The Volcanic and Magmatic Evolution of Volc N Ollag e A High K Late Quaternary Stratovolcano in The Andean Central Volcanic Zone
The Volcanic and Magmatic Evolution of Volc N Ollag e A High K Late Quaternary Stratovolcano in The Andean Central Volcanic Zone
The Volcanic and Magmatic Evolution of Volc N Ollag e A High K Late Quaternary Stratovolcano in The Andean Central Volcanic Zone
54 ( 1993 ) 2 2 1 - 2 4 5
22 l
ABSTRACT
Fecley, T.C., Davidson, J.P. and Armendia, A., 1993. ]-he volcanic and magmatic evolution of Volc~in Ollagfie, a high-K
iale Quaternary' stratovolcano in the Andean Central Volcanic Zone. J. ! blcanol. (i'eoH~erm. Res.. 54:221 245.
Volcfin Ollagfie is a high-K, calc-alkaline composite volcano constructed upon extremely thick crust in the :\ndean Central Volcanic Zone. Volcanic activity commenced with the construction of an andesilic to dacitic composite cone composed of n u m e r o u s lava flows and pyroclastic deposits of the Vinta Loma series and an ox erlying coalescing dome and
coulee sequence of the Chasca Orkho series. Following cone construction, the upper western flank of Ollagiie collapsed
toward the west leaving a collapse-amphitheater about 3.5 km in diameter and a debris avalanche deposit on the lowe~
western flank of the volcano. The deposit is similar to the debris avalanche deposit produced during the Ma? 18. 1980
eruption of Mount St. Helens, U.S.A., and was probably formed in a similar manner. It presently covers an area of 100
km 2 and extends 16 km from the summit. Subsequent to the collapse event, the upper western flank was reIbrmed via
eruption of several small andesitic lava flows from vents located near the western summit and growth of an andesitic dome
within the collapse-amphitheater. Additional post-collapse activity included construction of a dacitic dome and coulde ol
the La Celosa series on the northwest flank. Eield relatinns indicate that vents for the Vinta Loma and post-collapse series
were located at or near the s u m m i t of the cone. The Vinta Loma series is characterized by an anhydrous, two-pyroxenc
assemblage. Vents for the La Celosa and Chasca Orkho series are located on the flanks and strike N55 W, radial to the
volcano. The pattern of flank eruptions coincides with the distribution in the abundance of amphibole and biotite as the
main marie phenocryst phases in the rocks. A possible explanation for this coincidence is that an unexposed fracture or
fault beneath the volcano served as a conduit for both magma ascent and groundwater circulation. In addition to the lava
flows at Ollagfie, magmas are also present as blobs of vesiculated basaltic andesite and marie andesite that occur as inclusions in nearly all of the lavas. All eruptive activity at Ollagiie predates the last glacial episode ( ~ 11.000 a B.P. ), because
post-collapse lava flows are overlain by moraine and are incised by glacial valleys. Present activit~ is restricted to emission
of a persistent, 100-m-high fumarolic steam plume from a vent located within the summit andesite dome.
Sr and Nd isotope ratios for the basaltic andesite and marie andesite inclusions and lavas suggest that they haxe assimilated large a m o u n t s of crust during cDstal fractionation. In contrast, narrow ranges in 143Nd/144Nd and sTSr/~%r in the
andesitic and dacitic lavas are enigmatic with respect to crustal contamination.
Introduction
Previous discussions of magmatism in the
('orrcspondence to. T.C. Feeley, D e p a r t m e n t o f Earth artd
Space Sciences. Universit? o f California, Los Angeles. CA
90024, USA,
222
74"
I
--L
7(3
I
"\.
6,
I
~,~12
PERL
/
(
BOld\' i\
IJarirlacota l i ~
[ aclfic Ocean
CenozoicRecent
Volcanic
Series
-
International
frontier
0i ,
200 km
Capricorn
j'~, ~ ,.-
...,./,'
__
24
t CHILE ~i I :~
/'
)
ARGEN FINA
223
224
68:21'W
21 10'S
68 05'W
7y
\
Sahtr de ( Uhtgt7~"
Explanation
Salar
t ~
,s~
Chile
"x,
/ /
,:~i
[~
colluvitlm arid
alluviuna
salar
deposits
gtacial
moralrle
)
N
(
/ J
[] LaseriesCelosa
[] post-collapse
series
debris ax,alanchc
deposit
i[]
,
/
(
(
7J
r,)
(
Chasca Orkho
series
altered pyroclaslic
deposits
Vinta 1.oma
series
older volcanic
rocks
i/
ignimbrite
Salar
de
San Mart{n
exposed VCIII-,
Cerro
WS::5:SL#~
LkL=::;:W:2:kW:::: ~ -- . . . . .
................................
....................
.- i'J~77 7:777 ;7 7777777 "'(~
/
=================================================
1"
2126'S
68:~21'W
"~ t
!(i ~IT!L~!))
~
~[::
hnnarolc
)i
l
:0
~,,,a,~nch
........
