Introduction
Patagonian radar glacier zonesFig. 3. Map showing areas studied for determination of radar glacier zones
Fig. 4. Multi-component false-colored images showing seasonal changes in radar zones.
Fig. 5. Radar glacier zone decision tree classifier
Fig. 6. Classification results for SIR- C scenesEffects of weather events
British Columbia radar glacier zones
Monitoring glacier surface conditions with synthetic aperture radar (SAR) has the potential to detect surface and near-surface conditions related to melting and freezing, and thus has the potential for monitoring the effects of climate change on glaciers. For example, the four "glacial facies" developed by Benson [1961] and Williams et al. [1991] for Greenland, namely dry snow, percolation, wet snow, and ablation facies, have been successfully mapped in Greenland by SAR images from ERS 1 by Fahnestock et al. (1993). Monitoring changes in the boundaries of these facies could give an indication of a climate-related change in the mass balance of Greenland or other polar glaciers.
The definitions of the Greenland glacier facies are based on stratigraphic conditions that are established over an entire melt season, and the facies boundaries can be revealed by SAR only in the winter through a dry snow cover. These high latitude facies, however, do not usefully apply to temperate glaciers, which, in terms of the Greenland definitions, have mainly only two facies, "wet snow" and "ablation" (Paterson, 1994). In our studies of the South Patagonian Icefield, we are able to use the multi-parameter SAR data to define distinct, mappable and physically interpretable units which we describe as "glacier radar zones". These zones reflect processes that have operated during a timescale probably closer to days and weeks rather than the annual timescales implied by the glacier facies concept. We therefore refer to these as "radar glacier zones" rather than glacier facies.
Monitoring the location of these radar glacier zones through the yearly cycle may permit us to extend the facies monitoring strategy of Greenland to highly active temperate glaciers. The intraseasonal response of the glacier to meteorological conditions can be investigated through correlation of the spatial and temporal characteristics of the radar glacier zones with station and satellite meteorological observations.
Our delineation of radar glacier zones are developed in three papers, one published and two in press, as follows:
1. Forster, R.R., B.L. Isacks, and S.B. Das, 1996, Spaceborne imaging radar (SIR-C/X-SAR) reveals near-surface properties of the South Patagonian Icefield, J. Geophys. Res. (Planets), 101, 23,169-13,180.
2. Forster, R..R., L.C. Smith, and B. L. Isacks, 1997, Effects of Weather Events on X-SAR Returns from Icefields: A Case study of the Hielo Patagonico Sur, in press, Ann. Glaciology.
3. Smith, L.C., R.R. Forster, and B.L. Isacks, 1997, Seasonal climatic forcing of alpine glaciers revealed by orbital synthetic aperture radar, in press, J. Glaciology.
Abstracts from these papers are given in this report together with key figures from paper 1 listed above.
Forster, et al. 1996
Shuttle imaging radar C/X band synthetic aperture radar (SIR-C/X-SAR) views of the South Patagonian Icefield in southern Chile and Argentina demonstrate the ability of spaceborne multiparameter radar to detect climatically driven, intra-annual changes in the snow and ice conditions on glaciers. The SIR-C/X-SAR system aboard space shuttle Endeavor acquired images during two 11-day missions in April and October 1994. The radar signatures of differing snow and ice conditions are distinctive and homogeneous over large areas of the ice fields. The signatures are characterized mainly by (1) backscatter amplitude levels, (2) relative amplitudes of the C and L bands (wavelengths of 5.7 and 24 cm, respectively), and (3) polarization properties indicative of volumetric or surface scattering. The radar signatures are interpreted by correlating the radar characteristics with elevation of the snow and ice surfaces and with changes in the meteorological conditions. We are able to define four "radar glacier zones": (zone A) a relatively dry snow zone with dominant C band returns; (zone B) a moderately wet snow zone with dominant L band returns; (zone C) a wet snow zone with weak returns in all bands; and (zone D) bare ice and/or heavily crevassed surfaces with strong returns in all bands. The spatial changes in the radar glacier zones between April and October are consistent with colder temperatures recorded in October, producing drier snow conditions at lower elevations than in April.
