~

Journal of Sourh American Earth Scimc<.v, Vol. 9, Nos 1/2, pp. 91-I09, 1996
Pergamon Cl)pyright ti 1996 Elsevier Science Ltd & Earth Sciences & Rcsourccs InstitulC
Printcd in Greal Britain. Ali righl~ reserved
0895-9811/96 S 15.IK> + 0.IK>
PII: 80895-9811 (96)00030-2

Trace and rare-earth element characteristics of acidic tuffs from

Southern Pero and Northern Bolivia and a fission-track age for the
Sillar of Arequipa
1
N. VATIN-PERIGNON, 10. POUPEAU, 2 R.A. OLIVER, 3A. LAVENU, 1E. LABRIN, 1F. KELLER
AND 1·2 L. BELLOT-GURLET
1
Laboratoire de Géologie, URA 69 CNRS, Université Joseph Fourier de Grenoble, 15, rue Maurice
Gignoux, 38031 Grenoble Cedex, France
2
Institut Laüe-Langevin, BP 156, Avenue des Martyrs, 38042 Grenoble Cedex, France
3
0RSTOM, UR lH, 213 rue La Fayette, 75480 Paris Cedex 10, France et Laboratoire de
géodynamique et modélisation des bassins sédimentaires, IPRA, Université de Pau et des Pays de
l' Adour, avenue de l'Université, 64000 Pau, France
Abstract - Trace-element and REE data of glass and pumices of acidic tuffs and related fall deposits erupted in southem Peru
and northem Bolivia between 20 and 0.36 Ma display typical characteristics of subduction related continental are magmatism of
the CVZ with strong LILEIHFSE enrichment and non enrichment of HREE and Y. Geochemical variations of these tuffs are linked
to subduction processes and controlled by changes in tectonic regimes which occured with each Quechua tectonic pulse and
affected the astenospheric wedge and both the dowgoing and the overriding lithospheres. During Neogene - Pleistocene times,
tuffs erupted in northem Bolivia are typically enriched in Zr, Hf, Th, Ba, LREEs and other incompatible elements and incompat-
ible /Yb ratios are much higher relative to !hose erupted from southem Peru, ata given Si0 2 content (65-67 wt. for dacites, 72-73
wt.% for rhyolites). Zr/Hf ratios increase eastward from 27 to 30 and Ce/YbN ratios from 11 to 19 reflecting the variation of degree
of wedge contribution. Fractionation of the LREE over the HREE and fractionation of incompatible elements may be due to their
heterogeneous distribution in the magma source. More highly fractionated REE pattems of Solivian tuffs than Peruvian tuffs are
attributed to variable amounts of contamination of magmas by lower crust. After the Quechua compressional event at - 7 Ma, Sr/Y
ratios of tuffs of the same age, erupted at 150-250 km or 250-400 km from the Peru-Chile trench, increase from southem Peru
to northem Bolivia. These differences may be attributed to the subduction of a swarm oceanic lithosphere under the Solivian Alti-
plano, leading to partial melting of the sudbucted lithosphere. New FT dating of obsidian fragments of the sillar of Arequipa at
2.42± 0.11 Ma. This tuff dates the las! Quechua compressional upper Pliocene phase (-2.5 Ma) and confirms that the sillar is not
contemporaneous with the Toba 76 tuff or the Perez ignimbrite of northem Bolivia. Geochemical characteristics of tuffs erupted
before and after this last compressional phase remained the same and provide evidence that the upper Miocene (-7 Ma) compres-
sional deformations played the most importan! role on the variability of the geochemical characteristics of the southem Peruvian
and northem Solivian tuffs. Copyright© 1996 Elsevier Science Ltd & Earth Sciences & Resources Institute

INTRODUCTION (CVZ, Thorpe et al., 1984) between 15ºS and 18ºS.

In the Central Andes, large volumes of Neogene acidic These tuffs were emplaced during Neogene and early
tuffs cover large areas of the Altiplano of Bolivia and the Quaternary times (Early Miocene to Pleistocene) in the
Puna of northern Chile and Argentina. These have been Pacific piedmont and the south of the western Cordillera
extensively studied by numerous workers (see Schmitt- in Peru and in the northern part of the Altiplano and the
Riegraf and Pichler, 1988; de Silva, 1989, de Silva and eastern Cordillera piedmont (La Paz and Titicaca basin)
Francis, 1989 and refs. therein). Magmas of these volumi- in Bolivia. A detailed discussion of the geochronological
nous silicic pyroclastic flow deposits are considered as evolution of this part of the Bolivian Central Andes based
the response to uplift and thickening of the crust during on new K-Ar ages from volcanics including sorne of these
different tectonic pulses (Isacks, 1988; de Silva, 1989, de tuffs is given in a previous work (Lavenu et al., 1989). The
Silva and Francis, 1991). A variety of models have been specific goals of the present study are to provide a descrip-
proposed for the genesis of these silicic magmas, includ- tion of the physical and chemical characteristics of these
ing crustal anatexis and combined assimilation-fractional tuffs and to interpret their geochemical variations in terms
crystallization processes and (e.g. Thorpe et al., 1979, of the source region and their tectonomagmatic signifi-
Francis et al., 1980, Hawkesworth et al., 1982, James, cance in the context of changes in tectonic regimes of the
1984, Harmon et al., 1984), melting, assimilation, frac- southern Peruvian and northern Bolivian Andes between
tionation and homogenization (MASH) processes (David- 20 - 0.3 Ma. In addition, new fission track (FT) ages of
son et al., 1990, Davidson and de Silva, 1992) and large three obsidian fragments from the sillar of Arequipa are
scale of melting of middle to lower continental crust, per- presented and discussed here, because these tuffs have
haps in association with lithospheric thinning (Hawkes- important geotectonic implications.
worth and Clarke, 1994, Francis and Hawkesworth, 1994).
GEOLOGICAL BACKGROUND
Our study is focused on sorne tuffs and fall deposits
related to the subduction of the Nazca plate beneath west- Tuffs and related fall deposits of the CVZ, between
ern South America in the Andean Central Volcanic Zone 15ºS and 18ºS, are found along a west-east geographic

91
92 N. VATIN-PERIGNON et al.

traverse beginning in the Arequipa Province and continu- ered as a nonwelded ignimbritic tuff (Jenks and Goldich,
ing into the northern Bolivian Altiplano (Fig. 1). These 1956). This tuff infilled the valleys and was also depos-
tuffs are divided in four groups belonging to the main ited on relative flat surfaces. It is one of the most inter-
units of the morphostructural zoning of this part of the esting tuffs of the Arequipa Province due to its use as a
Central Andes. These paleogeographic domains are, from building stone. Lefévre (1979) recognized the high con-
west to east, the Pacific piedmont (coastal range) and the tent of juvenile fragments and the widespread vertical (up
western Cordillera (high volcanic are) in southern Peru, to 200 m thick) and lateral area of dispersa! of the SA tuff
the Altiplano and the eastern Cordillera in northern and observed the presence, but not everywhere, of com-
Bolivia (La Paz and Titicaca basin on the western flank of paction and welding of pumiceous fragments. A quarry
the eastern Cordillera). The study area is located 150 to located along the Quebrada Añas Huayco at the north-
300 km from the Peru-Chile trench axis over a moder- western edge of the town of Arequipa shows a partial
ately dipping segment of the subducted Nazca plate (aver- section of the SA tuff. The base is not exposed and the
age 25º-30º E from 25- to 100-km depth) and about 100- visible section consists at least of two ash-flow units with
130 km above the top of the seismic zone. The crustal a total thickness of about 10 m. The working of the
thickness is about 70 km beneath the western Cordillera quarry aids the recognition of a fine-grained )ayer at the
and the western part of the Altiplano and 50-55 km base of the upper flow unit suggesting different eruptive
beneath the eastern Cordillera. The subducted lithosphere pulses (Sparks et al., 1973) difficult. The SA tuff could
is separated from the continental crust by a well- have been an ~mplacement during a single volcanic epi-
developed wedge of astenospheric mantle (Grange et al. sode or a series of events in rapid succession (Fig. 2A
1984) and B).
A compilation of available K-Ar dates has provided The lower ftow unit (3-4 m thick) is a lithic-rich and
a geochronology upon which the tectonic evolution of well indurated tuff. At the base, both white and oxidized
these regions has been assessed by various authors (e.g. fibreous and vesicular glassy pumices and glass fragments
Tosdal, 1981; Bellon and Lefévre, 1976; Noble et al., all tend to be small. Towards the top of this unit, the
1985). More recen ti y, new KJ Ar dates on undeformed pumices increase in size ( ~ 10-15 cm) and relative abun-
or slightly folded tuffs from northern Bolivia after 1O dance. Lithic clasts (up to 25%) are concentrated particu-
Ma and their lithological position was discussed in larly at the base of the unit and consist mainly of
detail by Laven u (1986) and Laven u et al., ( 1989) and subrounded oxidized andesite fragments ( ~ 5 cm) of the
a revised stratigraphy for the SW Peru has been estab- underlying lower Barroso formations. Isolated crystals of
lished by Sébrier et al., (1988). Four phases of explo- plagioclase, biotite and magnetite are abundant in the
sive volcanism can be considered, using the geological fine- to medium-grained pumiceous and ashy matrix.
time scale of Odin ( 1994 ): 1) the early Miocene
sequences 20.3 - 15.8 Ma) from the Pacific piedmont The upper jiow unit (6 m exposure in the quarry) is an
and the western Cordillera, 2) the middle to late indurated, homogeneous, pumice-rich (30%) rhyolitic tuff
Miocene sequences (9.1 - 5.5 Ma) from the western which shows large columnar joints indicating emplace-
Cordillera, the Altiplano especially with the well- ment at a high temperature. The matrix is fine-grained,
known tuff so-called the Toba 76 tuff and the lower part crystal and lithic clast-poor ( ~ 10%, Fig. 2 C). Grey to
of the La Paz and Titicaca basin, 3) the Pliocene yellow glassy pumices are fragile, highly vesiculated and
sequences (5 - 2 Ma) which are present in all the chain their phenocryst content is low ( ~ 5% ). Coarse pumices
and its piedmont especially the Perez ignimbrite in occur at the top of this unit. The matrix is crystal-rich
Bolivia and 4) the Pleistocene sequences (1.6 - 0.4 Ma) with large plagioclase, biotite, Fe-Ti oxides as well as
with tuffs interbedded within formations of the upper apatite and rare hornblende. Obsidian glass fragments
Barroso Formation in Peru and contemporaneous of (SAI-90, SA2-90 and SA3-90) are abundant and petro-
glaciations in Bolivia. The chronology shows that this graphically identical to pumices.
explosive activity appears to have been significantly The top of the SA tuff is separated from an overlying
reduced between 15 and 10 Ma but continued to be salmon-pink tuff by an irregular erosion surface. This tuff
important from the late Miocene until recently (eruption in turn is overlain by glacial conglomerates, coarse fluvial
of the Huaynaputina volcano, east of Arequipa, on Feb- gravels and lahars which cap the mesa in the Arequipa
ruary 18, 1600; Torbio, 1899). are a.
