11 - 3

GDS

HEGGY ET AL.: MARTIAN GEOELECTRICAL MODELS

features formed by interactions of lavas with water as

shown in the upper part of Figure 1 [e.g., Allen, 1979].

Lavas are likely filled by interstitial ice at several hundred

meters deep because ground ice may not have been com-

pletely desiccated at the latitude of this volcano. Liquid

water still could exist in the subsurface of the volcano if the

thermal gradient is unusually high like in similar regions on

Earth and if the subsurface material is not too thermally

conductive to be frozen deeply due to the low surface

temperatures. Such conditions would imply a subsurface

stratigraphy such as in the bottom part of Figure 1: (1) 10

meters of dust covering the surface, (2) 50 m of eroded

basalt corresponding to the porous part of basalt eroded by

surface processes, (3) 100 m of basalt filled by interstitial

ground ice and (4) a layer of wet basalt above the melting

point. In this model we do not include magnetic minerals

like maghemite in the subsurface layers.

2.2. Outwash Plains

[9] This model can correspond to a typical situation where

outflow channels converge into the northern plains. A few

billion years ago, large bodies of water could have been

formed at the ends of the large outflow channels (Chryse and

Acidalia Planitiae, Utopia Planitia, East of Hellas Planitia).

This unit occupies the lowest areas within the channels and

may contain volatile materials. According to the study of

rampart craters, the thickness of the volatile rich layer of

sedimentary deposits is estimated to be less than 800 m

[Costard and Kargel, 1995]. In the proposed model, differ-

ent layers are interpreted to be fluvial sediments up to 500 m

in thickness. These sediments (both aeolian and fluvial

deposits) are considered as an uncemented ground with

porosity from 40% to 50%. These estimations are based on

the bulk porosities of Martian soil as analyzed by the Viking

Landers [Clark et al., 1976; Gooding, 1978], as well as from

model of the megaregolith proposed by Clifford [1993].

These outwash plains occupy a latitudinal band between

20° North and 45° South, which corresponds to a ground ice

thickness of several kilometers. According to theoretical

models [Squyres et al., 1992; Clifford, 1993] as well as

morphological analysis [Costard, 1989; Kuzmin et al.,

1988], a first zone extending down to 300 m is supposed

to be desiccated (sublimation process). It corresponds to

fluvial and volcanic episodes, as shown in Figure 2. The

second zone, starting at 300 m, is assumed to be basaltic and

filled with ground-ice down to 2500 m where the melting

point is reached and liquid water is present. This region of

fractured basaltic rock may persist to depths of 10 km or

more [Clifford, 1993; Clifford and Parker, 2001]. Several

investigators have emphasized magmatic activity in these

areas in relationship with the Tharsis activity [Tanaka and

Chapman, 1990]. It may have been responsible for the

generations of liquid water by melting ground ice trapped

in the underlying megaregolith [Zimbelman et al., 1992].

2.3. Ejecta Deposits

Figure 2. Top: the Outflow channel emerging from

[10] This model shown in Figure 3, corresponds to differ-

chaotic terrain (1 S, 43 W) (Viking image P-16983).

ent geological units overlaid by ejecta deposits from impact

Bottom: proposed geological model for this type of terrain

craters. The uppermost part of the stratigraphy is a dry

where the presence of outflow channels may be interpreted

material made of aeolian deposits (dust layer). A second

by a rapid release of water from buried aquifers or the

zone results from impact processes with a 50 m thick ejecta

melting of ground ice by volcanism.

deposit. This value takes into account the thickness (from 40