31%? «:0 WIUHHUN”WWWIHIIIWUIHIJIHIWHWHI

NIEV RS ILTY LIBRARI

Mas IIIIIIIIIIIIIIIIIIIIIIIIIIIIII IIII IIII IIIIIIIIII

I 3 1293 015948

 

 

 

 

 

 

 

 

 

 

This is to certify that the

thesis entitled

I féd’Lo/J'C [Own/167511071, pf I/d’nf O'S/lxlc/Z/i
1/7/07“, f/XC/ 4/Wtbllfg, G/lc“ {/ lit/(25’ Qx:

TW’J”[{»L"’.7;S* flif‘ HY 67/7 7br’1 5/: l2)(L_S‘(/\,
L (.4
presented by

061%“.er O

has been accepted towards fulfillment
of the requirements for

”7 5‘ degree in %

; Major ;ofessor
Date M. / f7

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

 

. _ , ,

 

LIBRARY

Mlchlgan State
Unlverslty

 

 

 

PLACE N RETURN BOX to roman this checkout from your record.
TO AVOID FINES Mum on or baton date duo.

    

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

II I -I

MSU In An Affirmative ActioNEqnl Opponunlty Institution

 

 

ISOTOPIC COMPOSITION OF VENT DISCHARGE FROM
THE MATANUSKA GLACIER, ALASKA: IMPLICATIONS
FOR THE ORIGIN OF BASAL ICE.

BY

Daniel D. Titus

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Geological Sciences

1997

ABSTRACT

ISOTOPIC COMPOSITION OF VENT DISCHARGE FROM
THE MATANUSKA GLACIER, ALASKA: IMPLICATIONS
FOR THE ORIGIN OF BASAL ICE.

BY

Daniel D. Titus

Basal ice at the Matanuska Glacier, Alaska is characterized
by elevated 3H concentrations and less negative 8mO and 8D values
relative to englacial ice. Recently it has been suggested that
basal ice is partially composed of modern meteoric water and
forms as subglacial discharge moves out of an overdeepening
resulting in supercooling of the discharge and nucleation of ice
crystals.

Values of SEQ, 8D and 3H were measured in subglacial
discharge collected during the summer (n? 1995 from vents along
the ice margin at the Matanuska Glacier. Application of a
simple open system freezing fractionation model indicates that
18O, D and3H in discharge are within the requisite ranges to form

basal ice, and that there is a genetic relationship between the
two. Additionally, the temporal variability of 8WD in the basal

ice can be attributed to annual deviations in relative amounts of

precipitation and meltwater present in subglacial discharge.

To my loving wife and confidant Suzanne, you are the joy in each

moment that is forever my source of motivation. Your strength

and devotion are unrivaled, and compel me every day of my life.
I Love you.

iii

ACKNOWLEDGMENTS

The work presented in this manuscript, and indeed the
completion of my graduate degree, would not have been possible
without the patient tutelage of my advisor Dr. Grahame J. Larson.
His confidence in me and my ability has fostered intellectual and
professional growth beyond my greatest expectations. Dr. Larson
has proven time and time again to be a consummate mentor and
professor, not to mention a great friend. Thank you for
everything Grahame.

I would also like to thank my parents Roger and Faye Titus.
Their contribution is not obvious in the text contained here in,
but it is pervasive in all that I do and have become. They
taught me how to strive to live up to my own expectations through
perseverance and hard work. To this day, I believe that is the

greatest lesson I have ever learned. Thanks Mom and Dad.

iv

TABLE OF CONTENTS

LIST OF FIGURES ................................................ Vi
INTRODUCTION .................................................... 1
DEVELOPMENT ..................................................... 4
Closed Meltwater Freezing Model ............................... 4
Open Meltwater Freezing Model ................................. 7
Freezing Envelope ............................................. 8
Open Meltwater/Precipitation Model ............................ 9
Tritium In A Open Meltwater/Precipitation Model .............. 12
FIELD SITE ..................................................... 13
Study Location ............................................... 13
Subglacial Vent Location ..................................... 13
METHODS ........................................................ 15
Sampling Strategy ............................................ 15
Stable Isotope Analysis ...................................... 16
Tritium Analysis ............................................. 17
RESULTS ........................................................ 18
Stable Isotope Values ........................................ 18
Tritium Values ............................................... 19
STABLE ISOTOPE DISCUSSION ...................................... 21
Model Application ............................................ 21
Homogenization Due To Sampling Strategy ...................... 22
Origin Of Subglacial Discharge ............................... 24
Basal Ice Accretion Rates .................................... 25
Variation In Stable Isotopes Relative To Precipitation ....... 25
TRITIUM DISCUSSION ............................................. 29
Historical Tritium In Precipitation .......................... 29
Simulated Tritium In Subglacail Discharge .................... 29
CONCLUSIONS .................................................... 31

Figure

Figure
Figure
Figure
Figure
Figure
Figure
Figure
Figure

Figure

Figure

Figure

Figure

Figure

10

11

12

13

14

LIST OF FIGURES

Aspect Ratio Of Surface To Bed

Required For Freeze-on ............................... 2
Closed Meltwater Reservoir ........................... 6
Open Meltwater Reservoir ............................. 6
Freezing Envelope ................................... 10
Open Meltwater/Precipitation Reservoir .............. 10
Regional Study Location ............................. 14
Local Study Location ................................ l4
Discharge Hydrograph ................................ 16
Stable Isotopes In Vent Discharge ................... 19

Tritium In Englacail Ice, Vent Discharge,
And Precipitation .................................. 20

Influence Of Rain On Stable Isotopes
In Vent Discharge .................................. 22

Theoretical Stable Isotopic Composition
Of Ice Derived From Vent Discharge ................. 23

Relationship Between Precipitation Deviations
And Variation In SRO Measured In Basal Ice ......... 27

Simulated Concentration Of Tritium In Vent

Discharge VS. Tritium In Precipitation
(1995-1958) ........................................ 30

vi

INTRODUCTION

Traditionally, formation of the debris rich basal ice zone
observed in many temperate and subpolar glaciers has been
attributed to the regelation process (Weertman, 1961, 1964;
Boulton, 1972; Iverson, 1993). Regelation occurs as the result
of phase changes caused by either variations in temperature at
the glacier sole or pmessure fluctuations 5J1 a glacier that is
near its pressure—melting point. These phase changes may occur
in response to: isolated cold patches or permafrost at the
glacier sole (Robin, 1976), as cold winter air penetrates to the
sole (Weertman, 1961; Clarke et al., 1984), or through melting on
the upstream side and subsequent refreezing on the downstream
side of an obstacle encountered by moving ice at the base of the
glacier (Kamb and IaChapelle, 1964; Iverson and Semens, 1995).
Irrespective of the particular cause of the phase change, clean
englacial ice is melted and then refrozen thereby incorporating
basal debris. Physical constraints (N1 the regelation process
dictate, however, that this nechanimn cannot incorporate 21 net
thickness of basal ice beyond 1cmy‘1 (Alley et al., 1996) Thus,
regelation alone cannot explain the >1m exposures of basal ice
observed at the NEtanuska Glacier, Alaska (Lawson et al. 1996;
Stasser et al., 1996).

