HR: 1340h
AN: C13B-1074 [Abstracts]
TI: Great lakes, cold lakes, and fuzzy lakes: An algorithm for characterizing diverse water environments
beneath the East Antarctic ice sheet.
AU: * Carter, S P
EM: sasha@ig.utexas.edu
AF: UT Austin Jackson School Geosciences, UTIG, 4412 Spicewood Springs Rd #600, Austin, tx 78759
United States
AU: Blankenship, D D
EM: blank@ig.utexas.edu
AF: UT Austin Jackson School Geosciences, UTIG, 4412 Spicewood Springs Rd #600, Austin, tx 78759
United States
AU: Peters, M E
EM: mattp@ig.utexas.edu
AF: UT Austin Jackson School Geosciences, UTIG, 4412 Spicewood Springs Rd #600, Austin, tx 78759
United States
AB:
Understanding subglacial lakes and their interaction with the underlying geology and overlying ice sheet requires the ability
to rapidly locate such bodies in the large data volumes of airborne radar sounding data. Historically, subglacial lakes in
airborne ice penetrating radar have been detected by identifying portions of the base of the ice interface which appear
nearly horizontal. Horizontality is defined as a slope nearly 11 times the surface slope but in the opposite direction of
surface slope, which is equivalent to a flat hydropotential surface. The character of the reflection should be bright on an
absolute scale as well as bright relative to reflections in the surrounding region. The reflection's amplitude should be
exceptionally consistent across the entirety of the reflection. It is also appropriate to use proxies for the ice sheet's
basal temperature to identify subglacial lakes. In fact, an ideal subglacial lake will meet all criteria, including hydraulic
flatness, brightness (absolute and relative), specularity, and high basal temperature. We have designed an algorithm that
automatically identifies portions of the flight lines that match all or at least some of these criteria. This algorithm has
now been applied to over 50,000 line kilometers of data from four regions totaling over 250,000 square kilometers. This data
was acquired in East Antarctica with the UTIG airborne geophysical platform which includes an ice penetrating radar. A
review of the results suggests that many of the lakes found in previous studies fail one or more of the subglacial lake
identification tests. It is known that many "lakes" in East Antarctica are in regions where the estimated basal temperature
is far below the pressure melting point of ice. We classify these as "cold lakes". Some areas are identified as lakes yet
show non-specular reflections and are referred to as "fuzzy lakes". More interestingly is that a number of these sub-par
lakes intersect more perfect lakes. The distribution and nature of subglacial lakes, their classifications, and
understanding gained from their intersections is the focus of this presentation.
DE: 0555 Neural networks, fuzzy logic, machine learning
DE: 0726 Ice sheets
DE: 0746 Lakes (9345)
DE: 0758 Remote sensing
DE: 0794 Instruments and techniques
SC: Cryosphere [C]
MN: Fall Meeting 2005