stream C grounding zone in West Antarctica, as well as over a hypothesized subglacial lake near the South Pole. These data were acquired using a uniquely configured coherent airborne radar system. Our focus has been to characterize the subglacial interface through radar echo analysis based on reflection and scattering theory. The radar system uses a programmable signal source linked to a 10 kW transmitter and a dual-channel coherent down-conversion receiver. The radar operates in chirped pulse mode at 60 MHz with 15 MHz bandwidth. High and low-gain channels allow for recording a wide dynamic range of echoes simultaneously and without range-dependent gain control. Data acquisition includes integrations of 16 returned radar signals about every 15 cm along-track. Pulse compression and synthetic aperture radar (SAR) processing were components of data analysis. Subglacial echoes are influenced by the physical properties of the interface such as the composition and roughness of the materials at the interface. Other important factors include dielectric losses and volumetric scattering losses from propagation through the ice as well as transmission and refraction at the air-ice interface. Unfocussed SAR narrows the along-track radar beam thus increasing resolution at the subglacial interface. Basal reflection coefficients are computed from these data and used for inferring materials at a smooth subglacial interface, most notably when significant quantities of liquid water are present immediately beneath the ice. Echo behavior based on reflection and scattering theory shows that diffuse scattering often dominates the echoes. Scattering analysis consists of Doppler frequency processing to determine the positions of effective scattering centers at the subglacial interface. The along-track offset from the radar to the scattering center indicates the extent of echo sources, thus relating to scattering from the subglacial interface. These results also provide roughness estimates ranging from less than 10 cm for smooth sea-water to 10 m for bottom crevasses. Additional scattering analysis involves imaging based on short integration distances to obtain a low-resolution wide-angle look at the interface. The combined results of the reflection and scattering analysis allows for the classification of distinct subglacial environments including smooth water, very smooth saturated sediments, accreted ice, rough interfaces with bottom crevasses (at a grounding line), and mixed conditions with partial liquid water or interstitial ice.
C41D-03 08:30h
Artificial Variability in Radar-Derived Accumulation Rates: Effects of Differential Deposition
Katabatic winds interact with surface topography to re-distribute snow at local scales on polar ice sheets. The historical record of snowfall is further modified by ice flow away from the original site of deposition. These processes complicate the use of ice cores and other point measurements for studying spatial patterns in accumulation rate and for identifying climate-related (temporal) changes in snowfall. We have been using GPR to obtain spatially-extensive estimates of accumulation rates across large portions of the West Antarctic Ice Sheet. The controlling factor for near-surface estimates of accumulation rate is surface slope, while at depth, historical averages are also controlled by ice dynamics. We discuss a high-resolution, 177 km-long radar profile recorded to 56 m depth in central West Antarctica, which contains strata that vary in depth by as much as 30 m. We present three profiles of accumulation rates vs distance. The rates become more variable with depth, and demonstrate the difficulty of deriving meaningful accumulation estimates for ice sheet mass balance studies.
C41D-04 08:45h
Application of Radio-Echo Sounding Internal Layers to Ice-Sheet Modeling
In recent years, ice-sheet internal layers derived from radio-echo sounding measurements are of particular interest to glaciologists because the layers are believed to contain information on ice flow and Paleoclimate records. Internal layers are considered to represent isochrones and can be dated by combining them with the associated ice cores where the age-depth relationships are known. The dated internal layers then can be used in ice-sheet modelling for model verification or model control. Here, we present the results from ice flow modelling by using dated internal layers in several regions of the Greenland and Antarctic ice sheets. Comparison of the dated internal layers with the isochrones calculated from a steady-state model is used to appraise the stability of the local ice-sheet flow, and to show the importance of ice rheology and stress configurations considered in the model. Using dated internal layers as the model control, ice flow is simulated by adjusting calculations so that the modeled isochrones match the dated layers. Our results also include using the internal layers to reconstruct the historic ice-sheet accumulation rates and thickness change.
C41D-05 09:00h
Ice Thickness and Basal Topography Near the Ross/Amundsen Ice Divide Revealed by Ground-based Radar and New Signal Processing
The U.S. Science Plan for Deep Ice Coring in West Antarctica calls for a new ice core from a site near the Ross Sea/Amundsen Sea ice divide. The region is attractive because very thick ice in the region promises recovery of a long climate record with relatively high time resolution during the last glacial period. Ice thickness (together with accumulation rate history) is needed to estimate depth-age relationships for candidate core sites. However bed echoes from both airborne (Morse et al., Ann. Glac., v 35, 2002) and our ground-based radar profiles are often faint or not detected, especially in regions of very thick ice (up to 3500 m) that are preferred for potential core sites. Here we apply a combination of matched filtering and lateral averaging, which improves the signal-to-noise ratio of both englacial and bed echoes, to our 1.0 and 1.5MHz ground-based radar data collected in the vicinity of the western divide. Matched filtering requires a reflector model, which we derive from a strong, deep radar layer that is observed in nearly all transects. The signal model is consistent with bed echoes at sparse locations where the latter are reliable. Lateral averaging consists of coherent averaging at each point in the echogram, along lines of several slopes, followed by selection of the average of largest magnitude (constrained by an estimate of the maximum reflector slope). Lateral averaging is analogous to performing more stacking during acquisition to reveal fainter reflectors. Compared with data that have been processed by previous standard methods (eg. Gades et al., J. Glac. v 46, 2000), the new processing improves detection of the bed. Of particular interest is a profile along a flow line that crosses the ice divide at 42 km. The new processing clearly illuminates the bed at candidate site E at 15 km on the Ross Sea side of the divide; ice thickness there is 3460m. The bedrock divide is displaced 8 km west from the ice divide. No additional englacial layers were revealed by the processing in this case, and the deep radar stratigraphy is consistent with that found in a 1.5MHz transect that intersects the Byrd core. The basal topography near Site E is rough on the scale of 2-3km but the bed reflector itself is fairly coherent.