The stratigraphic positions and age determinations of
tuffs and related fall deposits analysed in this study from
Fission-track Dating of Obsidian Samples of the SA
the western Cordillera piedmont in southern Peru to the
Tu.ff
eastern Cordillera piedmont in Bolivia (La Paz and Titi-
caca basin) are given in Table 1. The FT analytical procedure is described in the
appendix and analytical data are summarized in Tables
Stratigraphy ami New Fission-track (FT) Age of the
2 and 3. Track size analysis, as reported in Table 2 and
Sillar of Arequipa
illustrated in Fig. 3, shows that in the three samples
The well-known rhyolitic ash-flow tuff, so-called the measured, fossil spontaneous fission tracks are, on the
"sillar" of Arequipa (SA tuff) by Fenner ( 1948) is consid- average, 20% shorter than induced tracks, indicating
 ... _ ' ' ,... .....

~
~
\ / ----
e::
:¡:

- .1
. ® ~
~// -· ~ ~
;

,.
.I
.....,..~ ...
/
"
~
G

"'o..
::i

. 1. !or11~
,-•,:.. ·-~
,......---¡--
~'
-·. ~.......
,,...-/.

/ G

/
/· - · -.:t.fjº -
ª
G
<r
3
G
a
(")
:r
~o
::l.
~
;:;·
fVUAJAln l
"'o
....,
~
5.:
;:;·
e
~
/"~'
500km

~~~'~
!5' .z:_ .,. . . ~- ·~º!~
/
•¡'
~-~r.
\' '

\.

•!.
1
---- ,_.I

San,.;,~
~
s
U'.l
o
e
~ v··,_.-:\~ol 4 !,:'· \.'23 ~
G
....,
::i
,__ ~

1 ·;.\3
\,,/Ir,,. ...J / 18º
G
2
§
i\ r·....· 1 I o..
z
o
o 50 100Km

{
-.. ~
-- !!'r_-'.f¡"'
.
. I.
ª"
~
::l
to
2..
71º
l'vt'IFllatiof/
\
.69º :;·
73º ¡;;·
Fig.l. lndex map of Central Andes (southem Peru-northem Bolivia) between 68-73ºW and 15-18ºS. A: numbered stars indicate sample localities (see Table 1). B: morphostructural zonation of
Central Andes. 1: coastal Cordillera and Pacific piedmont; 2: western Cordillera; 3: Altiplano; 4: eastern Cordillera; 5: subandean zone. Bol: Bolivia; Per: Peru. C: location of sample 22, the
\O
Soledad tuff (S), east of Tirata (T) and south-east of Ayo Ayo (AA) in Bolivia after the simplified geological map of Lavenu et al. 1989. \;J
Table 1 Location and age determinations of acidic tuffs from southern Peru and northern Bolivia according to various authors. Letters and numbers quoted referred to samples 'f
dated by authors cited in reference.
-· -··-·-···---·-
Nº lsamp/a ILocatíon (lat.S, lonfl_. WJ Altilllde (m) Unít Mathod Aga (MaJ RafaranctJS
Peruvlon Paclfic Pledmont
-- ---------
186 Alto de la Alianza 17°59' - 70°15' _______J22 Huaylillas Fm. K/Ar 18.4±0.5 Sellan and Lef6vre, 1976
1 Pe3 Alto de la Alianza 17°59' - 70°15'
-----
770 Huaylillas Fm. K/Ar 20.3±0.8 Feraud In S6brler et al., 1988
186 Alto de la Alianza 17º57' - 70º 15' 920 Huayllllu Fm. K/Ar 18.35±0.5 Bellon and Lef6vre, 1976
-··
2 Pe2 Alto de la Alianza 17°57' - 70º15' 920 HuayllllH Fm. K/Ar 19.2±0.8 Feraud In S6brler et al., 1989
- --
3 Pa53 Rio Vitor,Qda del Impertinente 16º 27' - 71 °57' 1200 lower Barroso Gr. K/Ar 2.76±0.1 Bonhomme In Vatln-P6rlgnon et al, 1982
Qda de la Gloria 16º31' • 71 °46' 1800 lower Barro10 Gr. K/Ar 3.05. 3.4 Tauplnard In Laharle and Derruau, 1974
4
'"'"'
Pa32 Qda El Cuico, Arequipa 16º 19' • 71 º42' 2400 lower Barro10 Gr. K/Ar 2.9±0.1 M.G. Bonhomme, Grenoble, unpubllshed data
5 SA1-90 Qda Añas Huayco, Arequipa 16° 20' • 71°35'
- 2520 lower Barroso Gr. FT 2.32±0.21 thls 1tudy
6 SA2-90 Qda Allaa Huayco, Arequlpa 16º20' - 71°35' 2520 lower Barroao Gr. FT 2.43:1:0.24 thls 1tudy
7 SA3-90 Oda Añaa Huayco, Arequlpa 16°20' • 71°35' 2520 lower llatro10 Gr. FT 2.48±0.17 thl1 etudy
Peruvion western Cordillera
8 Pa29 Alto de Tala, Moquegua 17º05' • 70°41' 3$10 Huayllllae Fm. K/Ar 18.6±0.7 Feraud In S6brler et al., 1988
9 Pa17 Pampa de Tlntlnave, laguna Arlcota 17º 24' · 70°09' 2940 Tacaza Fm. K/Ar 15.8±0.6 Feraud In S6brler et al., 1988
10 Pe62 Callall, rlo Llapa, 15°30' • 71°25' 3900 Tacaza Fm. K/Ar 8.9±0.7 Feraud In S6brler et al., 1988
11 Pa44 Qda de Rata, Chuqulbamba 15°45' • 72°43' 3950 lower Barroto Gr. K/Ar 6.0±0.2 Feraud In S6brler et al., 1988
12 Pe23 Quellaveco Mine 17°06' • 70°30' 3900 lower Barro10 Gr. K/Ar 6.1 ±0.2 Feraud In S6bner et al., 1988
13 Pe331 =611 lnflemlllo, Laguna Salinas 16°16' • 71°04' 4300 lower Barroso Gr. K/Ar 4.45±0.3 Bellon and Lef6vre, 1976
14 Pe42 Co Pan de Azucar, lchuña 16º02' • 71 º54' 4600 upper Barro10 Gr. K/Ar 1.8±0.2 Feraud In S6brler et al., 1988
15 PH1-79c Co Hulñao, Cotahuaal 15°13' • 72º52' 2880 upper Barro10 Gr. K/Ar 1.9 ±0.2 Feraud In S6brler et al., 1988
1R111r Co Qulroz, Yanaqulhua 15°46' • 72°53' 2850 upper Barro10 Gr. K/Ar 1.8±0.5 Welbel et al., 1978
16 80-109 Campamento Slc¡ue, Nevado Corupuna 15°30' - 72°45' 4525 upper Barro10 Gr. K/Ar 1.8±0.2 Feraud In S6brler et al., 1988
lgnlmbrtta Acoypampa Grande 15°46' - 72°45' 3950 uppar Barro10 Gr. K/Ar 1.7±0.2 Welbel et al., 1978
17 80-104 Co Chachanll, Navado Solimana 15º31' - 72°55' 3800 upper Barroto Gr. Rb/Sr 1.46± L. Brlqueu, Montpalliar, unpubllahed data
18 Pe18 Yucamane Pampa, Candarav11, rio Callazaa 17°150' • 70°14' 3200 Sencca Fm. K/Ar 0.97±0.04 M.G. Bonhomme, Grenoble, unpublished data
19 Pea Estique Pampa, Pachia 17°32' • 70°02' 2900 Sencca Fm. K/Ar 0.36±0.01 Bonhomm11 In Vatln-P6rlgnon at al, 1982
Bolivian Altiplano
20 PH64 l=IJJ Upper Tlnajanl. Ulloma tuff 17° 29' • 68°30'
----
3800 Totora Fm. K/Ar 9.1 ± Evernden et al., 1966, 11177
21 PH611=111 Upper Tlnajanl , Callapa tuff 17° 27' • 68º 22'
---
3780 Maurl Fm. K/Ar 8.2± Evemden 'at 11 .. 1966, 11177
22 PH/fB Co Pucara, Soledad tuff 17°41' • 67º 19'
-·- -- 3760 Umala Fm. K/Ar 6.2±0.3 Lavenu et 11 .• 1989
SA504 Chillcanl Pampa, Soledad tuff 17°49' • 67°1 B' 3750 Umala Fm. K/Ar 6.8±0.2 Aadwood 1nd Maolntyre, 1989
23 809 1•ICJ Estancia Chlrlquiña, Toba 76 tuff 17°49' • 68° 22' 3940 Umala Fm. K/Ar 5.4± Evemden et 11., 1966, 1977
24 807 E1tancla Tirata, Pastara, Perez lgnimbrite 17°53' - 68º36' 4100 Perez Fm. K/Ar 3.3±0.3 Lav11nu et al •• 11189
25 LA82·2 Ayo Ayo and Villa Remedios, Ayo Ayo tuff 17º07' - 67°59' 3900 Umala Fm. K/Ar 2.8±0.4 Lavenu at al., 1989
La Poz and Titlcaca bosln.
26 LA82-1 Titicaca lak11, Chua tuff 16°1 O' • 68°44' 3900 Umala Fm. K/Ar 7.6±0.7 Lavenu et al .. 11189
27 M8161 Cota Cota tuff 16°32' • 68°03' 3520 La Paz Fm. K/Ar 6.5±0.2 Lavenu et al., 11189
28 PH64 Co Caballo, Chijinl tuff 16°08' - 6B 0 3B' 3970 La Paz Fm. K/Ar 2.8±0.07 Lavenu, 1986
29 MB153 Llmanpata, rio Kaluyo. Patapatanl tuff 16º 2~_:_~ 68º07' ·-.
4160 La Paz Fm. K/Ar 2.7±0.1 Lavanu et al., 1989
30 PH6311 Chuquiaauillo, Calvario Drift, Purapurani tuff 1Gº 27' - 68°06' 4100 Patapatanl Fm. K/Ar 1.6±0.1 Lavenu et al.. 1989
 Trace and rare-earth element characteristics of acidic tuffs from Southern Peru and Northern Bolivia 95

·.·~
- .".
.:

•.. ...
~- ·~.- ·~'
-.
. .,.
•
.• ¡

"·
•• .#; .
• "'
~
~ ;

·"' "
• ..
•
•. . . •·
.• - ....
•
"""........ .'
. ;í •#

.....
1 em ..
•
Ak.:
• •
'* ·:4.
., . • . i
--~ :
@' . . ...
Fig. 2. A: southeast wall of the Quebrada Añas Huayco near the Arequipa airport showing the two flow units of the "sillar", the
hand-cut building blocks and the southern slopes of the volean El Misti ( 16º l 8xS-71 º24xW, 5822 m) behind the quarry. B: detail of
the pink-colored flow unit beneath the light-colored pumice-rich and columnar jointed flow unit of the "sillar" (SA tuft). C: polished
section of a typical hand-cut building stone of the "sillar" of Arequipa.

natural track fading. The apparent fission track ages of consists, by submitting both an irradiated and an unir-
the sillar glasses are therefore meaningless and sorne radiated glass fragment to a convenient thermal proce-
correction procedure has to be applied. We used the dure before track etching, to reduce the fossil and
Storzer and Poupeau ( 1973) plateau technique, which induced track populations to the same degree of
96 N. VATIN-PERIGNON et al.

Table 2 Track diameter analysis in Sillar glasses. N and d: number of measured tracks and mean track diameters, respec-
tively. S and i refer to spontaneous and induced tracks. The reportad uncertainties on d are the standard errors of the
mean.