2
6805'W
Fig. 2. Simplifiedgeologic map of VolcfinOllagiie and the surroundingarea based on field workand photo interpretation,
the volcano is indicated by the pristine morphology o f the lavas and persistent fumarolic
emissions from a summit vent. The lava flows,
domes, and pyroclastic deposits of Ollagiie overlie the regionally extensive 5.9 to 5.5 Ma old
Carcote ignimbrite, which is exposed on its
eastern side (Fig. 2; Baker and Francis, 1978 ).
Older, extensively glaciated lavas and domes
(e.g., Cerro Huanaco, Cerro Chanchajapichina, Cerro Chijiliapichina; Fig. 2) not related to the main volcanic center are exposed
to the east and south of Ollagtie but are not
considered in this study.
O F VOLC,&N ( - ) L L A G O E
225
The lavas and pyroclastic rocks have been divided into four eruptive series. In ascending
stratigraphic order they are: (1) the Vinta
Loma series; (2) the Chasca Orkho series; (3)
the post-collapse series; and (4) the La Celosa
series. Compositional and modal data for individual eruptive series are illustrated in Fig-
EVOLUTION
D. Post-Collapse and
La Celosa stage
'
San Maiim
- - (~ll{)O i
\Vcsl
I ~
k,~,
Fig. 3. ( A - D ) Schematic cross-sections showing the eruptive history of Ollagiie. Note that the vertical scale is 2 times the
horizontal. Patterns are the same as in Fig. 2. Question marks indicate rocks erupted during the early stage of Ollagfie's
history, which are not exposed at the surface.
226
T . C FEELEY EI',At..
8
7
La Celosa series
c/3
<
:~
8
7
<
Post-Collapse series
ures 4 and 5, respectively. Grouping of the lavas into the eruptive series was based
principally on vent locations and contact relationships. Due to the difficulty in correlating
individual flows and domes where the vents are
covered or where stratigraphic position is not
obvious from contact relationships, a few lavas were assigned to eruptive series on the basis of flow morphology, mineral mode, and unpublished 4Ar/39Ar age determinations (G.
W~Srner, pers. commun., 1992). Andesites of
relatively uniform composition were the dominant magmas erupted, although significant
volumes of dacite vented on the flank of the
volcano (Fig. 4).
4
3
8
7
6
<
4
3
2
8
7
<
ca:
50 52 54 56 58 60 62 64 66 68 70
SiO 2 wt.%
~7
=~
60
q
40
30
{5~i{ Olivine
':!~ Clinopyr{}xenc
Orthopyroxcn{'
I'lagioclasc
i: Amphibole
ii Biotit{.'
[]
Oide
I
(b
-~ rJ
~
~l
i
3()
20
ll)
i
()0
(-/mSCa ()rkllo
50
e-
40
r,
4{)
-r
1. ,
5
r-
BO
u
5o
~C
I A C{'[oszl 5 e r i c 5
serie>
li
i
,-~
30
ca
20
I0
'o
O
o
2~
~0
q2
54
[~6 $8
7{3
SiO, wt/./~
Fig. 5. Volume% of phenocrysts versus wt.% SiO2 of representative Ollagfic lavas b.~ eruptive series. Data determined by point counting > 1000 points per thin section.
Phenocrysts are defined as larger than 0.3 mm.
228
II
,Amph
"
io yx
S~ale(km)
[~
Fig. 6. Spatial distribution of Ollagiie lavas in which amphibole and biotite are present in greater proportions than pyroxene and vice versa. Compare with Fig. 2.
229
( 1988 ). The debris avalanche deposit is presently only preserved west of Ollagiie in the San
Martin basin where it extends 16 km from the
summit and covers an area of approximately
100 km 2 (Fig. 2). Morphologically it forms a
h u m m o c k y terrain similar to the deposit produced during the May 18, 1980 eruption of
Mount St. Helens, and it was probably formed
in a similar manner (Fig. 7; Francis and Wells,
1988). Reconnaissance sampling within the
debris avalanche deposit suggests that most of
the megablocks are lithologically similar in
composition and phenocryst mode to the older
amphibole-bearing lavas of the Vinta Loma series on the northwest flank (G. W~Srner pets.
commun., 1991 ). The extent of these lavas was,
therefore, likely to have been greater than those
presently exposed. Polygonally jointed andesite bombs are occasionally found on top of the
debris avalanche deposit. It is unclear, however, if these bombs represent a juvenile magmatic component associated with the collapse
event.
Debris avalanche
Late-stage volcanic activity commenced with
sector collapse of the western flank of Ollagiie.