Alternatively it: has been shown (n1 theoretical grounds

1

2

when water moving through an open linked-cavity drainage system
flows out of an overdeepening, at or near the terminus of a
glacier, a net accretion of ice through freeze-on will occur at
the glacier bed if the aspect ratio of the bed to surface
gradient is between -1.2 and —1.7 (Figure 1) (Alley et al., in
press). This freeze-on process occurs because subglacial
discharge flowimg up gradient out (n? the overdeepening becomes
supercooled due t1) a rapidly increasing pressure melting point
and insufficient influx of tuxn: to compensate for the puessure
change (Strasser et al., 1996; Lawson, in review; Alley et al.,
in press). Supercooling of the discharge results in nucleation
of ice crystals in the water and thereby the release of latent
heat, thus allowimg the water to naintain thermal equilibrium.
According to this hypothesis, ice crystals formed in the
supercooled discharge are subsequently affixed to drainage cavity
surfaces resulting in a net accretion of basal ice (Strasser et

al., 1996; Lawson et al., in review; Alley, in press).

Surface

Glacial Ice

”W

Bed-rock

 

 

Figure l. Aspect Ratio Of Surface To Bed Required For Freeze—on

Strasser et al.(1996) and Lawson et al. (in review) recently

3

observed that 8180 values in frazzil ice and basal ice at the

Matanuska Glacier are virtually identical enui enriched greater
than 3% ifllBO relative to englacial ice. Frazzil ice forming
around. subgalcial discharge ‘vents at time Matanuska. Glacier .is
attributed to supercooling of discharge, thus Strasser et
al.(1996) and Lawson et al. (in review) conclude that frazzil ice
and basal ice are both formed by the freeze-on process.
Furthermore, because the maximum enrichment of 18O in response to
fractionation during freezing is only 3% (O’Niel, 1968; Lehmann
and Siegenthaler, 1991), they additionally conclude that the
source water for frazzil ice and. basal ice is not comprised
solely of englacial melt water, but of a mixture of meltwater and
isotopically heavier meteoric water (Strasser et al., 1996).
This CODCIUSICHI was supported. further In; the observation that
much of the basal ice contains appreciable amounts of bomb 3H, 5-
>95 TU, which is possible only if the source water for the basal
ice is partially derived from recent meteoric water (Strasser et
al., 1996; Strasser, 1995).

The hypothesized basal ice freeze—on theory requires,
therefore, that the isotopic composition of vent discharge at the
terminus of the Matanuska Glacier (corrected for fractionation
due to freezing) reflect the composition and variation of basal
ice with respect to it’s 18O, D, and 3H content. In this paper,
we present results of 18O, D and 3H measurements in vent discharge
collected along the ice margin at the Matanuska Glacier, and
attempt to show that there is a genetic relationship between vent

discharge and basal ice.

DEVELOPMENT

Based on the physical parameters of a simple closed
reservoir system, Jouzel and Souchez (1982) derived equations to
describe the fractionation path of 1%) and.I) during freezing of
subglacial water. In this closed system model the physical
parameters of the reservoir are defined as follows: Input to the
reservoir (I) is equal to zero; Output (O) from the reservoir is
equal to zero; The freezing rate of the reservoir (F) is less
than the isotopic diffusion coefficient (e.g. 10Tcmsfl)
eliminating the possibility of isotopic gradients in the water;

Finally, 8R1 is the initial SEQ and 5D values of the subglacial

reservoir. The SWO and 8D values of basal ice formed in this

conceptual closed reservoir system are given by the equations

(Jouzel and Souchez, 1982):

85180 = 10(1ooo+511"*o) ((1.1—K)°‘—(1—K)a)—1000 (1)

550 = 10(IOOO+8iD) ((1.1-K)B-(1—K)fi)-1000 (2)

where agfio and 840 are isotope values of basal ice formed at each
progressive fraction of the initial reservoir frozen (e.g. 0.1-1)
denoted by K, and where a and B are equilibrium fractionation

coefficients for 1%LFIO and D/H, 1.00291 and 1.0212 respectively

5

(Lehmann and Siegenthaler, 1991).

On a diagram where Sfifio and 84) values from equations 1 and
2 are plotted for corresponding values of K, and as the abscissa
and ordinate respectively, Jouzel and Souchez (1982) showed that
SWO and 8D values evolve co—linearly and fall on a freezing line
with a slope value less than that of the Global Meteoric Water
Line (e.g. 8). Therefore, Jouzel and Shouchez (1982) conclude
that when. basal ice sampled. front a :natural closed subglacial
system falls (NI a freezing line, the initial 5R1 composition of
the reservoir being frozen is defined by the intersection of the
freezing line and the Universal Precipitation Line. Thus, the
relationship between SR; and the SEC and 8D values of basal ice

can be determined by the slope of the line basal ice samples fall
on irrespective of the K value any individual ice sample
represents (Jouzel and Souchez, 1982). This closed reservoir
system freezing slope is given by the equation (Jouzel and

Souchez, 1982):

s = ((a-1)/<B-1>) ((1000+6iD>/<1000+5880)> (3)

where 5iD and Sileo are initial reservoir values and a and B are
maximum. equilibrium fractionation coefficients for 180/160 and
2H/lH respectively (e.g. a=1.00291 and B=1.0212). An example of
a closed reservoir system freezing line for 23 hypothetical 8RI

composition, within the range of'EVK) and 5D values Heasured at

the Matanuska Glacier, is given in Figure 2.