Sample thermal Ns ds ±la Ni di ±lcr ds/di 1 ±la

treatment µm µm

SA1-90 no 123 3,73 0,05 154 4,69 o.os 0,80 O,Oi

2h.205ºC 99 3,17 0,05 213 3,03 0,03 1,05 0,02
1\)

SA2-90 no 109 3,52 0,06 119 4,39 0,06 0,80 0,02

2h.205ºC 3 3,01 0,06 213 3,10 0,04 0,97 0,02

SA3-90 no 78 3,68 0,07 118 4,52 0,05 0,81 0,02

2h.205ºC 130 3,13 0,05 214 3,03 0,03 1,03 0,02

Un:urnoalcd snmplcs

..
•' ~o

~ SA1 SA2 SA3

f., 20
.l!
ñi
~

5 10
!rack diamelcr. pm

Sampl<!S annealed 2h at 205'C

r·;
1 1
r-1
1 1

"
1
1

1
1
1
L.
{\ 1 1 1 1

Fig. 3. Fossil (full lines) and induced (stippled lines) fission track diameter distribution in the 3 obsidian glass clasts dated, SAi, SA2
and SA3. Top: without any Jaboratory thermal treatment, fossil tracks appear as shortened relative to induced tracks due to natural
fading. Bottom: samples were heated for 2 hours at 205°C before track etching. Fossil and induced tracks were reduced in size,
although differentially and they now present similar track diameter distributions.

annealing, hence the same size distribution (Poupeau et are only lower estimates of their formation ages. After
al., 1994, Dorighel et al., 1994). It can be seen from a thermal treatment at 205ºC, fossil and induced track
Table 2 and Fig. 2 that a thermal treatment of 2 hours have reached similar size distribution. Electron micro-
at 205ºC was sufficient to reduce in the SA obsidian probe analyses revealed the glass in the SAtuff is
fragments the fossil and induced tracks to the same chemically homogeneous. The three samples analysed,
mean diameter and diameters distribution. SA 1-90, SA2-90 and SA3-90 present concordant "pla-
teau" ages, at respectively 2.32±0.21 Ma, 2.43±0.24 Ma
The apparent ages of the samples (Table 3) range
and 2.48±0.17 Ma ( 1s). We therefore adopt for the sil-
from l.19±0.10 Ma (SA2-90) to l.42±0.11 Ma (SA3-
lar of Arequipa the weighted mean age of 2.42±0. 11 Ma
90). The track size analysis having shown that fossil
(1 s).
tracks were affected by sorne natural fading, these ages
 Trace and rare-earth element characteristics of acidic tuffs from Southern Peru and Northern Bolivia 97

Table 3 FT dating of Sillar glasses. Analytical data.

Sample Mic. Thermal Ns ns Os ±la N(s+i) n(s+i) Di ±lo t app. ±la t pi ±lo
treatment 10E3t/cm2 1OE3 t/crri2 Ma. Ma

SA1·90 1 no 198 990 1,51 0,11 2080 170 91,0 2,0 1,36 0,10
2 2h. 205ºC 139 920 1,42 0,12 1094 200 50,2 i,5 2,32 0,21

SA2·90 1 no 151 800 1,43 0,12 2372 180 98,2 2,0 1,19 0,10
2 2h.205ºC 123 855 1,36 0,12 1305 260 45,9 1,3 2,43 0,24

SA3-90 1 no 177 880 1,52 0,11 2126 180 87,8 1.9 1,42 0,11
2 2h.205ºC 157 1040 1,42 0,11 2415 470 47,0 1,0 - - 2,46 Qg1
-
1,32 . 0,06 2,40 0,13

míe: rrücroscopes. 1: Orthoplan Lei~ 2: Jenavert Zeiss Jena

N, n and D, respectively total number of tracks counted, of counts, and track densities. The s and i refer to
respectively spontaneous and induced tracks.

tapp and tpl, appai:ent and plateau ages. FT ages were calculated using the equation t=[D 5f(D(s+i)·D 5)](cr!<l>/A),
where D(s+i) is the track density measured in the irradiated fragment. ·cr=sso.2x10-24 cm-2; 1=7.253xl0-3;
A.=7.03xIO·l7 years-1.

The precision on a FT age depends primarily on the precision on the density ratio of the above equation. It has
been derived as (Biga.ui 1986):
ot/t = [1/N5 + IIN(s+i)Jll2 X [D(s+iy'Di]

Weighed ages and uncertainties in the last line ofthe table were computed as:
cr2 = 1/L[llcrtj]2 and t =l:{t/(crtj)2]/L(l/cr1j2]

Brief Description of Other Tuffs in the rio Llapa valley. The most complete section is com-
posed of three units separated by layers of river gravels.
Southern Peru. In the Arequipa area, two rhyolitic
The Iowest unit is a 20 m thick tuff with abundant lithics
pumice ftow deposits (Pe32 and Pe53) are sandwiched
and basement clasts in a coarse-grained matrix containing
between Pliocene conglomeratic layers of the Pacific
pumices and crystals of quartz, plagioclase and biotite.
Piedmont. These tuffs developed columnar joints and
The other two dacitic pyroclastic flow units consist of a
filled river valleys. Pe53 has given a feldspar age of
well-indurated pumiceous tuff with well-developed
2.76±0.l Ma (Vatin-Pérignon et al., 1982) which disagree
columnar jointing (50 m thick) underlying a more vesicu-
with ages of 3.05 reported by Laharie and Derruau
lated upper ftow unit (30 m thick) containing abundant
(1974).
dacitic fragments. These three units are separated by thin
The Huaylillas Formation occurs along the Pacific layers of fluvial gravels. Tuffs of the Plio-Pleistocene
Piedmont and in the western Cordillera and consists group (Barroso Group) are well represented by dacitic
mainly of tuffs that have yielded lower Miocene ages. tuffs near Chuquibamba (Pe44). At Quellaveco Mine, rhy-
Sequences of rhyolitic flows and falls erupted between olitic tuffs (Pe23) consist of two flow units and an over-
18.6 and 18.3 Ma have been sampled in various localities: lying rhyolite lava ftow. The lower unit corresponds to a
Alto de la Alianza (Pe3 and Pe2) and Alto de Tala, NE of ground surge deposit and the upper unit is a crystal-rich
Moquegua (Pe29). The oldest Miocene dacitic tuffs occur rhyolitic tuff with plagioclase, bipyramidal quartz and
in the area of the laguna Aricota (Pampa de Tintinave, biotite in a pumiceous matrix. In the Laguna Salinas
Pe 17) and contain large quartz, and plagioclase, biotite, region, tuffs (Pe33) are very similar to those of the
amphibole and Fe-Ti oxides in an abundant grey fine Moquegua region but the upper rhyolite lava ftow is bet-
matrix of glass and feldspar. Near Callalli, the 8.9 Ma-old ter developed. The young tuff of the Pan de Azucar
Sibayo tuffs (Pe52) of the Tacaza Group are well exposed (Pe42) consists of two pyroclastic flow units separated by
98 N. VATIN-PERIGNON et al.

thin (10-20 cm) laminated pyroclastic surge deposits. The rich leve! with a fines-depleted matrix. Well-vesiculated
lower member is a single, thick (over 30 m), nonwelded, crystal-rich rhyodacitic pumices contain dominant quartz,
homogeneous unit and the upper member, with a thick- biotite and plagioclase with small amounts of arnphibole
ness of about 10 m, is a indurated unit with numerous and Fe-Ti oxides. The Toba 76 tuff and the Perez ign-
vitrophyric fiarnme. The large tuff deposits of the upper imbrite are separated by about 10-30 m of clastic sedi-
Barroso Group are an extensive and stratigraphically ments (Pliocene Umala Formation) at the Estancia Tirata.
important horizon of at least 500 m thick in the Cotahuasi The Toba 76 tuff covers large areas of the western Cor-
graben (PH1-79c). The formation consists of a lower dillera, the central Altiplano and possibly the La Paz
welded flow unit which varies in thickness from 55 to 75 basin (Martinez, 1980). The Perez ignimbrite (B07, Fig.
m. Basal vitrophyre and fiamme are comrnon with small 4 C) is more diverse vertically than the Toba 76 tuff. The
lithic fragments and rare pumices. The matrix is fine- depositional sequence is characterized by a typical plinian
grained and contains plagioclase, amphibole and ubiqui- thinly bedded ash-fall deposit (3m thick) overlain by a
tous oxides. The upper unit is typically a pink welded tuff thick ( ~ 1 m) surge !ayer with sandwave bed forms and
with small white pumices and abundant lithics. The two two recognizable flow units. The 20m-thick lower flow
units are columnar jointed . The uppermost member of deposit consists of a nonwelded fine-grained basal part, a
this series is visible near Allahuay and consists of a pumi- poorly sorted pumices and lithics main body and a
ceous nonwelded flow unit underlying Plinian airfalls. smaller upper part reversely graded with a coarse pumice-
The 0.97 Ma-old tuff (Pe 18) is found filling in a river rich zone. A black-brown vitrophyre of 2 m thick is
valley near the Yucamane volcano. This nonwelded single present at the base of the 20-30m-thick upper flow unit.
flow unit of about 30-50 m thick contains abundant pum- This densely welded, typically light-gray-colored, pumice
ice blocks. The youngest Pleistocene tuff (Pe8) is an flow deposit exhibits columnar joints and coarse pumice
homogeneous, grey andesite of limited extent with rela- concentrations are common, particularly at the top. The
tively dense pumices and well-developed columnar joints. pumiceous matrix contains coarse pumices, lithics and
quartz, biotite, plagioclase, amphibole, Fe-Ti oxides and
Northern Bolivia. The Altiplano pyroclastic forma- rare small crystals of clinopyroxene. The Ayo Ayo tuff
tions correspond to tuffs of the Totora and Umala Forma- (LA82-2) is a single, non-welded, homogeneous, col-
tions (Laven u, 1986). The Ulloma tuff (PH64) is an oured, crystal-poor ( ~ 20%) fiow deposit. The matrix is
homogeneous, pumice-rich, with pumices of ali sizcs dis- fine-grained with lithics and small dacitic pumices con-
persed in a fine-grained matrix containing dominant taining plagioclase, quartz, biotite, amphibole and Fe-Ti
quartz and plagioclase with small amounts of biotite and oxides.