Figure 3. The northern part of the South Patagonian Icefield with the locations of the SIR-C/X-SAR scenes used to define the radar glacier zones.
Figure 4. SIR-C image of a portion of the South Patagonian Icefield acquired from orbit 77 on April 14, 1994 (top), and October 5, 1994 (bottom), with red C- band horizontal transmit and receive; green, L- band horizontal transmit; and blue, L-HV. The image locations are shown in Fig. 3, and the size is 100 km x 57 km with radar illumination from the top. Click on the image below to see the higher resolution version (242K GIF file).
The most remarkable characteristics of these images is the striking division of the ice fields and glaciers into large, homogeneous regions of similar backscatter characteristics. These zones change dramatically between the April and October images. Detailed analyses of the multi-parameter radar backscatter signals in relation to their spatial distribution on the icefields leads to the definition of radar glacier zones illustrated in Figs. 5 and 6.
Figure 5. Radar glacier zone decision tree classifier based on April and October SIR-C scenes from orbit 57. The symbols are as follows: LHH or CHH is L-band or C-band horizontal component transmit and horizontal component receive; and CHV or LHV is C-band or L-band horizontal transmit and vertical receive. This decision tree classifier is applied to the images depicted in Fig. 6.
Zone |
radar signature |
interpretation |
A |
strong CHH, C dominant, low CVV/CHH | drier snow conditions with snow grain scattering of C. |
B |
strong LHH, L dominant, C decreased from zone A | wetter snow conditions with L backscatter from subsurface inclusions |
C |
weak L and C, high CVV/CHH | thick wet snow absorbing all radar bands |
D |
C and L comparable and strong for HH and HV | rough bare ice or heavily crevassed regions |
Figure 6. Classification results for SIR-C scenes, based on decision tree classifier in Fig. 5 for orbit 57, April 12 (a) and Oct. 3 (b); orbit 73, April 13 (c) and Oct. 4 (d); and orbit 61, April 13 (e) and Oct. 4 (f). The locations of these scenes are indicated in blue in Fig. 3. Click on figure below to see higher resolution image (253K GIF file). Note the consistent seasonal changes in zones.
Forster, et al., 1997
The space shuttle based SIR-C/X-SAR synthetic aperture radar (SAR) imaged a portion of the South Patagonian Icefield for five successive days during missions in April and October 1994. A significant meteorological event occurred during each mission, including a major storm in April and a sharp temperature decrease in October. Changes in backscatter are observed for both episodes in X-SAR returns from the mid-portions of one of the two large outlet glaciers in the study area. Meteorological and hydrological data are combined with the daily X-SAR images to interpret changes in glacier surface conditions caused by meteorological events. Effects interpreted from the April storm are (1) wind and precipitation influenced surface roughening of a wet snow pack and (2) the deposition of new wet snow at lower elevation and its subsequent retreat up-glacier. An abrupt decrease in regional temperature during October is thought to reduce the snow wetness and increase grain size. The changes in the radar-defined glacier zones due to the April precipitation event are subtle while the October temperature drop causes significant backscatter increases. Our results suggest that trends in glacier surface and near surface conditions on the South Patagonian Icefield observable from spaceborne SARs are not significantly masked by precipitation events.
Smith et al., 1997
The evolution of four dynamic radar glacier zones at the surface of an alpine ice field in British Columbia is monitored using a time series of thirty-five First European Remote Sensing Satellite (ERS-1) synthetic aperture radar (SAR) C-band images acquired from 1992 to 1994. These zones result from changing wetness and textural properties and appear to represent: (1) cold snow with no liquid water present; (2) an initial melt front with an upper boundary near the elevation of the zero degree isotherm; (3) metamorphosed, rapidly melting first-year snow with a rough or pitted surface; and (4) bare ice. This interpretation is aided by temperature and runoff data, air photos and field measurements of snowpack properties acquired simultaneously with two ERS-1 SAR scenes, ice surface elevations derived from 1:50 000 topographic maps, and simulations of radar backscatter from a geometric optics model of surface scattering. Meltwater production is affected by the development of zones (2), (3), and (4), which form, migrate up-elevation, and disappear each year between April and September.