6

 

 

 

18
8 O
.3000 mm
. -25.00
I , l
Glacier Ice Global Meteoric Water Line I
I I .
I = o \\\_R_____9!L/ O = o /'./ I— 140m
F < Diffusoon’ Coefficient /,- I
Of Watet ’ I
”I «=04 I— -160 00
C" I" l
2/ I— -180.00
.m 8Ri
I SD
(D, (I I
Slope = 5.84 *— -200.00
KT: 1 I ~220.00

Figure 2. Closed Meltwater Reservoir

 

 

 

 

1
5 80
—30.00 -20.00
_ own
I L I I
I Glacier Ice I Global Meteoric Waller Line I
I > o K‘wdl/‘j o g o L— _1 40‘ 00
F <l . A I
6' 8 8R. V/, I
2:) K=O.1 I—‘— '160.00
:* l
. c Slo = 5.95
O“ m
/ 5” 4mun
80
I— 20000
I
'- -220.00

Figure 3. Open Meltwater Reservoir

7

Expanding on the conceptual closed reservoir system freezing

model, Souchez and Jouzel (1984) developed a slope equation for

interpretation of the relationship between 5R1 composition and

STO and (K) values of basal ice formed in an open meltwater
subglacial reservoir. In this open meltwater system model the
values of I, O, and F can vary according to the physical
characteristics of the system being described, plus the isotopic
value of the input (SI) is incorporated to account for changes in
the 18O and D content of I during freezing of the reservoir.

Souchez and Jouzel (1984) showed when SI was not significantly

different from.8RI, which is reasonable if 8R1 and SI are composed
primarily of englacial ice melt, that irrespective of the values
of I, O and F the freezing slope in an open meltwater subglacial

reservoir is given by the equation:

8 = (oz/I3) ((a—1)/(B-1)) ((1000+8iD)/(1000+511I“IO)) (4)

again where SJ) and SEBO are initial reservoir isotope values and
a aumi B are maximum equilibrium fractionation coefficients for

180/160 and 2H/ln receptively (e.g. (1:1.00291 and B=1.o212).
This open. meltwater reservoir system slope equation yields a
value that is virtually identical to that of the closed reservoir
system freezing slope of Jouzel and Souchez (1982). Thus, for
these open and closed system models the only significant
difference in the freezing process is that ice produced in an

open meltwater system is always enriched in 1PO and D relative to

8
SRIdue to continuous reconstitution of the reservoir by I, where
SI and.SR¢ are equal and I is not less than F (Souchez and Jouzel,
1984). Alternatively, in an open meltwater system ice depleted
in heavy isotopes relative to SRJ may be produced if the value of

F is greater than I. This situation, however would be limited
to systems that produce little meltwater such as high elevation

cirque glaciers. An open meltwater reservoir system freezing
line, where I is not less than F, is given in figure 3. The SmO
and SD values used in this simulation are the same as those used

previously in the closed reservoir system freezing line (Figure
2).
Hubbard and Sharp (1995) identified that single co—isotopic

freezing slopes (Jouzel and Souchez, 1982; Souchez and Jouzel,

1984) were not adequate for interpretations of basal ice SRO and
SD values if that ice was formed during discrete freezing

episodes from reservoirs with different SR; compositions.

Furthermore, Hubbard and Sharp (1995) suggested that field

sampling of basal ice often crosses discrete basal ice layers
resulting in mixed SmO and SD values of ice formed from
distinctly different SR4 compositions. Thus, the combination of
natural variation in values of SR1 and the recovery of samples
across discrete layers results in a scatter of SEC and SD values

around poorly defined freezing slopes (Hubbard and Sharp, 1995).

Therefore, Hubbard and Sharp (1995) propose the use of freezing

envelopes where both SR; and K can vary to include all possible

9

SRO and SD values of basal ice formed in closed and open
meltwater reservoir system freezing events. The height of such
an envelope is defined by the range of variation in values of SR;
along the Universal Precipitation Line, while the width of the
envelope is dependent on the sensitivity of K included in
individual basal ixxa samples (Hubband and Sharp, 1995). A
theoretical freezing envelope for a representative range of SR1
values at the Matanuska Glacier is given in Figure 4.

The co-isotpic freezing slopes (Jouzel and Souchez, 1982;
Souchez and Jouzel, 1984) and envelopes (Hubbard and Sharp, 1995)
previously described are designed for interpretations of basal
ice samples formed in subglacial reservoirs with either no
variation or well constrained variation in both the value of K
and the composition of SR1 along the meteoric water line. In
reality these forms of analysis may not be appropriate for
dynamic open subglacial reservoirs where large volumes of water
are moving through the system, which is characteristic of low
elevation mountain valley glaciers such as the Matanuska. Due
to the large volumetric flux of water relative tx) the rate of
basal ice formation the value of F in this conceptual open system

is many orders of magnitude less than I, subsequently requiring

that O and I are essentially equal. The S”O and SD values of

basal ice formed from discrete SR; compositions in this system,
therefore, will always correspond to values of K much less than
0.1. Thus, the pmoduction (n5 basal ice anJ. not measurably

change the value of'SRq and the ice will reflect the SMO and SD

 

 

 

 

 

 

 

18
S O
-30.00 -20.00
~25.(X)
| I I
I I
_ Global Meteoric Water Line I
. GlaCIer Ice fiI
:5 “\RLWLV’ 5:0 3:," K = 01 -140-00
F < l '
8R) = variable alone water line {at I I
SI = variable along waterline .4 *" ’ .'
RangeOf _ '1 60.00
SRI
L-mom
SD
:— -2oo.oo
I
I
I— 22000
K=1d
Figure 4. Freezing Envelope
18
S O
-30.00 -25.00 -20.00
-27.50 -22.50 - 7.50
I
L I I L' I I j
I were.
I=mmr\\JEEQL/Io~l
F << I or O
K << 0.1
SI = variable
0 ‘3 I
CBS: ~\ I OD
‘I {,A) .V, I
«’l Range of (RI I
»— 480(1)
I
I
L- -200.00

Figure 5. Open Meltwater/Precipitation Reservoir

11

value of 0 plus the maximum fractionation coefficients 2.91% and
21.2% for SMO and SD, respectively. Additionally, during
discrete time intervals recent meteoric water (e.g., rain) mey
comprise a significant portion of the subglacial reservoir in low
altitude glaciers. Contributions of precipitation to the
subglacial reservoir may result in variation of SR1 both on and
off the meteoric water line. The degree to which precipitation

affects the SR1 of a reservoir is dependent on the volumetric

contribution of precipitation and its SEC and SD values.