amphibole. The Callapa tuff (PH61) is a pumice pyroclas-
A series of dacitic and rhyolitic tuffs is interbedded at
tic !ayer, reworked and/or weathered, with a crystal-poor
different levels with the lacustrine or glacial formations of
matrix. The Soledad tuff (PH48), located northwest of
the La Paz and Titicaca basin (Lavenu, 1986, Lavenu et
Oruro, has been described in detail by Redwood (1987)
al., 1989). The Chua tuff (La82-1 ), near lake Titicaca, is
and interpreted as a sequence of interbedded Plinian air-
a coarse-grained, crystal-rich, densely welded, dacitic tuff
fall deposits and ash-flow tuffs accompanying a large cal-
showing numerous oxidized Jenses. Crystals are domi-
dera subsidence and overlain by dacitic lavas. These tuffs
nantly quartz, biotite, plagioclase and small amounts of
provided K/Ar ages of 5.8±0.2 on biotite in dacitic clasts
amphibole and Fe-Ti oxides. The Cota Cota tuff (MB 161)
(Redwood and Macintyre, 1989) and 5.2±0.3 and 5.0±0.7
from the lower part of the La Paz Formation, is a pale
on two fractions of plagioclase from the crystal-rich ashy
green-colored, fine-grained, welded deposit with rare
matrix (Bonhomme in Lavenu et al., 1989). Unconsoli-
mm-sized crystal fragments: plagioclase, biotite and
dated ashy pyroclastic deposits contain abundant lithics
amphibole in a devitrified glassy matrix. The Chijini tuff
and pumice blocks in a matrix which contains ~ 35%
(PH54) from the upper part of the La Paz Formation,
crystals of quartz, plagioclase and sanidine, biotite and
consists of a soft white rhyolitic deposit with sparse
small amounts of amphibole and magnetite. The stratigra-
quartz and biotite in a very abundant ashy matrix. The
phy of the Toba 76 tuff (that locally total 80 m thick) and
Patapatani tuff (MB 153) is also a white rhyolitic deposit
the Perez ignimbrite (up to 50 m thick) are best observed
which resembles the Chijini tuff and a correlation has
along the rio Tirata (Fig. 4 A, Lavenu et al., 1989). At the
been established (Lavenu et al., 1989). The Purapurani
Estancia Chiriquiña where it is 15 to 30 m thick, the Toba
tuff (PH53a) is a rhyolitic salmon-pink fall deposit with
76 tuff (B09, Fig. 4 B) lies directly above Miocene clays
quartz, biotite, feldspar and glassy fragments.
and sandstones (Crucero Formation) and is made up a
complete eruptive episode. This single cooling unit con-
sists of at least 3 distinct layers. The !ayer 1 is a thinly GEOCHEMISTRY
bedded Plinian ash and pumice fall deposit up to 2m
Analytical Methods
thick. The !ayer 2 is a poorly sorted ash-fall deposit (0.20
to 2 m thick). The !ayer 3 is a salmon pink ash flow The material analysed in this study consist of 23 frag-
deposit with a biotite-rich base and minor amounts of ments of vitrophyres, obsidians, pumices and glassy frac-
lithics mixed with small pumices in a fine ashy matrix. tions of tuffs and fallout deposits from the Pacific
The top of the flow unit is eroded in many places and piedmont and the western Cordillera in southern Peru, the
consists of a reversely graded nonwelded coarse pumice- Altiplano and the La Paz and Titicaca basin in northern
 Trace and rare-earth element characteristics of acidic tuffs from Southern Peru and Northern Bolivia 99

Pumice llow deposit

columnar joints

Perez lgnimbrite

60 7 basal vitrophyre

Umala Fm. .·. J. . ash llaw deposit

:.,·¡: ..... ": minar amounl of lithics

mixed 'w'Vilh small pumices
in a fine ashy malrix
coarse pum1ce concentralion

:··\ ........
.....-1 ¿..
;
QJ
e
nonwcltlcd pumiccous
8 •.~· •·• • .. la'· ••• llowdeposit
.Q . : •=- ~ .•.
a: vvvvv
vvvvv Toba 76 tuff ·~~~·::.~·:¡ i ·. .
....
. . . .·
•.t:- •
• d: .. ..., -"\.,.
vvvvv unsorled Jine-ash fall deposit ~~~f.~~-
\/V V V\/
vvvvv
vvvvv
vvvvv
I 1om ~ bedded while ash
and purnice fall deposit
.-;_
~:~-Y~~~_ ~ .
';•' t >. ·._·:'".:;
QJ
e Miocene clays
Crucero fm surge layer
B
o
and sandstones

~ ash fall deposil

Toba 76111ff
Plioccne claslic sedirnents
Perez iqnimbrile
rio Tirata
Estancia Chiriquiña
Pastara

A B e
Fig. 4. Schernatic stratigraphic colornns of the Toba 76 tuff and the Perez ignirnbrite for two locations in the Tirata area. A: section
atrio Tirata frorn Lavenu et al. (1989). B: section of the Toba 76 tuff at Estancia Chiriquiña. C: section of the Perez ignirnbrite at
Pastara.

Bolivia. In addition, 7 hydrated glasses from tuffs of the Blaise Pascal of Clermont-Ferrand and on other samples
Bolivian Altiplano and the La Paz and Titicaca basin have by energy dispersive X-ray fluorescence spectrometric
also been analysed because their stratigraphic position is method (EDXRF) al the University Joseph Fourier of
well-known (Table 1). Ulloma and Callapa tuffs are inter- Grenoble.
calated at different levels in the late Miocene sequences
and are the only tuffs to be dated (Evernden et al., 1977). Geochemical Characteristics
Recent K/Ar dates have been obtained on the other tuffs
(Ayo Ayo tuff and tuffs intercalated in La Paz Formation, Tuffs and associated fall deposits range in composition
Lavenu, 1986, Lavenu et al., 1989). Nevertheless devitri- from high-silica rhyolite to andesite (78.8-57.6% Si02 )
fication is a serious problem, therefore results on these and are subalkaline (Na20 + K2 0 between 5.3-9.1, Table
tuffs are given for comparison. Samples were analysed for 4).
Zr, Nb, Y, Ba, Sr and Rb at the University Claude Ber- Representative minor and trace-element MORB-
nard of Lyon using X ray fluorescence (XRF) method normalized patterns for tuffs from the four paleogeo-
(Germanique and Briand, 1985). Rare earth elements graphic domains show differences which are illustrated by
(REE) plus Ta, Hf, Th, U, Se and Cs contents were meas- the multi-element diagram in Fig. 5. Plots of individual
ured by instrumental neutron activation analysis (INAA) tuffs exhibit typical features for continental-margin vol-
at the Institut Laüe-Langevin of Grenoble following the canic ares with marked enrichment mainly in the large ion
procedures of Oliver et al., (1990). Geochemical data are lithophile (LIL) elements such as K, Rb, Ba and Th rela-
presented in Table 4. Table 4 also includes Si02 and alka- tive to the incompatible high-field strength (HFS) ele-
lies which have been recalculated to 100 wt% major ele- ments, notably Nb and Ta, and to ali transitional
ment data on volatile-free basis. Seven samples with loss elements. The LIL enrichment and HFS-depleted nature
of ignition (LOI) > 3 wt% have not been excluded but are of Nb (and Ta geochemically similar to Nb) relative to Ce
noted "hydrated glass" (HD) in Table 4. Other samples in subduction zones are well-known and generally attrib-
have between 0.5 and 3 wt% LOI. Major element compo- uted to a subduction component (McCulloch and Gamble,
sitions were determined on 6 glasses (Sal ,2,3-90, PH1- 1991; Smith and Leeman, 1993). Ce shows a same small
79c, 80-109, 80-104) by electron microprobe at the positive anomaly for ali samples. Nb and Ta show a dis-
Centre Régional de Mesures Physiques of the University tinct depletion which may be an indication that the source
 Table 4 Trace-element compositions of obsidian fragments, pumices and glassy matrices of tuffs from southern Peru and northern Bolivia. Si0 2 and total alkali values correspond
to major-element analyses recalculated to 100% on volatile-free basis.
---- -·------- -···-
-8
Peruvian Pacific piedmont - - Peruvian western Cordillera
Nº 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
--
Samp/e--:~ P-Pe3 P-Pe2 P-Pe53 P-Pe32 OF-SA1-90 OF-SA2-90 OF-SA3-90 G-Pe29 G-Pe17 P-Pe52 O-Pe44 P-Pe23 P-Pe33 BV-Pe42 BV-PH1-79
-
---
Trace eloment abundances
Ba 720 616 434 928 993 952 586 828 497 936 496 806 520
e---
Rb 116 95 211 105 386 153 123 133 54 143 96 194 243
- ·--- --
Sr 233 179 79 216 173 65 338 334 768 225 235 79 45
y 14 15 17 14 19 18 18
---·- ,_ --
19 16 16 16 15
Zr 94 73 122 134 84 122 123 147 119 156 107 112 119
Nb 10 11 13 11 14 10 10 8 9 12 11 12
Se 4.62 3.48 1.84 2.97 2.06 2.24 2.12 11.72 6.54 1.26 3.03 4.32 5.26 17.30 3.10
Hf 3.07 3.08 4.48 4.10 3.67 4.30 3.90 4.18 5.12 2.61 3.81 4.69 4.25 6.37 8.60
Ta 1.23 1.22 1.12 0.85 1.29 1.32 f :10 0.66 0.73 0.81 0.84 1.25 1.09 1.36
Th 13.53 13.80 43.13 8.65 14.35 15.41 14.70 5.85 15.47 7.16 18.77 17.7f 9.28 35.00 6.10 ;z:
u
~
2.2 2.0 14.8 1.4 3.8 4.1 3.8 1.3 4.6 2.7 1.9 2.9 2.8 7.0
Zr/Hf 30.7 23.6 27.1 32.7 22.9 31.3 29.4 28.8 45.4 41.0 22.8 26.3 18.6 tJ
Th!Ta 11.0 11.3 38.5 10.2 11.1 11.7 13.4 8.9 21.2 8.8 22.3 14.2 8,5 25.7 z
Sr/Y 16.5 12.2 4.5 15.5 9.1 3.6 17.5 20.6 48.9 12.4 14.6 5.2 ot
-
~
REE co11centratlons
La 34.68 28.29 45.53 39.76 29.61 35.43 33.60 33.00 31.82 21.77 41.99 37.85 33.04 40.10 78.10 ~
Ce 63.27 55.29 72.40 83.44 68.93 75.71 68.93 65.84 57.35 51.13 74.36 74.50 64.99 79.27 161.80 ~
Nd
Sm
Eu
25.44
3.12
0.72
21.65
2.83
0.58
25.66
3.90
0.40
32.77
4.08
0.78
26.05
3.85
0.67
26.53
4.30
0.74
25.35
4~ 13
0.70
31.70
5.36
1.54
21.97
3.83
0.93
17.80
4.20
0.32
31.93
4.82
0.68
26.48
4.60
0.70
25.79
3.60
0.85
40.06
4.48
1.92
56.10
9.89
0.76
-
~

:::,
:-

Gd 3.72 3.26 4.11 3.90 5.11 5.75 4.84 4.83 3.66 4.05 5.08 5.06 3.78 4.23 8.31
·-
Tb 0.70 0.60 0.72 0.64 1.01 1.14 0.90 0.76 0.60 0.67 0.89 0.91 0.66 0.69 1.23
Yb 1.86 1.48 1.47 1.61 1.31 1.46 1.45 1.20 1.47 1.45 1.96 2.04 2.23 1.67 2.80
Lu 0.32 0.22 0.19 0.21 0.20 0.20 0.19 0.19 0.21 0.23 0.32 0.30 0.30 0.26 0.38
SREE 133.83 114.20 154.38 167.19 136. 74 "151.26 ~110.09 144.42 121.84 101.62 162.03 152.44 135.24 172.68 319.37
La/Yb(N) 12.47 12.78 20.71 16.51 15.11 16.23 15.50 18.39 14.47 10.04 14.33 12.41 9.91 16.06 18.65
La/Sm(N) 6.86 6.17 7.20 6.01 4.75 5.oa 5.02 3.80 5.13 3.20 5.38 5.08 5.66 5.52 4.87
Ce/Yb(N) 8.65 9.50 12.53 13.18 13.38 13.19 12.09 13.95 9.92 8.97 9.65 9.29 7.41 12.07 14.70

mafor elements reca/cu/at•d to 100 wt% on a vo/at//e-free bas/s

Sí02 72.7 72.8 74.7 73.6 75.5 76.4 76.5 73.5 67.2 65.6 61.3 70.7 72.9 75.9 74.7
Na20 4.19 3.67 3.37 4.57 4.57 3.64 3.85 3.88 4.10 2.75 4.15 4.35 3.75 4.00 2.83
K20 4.25 4.99 5.10 3.97 4.51 4.62 4.54 4.98 3.58 3.03 2.39 3.93 3.70 4.92 5.56
Na20+K20 8.44 8.66 8.47 8.54 9.08 8.26 8.39 8.86 7.68 5.78 6.54 8.18 7.45 8.92 8.39
K/Rb 303 436 200 313 97 269 242 189 367 227 321 210 190
- -..... · - ------·---
~<_!_numbr.r Qrclixcs
---- --------·- --~--------- ---------··-
are as follows: BV, b¡¡snl vitroQhyrc, F, fiammc, HG, hydrat!!.q__glass G, alass, OF, obsidian fraament, P, oumice. Samole locations and a¡¡es are indicated in Table 1.