Although the sector collapse event did not destroy the actual summit, it probably resulted in
formation of a collapse-amphitheater about 3
km in diameter on the upper western flank
(Figs. 2 and 3). The collapse event also produced a debris avalance deposit, which was first
recognized and described by Francis and Wells
Post-collapse series
Contemporaneous with a n d / o r following
sector collapse of the upper western flank,
eruptive activity continued with extrusion of
numerous short flows of grey andesitic lava,
which are well preserved on the western flank
of the volcano, and with growth of an andesitic
dome within the inferred collapse amphitheater (Figs. 2 and 3). All of these extrusions are
compositionally very similar (Fig. 4). The lava
flows erupted from the western summit area
and flowed toward the west over the debris avalanche deposit. Young morphological features such as well-developed lev6es and pressure ridges on upper flow surfaces, and Iowalbedo (Rothery et al., 1986) on satellite imagery suggest that these lavas are younger lhan
Chasca Orkho series lavas, although this evidence is not conclusive.
The presently exposed volume of the summit dome is about 0.35 km 3. At its base are
F.C. FEELEY El A I .
~30
Fig. 7. View of the debris avalanche deposit from the western flank of Ollagfie. The deposit forms the hummock) terrain
in front of the white salar deposits in the San Martin basin. Note Cerro La Porufiita, a phreatomagmatic tephra cone
constructed on top of the deposit, on the right. Peaks in the background are Volcans Chela (right) and Palpana (left).
piles a n d t o n g u e s o f a n g u l a r b l o c k s as large as
l 0 m in d i a m e t e r t h a t r e s u l t e d f r o m s m a l l r o c k
a v a l a n c h e s d u r i n g d o m e g r o w t h (Fig. 8; c.f.,
S w a n s o n et al., 1987 ). T h e d o m e p r o b a b l y fills
Fig. 8. View of summit andesite dome from the north. Note the small rock avalanche at the right base of the dome and
the active summit fumarole. The arrow points to a young post-collapse lava flow that underlies the dome.
?_31
232
233
Fig. 9. Photomicrographs of inclusions found in Ollagiie andesitic and dacitic lavas. Width of both photomicrographs is
2.5 ram. Crossed-polarized light. (A) Clinopyroxene phyric andesite inclusion OLA9021i. Euhedral clinopyroxene phenocrysts (center and right) are surrounded by reaction rims of amphibole and are set in a vesiculated glassy groundmass
with acicular plagioclase and amphibole, and more equant Fe-Ti oxide microphenocrysts. To the lower right of center is
a plagioclase xenocryst with sieve-textured core and euhedral overgrowth. (B) Gabbroic inclusion (center) in post-collapse lava OLA9054. Early grown plagioclase and Fe-Ti oxide phenocrysts are partially and completely surrounded by
large clinopyroxene oikocryst.
234
Vinta Loma series. Through time, eruptive activity migrated to the southeast flank where a
sequence of olivine basaltic andesites to amphibole and biotite-bearing andesites and dacites (Chasca Orkho series) was erupted on top
of Vinta Loma flows. The trend toward eruption of more silicic compositions with time in
the Chasca Orkho series implies that crystal
fractionation was operating within a magma
chamber to produce the intermediate composition lavas (c.f. Volc~n Colima; Luhr and
Carmichael, 1980). In this model the tavas
represent leaks from the magma chamber during progressive degrees of differentiation. The
culmination of magmatism to date at Ollagiie
occurred following collapse of the upper western flank. At this stage, eruptive activity largely
returned to the summit area and modal compositions of the lavas became highly variable,
although bulk compositions of erupted magmas were relatively restricted (Figs. 4 and 5 ).
During all of the stages, andesites were the
dominant magmas erupted, although flank
vents produced basaltic andesites and a higher
proportion of dacite than summit eruptions.
Vesiculated inclusions are found in nearly all
of the lavas suggesting that a subvolcanic
magma chamber, from which the andesitic and
dacitic lavas were derived, was repeatedly
fluxed from below with parental mafic lavas.
Vents of the Chasca Orkho and La Celosa series are aligned and strike radial to the volcano
(Fig. 2). In particular, those from the La Celosa series and the southern part of the Chasca
Orkho series are aligned with the summit vent
and strike N55W. The alignment of the La
Celosa, Chasca Orkho, and summit vents suggests that an unexposed NW-striking structure
beneath Ollagiie may have served as a principal conduit for magma ascent and surface
eruption. Similar linear zones of flank vents at
Medicine Lake volcano (Fink and Pollard,
1983), Mount Mazama (Bacon, 1983), and
South Sister volcano (Scott, 1987) have also
been interpreted to result from magma ascent
along fractures radial to the volcanoes. Lavas
0.40
'
'
'
'
'
'
'
'
'
235
10
'
0.35
~ ,
,T-
o"!.
0.30
o*" d
0 ~
i::L~ 0.25
t3~
0.20
mE]
0.15
20
2,0
"~""~
19
T r ' l , , , [ , , ' ~ - , , ~ ! , ~ r
OLA9027i
1,6
-,-:18
1.2
O e
--~17
<
0.8
16
0.4
O []
0" D
15
@.0
1o
0
orqo
o~
Ii
. . . .