Therefore, we propose an additional model for the
interpretation of the co-isotopic composition of basal ice formed

in dynamic open subglacial systems. In this conceptual open
meltwater/precipitation system, since tjma S values of C) and SR1

are equal, the values of SRO and SD for basal ice produced can be

calculated from the equations:

SD = SI)+ 21.2% (5)
and
5190 = 5,180 + 2.191,. (6)

where SJD and SfEO are initial values of SRI, and 21.2% and 2.91%

are the isotopic enrichment of ice relative to SRI (Lehmann and
Siegenthaler, 1991). This form (n? analysis can kxa conducted
irrespective of freezing slopes and allows for maximum variation
in the values of SRL and thus basal ice values of 5”o and SD, on

and off the meteoric water line. A series of predicted basal

12
ice SRO and SD values produced from a range of SR1 values,
represented by vent discharge at the Matanuska Glacier, is given
in Figure 5.

The presence of liquid precipitation in an open
meltwater/precipitation subglacial reservoir allows for the
incorporation of anthropegenic :II into the interpretation of
basal ice formed in this system (Strasser et. al., 1996; Lawson
et. al., iJI press). 3H II} basal ixxa produced In! freezing
subglacial water will be fractionated due to differences in
spacing of vibrational levels of individual atoms in a water
molecule. This spacing is proportional to the square root of
the mass ratio of the heavy to light isotopic species. Thus,
fractionation of 3H/IH is 1.4 times the fractionation of EVIH, and
the ratio of D/lH is 8 times that found in the mass ratio of
18O/I‘IO (Strasser et al., 1996). In an open drainage system
freezing event, the maximum fractionation of 1R) from the liquid
to solid phase is 2.91% (Lehmann and Siegenthaler, 1991).
Enrichment of 3H, therefore, is approximately 4% from the liquid
to solid phase (Strasser et al., 1996). Concentration
variations in PP during single precipitation events and
analytical error associated with lower detection limits are much
greater than the 4% enrichment of 3H due to freezing. Thus, 3H
concentration :Ul basal ice, produced from. variable subglacial

reservoir concentrations of 3H, will reflect the 3H content of O.

FIELD SITE

The Matanuska Glacier is located in south-central Alaska
where it flows northwest out of icefields in the Chugach
Mbuntains. 'Fhis large valley glacier is approximately 45km in
length and ranges in width from 2.2km, at its source in the
icefields, tx> 5Mn at the terminus (Figure 6). Over its 45km
length, the <glacier‘ realizes ea 3000m1 change i1) elevation from
3500m at the icefields to 500m at the terminus where, at it’s
present position, it has remained relatively stable for the last
200yrs (Williams and Ferrians, 1961). The northern portion of
the terminus is stagnant and covered with debris, while the
western portion is characterized by active ice and a lack of
debris cover (Figure 7).

Subglacial hydrologic activity is most evident on the
western portion of the terminus where ice flows through a small
basin or overdeepening (Arcone et al., 1995). Along this margin
numerous discharge vents can be observed in the form of
upwellings or fountains. Many' of these vents emanate from
directly under basal ice exposures that range in thickness from
1m1 to more than 5m. Collective discharge from the vents
coalesces into 23 proglacial catchment which subsequently
discharges into a series of streams that flow west eventually

forming' the .Matanuska River (Figure 7).

13

14

 

’ ’ I ’
“WV,“

Mount-Ins,

 

qukeetnI

I

 

I

I

  

            

 

Figure 6. Regional Study Location

 

 

 

trawl-Ice on lap-glacier tlde

gum“:
gwmeormm
fiflgowmemuwru

I
I

+ Subghduly fed dbelume vent

 

 

Local Study Locatrion

Figure 7.

Methods

Discharge vents along the ice margin (Figure 7) were sampled
because they represent the distal portion of the subglacial
drainage system (Lawson, 1986; Strasser et al., 1992). This
assertion is supported by observations of acute levels of
suspended sediment in the vent discharge, which is characteristic
of subglacially derived water and not water transported through
clean englacial conduits. Vent samples were extracted either
manually or through an ISSCOTm water sampler with a weighted
sampling tube inserted into the vent throat. Samples were
subsequently transferred to, and sealed in, 275ml NalgeneTm
bottles for shipping.

Vent sampling was conducted on a diurnal interval based on
the times of daily high and low discharge determined at a gauging
station installed 300m down—stream from the vent catchment area
(Figure 7). At this location, the stream is fed predominantly
by discharge vents along with a small contribution from englacial
ice-melt in the form of surface runoff from the glacier. The
hydrograph (Figure 8) recorded at the gauging station shows a
strong diurnal variation in stage, and thus subglacial discharge,
in addition to a steady increase in flow during the spring
months. Peak discharge occurs during June and July followed by

a steady decrease in flow through the fall (Strasser, 1995; D.E.

15

16
Lawson unpublished data). Daily inspections of the gauging
station hydrograph were made in order to adjust the sampling

strategy to exact times of high and low discharge.

300 a

275 ~

I
250 a

Q(ft3/sec)

 

 

IOOIO I08“)
022202-93
(O w
v- ‘-
JulIanDays

140
145
150
155
200 4 .

Figure 8. Discharge Hydrograph

Approximately 10mls of each vent sample was allocated for
18O/I“O and jH/I‘I-l ratio measurements. Sample splits were sent to

Coastal Labs Inc. for analysis where the ratios of :H/ZH and

17

18O/“IO were determined using the standard calculation method:

5(1’0) = IRI'sample) — RIstandard) / Rlstanxiard)I * 1000 (7)

where R is equal to D/H and IKLCIO for the experimental sample or
the laboratory standard respectively (e.g. SMOW). Analytical
accuracy was reported as 0.1% and 1.0% for SmO and SD
respectively.

Analysis for 3H was conducted at the Michigan State
University low level 3H lab using a standard scintillation
counting procedure (Kessler, 1988). Abundance of 3H in a sample
volume is expressed in 3H units, where one 111 is equivalent to
one 3H atoms In) every 1H*1018 atoms. Concentrations of 3H ijl
discharge have consistently been observed to be less than 20 TU
at the Matanuska Glacier (Strasser et al., 1996; Lawson et al.,

in review). Low level 3H analysis has been shown to be accurate
to :1 TU if electrolytic enrichment is applied to samples

(Ostlund and Werner, 1962). 250mls of each sample were

allocated for electrolytic enrichment and subsequent 3H analysis.