Table 4 Continuad
-•L ••••• • ••••• '• • --- -~

Peruvian western Cordillera (cntd) Bolivian Altiplano La Paz 11nd Tltloaoa basin
Nº 16 17 18 19 20 21 22 -· 23 24 25 26 27 28 29 30
---·
Sample F-80-109 F-80-104 P-Pe18 F-Pe8 HG·PH64 HG-PH61 G-PH48 P-809 8V-807 HG-LA82-2 G·LA82-1 HG-M8161 HG-PHM HG-MB1S3 HG-PHS3a
-------- =;3
~
('!¡
Trace element •bundanc•
Ba 1038 838 918
§
1 760 608 540 1101 863 721 719 1035 467 388 639 Q..

Rb 11 104 98 76 114 244 158 124 103 103 136 118 127 126 102 ....
Sr 2se¡ 65 565 626 225 246 487 516 338 312 684 295 209 371 427
~1

y
Zr 128
15 17
164
11
107
20
183
18
199
14
116
10
152
13
155
23
190
7
119
12
129
13
127
13
134
~
('!¡

Nb
Se 1. 0 0 4.10
9
8.54
8
15.27
13
3.60
13
5.J7
16
7.21
1i
3.78
13
2.65
13
2.34
21
10.BO ae
11 14
1.65
13
1.93
12
2.41
[
Hf 2.80 9.83 4.51 5.00 3.67 6.21 6.12 4.13 4.53 4.77 6.13 3.53 3.96 4.17 4.99
a
a
Ta 1.15 0.86 0.68 1.35 2.58 1.17 1.22 1.40 1.04 2.27 0.68 1.29 1.25 0.91 ::r
Th 1.30 13.58 13.68 9.92 16.89 18.97 16.28 --15.Hi - .. -·--1"2.40 --· -- 14.57
·-·4.6 ----3:5 ---------
13.42 15.59 14.23 11.66 12.85 ~
u
Zr!Hf
2.6 3.4
28.4
2.6
32.7
4A
29.1
10.6
29.4
3.5
32.5 28.1 33.6
-
3.8
32.4
3.0
30.9
4.4
33.8
4.2
32.6
3.6
30.4
2.8
26.8
a.
"'a
::::-.
Th!Ta 11.8 15.9 14.6 12.5 7.4 13.9 12.5 8.9 14.0 5.9 22.9 11.0 9.3 14.1
Sr/Y 38.1 37.9 20.3 12.2 27.2 37.4 34.2 24.0 30.3 44.7 17.2 29.2 32.6 "'o
.....
f5
REE concentratlor.s s.:
()•
La 25.20 42.93 33.30 29.77 39.67 63.88 59.93 37.66 54.91 45.90 54.36 27.98 47.69 46.57 55.18
Ce 49.70 96.18 65.71 66.24 67.42 138.19 114.60 59.39 92.13 81.19 108.52 41.46 90.41 90.07 105.07 2
Nd 16.4') 45.56 25.82 27.16 27.33 30.73 ~
29.58 20.64 57.41 46.31 20.02 28.41 26.74 40.41 11.39
Sm 2.80 6.86 4.63 5.05 3.39 10.18 7.63 3.50 4.18 4.01 7.36 1.99 4.08 3.94 4.31 8'
Eu 0.53 1.58 1.19 1.42 0.81 1.44 1.66 0.94 1.02 0.91 1.81 0.42 0.71 0.98 1.20 s
Cl:l
Gd 2.04 6.78 4.72 4.99 3.87 8.64 5.42 2.25 2.66 2.48 7.05 1.72 4.04 3.91 4.24 o
i::
Tb 0.26 1.14 0.81 0.84 0.71 1.29 0.67 0.24 0.28 0.28 1.16 0.26 0.68 0.66 0.71
Yb 1.00 2.71 1.49 1.60 1.22 1.73 1.53 0.90 1.03 1.16 2.14 0.61 0.98 1.08 1.00 ~
3
Lu 0.14 0.36 0.21 0.25 0.22 0.26 0.22 0.14 0.15 0.17 0.31 0.08 0.14 0.14 0.15
SREE 98.07 204.1 o 137.88 139.74 137.95 283.02 237.97 125.04 184.77 ·; 62.84 223.12 85.91 175.89 174.68 202.59 a'
2
La!Yb(N) 16.85 10.59 14.94 12.44 21.-4 24.69 26.19 27.98 35.65 26.46 16.99 30.67 32.5'4 28.83 36.90
La/Sm(N) 5.55 3.86 4.44 3.64 7.22 3.87 4.85 6.64 8.11
-
•..
7.06 4.56 8.68 7.21 7.29 7.90 "'o..
::s
Ce/Yb(N) 12.64 9.03 11.22 10.53 14.0o 20.32 19.05 16.78 22.75 17.80 12.90 17.29 23."46 21.21 26.72
~
majar elements recalcul•t•d to 100 wt% on a volat/le-free basls ..,~
Si02 66.3 74.5 75.0 ::s
78.8 73.5 61.6 57.6 73.3 70.7 64.9 70.4 74.3 72.8 66.1 74.2
3.36 lli
Na20 1.66 2.77 3.89 3.76 5.08 4.63 2.70 3.69 3.77 2.63 3.56 3.73 3.32 3.33 2..
- 5.59 :;:·
K20 3.67 5.59 2.43 2.24 2.31 2.82 3.96 4.12 4.54 5.76 3.47 4.00 5.25 5.42
8.57 8.75 8.95 ¡;;·
Na20+K2 5.33 8.36 6.32 6.00 7.39 7.45 6.66 7.81 8.31 8.39 7.03 7.73
K/Rb 170 446 206 244 168 96 209 275 3q7 463 211 282 343 356 453

samole number orefixes are as follows: BV, basal vitroohvre, F fiamme, HG, hvdrated alass G. alass, OF, obsidian fraamcnt, P oumice. Samole locatiQn$ tind aaes are indicated in Table 1. -
o
102 N. VATIN-PERIGNON et al.

• Pe3 western Cordillera • PeZ9

B
o Pe2 A o Pe17
• Pe52
• Pe53
100 v Pe32 100 v Pe44
• SAt-90 • Pe23
• • • • -Cenozoic l¡¡nimbriles o Pe33
• Pe42
" PelB
10 10 • Pe8

"':::;:e::o
'""uo
c::

0.1 0.1

~
o
~
~ o
o ~ ~ o ... ..;"' ~
~ o ...
~ o N
"' "'o o "'
..; ~ ci
"' o.i e? e?
O.O!
Sr K Rb Ba Th Ta Nb Ce p Zr Hr Sm Ti y Yb Sr K Rb Ba Th Ta Nb Ce P Zr Hf Sm Ti Y Yb

Altiplano •
o PH61
PH64
e La Paz
and Titicaca basin
• La82-1
o PH54 D
• PH46 • MB153
100 V B09 100 v PH53a
• 807
o La62-2

10 10

.,c::
o "'oc::
~u ~u
o o
e:: c::

0.1 0.1

Sr K Rb Ba Th Ta N1' Ce P Zr Hf S.. Ti Y Yb Sr K Rb Ba Th Ta Nb Ce P Zr Hf Sm Ti Y Yb

....... Pacific piedmonl

E
.'''•
'
·---
-
western Cordillera
Altiplano
- La Paz al'Od
Titicillca basin

Sr K RbBli ThTa Nb Ce P Zr HIS... TI· Y Yb

Fig. 5. MORB-normalized multi-element variation diagram for selected tuffs from A: the Pacific piedmont, B: the western Cordillera,
C: the Altipolano and D: the La Paz and Titicaca basin. E: comparison of trace element distributions for samples respectively the
most evolved and least evolved compositions of each group. Normalizing factors shown on X axis of A taken from Pearce (1983).
Cenozoic ignimbrites from Schmitt-Riegraf and Pichler (1988).
 Trace and rare-earth element characteristics of acidic tuffs from Southern Peru and Northern Bolivia 103

of at least sorne of the magma is of mantle origin. The

trace element patterns of these tuffs are very similar to Hf ppm
those of Cenozoic ignimbrites plotted in Fig. 5A
(Schmitt-Reigraf and Pichler, 1988). Compositional dif- Zr/Hf=25

ferences between the obsidian glass (SAl-90) and the

pumices (Pe32) of the SA tuff are small depletions of Rb o
and Ba in the pumice with respect to the glass, other ele- "'v"f'
6
ments as Hf and Ta having broadly similar concentra- •
tions. The relative Sr enrichment for andesitic and dacitic
tuffs from the western Cordillera (Fig. 5B) indicate plagi- 'V
o+•o Zr/Hf=50
oclase control and is coherent with subduction-related o ....
magmas (Luhr, 1992). The anomaly of Ba may be an
indication of refractive mineral phases, possibly
phlogopite and amphibole, in the source or the fractiona-
tion of the plagioclase. Decreases of Zr from dacites to
4
• ....
eQ +
,,07.
'V
"V
T

o
/
rhyolites may be interpreted as the fractionation of biotite
(Davidson et al., 1990). Positive anomalies for Rb and Th
• •
relative to Ba and negative anomalies for Sr, P and Ti
o
increase from dacitic tuffs to high-silica rhyolitic tuffs • Pacific piedmont
2 O western Cordillera
(Fig. 5B and C). Sorne obsidian glasses and homogeneous
.., Altiplano
vitrophyres (SAl-90, Pe42) are distinguished by their v La Paz and
higher abundances of Rb and Th relative to pumices and Titicaca basin
glass matrices. Strong negative anomalies of Sr, P and Ti • northern Chile
are very typical for are magmas (Pearce, 1983) and reflect
the importance of plagioclase, apatite, ilmenite or
Zr ppm
Ti-magnetite in the fractionating assemblage of these
100 200
tuffs, independently of their age and their position from
the trench. Trace element compositions of hydrothermally Fig. 6. Hf-Zr correlation diagram for selected tuffs from south-
altered tuff deposits from the Altiplano and the La Paz em Peru and northern Bolivia. lgnimbrites of northem Chile
taken from de Silva and Francis (1989) and Davidson et al.
and Titicaca basin (Fig. 5D and E) approach those found
(1990).
in nonaltered tuffs from the western Cordillera and the
Pacific piedmont and in sorne cases are indistinguishable
except for the Cota Cota tuff (MB 161, not shown in Fig.