'
'
'
'
'
'
'
m_
X [~a r i n a c o t a
[~
+ Sail Pedro
I
+4-+
~ 3
0
[]
~ ~i[:]
X
' - ~
Cumulate inc!uston
zk
Magmatic Inclusions
La Celo sa
Post Collapse
[]
Chasca Orkho
Vinta Loma
~2
mK
basaltic
and~sit~
50
Low-K
andesite
dacite
58
SiO
62
2
wt%
66
;'~
i
i
54
70
50
54
, =J
58
SiO
62
2
L ~_J_~
66
70
wt.%
Fig. 10. MaJor-element oxide variation versus SiO2. K2O classification b o u n d a r i e s ( italics ) are from Peccerillo and Taylor
(1976). N o m e n c l a t u r e for the volcanic rocks is indicated along the b o t t o m of the diagram. Circled field indicates the
range of inclusion compositions. On the K 2 0 versus SiO2 diagram, represents Parinacota lavas from I)avidson et al.
( 1 9 9 0 b ) and X represents San Pedro lavas from O'Callaghan and Francis ( 1 9 8 6 ) for comparison with Ollagtie data.
236
TABLE 1
Representative major (wt.%) and trace (ppm) element analyses of Ollagiie rocks
Vinta Loma series
Sample:
SiO2
TiO2
AI203
FeO*
MgO
MnO
CaO
Na20
K20
P205
LOI
Total
Rb
Sr
Cr
Ni
OLA9014
60.9
0.89
16.9
5.2
2.6
0.07
5.1
4.0
2.7
0.32
1.0
99.6
77
530
25
8
OLA9023
OLA9031
63.0
0.82
16.6
5.0
2.4
0.07
4.9
3.9
3.0
0.23
1.1
101.0
91
527
17
5
58.6
0.95
16.7
5.7
3.5
0.11
6.0
3.7
2.5
0.25
0.9
99.0
OLA9032
64.4
0.86
16.7
5.1
1.4
0.10
4.0
4.1
3.2
0.24
0.7
100.8
OLA9056
62.2
0.85
16.4
4.9
2.3
0.09
4.9
3.5
3.2
0.28
1.2
99.9
OLA9058
~0,2
ii.84
i(~.7
~.8
3.2
~LII
L5
',.8
2.8
i).22
i~2
lo0.?,
69
547
43
8
96
494
6
4
95
722
11
6
,';4
>~8
B~
:,
OLA9020
OLA902 t
OLA9026
Ol, A9027
OLA9015
Na20
K20
P20~
LOI
Total
67.0
0.55
15.4
3.3
1.2
0.05
3.1
4.1
3.8
0.17
1.2
99.7
Rb
Sr
Cr
Ni
118
422
6
3
CaO
OLA90 t 6
52,7
1.33
16.1
8.6
6.6
0.12
8.5
3.4
1.6
0.24
0.1
99.3
37
702
221
75
65.6
0.57
15.5
3.5
1.4
0,06
3.5
4.1
3.7
0.18
1.3
99.5
126
441
10
25
63.5
0.77
16.2
4.6
2.1
0.06
4.4
4.4
3.2
0.25
0.7
100.0
117
451
14
6
60,1
0.91
16.8
5.4
2.9
0.08
5.2
3.6
2.9
0.23
1.4
99.7
89
485
30
7
62 7
(!.80
i6.3
47
2.3
007
45
L0
~i
0 22
i.3
q99
1~5
436
,
elevated and elements not compatible in plagioclase are depleted relative to other samples
(Fig. 10).
The compositional variations highlight the
importance of crystal-liquid fractionation as a
petrological process in the evolution of the
Ollagiie rock suite. For example, the strong decrease in A1203 and P205 in samples with
~3 7
.._
La Cclosa series
Post-collapse series
Sample:
OLA9037
OLA9047
OLA9054
()LA31
OLA33
Sit),
TiO2
AI:O?.