RESULTS

The S3T>emmi SD values measured III vent discharge samples
collected fromI the Matanuska glacier after June 9, 1996 are
presented in Figure 9. This time period is characterized by a
rapid increase :hi meltwater production (Figure EH, which is aa
condition required for expansion of the basal drainage system and
the subsequent accretion of basal ice (Alley et al., in press)
Also included in Figure 9 are: Previously published SEC and SD
values for basal ice samples collected in 1995 from site 95-1
near the glacier terminus (Strasser, 1996), average isotope
values for basal ice (Strasser, 1996), and average isotope values
for englacial ice based on an extensive unpublished data base of
samples collected since 1978 from cufferent locations near the
terminus by D. E. Lawson (unpublished data). The local meteoric
water line and the average value for precipitation in Figure 9
are also derived from the unpublished data base.

The vent discharge isotope values show a considerable amount
of scatter about the local meteoric water line and vary from a

maximum of —21% to a minimum of -24.6% with an average of —22.8%
for SEC, and from a maximum of -l72% to a minimum of —189% with

an average of ~181.8% for SD. It is evident from the data that

the average ikn: vent discharge is more negative than average
18

19
5180

.0
-25.00 -22.50 -17.50 ~15.00 42.50

I I . .
I l . L I I
j

Global Meteoric WaterLine I

 

/ / I— 43000
// ,/ I
' Matanuska Water Line I
I -14o.oo
, I
/// I7 -150 00
I!” .I/ I
t/ I
I -1 60.00
£41 I . I -170 00
} / h . I
I7) I ) o VentDIscha
(A) /+({+ V I To. ' I I
e. /,+ A + Basal loeFrom Site 95-1 Strasser(1996) .
A .v’; L,“ A Average Precipitation Unpublished ~ I.
a: k Dan-From D-E-L-wson I I‘ '180-00
("I [fiINT'I‘ J ,A. 0 Average Englacaillernpubllahed '
if" .. I‘"\_-i‘ '-.._,I Datatrom D.E.Lawaon I
/ 1;:‘IVI I"; : CI . Average Baaalloefrom I
I.’ ,' —' 5m (19%) I ’7“ -190.00
I/I/I 9 Average Groundwater Unpublished I I
I

DetafromD.E.Lawson I
I

Figure 9. Stable Isotopes In Vent Discharge

basal ice cm: precipitation. Vent. discharge, furthermore, is

only slightly less negative with respect to the average STO and
SD values of englacial ice.

Vent discharge3H concentrations, much like SEC and SD, also

A

show considerable variation (Figure 10). Vent discharge HI
values were found to range from a maximum of 8.3 TU to a minimum
of 2.3 TN} with EH1 average value (n? 4.6 TU. Compared to 3H
concentration measured in summer precipitation in Anchorage,
Alaska (unpublished data from R. L. Snyder, U.S. Geological

Survey), which is within 150Km. of the .Matanuska Glacier and

20
generally experiences the same weather systems, it is apparent
that vent discharge is diluted with respect to it's 3H content.
Conversely, discharge is enriched in 0H relative to the
concentrations of 3H measured in englacial ice at the Matanuska
Glacier (Figure 10) (Strasser et. al. 1996; unpublished data from

D.E.Lawson).

 

    

 

III II -4-—ve—474——v—7
600 III) III I um. I
IIIIII IIIII I : VentDischarge I
5.00 a III "'II I _ PI°°IPIm°II I
I IIII IIII . “32$“: Engleciallce I
... I III III L_¥‘
g 400 AI III III
8 I i III WI
3.00 J I II I I
I
2.00 i
1.00 “I
0.00 4 .
2.50 7.50 I 12.50
0.00 5.00 10.00 15.00

Tritium Concentration Range

Figure 10. Tritium in Englacial Ice, Vent
Discharge, and Precipitation

STABLE ISOTOPE DISCUSSION

Only the Open meltwater/precipitation reservoir system

freezing model is appropriate for interpretation of a genetic
relationship between values of SEC and 8D measured in vent

discharge and basal ice at the Matanuska Glacier for two reasons:
First, Alley et al. (in press) has calculated a gross basal ice

accretion rate of 1.69myr'1

in the overdeepening at the Matanuska
when the volumetric flux of water through the drainage system is
3.2xlO%fiyr‘K Thus, in this system the value of K (e.g.

fraction of the reservoir frozen) is orders of magnitude less
than 0.1 requiring, therefore, that I and O are essentially equal
and that the 5:80 and 5D values of O and 5R1 are equivalent.
Second, given that vent discharge is comprised of a variable mix
of englacial ice melt, groundwater, and recent liquid

precipitation, Figure 11 shows that only the variation in SEC and

8D values of precipitation are great enough to explain variation
of these same isotopes in vent discharge (Strasser, 1995). This
condition, in addition to the elevated levels of 3H measured in
vent discharge, suggests that recent liquid precipitation is
present III significant volumes III the subglacial systenlenui is
responsible for the variation of SEQ and 6D values both on and
off the water line in vent discharge.

21

 

 

     

 

 

 

-20.00
-35.00 -30.00 -25.00 | -15.00 -10.00 -5.00
I o Vent Dacha,” '. Global Meteoric Water Line T
I z m? $951.00...“ I / / I
byStrasser(1996) I I——— ‘100.00
A Average Precipitation I/ ‘
. ¥ Average Snatch-Hoe I I I Matanuska Water Line 2
L1 m... . I -125...
. I— 45000
, / I I
|
I— -175.00 80
I— -200.00
/ ‘.
/ I
2 I L ~225.00
I
—~ I
I— -250.00

Figure 11. Influence Of Rain On Stable
Isotopes In Vent Discharge
Application of equations 5 and 6 from the open

meltwater/precipitation reservoir system model, to vent discharge
samples, results in theoretical basal ice SRO and 8D values that

display' a considerable degree of scatter about the Matanuska
water line (Figure 12), and in this way are inconsistent with
those values measured in basal ice by Strasser et al. (1996) at
time Matanuska. Glacier. It. is instructive at this point to
investigate the sampling method used by Strasser et al. (1996) to
collect basal ice for 103, D and 3H analysis. Core sections 5—

10cm thick were collected and melted to provide adequate sample

23

volumes for enriched 3H measurements (Strasser et al., 1996).
This process physically homogenizes the 5:10 and 8D values of
discrete basal ice horizons and therefore masks the variability
of these values in the ice. Thus, the true 8150 and SD range of
discrete basal ice layers is not represented, and it is implicit
that comparison of theoretical frozen discharge to basal ice is
only reasonable if the theoretical values are averaged thereby
representing a homogenized basal ice 8:30 and 5D value produced

from discharge (Figure 12) .