C41D-06 09:15h
Inferred accumulation and thickness histories near the Ross/Amundsen divide, West Antarctica
The United States ice coring community has identified the ice divide between the Ross and Amundsen Seas as the next site for a deep U.S. coring effort (Inland WAIS). As part of this program, we have used ground based radar-detected internal layers (assumed to be isochrones) to derive depth-age relationships for a prospective ice core site in the Inland WAIS area (Site E, Morse et al. 2002). Site E is located approximately 30km from the current divide on the Ross Sea side. Our initial depth-age relationship is based on radar layers tracked from the dated ice core at Byrd to the prospective site. The depth of the oldest distinct radar layer (observed at 1.5 MHz) corresponds closely to an acidity event in the Byrd core, which has been attributed to `excessive volcanism' 17,500 yrs BP (Hammer et al., 1994). The traced-radar-derived time scale for Inland WAIS is compared with results from a one-dimensional flow model to infer possible combinations of past thinning, accumulation, and ice sheet geometry at the site since the last glacial maximum (LGM). We calibrate the model and generate an initial accumulation rate profile using the measured layer-thickness and depth-age records from the Siple Dome and Byrd ice cores. Our model is based on the formulation of Dansgaard and Johnsen (1969) and includes the effects of strain-thinning, variations in accumulation rate and ice thickness in the past and basal melting. This approach also allows us to extend the depth-age relationship beyond the oldest distinct radar layer. Our results suggest that the thick ice in the region (3,300m) makes the depth-age relationship relatively insensitive to small changes in ice thickness. Variations in layer thickness on short time scales (relative to the characteristic response time of ice thickness divided by accumulation rate) arise primarily from variations in accumulation rate. Model results also suggest relatively low accumulation ($\sim$50% present) during the LGM but slightly higher than present ($\sim$110%) during the early to mid-Holocene.
C41D-07 09:30h
Extracting Accumulation And Vertical Strain Rate Information From Internal Layering in Ice Penetrating Radar Profiles Over the Greenland Ice Sheet: Basic Assumptions, Recognition of Potential Errors, and Techniques for Large-Area Retrievals
Tracing of internal layering in aircraft-based-radar profiles of the upper 1500 meters of the ice in the Greenland Ice Sheet allows extension of the age-depth information at deep ice core sites to other parts of the ice sheet. The fitting of a simple model, based on continuity and the assumption of constant rates of strain and accumulation (but not knowledge of ice thickness), allows an estimation of both of these parameters at any point along these profiles where internal layering is well developed. This has been reported on in several places. In this paper, we address the issues behind the use of the relationship between accumulation rate and vertical strain rate at a site to constrain the ice flow responsible for the vertical strain, with a view toward both the potential and limitations of the technique. For example, with the addition of information about the ice thickness from the same radar data, it is possible to show that large areas of the northeast quadrant of the ice sheet are sliding over the bed, and that there are localized areas of sliding in the northwest as well. This method provides a constraint on basal conditions that is useful in modeling the evolution of the ice sheet over time. While there are complications to the layer-tracing approach related to the assumptions of constant accumulation and strain rates, the impact of violations of these assumptions can be constrained from the data. A comparison of 10,000 line kilometers of internal-layer-derived accumulation rate with the NASA PARCA program surface-based accumulation map shows strong agreement in both accumulation rate and spatial variations in that rate, in spite of the longer sampling interval in the internal-layer based estimates.
C41D-08 09:45h
Dating Ice Divides in West Antarctica by the Operation of the Raymond Effect
Once an ice-divide forms, isochronic layers become locally upwarped through the operation of the well-known Raymond Effect. The distribution of bump-amplitude with elevation (BAE curve) evolves with time to a steady configuration that depends on the ice rheology, temperature, the accumulation rate and the chaning geomety of the ice mass. Unsaturated BAE curves can be used in principle to determine the age of the divide. We apply these techniques to several locations in West Antarctica, where isochrones have been located using radar echo sounding. Firstly, we repeat the investigations into Roosevelt Island reported by Conway et al (1998), using a more sophisticated set of calculations which include thermo-mechanical coupling. Their dating results can be obtained, but only with a relatively high rheological index. We next consider the divide in Marie Byrd Land between the Siple Coast and the Amundsen Sea Sector. Here the bump amplitudes are less than their theoretical saturated value, and we explore rheological and thermal hypotheses for this related to the likely presence of sliding. Finally we consider the Hercules dome. Here the bumps are shown to be comparable in size to their saturated steady state value.