Discrimination of glacier facies using multi-temporal SAR data

Abstract

ABSTRACT. Glacier facies from the Greenla nd ice sheet a nd the Wrangell St EliasMounta ins, Alas ka, are ana lyzed using multi-temporal synthe ti c aperture ra d a r (SAR) data from the Europea n Space Agency ERS-l satellite. Distinct zones and facies a r e visible in multi-temporal SAR d a ta, including the dry-snow facies, the combined p ercolation and wet-sn ow facies, the ice facies, transient m elt areas and m or a ine. In Greenl a nd and so uth-centra l Alas ka, ver y simil ar multi-temporal signatures a re evident fo r the same facies, a ltho ug h these fac ies a r e found at lower a ltitude in ' Nes t Greenland where the equilibrium lin e appea rs to be fo und at sea level a t 71°30' N during the yea r ana lyzed , proba bl y because of the cooling effect of the eruption of M ount Pinatubo. In Greenland, both the percolatio n a nd dry-snow facies a re excellent distributed ta rgets fo r sensor calibration, with backscatter coefficients stable to within 0.2 dB. H owever, th e p ercolation facies near the top orMount Wrangell are m or e complex a nd less easily delineated th an in Greenland , a nd at high altitude the glacier facies have a multi-temporal signa ture which depends sen sitively on slop e orientation.

Detection of snowmelt regions on the Greenland ice sheet using diurnal backscatter change

“…With radiometer data from the Special Sensor Microwave/ Imager (SSM/I), Mote and others (1993) used a singlechannel threshold method, Steffen and others (1993) applied the horizontally polarized gradient ratio technique, and Abdalati and Steffen (1995) employed the crosspolarized gradient ratio approach. C-band synthetic aperture radar (SAR) data have been used to classify different snow facies, including dry-and wet-snow zones, on Greenland ice (Drewry and others, 1991;Fahnestock and others, 1993;Jezek and others, 1993;Partington, 1998). Absolute backscatter data from Ku-band scatterometers such as Seasat-A Satellite Scatterometer (SASS) and NASA Scatterometer (NSCAT) are also used to map Greenland melt areas with an imaging period of 2 weeks (Long and Drinkwater,1999).…”

Changes in glacier facies zonation on Devon Ice Cap, Nunavut, detected from SAR imagery and field observations

“…2; Bardel et al, 2002;Engeset et al, 2002;Wolken et al, 2009). As discussed in more detail in the Results (Section 5.1), the glacier ice and superimposed ice zones typically have the lowest σ 0 of all facies due to the lack of reflectors at depth, the saturation and percolation zones have high σ 0 due to extensive ice layers and pipes within the firn, while the dry snow zone has low σ 0 due to the lack of internal reflectors (Partington, 1998;Langley et al, 2008). The presence of fresh dry snow above the last summer surface has a negligible impact on the reflected signal due to its low density and small crystal structure, so winter imagery is usually assumed to represent facies formed during the previous summer (Wang et al, 2007).…”

“…The dry snow zone occurs across the highest elevations of the ice cap in years when summer warmth is not sufficient to produce melt and refreezing of liquid water within or on top of the winter snowpack. This surface is a poor reflector of radar energy due to the lack of surface or internal scatterers, low snow density and small grain size (Wang et al, 2007;Partington, 1998), and is thus characterized in the Envisat ASAR imagery as areas near the ice cap summit with negative σ 0 values. Over the study period, a dry snow zone only existed on Devon Ice Cap between…”

“…The lower elevation limit of the dry snow zone transitions into the percolation and saturation zones, which are characterized by high SAR backscatter due to vertical ice pipes and horizontal ice layers which result from meltwater that refreezes at depth within the previous winter's snowpack (Partington, 1998). Most of the scattering appears to occur at the interface between the last summer surface and base of the winter snowpack, as well as within the upper 6-11 m of firn in the form of volume scattering (Davies and Poznyak, 1993).…”

Radar Backscatter Modeling Based on Global TanDEM-X Mission Data

“…Two characterizing radar quantities are considered for classifying snow facies: radar backscatter and volume correlation factor, which represents the coherence contribution due to volume decorrelation. The choice of an unsupervised classification method resides in the fact that a gradual transition of backscattering intensity between different snow facies on the Greenland Ice Sheet was observed by K. C. Partington in [125], impairing the use of a manual partitioning approach, which would strongly depend on the subjective choice of the decision thresholds. Moreover, the c-means fuzzy clustering algorithm has already been used in the literature for discriminating snow facies using Envisat active and passive microwave observations, showing it to be a promising approach for clustering similar regions of the Greenland Ice Sheet, as presented by Tran et al in [126].…”