 

 

 

18
6 O
-20.00
-25.00 -22.50 -17.50 '1 5.00 -1 2.50
_L_hlik_~ _ _L_,, _—J___L
W0"! f” 1‘ Global Mohodcflluhr Line] '_ ‘130 w
K 9 MW I / / I I '
l+ auulooFmsn-esl /' wan-mm I
I '3} Mean W Fm Ditch-mo ‘ // / I,
IOflwmmuanmflMmb ‘ ,’ // ” *wno
-1 50.00
P -160.00
l 6D
J» -170.00
‘9 430.00
I 490.00
,« |
l
/ l
L. -200.00

Figure 12. Theoretical Stable Isotopic Composition Of Ice Derived
From Vent Discharge

24
Homogenization of basal ice 8:90 and 5D values may also be
caused by two additional mechanisms: First, if basal ice is

forming in an open lattice framework (Strasser et al., 1996) and

the 6R; of the reservoir is changing during that formation period,
then basal ice SRO and 8D values will reflect the averaged SRO

and 8D values of the two reservoirs. Second, Alley et al.
(1996) theorizes that freeze-on and regelation may be occurring
III the overdeepening either simultaneously (n: during different
time periods during the year. Regelation through partial
melting and subsequent refreezing of cfiscrete basal ice layers
may also homogenize SRO and 5D values of basal ice.

Regardless cm? the homogenization. effects of time sampling
strategy, 21 strong linear trend ii; still evident le basal ice
profile 95—1 cxfllected In! Strasser et EN” (1996) (Figure 12).
Simplistcally, all water at the FBtanuska originated as
precipitation and therefore high elevation samples will fall on
the light end of the water line while low elevation samples will
fall on the heavy end with respect to 6130 and 8D. If, then,
subglacial discharge is a nuxture of isotopically depleted high
elevation precipitation (represented by englacial ice melt) and
isotopically enriched lower elevation precipitation (e.g., rain),
then discharge SmO and 6D values will fall on the heavy or light
end of the water line as a function of the relative contribution
of either end member in a sanmle. Furthermore, if individual

basal ice samples collected by Strasser et al. (1995) are

approximately a seasons worth of ice production, then the SRO and

25
8D values of those samples will represent the dominance of the
heavy or light end-member in vent discharge over the course of a
freeze—on season.
Using a one dimensional model Alley et al. (in press)

calculated theoretical rates of basal ice accretion in an

overdeepening similar to that of the Matanuska Glacier. .Alley
et al. (in press) calculates that when 10% of the Matanuska
glacial discharge (e.g., 3.2x106 Hfiyr*) is routed through the

overdeepening for a time interval of 0.1yrs, "'2mmy'1 of basal ice
will be accreted. Net accretion rates of 16cmy'1 may be realized
if 100% of the discharge is routed through the overdeepening for
the same time period (Alley et al., in press). Given the aerial
extent of the overdeepening at the Matanuska Glacier, and the
density of discharge vents in that area (Figure 7), a reasonable
estimate of discharge moving through the overdeepening may be 50%
at any point in time. Following that assumption, a net
accretion rate of ~8cmy’1 will be realized according to the Alley
et al. (in press) one-dimensional model.

Theoretical accretion rates of basal ice can be used for
time series analysis of variations in.€?%) and 8D values when a
reference datum in the ice can be identified. Strasser et al.
(1995) labeled the 3H peak in basal ice profile 95—1 as
correlative with peak 3H concentrations found in precipitation in
1963. Using the 3H peak as a datum and an assumed annual
accretion rate of 8cmyfl, a total time of ~16yrs would be required
to accrete the 127cm of basal ice below the 3H peak in profile

95-1. Thus, the EMS samples collected below the 3H peak datum

26

would represent 0.62yrs of ice production each from 1963 to 1979.

A comparison of 890 values taken from basal ice profile 95—
1 and precipitation deviation from normal, both as a function of
time, is shown in Figure 13. The basal ice profile provided
includes samples taken in the englacial ice as a means of
graphically illustrating that there appears to be an isotopic
dispersion zone between englacial ice and basal ice in the
profile (Figure 13). An investigation as to the precise nature
of possible dilution in this zone is beyond the scope of this
paper. Contamination of 5150 values in this dispersion zone
precludes any reasonable comparison with precipitation deviation
values, yet possible dilution. of the peak :01 value does not
obfuscate its utility as a reference datum to 1963. A lower
stratigaphic limit of the dispersion zone is included on the
plot, and comparison of SRO deviation from. normal to
precipitation deviation is conducted only for samples beyond this
limit.

The variation in SEQ of basal ice samples and precipitation
deviation as a function of time exhibits a series of peaks, in
the positive direction, that are nearly in phase with each other
(Figure 13). The 6”o value of theorized basal ice from vent
discharge for summer 1995, additionally, shows a strong positive
deviation from normal, correlative with pmecipitation deviation
from normal. Large negative deviations in precipitation,
conversely, are not reflected in the basal ice 5E0 values, nor

should they be. Generally the majority of subglacial discharge

27
is comprised of ice melt, and therefore vent discharge 5150 values

are always at or near the lightest end member of the system,

Depth Below Englacial Ice
in (cm)

0. 0025 050. 00 100 00 150 00 200 00 250. 00 300. 00 350. 005 4000. 00
7500 125001.7500 2250 27500 32500

I lili. ii“ iii; V; l, A, ,, 1:,71:,I

Average Theorized Basal Ice
From Discharge l

 

 

 

 