FeO*
MgO
MnO
CaO
NaeO
K20
P205
k()l
Tolal
62.6
0.79
16.6
4.9
2.3
O. 11
4.6
4.3
2.9
0.27
1.2
100.8
60.5
0.90
16.6
5.4
3.1
0.09
5.5
3.6
2.8
0.22
1.4
100.1
60.8
0.94
16.8
5.4
2.7
O. 10
5.2
4.1
2.7
0.26
1.1
100.2
64.7
0.67
16.2
3.~I
1.8
0.06
3"
4.0
3.6
0.19
0.9
100.1
66.4
0.57
15.7
3.3
1.2
0.OS
3.0
4.1
3.8
0.1 ~
1.3
100.0
Rb
Sr
Cr
Ni
93
506
15
5
89
491
23
6
82
490
30
8
116
460
25
7
123
424
14
4
Inclusions
Sample:
SiO2
TiO2
AI20:~
FeO*
Mg()
MnO
CaO
Na20
K20
P2Os
L()I
Total
Rb
Sr
(7I'
Ni
OLA901 li
59.7
0.90
16.5
5.5
3.0
0.09
5.3
3.3
2.8
0.24
1.8
99.2
71
652
78
9
OLA9015i
56.2
1.16
17.1
7.5
3.3
0.10
6.0
3.6
2.3
0.36
1.4
99.0
63
610
7
8
OLA9021i
57.1
0.94
16.8
5.7
3.8
0.08
7.6
3.5
1.9
{).26
1.9
9~).6
43
650
5 "v
27
OLA9025i
59.2
0.97
16.8
5.9
3.5
0.08
6.1
3.9
2.3
0.27
0.7
99.9
54
600
53
13
OLA9026i
57.1
1.18
16.6
6.6
4.0
0.09
6.7
3.1
2.0
0.26
1.5
99.2
59
520
79
13
OLA14
()LA32
56.4
1.07
17.4
7.3
3.9
0 1
6.8
3.7
2.2
0.28
0.5
99.7
53.1
1.42
173
8.3
5.2
0.11
8.7
3.5
1.5
0.27
1.4
100.8
55
603
31
14
45
641
69
15
OLA91)27i
53.1
1.03
19.7
7.8
2.1
0.13
8.5
3.9
1.0
(}.37
1.5
99.1
32
813
7
11
All data collected by standard XRF techniques on dried rock powders using a Rigaku 3070 X-rat spectrometer at the University
of Southern California. Precision on major and trace elements is estimated at 1% (one-sigma standard deviation ) except for Cr
( < 3.5%). Analyses of Ni less than 30 p p m are regarded as semiquanlitative. LOI determined by igniting at 900:C a separate
aliquot of powder. FeO* is total Fe as Fe 2+. Samples without -90- prefix are from W6rner et al. ( 1992 ).
238
T.C.
140
o
120
~m D
IOO
E
O.
o. 80
J~
rr
60
40
20
9O0
800
E 700
El.
C~ 600
500
400
300
g
250
D
[]
[]
200
E
(D..
Q-150
c3
IO0
50
lOO
[]
8o
E6o
Q..
c~
FEELEY
El
A1.
Isotopes
Z40
[]
20
0
5O
54
58
62
66
SiO wt.%
2
Fig. 11. Trace-element variation versus SiO2. Data symbols are the same as in Fig. 10. Circled field indicates the
range of inclusion compositions.
239
0.5125
0.5124
[T[~
0.5123
0.51300
0.51280
0.5126{)
,
//
/L 5 1 2 2
0.51320
~ ~ A v
7 --~
i
.~_~__-
~'~F Ax~k.
/
NVZ
(0-2S)
0,706
_ 0.707
0,708
0.709
0.71 I)
g33_42S)
" ~ 0.51240
~Z
~
0.31220
cvz
0.7060
(17.5-26%)
0.7100
0.7140
0.7180
",~Sr/~',Sr
Fig. 12. Isotope data for select Ollagtie lavas and inclusions compared with island-arc tLlt') data and data from the
Andean Northern (NIT), Central (C17), and Southern (SIT) volcanic zones. Diagonall? ruled field in inset sho~s the
compositions of rocks of the Altiplano-Puna volcanic complex from de Silva (1987). Symbols are the same as in Fig. 10.
Arrows point to assumed composition of bulk Earth (BE). After Davidson et al. {1990b ).
240
T . c , F E E L E Y E T AL.
0.709
A
0.708
{)
r.g3
~.: ~
(I.707 -
r./3
/....x
{17
~ 1/ / ~ 0795
0.706 -
0.00
[]
o6
/ .-" ;-'--~1
~"- II{}{}
"
AFC
~- 7/ /"./.- - " / /
Model
.'/-.,Z tr~-~o 95
/
1
2
Ds,
I ~2 5 0.3
1.25 {}.3
0.3
0.8
{}.5
{}.5
0.05
0.10
0. t 5
0.20
0.25
{}.3{}
Rb/Sr
0.709
0.708
--
,c~./~
g
0.707
A2ZX
0q
....F C ' a
[]
t-CoU~p,~
<>
st-iie ~
r,3
0.706 -
0.00
0.05
0.10
O. 15
0.20
0.25
0.30
Rb/Sr
Fig. 13. (A) Sr isotope constraints on bulk mixing and assimilation-fractional crystallization (AFC; DePaolo, 1981 )
models for Ollagiie inclusions and basaltic andesite lavas. The assumed contaminant in all of the models has aTSr/
S6Sr=0.725 and Rb/Sr ~ 0 . 4 (Kntiver and Miller, I981 ). The legend shows the bulk distribution coefficients (Ds, and
DRb ) and r values used to calculate the model curves. Tic marks on Bulk Mixing curve indicate the percentage of silicic
endmember in the mixture. Tic marks on AFC curves indicate the amount of original magma remaining (F). Symbols
are the same as in Fig. 10. See text for discussion. (B). Sr isotope constraints on differentiation models for Ollagtie
andesitic and dacitic lavas. See text for discussion.
~4 ]
(the rate of the mass of crust assimilated relative to the mass of crystals fractionated; DePaolo, 1981 ). The dashed model curve (AFC
3 ) shows the effect of simultaneously decreasing r and increasing the bulk distribution coefficient for St.