5D). We conclude that in these cases, the compositions of
Ta ppr.'1
tuffs in minor and trace elements are relatively little Th/Ta:\0 +
• Pacific piedmont

•..
affected by post-magmatic hydrothermal alteration. O western Cordillera
,. Altiplano
•
Among ali incompatible elements, Zr and Hf are espe- ~

.
'V La Paz and
cially interesting elements with respect to their excellent Titicaca· basin o ,. .... º•
concentration in late forming mineral as ilmenite. Figure • norlhern Chile O ....• Th/Ta•20

6 is a plot of Zr vs Hf for ali the tuffs and highlight sorne 'V o

o o
differences in Zr and Hf contents. The abundances of Zr
o o
and Hf are generally greater in tuffs of the Altiplano than
those of the western Cordillera at comparable Si02 con-
tents. Zr/Hf values range from 18 to 45 with the majority
of values varying between 25-35, as commonly observed Th ppm
in rhyolites (Hildreth and Moorbath, 1988). The mean
Zr/Hf ratio is 27 for the rhyolitic tuffs of the Pacific pied- 10 15 20 25

mont and the western Cordillera and 30 for the rhyolitic Fig. 7. Th-Ta diagram for selected tuffs from southern Peru and
tuffs of the Altiplano and La Paz and Titicaca basin. This northern Bolivia. lgnimbrites of northern Chile as in Fig. 6.
last value is close to the value of 30.4 for the northern
Chilean ignimbrites (de Silva and Francis, 1989; David-
son et al., 1990). Howevcr, Hf contents are more scat-
which have anomalously high Th or Ta contents and plot
tered in tuffs from the western Cordillera and the Zr/Hf
just outside this field. For the other tuffs, Th and Ta con-
ratios vary from 18 for a high-silica vitrophyre (Pe42) to
tents are compared with those from northern Chile. Th/Ta
41 and 45 for two dacitic tuffs (Pe44 and Pe52): values
ratios are high and variable and range from 7 to 15 in
which are also found in other pyroclastics in thc CVZ (de
rhyolites with a mean value of 11 for all groups. This
Silva et al., 1993 ).
value is close to 1O which is a typical value for continen-
Figure 7 íllustrates the range of Ta and Th contents of tal are magmas. Other samples plot in the same area as
the majority of tuffs, with the exception of four samples ignimbrites from the CVZ (Davidson et al., 1990) and
104 N. VATIN-PERIGNON et al.

close to the Th/Ta value of 20, reflecting a strong crustal Paz and Titicaca basin). Abundances of the LREEs are
influence. well correlated with those of other incompatible elements
such as Ba, Zr and P (e.g. for the tuffs from the Bolivian
Altiplano: PH48 and La82-1 or from the western Cordill-
Rare Earth Elements
era: Pe44). The highest total REE contents (>200 ppm)
Chondrite-norrnalized REE distribution patterns for ali correspond to the higher Y values (>20 ppm, e.g. the
tuffs (Table 4, Fig. 8) show moderate to high REE con- Chua tuff LA82-l). As reported by Cameron and Hanson
tents (98-319) with the exception of the Cota Cota tuff (1982) and Wark (1991), Y behaves similar to a HREE
(MB161) which have unusually low REE contents (-86) and its variations are consistent with the decreasing
and fractionated patterns. Relations are similar to those of importance of hornblende fractionation.
high-K calc-alkaline ash-flow tuffs in convergent margins
with large light REE concentrations and light/heavy REE Rhyolitic tuffs of the Pacific piedmont (Fig. 8A) show
(LREE/HREE) ratios. These tuffs are enriched in LREE moderate LREE enrichments, depleted HREEs and pro-
(60-180 times chondritic abundances) with La/YbN vary- nounced negative Eu anomalies. Their distribution pat-
ing from 10 to 36 (between 10-20 for tuffs from the terns are parallel but slightly depleted to other CVZ
Pacific piedmont and the western Cordillera and between ignimbrites (Schmitt-Riegraf and Pichler, 1988). The
15-36 for those from the Bolivian Altiplano and the La REEs of obsidian glasses and pumices of the SA tuff

Pacific piedmont
• Pe3
A western Cordillera • Pe29 E
o Pe2 o Pe17
• Pe53 .,. Pe52
v Pe32 v Pe44
• -.... • SA3-90 • Pe23

~-<--~
100 100 a Pe33
.. Pe42
"' PH1-79c
• Pe!B

·._...

Dacite Al

La C" Nd Sm Eu Gd Tb Yb Lu La Ce Nd Sm Eu Gd Tb Yb Lu

Altiplano e La Paz
and Titicaca basin • LAB2-1
D
• PH64 o PH54
o PH61 .,. MB153
• PH4B v PH53a
., B09
100 • B07 100
a LAB2-2

Cenozolc ''lgnimbite" Fm

La Ce Nd Yb Lu La Ce Nd Sm Eu Gd Tu Yb Lu

Fig. 8. Chondrite-normalized REE abundances for tuffs from A: the Pacific piedmont, B: the western Cordillera, C: the Altiplano and
D: the La Paz and Titicaca basin. Normalizing values after Nakamura (1974). "Ignimbrite" field of Central Andes from Schmitt-
Riegraf and Pichler (1988). CVZ field from Thrope et al. (1984). Dacite Al and Dacite A2 from Lefévre (1979).
 Trace and rare-earth element characteristics of acidic tuffs from Southern Peru and Northern Bolivia 105

(SA3-90) exhibit similar patterns of typical calc-alkaline tuffs (Fig. 8D). These tuffs interbbeded with lacustrine
high-silica rhyolite and show enrichment in the LREEs formations show almost identical and parallel REE pat-
relative to the HREEs (La/YbN = 15-16). From light to terns with minor Eu anomalies for the Chua and Chijini
heavy elements, concentrations decrease from about 100 tuffs. REE pattern of the Chijini tuff closely resemble
times for La to about 5 to 8 times chondritic abundances those of Miocene tuffs. Three Plio-Pleistocene tuffs dis-
for Yb. Compared to the basal vitrophyre of the Cota- play LOI > 3 wt%, but have high Ce abundances and the
huasi tuffs (PH1-79c, Fig. 8B) in the western Cordillera highest Ce/YbN ratios (21-27), therefore their alteration
having similar Si02 contents, typical glasses of the SA is considered, as for the Miocene tuffs, without inftuence
tuff are relatively less enriched in LREEs (La = 90-108 on their REE distribution patterns.
times chondrites) than the Cotahuasi vitrophyre (up to
200 times for La). This vitrophyre with the highest REE
DISCUSSION AND CONCLUSIONS
contents show a Jarger negative Eu anomaly similar to
this of another tuffs of the Pacific piedmont (Pe53). Ages of tuffs and plinian fall deposits from southern
Tuffs of intermediate to dacitic compositions of the Peru and northern Bolivia between 15ºS and 18ºS are
western Cordillera (Fig. 8B) display typical characteris- well-established and new geochemical data of these tuffs
tics of the CVZ. They have nearly linear trends but permit to compare trace-element and REE contents for
slightly depleted in HREEs (La/YbN = 10-14) and exhibit dacite-rhyolite tuffs from the western Cordillera and its
a weak Eu anomaly that might indicate major plagioclase piedmont to those from the northern Bolivian Altiplano
fractionation. The more evolved rhyolitic tuffs and obsid- and the eastern Cordillera piedmont. As a starting point
ians display moderate to pronounced negative Eu anoma- for discussion, we emphazise that geochemical variations
lies (e.g. Pe52 and PH1-79c) which suggests extensive of these tuffs may be a consequence of the tectonic com-
plagioclase fractionation except for one rhyolitic vitro- plexity in these areas and may provide evidence of the
phyre (Pe42) characterized by a small positive Eu heterogeneity of the magma source regions. Most authors
anomaly inside the precision of the analytical technique. agree on a crustal thickness slightly more than 50 km
Similar REE patterns in dacites are also found in many beneath the Bolivian Altiplano and a 30º east dipping
other calc-alkaline Cenozoic andesite-rhyolite associa- angle of the subducted slab between 17ºS and 20ºS.
tions. The REE distribution patterns of these dacitic tuffs Cahill and Isacks (1992) argued that the presence of an
is parallel to the Dacite Al trend defined by Lefévre asthenospheric wedge overlying the subducting Nazca
(1979) for dacites of southern Peru and reported to the plate is presumed to govern the process of magma gen-
CenozoYc volcanic activity between 150-250 km from the eration. Dorbath et al., (1993) predicted that the subduc-
Peru-Chile trench. tion of young oceanic lithosphere at a high speed for a
substantial time could produce hot mantle under the Alti-
More important for later discussions are the observa- plano and evidenced a major suture at the eastern border
tion that dacitic tuffs of the Altiplano (Fig. 8C) are com- of the La Paz and Titicaca basin (eastern Cordillera pied-
positionally distinct. The REE patterns for three Bolivian mont). In southern Peruvian and northern Bolivian Andes,
tuffs (the Toba 76 tuff, the ignimbrite Perez and the Ayo unconformities are overlain by tuffs and only in sorne
Ayo tufi) differ from the others and are weakly concave cases these tuffs are slightly folded and faulted (e.g. the
upward (115-165 times chondrites for La) and flat for Toba 76 tuff and the Chijini tuff in northern Bolivia). This
HREEs (about 5 times chondrites). This could be due to explosive volcanism may be interpreted in response to
source differences given that the Bolivian tuffs are from upper-crustal shortening and thickening during Quechua
-350-400 kms further from the Peru-Chile trench. None compressional events at -25, -17, -10, -7 and -2.5 Ma
of these rhyolitic tuffs exhibit Eu anomalies, which would (Lavenu, 1986; Sébrier et al., 1988; Mercier et al., 1992)
confirm that either little plagioclase was extracted during During Miocene times (7-8 Ma), a major compressional
crystallization. Miocene tuffs of the Bolivian Altiplano deformation induced a strong NE-SW shortening in the
which display little evidence of alteration with LOI > 3 High Andes of Bolivia (northern Altiplano and eastern
wt. %, exhibit distribution patterns similar for light and Cordillera, Laven u and Mercier, 1991 ). This tectonic
middle REE to the Dacite A2 trend reported by Lefévre event could be the paroxismal expression of a compres-
(1979) to Cenozoic volcanic activity between 250-400 sional period which affected the Andean Chain during the
km from the Peru-Chile trench. REE distribution pattern late Oligocene-Miocene times (Baby et al., 1990; Sem-
of the Ulloma tuff (PH64) is parallel but slightly depleted péré et al., 1990, 1991). The orogenic structuration of the
relative to this of the Callapa tuff (PH61 ). These samples Bolivian Altiplano and the western part of the eastern
show no evidence for any negative Ce anomalies result- Cordillera has been acquired through thrusts and strike-
ing from incipient alteration due to weathering in vol- slip faults during this period (Hérail et al., 1993) and the
canic rocks (Price et al., 1991). The Callapa tuff (PH61) Andean Chain reached its present topographical evolution
has high Ce and lower HREE contents that yields a high at this time. Thickening of the crust during uppermost
Ce/YbN ratio of 20. lt is concluded that the weathering of Miocene (-7 Ma) and upper Pliocene (-2.5 Ma) compres-
these tuffs is without influence on their REE distribution sional pulses was directly associated with uplift of the
patterns. High Andes in Peru and Bolivia (Sébrier et al., 1988).