-20.00 a I
I . fl ° ~— 20.00
I 19 e g
l I I
«H 9. 9 I 8
4‘77I , “3%; I Z
I T I . ~ \’ Average Basal Ice -21.09 I
I I? ' 0 I g
I. ' ‘— 1000 '-
II ‘I l? ’ 7 {‘I I u-
2200 I ’I‘II ' I I 8
_ . __ I III» I I .—
8 ‘I ‘I‘ mm" “’1‘ >95 TU I Precipitation Devlalion1995 l ‘5
‘— I LI A I .03,
I . “I II [In imIl I D
‘ “‘3 If; I / , 1:
V, I A" .I‘ , 3.35 ‘ 0 00 '9
" .5. Q) 4—0
5 II JIIFII I g
I I I I ’ I .9
‘1
I 1 (+250 I ,. p ' o '5 F N l I 8
-24 00 -—I y ‘ _, ‘ .Vv reap eVla on rom orma ‘-
- l e Basal lce180 95-1 collected l (L
I ‘I Lflflbe 40.00
‘ . I“ Diffusion [me I
If" I . 7 Average Englacial Ice -24.8_I
I I I I 1 7 I l I l f T 7 I n I
. I . . I
8 8888 8888 8888 8888
o o Q 75 I2 :2 o “‘ “° 0 ° 0
fidgééfiédeeeedsfiééédgéhd
(D Ix co 0) o
a: m m m o
s- ‘— 1— F N
Year

Figure 13. Relationship Between Precipitation Deviations And
Variation In 81:10 Measured In Basal Ice

englacial ice. Only a large annual increase in the amount of

low elevation precipitation (e.g., rain) can shift average annual

28

SRO values of vent discharge, and basal ice produced from vent
discharge, away from. the englacial ice end member. Thus,
variation in the basal ice produced from discharge, regardless of
the homogenization due to sampling, must reflect averaged
variation in contributions of different end members in the
subglacial reservoir. Oscillations in the relative amounts of
precipitation present in vent discharge from year to year seems
to be the only plausible explanation for the observed variation
in SWO.

It should be restated that accretion rates for basal ice
samples presented in Figure 13 are assumed to be constant through
time. Annual variation in subglacial hydrologic activity, due
to a myriad of possible changes in atmospheric and subglacial
conditions, require that accretion rates from year to year are
not constant. Increased or decreased accretion rates between
years could explain the peak phase shift in 8190 observed in
Figure 13. The sampling strategy of Strasser et a1. (1996),
alternatively, may have inadvertently crossed stratigraphic
boundaries of basal ice produced during single freeze-on seasons

resulting in homogenization of two or more years worth of basal

ice production.

TRITIUM DISCUSSION

It is possible to approximate average seasonal 3H content in
discharge back tx> 1958 using time 1995 subglacial discharge 3H
concentration average of 4.657NJ and the historical record of 3H
concentration in precipitation for Anchorage AK (unpublished data
from R.L. Snyder, U.S. Geological Survey). Long term summer
month records (e.g. May-Sept.) of 3H concentrations in
precipitation for .Anchorage .AK. are presented in (Figure 14)
(unpublished data from R.L. Snyder, U.S. Geological Survey).
The 4.6 TU measured vent discharge average is 35% of the 13.2 TU
average measured concentration in precipitation for Anchorage AK
in the summer of 1995. Assuming this percentage represents an
average volume of recent meteoric water in the subglacial system
over the course of an ablation season, the range of ITI in
discharge from 1963 to 1995 corrected to a reference date of Dec.
31, 1995 would be 208.2 TU to 2.2 TU, respectively (Figure 14).

Estimating the range of 3H in basal ice by this percentage
method is a crude approximation. Strasser et al. (1996)
reported.ai 6 TU measurement in subglacial discharge immediately
following a precipitation event whose 3H concentration was
reported as 8 TU. A minimum value of 2.3 TU, which is close to
background levels, was measured. in discharge during the same

summer over the course of a protracted period (n3 dry sunny

29

30

Thus, a wide and variable range of 75% to nearly 0% meteoric
water was present in the subglacial drainage system at any given
time during the summer of 1995. Nevertheless, using the 75%

meteoric: water concentration III vent discharge as 6M1 average
annual value, basal ice 3H concentrations of >95 11] reported by
Strasser‘ et al (1996) could. have only' been formed froni vent
discharge during the time interval 1962-1966 (Figure 14).
Based on the more reasonable estimate of 35% average meteoric
water concentration in discharge, the >95 TU basal ice could have

only been accreted during the years 1963 or 1964.

100000

I iii ILLLI

l
nib-—

 

IO Tritium in Precipitation
50 Tritium in Discharge

  

 

 

 

u-L
O
O
O
O
I
£%”’

 

 

_J
__J

E : ,-

.2 ‘I K !I\

u _ I,‘

'i': \ If R

I" ‘8 ‘\ I'

8’ “I II

_I J I ‘II

“100 \
j \I I .
4 u ngik
j (V? I ’I \. f V I jj-l Q Q
I I.’ I)! I "U“ >‘ /
4' d) IVI J u‘;z‘«:‘j&:€3 2“)
_I mmmmmmomnms v .3
100 I I
I i T I I I I I I j
o O O O O C
o, 0. o 0. o 0.
O O O o O 0
IO (0 N Q (D o
O: O) O) a: O) O
‘- ‘- v- 1- !- N
Year

Figure 14. Simulated Concentration Of Tritium In Vent Discharge
VS. Tritium In Precipitation(1995-1958)

CONCLUSIONS

The SEO, 5D and :31 measured in vent discharge at the
Matanuska Glacier Alaska are within the requisite ranges to
account for variations in these same isotopes in basal ice. A
simple open meltwater/precipitation reservoir fractionation model
shows that the average theorized basal ice fink) and. SD *value
produced from vent discharge is similar to those values measured
in basal ice by Strasser (1996), and that there is a genetic
relationship between vent discharge and basal ice. This result
supports the hypothesis of Strasser et al. (1996) and Lawson et
al. (in review) that basal ice is being accreted in a top down
fashion in the overdeepening at the Matanuska Glacier from
supercooled subglacial vent discharge. Alternatively,

regelation may also be responsible for accretion of this basal

ice. However, basal ice 8mC> values fractionated. beyond 3%

relative to Smo values of vent discharge, which is characteristic
of a melting-refreezing process, have not been observed.
The range of 8WD and. 6D ‘values of theorized basal ice

produced from vent discharge, furthermore, encompass all but the
most negative samples from basal ice profile 95-1. The summer
of 1995 was characterized as ea season with greater than normal

precipitation amounts, thus increasing the average volume of

31

32

deviations in SEQ values from basal ice with respect to
deviations in precipitation, additionally, indicate further that
basal ice formed in the overdeepening at the Matanuska Glacier is
directly influenced by the percentage of recent meteoric water
present in vent discharge. Depleted SEC and 8D values of basal
ice profile 95-1 were presumably accreted during dry years when
less meteoric water existed jmi the subglacial drainage system.
The sampling strategy of Strasser et al. (1996), however,
homogenized discrete basal ice layers inadvertently obfuscating
the true range and variability of ”O, D and UL