Model curves labeled Bulk Mixing and AFC
1 were constructed to simulate the effect that
differentiation under deep crustal conditions
has on isotopic compositions and trace element ratios. In the deep crust we assume that
ambient temperatures are high, allowing large
amounts of assimilation relative to crystallization. Although a bulk distribution coefficient of 1.25 for Sr is moderate to somewhat
high for basaltic andesite to andesite systems,
Sr decreases with increasing SiO2 and was
therefore compatible during differentiation
(Fig, 11 ). The curves AFC 2 and AFC 3 were
constructed to simulate the effect of AFC at
shallower crustal levels where temperatures are
lower so the a m o u n t of crust assimilated for a
given a m o u n t of crystallization (r) is less, and
plagioclase constitutes a larger percentage of
the crystallizing assemblage. We interpret the
data for the magmatic inclusions and mafic lavas to be more compatible with deep-crustal
differentiation involving large amounts of
crustal assimilation.
In contrast to the data trends for the inclusions, Figure 13B illustrates that when viewed
on the scale of individual eruptive series where
field evidence indicates that the rocks are comagmatic, the andesitic and dacitic lavas have
little isotopic variability. It is, therefore, possible that these lavas have undergone little to
no crustal contamination during differentiation. It is especially difficult to demonstrate
AFC trends for Ollagiie andesitic and dacitic
lavas on Figure 13B because at any point along
the data array of the inclusions, which we inter
to be parental magmas to the andesite and dacire lavas, it is possible to begin a horizontal
fractionation trend through the compositions
of the lavas.
In addition to a major change in the a m o u n t
of assimilation, there is a change in the type of
242
tory of these and other CVZ volcanoes probably results from oversteepening of the edifices
due to eruption of viscous andesitic and dacitic magma, as discussed by Francis and Wells
(1988).
Lavas preceding the debris avalanche at San
Pedro are predominantly basaltic andesites,
whereas those erupted after the debris avalanche are mainly andesites and dacites
(O'Callaghan and Francis, 1986). At Parinacota, lavas erupted prior to the collapse event
are generally more silicic (andesites to rhyolites) and have a larger compositional range
relative to post-collapse lavas (andesites to basaltic andesites; W6rner et al., 1988). O'Callaghan and Francis (1986) found that postcollapse magmatism at San Pedro produced a
succession of four eruptive groups. Within each
group increasingly more silicic compositions
were erupted with time, similar to the pattern
for the Chasca Orkho series at Ollagfie. These
field observations are consistent with chemical
data suggesting that crystal fractionation, with
or without crustal contamination, is an important petrologic process in upper crustal magma
chambers beneath stratovolcanoes in the CVZ
and elsewhere. The significance of amphibole
in CVZ andesites and dacites is presently unclear and requires additional work. At Parinacota, amphibole is an abundant phenocryst
phase in pre-collapse lavas, yet it is virtually
absent in post-collapse lavas (W6rner et al.,
1988 ). Data on the abundances of phenocrysts
in San Pedro lavas are not available.
A persistent problem in CVZ magmagenesis
is the extent to which the andesitic and dacitic
lavas have been affected by crustal contamination. Thorpe et al. (1976), Francis et al.
( 1977 ), and O'Callaghan and Francis ( 1986 )
published rare earth element, Sr isotopic, and
major- and trace-element data, respectively, for
San Pedro lavas. 87Sr/S6Sr ratios of San Pedro
lavas are high (0.7055 to 0.7070) although they
are systematically lower than those of Ollagfie
lavas (Fig. 14; Francis et al., 1977). Like
Ollagiie andesites and dacites they also do not
exhibit a clear correlation with any index of
0.7080
bhi
0.7070
O.7060
0.7050
~
50
I
54
I
58
~_ ,
62
I
66
.......
71)
SiO2 w t . %
Fig. 14. Comparison of Sr isotopic compositions of lavas
from Parinacota, San Pedro, and Ollagtie. Data from
Davidson et al. (1990b) and Francis et al. ( 1977 ).
differentiation (Francis et al., 1977). Nonetheless, Francis et al. (1977) infer that San
Pedro lavas were contaminated by lower continental crust because Sr isotopic ratios are elevated relative to Sr isotopic ratios of lavas
from Ecuador (87Sr/86Sr "~ 0.7044), where the
crust is 20-30 km thinner than in northern
Chile. O'Callaghan and Francis (1986) successfully duplicated major- and trace-element
trends of San Pedro lavas by fractional crystallization calculations without accompanying
crustal assimilation. They also infer crustal assimilation, however, mainly on the basis of
disequilibrium phenocryst textures. At Parinacota, 87Sr/86Sr ratios are very similar to those
from San Pedro and also show little correlation with differentiation (Fig. 14). Davidson
et al. ( 1990b, 1991 b ) explained this feature as
a result of the establishment of "baseline" isotopic compositions in parental mafic magmas
during contamination in deep-crustal magma
chambers, followed by rise and further differentiation of magmas in shallower crustal
magma chambers with or without subsequent
contamination. This model is very similar to
the one proposed here, except that parental
mafic magmas that fed shallow crustal magma
chambers at Oltagiie were not isotopically homogeneous. It thus appears that shallow-level
magma-chamber processes do not result in significant or systematic changes in radiogenic
isotopic compositions. This may result from
Conclusions
Field relations indicate that cone growth at
Volc~in Ollagiie evolved during at least four
main eruptive stages with an intervening sector collapse event between the second and third
stage. During all of the stages, andesitic magma
of relatively uniform composition was the
dominant eruptive product, although magmas
that vented on the flanks included more mafic
types and a higher proportion of dacite.