Tuffs of the La Paz and Titicac basin are strongly Crustal shortening after 10 Ma have played an important
LREE enriched and HREE depleted as other Bolivian role in controlling the magmatic processes initiated by
106 N. VATIN-PERIGNON et al.

subduction. Tuffs took place at successive extensional worth and Clarke (1994 ), the high Nd values could reflect
periods during which thermal weakening and thinning of magma differentiation processes, perhaps associated with
the crust related to lengthening and increase in dip of the crustal contamination, smaller degrees of partial melting
slab probably occurred (Mercier et al., 1992). Voluminous and/or higher trace element abundances in the source.
peraluminous ash-ftow tuffs in the Morococala and Los Highly fractionated REE patterns of Bolivian tuffs may
Frailes volcanic fields (Ericksen et al., 1990) are of late be attributed to intensive crustal contamination. These
Miocene age (Morococala tuffs, 6.3-5.8 Ma; Lavenu, results are consistent with the interpretation of isotopic
1986 and Los Frailes tuffs, 8.5-6.4 Ma; Koeppen et al., variations of Quaternary volcanics in Central Andes
1987). These geophysical and tectonic data suggest that between 17ºS and 22ºS which allow Worner et al, (1992)
changes in tectonic regimes of the southern western Cor- and Davidson and de Silva (1992) to provide evidence for
dillera and northern Bolivian Altiplano which occured major crustal contribution to are magma genesis, surim-
with each Quechua tectonic pulse and which affected the posed on variations in the mantle source composition.
astenospheric wedge and both the dowgoing and the over-
riding lithospheres may be responsible for the trace ele- Especially after the upper Miocene compressional
ment diversity observed in tuffs along the east-west phase which was a time of major crustal growth in the
CVZ, tuffs erupted between 5 and 2.8 Ma have much
traverse from the Pacific piedmont in Peru to the eastern
lower Sr/Y ratios (4.5-15.5) in southern Peru than those
Cordillera piedmont in Bolivia.
erupted in northern Bolivia (24-37.5). These differences
With the exception of vitrophyres from the Cotahuasi in tuffs of the same age erupted at 150-250 km or 250-
region in the western Cordillera and the Cota Cota tuff of 400 km from the Peru-Chile trench may be interpreted as
the La Paz and Titicaca basin, the incipient hydratation reftecting variable contributions from the mantle wedge
observed in glasses of sorne tuffs and due to weathering and partial melting of the subducted lithosphere. The
might have relatively no inftuence on their REEs and Bolivian tuffs have a large degree of LREE enrichement
other trace element concentrations. These concentrations, relative to HREE and high Ce/Y ratios. These data pro-
similar to those of nonaltered tuffs, may be considered to vide additional evidence for the importance of the com-
reftect the magmatic processes and compositions of tuffs. pressional event at - 7 Ma and support our interpretation
Trace-element compositions and REE concentrations in that these differences in magma chemistry may be attrib-
tuffs correspond closely with the composition of the lava uted to the subduction of a warm oceanic lithosphere
flows for a given whole-rock Si02 content. Tuffs have under the Bolivian Altiplano (Dorbath et al., 1993). Such
trace element signature typical of subduction-related con- a phenomenon might have led to partial melting of the
tinental are magmatism, showing a strong LILE/HFSE subducted lithosphere rather than of the mantle wedge (de
enrichment and nonenrichment of HREE and Y, as Boer et al., 1991)
pointed out by Gill (1981). It is difficult to evaluate the
FT dating of obsidian glass fragments of the SA Tuff
effects of fractionation and assimilaltion in these tuffs
have yielded an age of 2.42±0.11 Ma similar in age to the
which exhibit a restricted range in Si02 (the majority
Vitor tuff at 2.76±0.10 Ma. The SA tuff is not contempo-
have 61-76%Si0 2 ). However, sorne space-related obser-
raneous with the Toba 76 tuff or the Perez ignimbrite
vations can be made on the basis of their geochemical
from the Bolivian Altiplano as it has been proposed.
characteristics and differences in age. During Neogene-
These tuffs typically date the last Quechua compressional
Pleistocene times it appears two distinct groups of tuffs.
phase during the lower Pleistocene (ca. 2.5 Ma) and the
Tuffs erupted in northern Bolivia are typically enriched in
characteristics of magmas erupted before and after this
Zr, Hf, Th and LREEs relative to those erupted in south-
phase essentially remained the same. Geochemical data
ern Peru, ata given Si02 content (65-67 wt.% for dacites
record the effects of changes in the subduction beneath
and 72-73 wt.% for rhyolites). These two types of enrich-
the southern Peru and northern Bolivia during the late
ments may be related to differences in subduction proc-
Miocene compressional events. This would appear to con-
esses and could be consistent with their derivation from
firm that from the last compressional event, the geody-
different magma sources. The higher values in Bolivian
namical situation remained one of high convergence.
tuffs suggest an enriched mantle source. Ratios of Ce/YbN
Geochemical data of Bolivian tuffs yielding an age
increase eastward from 7-13 (mean value of 11) for Peru-
between 5 and 2.8 Ma provide evidence that the upper
vian tuffs to 13-27 (mean value of 19) for Bolivian tuffs,
Miocene (-7 Ma) compressional deformations were a
values which could indicate variable degrees of melting
major event prior to these tuff emplacements and played
of mantle source. Along the Central American are, such
the most important role on the variability of the geo-
variations are linked to the subduction geomctry and vol-
chemical characteristics of the southern Peruvian and
ume of the astenospheric wedge (Carretal., 1990). In the
northern Bolivian tuffs.
case of Neogene and Pleistocene tuffs of southern Peru
and northern Bolivia, higher Ce/YbN ratios could reftect
the variation of degree of wedge contribution toward the Acknowledgements - This work is part of a continuing program on the
east. The late Miocene to Pliocene Bolivian tuffs have dating and characterization of Peruvian Neogene silicic tuffs. Two of us
(G.P. and N.V.-P.) thank "Multiciencias", Lima, for partial support in
high Th/Ta ratios (10-15), high Th/Nb ratios close to field work in Peru. G. Poupeau is indebted to Dr. Victor Latorre, "Mul-
those of other CVZ tuffs (de Silva and Francis, 1989) and ticiencias" Director, and the International Center for Theoretical Physics
high Nd contents (20-46 ppm). As shown by Hawkes- (Trieste) for pluri-annual support in Peru. We are grateful to the Institut
 Trace and rare-earth element characteristics of acidic tuffs from Southern Peru and Northern Bolivia 107

Fran~ais de Recherche Scientifique pour le Développement en Coopéra- Francis, P.W. and Hawkesworth, C.J., 1994. Late Cenozoic rates of
tion (ORSTOM) for bis logistical support and assistance during field- magmatic activity in the Central Andes and their relationships to
work in Bolivia, Christian Picard for a critica! reading of an earlier draft continental crust formation and thickening. Journal of the Geologi-
and C.W. Naeser and S. de Silva for thorough, constructive reviews cal Society, London, 151, 845-854.
which resulted in substantial improvement of this paper. This paper is a
Germanique, J.C. and Briand B., 1985. XRF determination of Zr, Nb, Y,
contribution to I.G.C.P. project 345 "Lithospheric evolution of the
Andes". Sr, Rb, Zn and Pb in fifteen Intemational geochemical reference
samples. Geostandards Newsletters, 9, 31-34.
Gill, J.B., 1981. Orogenic andesites and Plate tectonics. Springer Ver-
REFERENCES lag, Berlin, 310 p.

Baby, P., Sempéré, T., Oller, J., Barrios, L., Hérail, G. and Marocco, R., Gleadow, A.J.W., 1981. Fission track dating methods: what are the real
1990. Un bassin en compression d'age Oligo-Miocéne dans le sud alternatives? Nuclear Tracks, 5, 3-14.
de l'Altiplano bolivien. Compte Rendus de l'Académie des Sciences,
Grange, F., Cunningham, P., Gagnepain, J., Hatzfeld, D., Molnar, P.,
Paris, 311, 341-347.
Ocola, L., Rodrigues, A.L., Roecker, S.W., Stock, J.M. and Suarez,
Bellon, H. and Lefévre, C., 1976. Données géochronologiques sur le G., 1984. The configuration of the seismic zone and the downgoing
volcanisme andin dans le sud du Pérou. Implications volcano- slab in southem Peru. Geophysical Research letters, 11, 38--41.
tectoniques. Compte Rendus de l'Académie des Sciences, Paris, 283,
Hawkesworth, C.J. and Clarke, C.B., 1994. Partial melting in the lower
1--4.
crust: new constraints on crustal contamination processes in the cen-
Bigazzi, G., Dompnier, E., Hadler Neto, J.C., Poupeau, G. and Vuilliez, tral Andes. In: Tectonics of the Southern Central Andes (edited by
E., 1988. A reactor rule calibration for fission-track dating and Ura- K.-J Reutter, E. Scheuber and P.J. Wigger), pp. 93-101. Springer-
nium micromapping. Nuclear Tracks Radiation Measurement, 15, Verlag, Heidelberg, Germany.
755-757.
Hérail, G., Soler, P., Bonhomme, M. et Lizeca, J.L., 1993. Evolution
de Boer, J.Z., Defant, M.J., Stewart, R.H. and Bellon, H., 1991. Evi- géodynamique de la transition entre l' Altiplano et la Cordillére Ori-
dence for active subduction below western Panama. Geology, 19, entale au nord d'Oruro (Solivie). Implications sur le déroulement de
649-652. I'orogenése andine. Compte Rendus de l'Académie des Sciences,
Paris, 317, 515-522.
Cahill, T. and lsacks, B.L., 1992. Seismicity and shape of the subducted
Nazca plate. Journal of Geophysical Research, 97, 17503-17529. Hildreth, W. and Moorbath, S., 1988. Crustal contributions to are mag-
matism in the Andes of Central Chile. Contribution.f to Mineralogy
Cameron, K.L. and Hanson, G.N., 1982. Rare Earth element evidence
and Petrology, 98, 455--489.
conceming the origin of voluminous mid-Tertiary rhyolitic ign-
imbrites and related volcanic rocks, Sierra madre Occidental, Chi- Hurford, A.J., 1990. lntemational Union of Geological Sciences, sub-
huahua, Mexico. Geochimica et Cosmochimica Acta, 46, 1489- commission on Geochronology recommandation for the standardiza-
1503. tion of fission track dating calibration and data reporting. Nuclear
Tracks Radiation Measurement, 17, 233-236.
Carr, M.J., Feigenson, M.O. and Bennett, E.A., 1990. Incompatible ele-
ment and isotopic evidence for tectonic control of source mixing and Hurford, A.J. and Green, P.F., 1983. The zeta age calibration of fission-
melt extraction along the Central American are. Contributions to track dating. Isotope Geosciences, 1, 285-317.
Mineralogy and Petrology, 105, 369-380.