The 4.6 TU average measured in vent discharge demonstrates
that an average of 35% meteoric water was present in vent
discharge during the summer of 1995. Based on this percentage,
which may be an over estimate due to the greater than normal
precipitation volumes of summer 1995, the range of 3H in tesal
ice produced froml vent discharge through the years 1958-1995
would be 208.2 to 2.2 TU. These values are consistent with
those measured in profile 95—1 collected by Strasser (1996) and
indicate that time peak 3H concentration measured iJI basal ice
could. only' have been accreted. during 1963 or 1964 from1 vent

discharge.

LIST OF REFERENCES

LIST OF REFERENCES

Arcone, S.A., Lawson, D.E., Delaney, A.J. 1995. Short-pulse
radar wavelet recovery and resolution of dielectric
contrasts within englacial ice of the Matanuska Glacier,
Alaska, U.S.A. J. Glaciol., 41(137), 68—86.

Alley, R.B., Cuffey, K.M., Evenson, E.B., Strasser, J.C., Lawson,
D.E., Larson, G.J. 1996. How glaciers entrain and
transport sediment at their beds: physical constraints.
Quat. Sci. Rev.

Alley, R.B., Lawson, D.E., Evenson, E.B., Strasser, J.C., Larson,
G.J. In press. Super Cooling and debris—rich ice
accretion from water flowing in a distributed subglacial
system within an overdeepening: 2. Theory. J. Glaciol.

Boulton, G.S. 1972. The role of thermal regime in glacial
sedimentation. Ins. of British Geo., Special Bulletin no.
4, 1-19.

Clarke, G.K.C., Collins, S.G., Thompson, D.E. 1984. Flow,
thermal structure, and subglacial conditions of a surge-type
glacier. Can. J. Earth Sci., 21(2), 232-240.

Hubbard, 8., Sharp, M. 1995. Basal ice facies and their
formation in the Western Alps. Arc. and Alp. Res., 27(4),
300-310.

Iverson, N.R. 1993. Regelation of ice through debris at glacier
beds: implications for sediment transport. Geology. 21(6),
559-562.

Iverson, N.R., Semmens, D.J. 1995. Intrusion of ice into
porous media by regelation: a mechanism of sediment

entrainment by glaciers. J. of Geo. Res., 100(B7), 10,219-
10,230.

Jouzel, J., Souchez, R.A. 1982. Melting-refreezing at the

glacier sole and the isotopic composition of the ice. J. of
Glaciol. 28(98), 35—42.

Kamb, B., LaChapelle, E.R. 1964. Direct observation of the
mechanism of glacier sliding over bedrock. J. Glaciol.,
5(38), 159—172.

33

34

Kessler, M.J. 1988. Effective use of low level liquid
scintillation analysis. The Second International Seminar
for Liquid Scintillation Analysis, Proceedings, June 8,
1988, Tokyo, Japan, 256-301.

Lawson, D.E., Kulla, J.B. 1978. An oxygen isotope
investigation of the origin of the basal zone of the
Matanuska Glacier, Alaska. J.Geol., 86, 673—685.

Lawson, D.E., Evenson, E.B., Strasser, J.C., Alley, R.B., Larson,
G.J. 1996. Subglaical supercooling, ice accretion, and
sediment entrainment at the Matanuska Glacier, Alaska.
Geological Society of America, Abstracts with Programs, 28.

Lawson, D.E., Strasser, J.C., Evenson, E.B., Alley, R.B., Larson,
G.J., Arcone, S.A. In review. Super cooling and debris-
rich ice accretion from water flowing in a distributed
subglacial system within an overdeepening: 1. Field
Evidence. J. Glaciol.

Lehmann, M., Siegenthaler, U. 1991. Equilibrium oxygen and
hydrogen isotope fractionation between ice and water. J.
Glaciol., 37(125), 23-26.

O’Neil, J.R. 1968. Hydrogen and oxygen isotope fractionation
between ice and water. J. Phys. Chem., 75(36), 7628-7633.

Osterlund, H.G., Werner, E. 1962. The electrolytic enrichment
of tritium and deuterium for natural tritium measurements.
Tritium in the physical and biological sciences, IAEA, 1,
95—104.

Robin. G.deQ. 1976. Is the basal ice of a temperate glacier at
the pressure melting point? J. Glaciol., 16(74), 183-196.

Souchez, R.A., Jouzel, J. 1984. On the isotopic composition in

8D and 8WD of water and ice during freezing. J. Glaciol.,
30(106), 369-372.

Strasser, J.C., Lawson, D.E., Evenson, R.B., Larson, G.J., Alley,
R.B. Submitted. Rapid basal accretion of ice by glacial
hydraulic super cooling Matanuska Glacier, Alaska. J. Geol.

Strasser, J.C., Lawson, D.E., Larson, G.J., Evenson, E.B.,
Alley, R.B. 1996. Preliminary results of tritium analysis
in basal ice, Matanuska Glacier, Alaska, U.S.A.: evidence
for subglacial ice accretion. Annals of Glaciol., 22, 126-
132.

Strasser, J.C. Unpublished. Processes of subglacial ice growth
and debris entrainment at the Matanuska Glacier, Alaska.
[Ph.D. thesis, Lehigh University, 1996].

35

Weertman, J. 1961. Mechanism for the formation of inner
moraines found near the edge of cold ice caps and sheets.
J. Glacio., 3(30). 965—978.

Weertman, J. 1964. Glacier sliding. J. Glacio., 5(39), 287—
303.

Williams, J.R., Ferrians, O. 1961. Late Wisconsinan and recent
history of the Matanuska Glacier, Alaska. Arctic, 14, 82-
90.

HICHIGQN STR

(III I

I)
53

TE UNIV. LIBRARIES
4IIIIII!III!HIIMIIHIIIIMHI
15948163

)3