Quenched mafic inclusions in nearly all lavas
preserve evidence that the magmatic system
beneath Ollagiie was repeatedly fluxed from
below with parental basaltic andesites and
mafic andesites. Petrographic features such as
the large proportion of xenocrysts in the inclusions indicate that crustal contamination was
an important process in their petrogenesis.
Whole-rock geochemical and isotopic trends of
the inclusions indicate they are not simple twocomponent mixtures between a more mafic
magma and exposed andesite and dacite lavas,
however.
It is possible to explain geochemical trends
of the basaltic andesites and mafic andesites by
differentiation at deep crustal levels where
thermal conditions permit large degrees of assimilation relative to factionation. The restricted range in isotopic compositions of the
andesitic and dacitic lavas suggests that these
rocks may have undergone smaller amounts of
crustal assimilation during differentiation at
shallow (cool) crustal conditions. Large
amounts of crustal assimilation are possible if
the upper crust is chemically and isotopically
similar to the andesites and dacites. The pres-
243
Acknowledgements
Supported by National Science Foundation
grant EAR-8915808 to Davidson. We thank the
Servicio Geologico de Bolivia (GEOBOL) for
arranging field logistics, Peter Holden for assistance in the isotope lab, Dave Mayo for help
with the XRF analyses, and Anne Loi for computer support. Wendy Bohrson assisted during
field work. This work benefitted from discussions with Gerhard WiSrner, Shan de Silva, and
Mary Reid. Unofficial reviews by Anita Grunder and Gerhard W6rner and official reviews
by Jim Luhr and Bill Rose resulted in significant improvements to this manuscript.
References
Bacon, C.J., 1983. Eruptive History of M o u n t M a / a m a
and Crater Lake caldcra, Cascade Range, I;SA. J. Volcanol. G e o t h e r m . Res., 18:57-115.
Bacon, C.J., 1986. Magmatic inclusions in silicic and intermediate volcanic rocks. J. Geoph3 s. Rcs., 9 I: 6!)91 6112.
Bacon. C.R. and Metz, J.. 1984. Magmatic inclusions in
rhyolites, c o n t a m i n a t e d basalts, and compositional
zonation beneath the ('oso ~olcanic field. ('alifi)rnia.
Contrib. Mineral. Petrol., 85: 346-365.
Baker, M.C.W. and Francis, P.W., 1978. U p p e r Cenozoic
~olcanism in the central Andcs--~\ges and volumes.
Earth Planet. Sci. Lctt., 41: 175-187.
Baranzangi, M. and lsacks, B.L., 1976. Spatial distribution of earthquakes and subduction of the Nazca Plate
below South America. Geology, 4: 686-692.
Coira, B., Davidson, J., Mpodozis, C. and Ramos, V..
1982. Tectonic and magmatic c~olution of the Andes
of n o r t h e r n Argentina and Chile. Earth-Sci. R e \ , 18:
303-332.
Davidson, J.P., 1988. U p p e r plate controls on the c o n >
positions of A n d e a n magmas. Geol. Soc. A.m., ,\bstr.
Progr., 20: A 10.
Davidson, J.P.. de Silva, S i . , Holdcn, P. and Hallida?.
A.N., 1990a. Small scale disequilibrium in a magmatic
inclusion and its more silicic host. J. Geophys. Res.,
95: 17,661-17,675.
Davidson, J.P., McMillian, N.J., Moorbath, S., W~rner,
G., Harmon, R.S. and Lopez-[!scobar, l.., 19%~()b The
244
Francis, P.W., McDonough, W.F., Hammill, M., O'Catlaghan, L.J. and Thorpe, R.S., 1984. The Cerro Purico
shield complex, north Chile. In: R.S. Harmon and B.A.
Barreiro (Editors), Andean Magmatism, Chemical and
Isotopic Constraints. Shiva Publishing, Nantwich, pp.
106-123.
Galli-Oliver, C., 1967. Pediplan in northern Chile and the
Andean uplift. Science, 158: 653-655.
Gans, P.B., Mahoood, G. and Schermer, E.R., 1989. Synextensional magmatism in the Basin and Range Province: A case study from the eastern Great Basin. Geol.
Soc. Am., Spec. Pap. 233.
Gill, J.B., 198t. Orogenic Andesites and Plate Tectonics.
Springer-Verlag, Berlin, New York, Tokyo.
Grove, T.L. and Donnelly-Nolan, J.M., 1986. The evolution of silicic lavas at Medicine Lake Volcano, California: Implications for the origin of compositional gaps
in the calc-alkaline series lavas. Contrib. Mineral. Petrol., 92: 281-302.
Harmon, R.S., Barreiro, B.A., Moorbath, S., Hoers, J.,
Francis, P.W., Thorpe, R.S., Deruelle, B., McHugh, J.
and Viglino, J.A., 1984. Regional O-, Sr-, and Pb-isotope relationships in late Cenozoic calc-alkalinc lavas
of the Andean Cordillera. J. Geol. Soc. London. 141:
803-822.
Hawksworth, C.J., Hammill, M., Gledhill, A.R., van Calsteren, P. and Rogers, G., 1982. Isotope and trace element evidence for late stage intra-crustal melting in the
high Andes. Earth Planet. Sci. Lett., 58: 240-254.
lsacks, B.L., 1988. Uplift of the Central Andean Plateau
and bending of the Bolivian Orocline. J. Geophys. Res.,
86:10,153-10,192
James, D.E., 1971. Plate tectonic model for the evolution
of the Central Andes. Geol. Soc. Am. Bull., 82: 33253346.
James, D.E., 1981. Role of subducted continental material in the genesis of calc-alkaline volcanics of the central Andes. Geol. Soc. Am. Mem., 154: 769-790.
James, D.E., 1982. A combined O, Sr, Nd, and Pb isotopic and trace element study of crustal contamination
in central Andean lavas, I. Local geochemical variations. Earth Planet. Sci. Lett., 57: 47-62.
James, D.E., Brooks, C. and Cuyubamba, A., 1976. Andean Cenozoic volcanism: Magmatism genesis in the
light of strontium isotopic composition and trace-element geochemistry. Geol. Soc. Am. Bull., 87: 592-600.
Jordan, T.E.M. and Gardeweg, P., 1989. Tectonic evolution of the late Cenozoic Central Andes (20 -33 S).
In: Z. Ben-Avraham (Editor), Cenozoic and Mesozoic
Evolution of the Pacific Margin. Oxford Univ. Press,
London, pp. 193-207.
Klerkx, J., Deutsch, S., Pichler~ H. and Zeil, W., 1977.
Strontium isotopic composition and trace element data
bearing on the origin of Cenozoic volcanic rocks 0fthe
central and southern Andes. J. Volcanol. Geotherm.
Res., 2:49-71.
245
Siegers, A., Pichler, H. and Zeil. W., 1969. Trace element
abundances in the '%ndesite'" formation of northern
Chile. Geochim. Cosmochim. Acta. 33:882-887
Stern, C.R., 1991. Role of subduction erosion in the generation of Andean magmas. Geology. 19:78-81.
Swanson, D.A., Dzurisin, D., Holcomb, R.T., lwatsubo,
E.Y., Chadwick, Jr, W.W., Casadevall, T.J,, Ewert, J.W.
and Heliker, C.C.. 1987. Growth of the lava dome at
Mount St. Helens, Washington, (LISA}, 1981-1983.
In: J.H. Fink (Editor), The Emplacement of Silicic
Domes and Lava Flows. Geol. Soc. Am.. Spec. Pap.
212.
Thorpe, R.S., Francis, P.W., Hammill, M. and Baker.
M.C.W., 1982. The Andes. In: R.S. Thorpe (Editor),
Andesites. Wiley, New York, NY, pp. 187-205.
Thorpe, R.S., Potts, P.J. and Francis. P.W., 1976. Rare
earth data and petrogenesis of andesites from the N.
Chilean Andes. Contrib. Mineral. Petrol., 54: 65- 78.
Tosdal, R.M., Clark, A.H. and Farrar, E.. 1984. Cenozoic
polyphase landscape and tectonic evolution of the
Cordillera Occidental. southernmost Peru. Geol. Soc.
Am. Bull., 95: 1318-1332.
W6rner, G., Harmon, R.S., Davidson, J., Moorbath, S.,
Turner, D.L., McMillian, N. and Nyc, C., 1988. The
Nevados de Payachata volcanic region ( 18 S / 6 t ) W .
N. Chile ). 1. Geological, geochemical, and isotopic observations. Bull. Volcanol., 50: 287--303.
WOrner, G., Moorbath, S. and Harmon, R.S., 1991. Alongstrike geochemical variations in the Andean central
volcanic zone (CVZ) ofN. ('hile ( 17-22 S). EOS, 72:
520.
W6rner, G., Moorbath, S., Entenmann. J., gopez-Escobar, L., Horn, S., Harmon, R.S. and Davidson, J.P.,
1992. Variciones Geoquimicas, locales x regionales, en
el frente volcanico (vaternario de los Andes Centrales
(17 30'-22 00'S, Nortede Chile. Rex'. Geol. ('hiD,
19: 37-56.