Isacks, B.L., 1988. Uplift of the Central Andean plateau and bending of
Davidson, J.P., McMillan, N.J., Moorbath, S., Womer, G., Harmon, R.S. the Solivian orocline. Joumal of Geophysical Research, 93, 3211-
and Lopez-Escobar, L., 1990. The Nevados de Payachata volcanic 3231.
region (18ºS/69ºW, N. Chile) II. Evidence for widespread crustal
involvement in Andean magmatism. Contributions to Mineralogy Jenks, W.F. and Goldich, S.X., 1956. Rhyolitic tuff flows in southem
and Petrology, 105, 412--432. Peru. Journal of Geology, 64, 156-172.

Davidson, J.P. and de Silva, S.L., 1992. Volcanic rocks from the Soliv- Koeppen, R.P., Smith, R.L., Kunk, M.J., Flores, A., Luedke, R.G. and
ian Altiplano: insights into crustal structure, contamination and Sutter, J.F., 1987. The Morococala volcanics: highly peraluminous
magma genesis in the Central Andes. Geo/ogy, 20, 1127-1130. rhyolite ash flow magmatism in the Cordillera Oriental, Bolivia
(abs.). Geological Society of America Abstracts with Programs, 19,
Dorbath, C., Granel, M., Poupinet, G. and Martinez, C., 1993. A tel- 731.
eseismic study of the Altiplano and the Eastem Cordillera in north-
em Bolivia: new constraints on a lithospheric model. Journal of Laharie, R. and Derruau, M., 1974. La morphogenése des Andes du Sud
Geophysical Research, 98, 9825-9844. du Pérou. Revue de Géographie Alpine, 70, 479-505.

Dorighel, O., Poupeau, G., Bouchard, J.-F. and Labrin, E., 1994. Data- Lavenu, A., 1986. Etude néotectonique de /'Altiplano et de la Cordi/lére
tion par traces de fission et étude de provenance d'artefacts en obsi- orienta/e des Andes Boliviennes. Unpublished Thesis, University of
dienne des siles archéologiques de La Tolita (Equateur) et lnguapi Paris-Sud, Orsay, 434 p.
(Colombie). Bulletin de la Société de Préhistoire de Frunce, in Lavenu, A., Bonhomme, M.G., Vatin-Pérignon, N. and De Pachtére, P.,
press.
1989. Neogene magmatism in the Solivian Andes between 16ºS and
Ericksen, G.E., Luedke, R.G., Smith, R.L., Koeppen, R.P. and Urquidi ISºS: stratigraphy and K/Ar geochronology. Journal of South
F., 1990. Peraluminous igneous rocks of the Solivian tin belt. Epi- America Earth Sciences, 2, 35--47.
.wdes, 13, 1-7.
Lavenu, A. et Mercier, J.-L., 1991. Evolution du régime tectonique de
Evemden, J.F., Kriz, S.J. and Cherroni M, C., 1966. Correlaciones de las l' Altiplano et de la Cordillére Orientale des Ande de Solivie du
formaciones terciarias de la cuenca altiplanica a base de edades Miocéne supérieur a l' Actuel. Un effet des forces de gravité et des
absolutas determinadas por el metodo Potasio-Argon. Servicio forces aux limites. Géodynamique, 6, 21-55.
Geologíco de Bolivia, Hoja inf. nº 1.
Lefévre, C., 1979. Un exemple de volcanüme de marge active dans les
Evemden, J.F., Kriz, S.J. and Cherroni M, C., 1977. Potassium-Argon Andes du Pérou (Sud) du Miocéne a l'Actuel (Zonations et pétro-
ages of sorne Solivian rocks. Economic Geology. 72, 1042-1061. ge11ése de.f andésite.f et des shoshonites). Unpublished Thesis, Uni-
versity of Montpellier, 555 p.
Fenner, C.N., 1948. lncandescent tuff flows in southem Peru. Bulletin of
the Geo/ogical Society <!fAmerica, 59, 879-893. Luhr, J.F., 1992. Slab-derived fluids and partial melting in subduction
108 N. VATIN-PERIGNON et al.

zones: insights from two contrasting Mexican volcanoes (Colima Sempéré, T.. Baby, P., Oller, J., and Hérail, G .. 1991. La nappe de Cala-
and Ceboruco). Journal of Volcanology and Geothermal Re.vearch, zaya: une preuve de raccourcissements majeurs gouvemés par des
54, 1-18. éléments paléostructuraux dans les Andes boliviennes. Compre Ren-
dus de l'Académie des Sciences, Paris, 312, 77-83.
Mercier J.L., Sébrier, M., Lavenu, A., Cabrera, J., Bellier, O., Dumont,
J.-F. and Macharé, J., 1992. Changes in the tectonic regime above a de Silva, S.L., 1989. Geochronology and stratigraphy of the ignj¡nbrites
subduction zone of Andean type; the Andes of Peru and Bolivia dur- from the 21 º30xS to 23º30xS portion of the Central Andes of north-
ing the Pliocene-Pleistocene. Journal of Geophysical Re.vearch, 91, ern Chile. Journal of Volcanology and Geothermal Research, 37,
11945-11982. 93-131.
McCulloch, M. T. and Gamble, J .A., 1991. Geochemical and geodynami- de Silva, S.L., Davidson, J.P., Croudace, A.W. and Escobar A., 1993.
cal constraints on subduction zone magmatism. Earth and Planetary Volcanological and petrological evolution of Volean Tata Sabaya,
Science Letters, 102, 358-374. SW Bolivia. Journal of Vocanology and Geothermal Research, 55,
Martinez, C., 1980. Structure et évolution de la chaine andine dans le 305-335.
nord de la Cordillére des Andes de Bolivie. ORSTOM, Travaux et de Silva, S.L. and Francis, P.W. 1989. Correlation of large ignimbrites-
documents, 119, 352 p. two case studies from the Central Andes of northem Chile. Journal
Nakamura, N., 1974. Determination of REE, Ba, Fe, Mg, Na and k in of Volcanology and Geothermal Research, 37, 133-149.
carbonaceous and ordinary chondrites. Geochimica et Cosmo- de Silva, S.L. and Francis, P.W. 1991. Volcanoes of the Central Andes.
chimica Acta, 38, 757-775. Springer-Verlag, Heidelberg, Germany, 216 p.
Noble, D.C., Sébrier, M., Mégard, F. and McKee, E.H .. 1985. Demon- Smith, D.R. and Leeman, W.P., 1993. The origin of Mount St. Helens
stration of two pulses of Palaeogene deformation in the Andes of andesites. Journal of Volcanology and Geothermal Research, 55,
Peru. Earth and Planetary Science Letters, 73, 345-349. 271-303.
Odin, G.S., 1994. Geological time scale. Compre Rendus de /'Académie Sparks, R.S.J., Self, S. and Walker, G.P.L. (1973). products of ign-
des Sciences, Paris, 318, 59-71. imbrite eruptions. Geology, 1, 115-118.
Oliver, R.A .. Robinson, S., Vittoz, P. and Vivier, G., 1990. The method
Storzer, D. and Poupeau, G. 1973. Ages-plateaux de verres et minéraux
of Instrumental Activation Analysis as practised at the lnstitut Laüe-
par la méthode des traces de fission. Compre Rendus de l'Académie
Langevin (ILL), together with the abundance of 16 elements in
des Sciences, París, 276, 137-139.
twenty one International standard reference materials measured by
this method. Interna/ ILL report, 11 p. Thorpe, R.S., Francis, P.W. and Moorbath, S., 1979. Rare earth and
strontium isotope evidence conceming the petrogenesis of north
Pearce, J.A. 1983. Role of the sub-continental lithosphere in magma
Chiliean ignimbrites. Earth and Planetary Science Letters, 42, 359-
genesis at active continental margins. In: Continental basalts <md
367.
mantle xenoliths (edited by C.J. Hawkesworth and M.J. Norry), pp.
230-249. Shiva Publishing, Nantwich, UK, 262 p. Thorpe, R.S., Francis, P.W. and O'Callaghan, L. 1984. Relative soles of
source composition, fractional composition and crustal contamina-
Poupeau, G., Labrin, E .. Sabil, N., Bigazzi, G., Arroyo, G. and Vatin-
tion in the petrogenesis of Andean volcanic rocks. Philisophical
Pérignon, N., 1994. Fission-track dating of 15 macusanites glass
Transactions of the Royal Society of London, A310, 305-320.
pebbles from the macusani volcanic field (SE Peru). Nuclear Trace
Radiatio11 Measurements, in press. Toribio, P. 1899. Sinopsis de los tremblores y volcanes del Perú. 111.-
Price, R.C., Gray, C.M., Wilson, R.E., Frey, F.A. and Taylor, S.R., 1991. Erupción del Huaina-Putina en 1600. Boletín della Societá
The etfects of weathering on rare-earth elements. Y and Ba abun- Geografíca de Lima, 9, 67-68.
dances in Tertiary basalts from southeastern Australia. Chemical Tosdal, R.M., Farrar, E. and Clark, A.H., 1981. K-Ar geochronology of
Geology, 93, 245-265. the late cenozoic volcanic rocks of the Cordillera Occidental, South-
Redwood, S.D., 1987. The Soledad caldera, Bolivia: a Miocene caldera emmost Perú. Journal ofVolcanology and Geothermal Research, 10,
with associated epithermal Au-Ag-Cu-Pb-Zn mineralization. Geo- 157-173.
logical Society of America Bulletin, 99, 395-404. Vatin-Pérignon, N., Vivier, G., Sébrier, M. and Fomari. M. 1982. Les
Redwood, S.D. and Macintyre, R.M., 1989. K-Ar dating of Miocene demiers événements andins marqués par le volcanisme cénozolque
magmatism and related epithermal mineralization of the northeastem de la Cordillére occidentale sud-péruvienne et de son piémont paci-
Altiplano of Bolivia. Economic Geology. 84, 618-630. fique entre 15º45x et 18ºS. Bulletin de la Société Géologique de
France, (7) 24, 679-650.
Schmitt-Riegraf, C. and Pichler, H., 1988. Cenoaoic ignimbrites of the
Central Andes: a new genetic model. In: The Southern Central Wark, D.A., 1991. Oligocene ash tlow volcanism, northem Sierra Madre
Andes (edited by H. Bahlburg, Ch. Breitkreuz and P. Giese), Lecture Occidental: role of mafic and intermediate-composition magmas in
Notes in Earth Sciences, 17, pp. 183-197. Springer-Verlag, Berlín, rhyolite genesis. Journal of Geophysical Research, 96, 13389-
Germany, 261 p. 13411.

Sébrier, M., Lavenu, A., Fomari, M. and Soulas, J.-P. 1988. Tectonics von Weibel, M., Frangipane-Gysel, M. and Hunziker, J., 1978. Ein bei-
and uplift in Central Andes (Peru, Bolivia and Northem Chile) from trag zur Vulkanologie Süd-Perus. Geologische Rundschau, 61, 243-
Eocene to Present. Géodynamique, 3, 85-106. 252.

Sempéré, T., Hérail, G., Oller, J. and Bonhomme, M.G., 1990. Late Wiimer, G., Moorbath, S. and Harmon, R.S., 1992. Andean Cenozoic
Oligocene-early Miocene major tectonic crisis and related basins in volcanic centers reftect basement isotopic domains. Geology, 20,
Bolivia. Geology, 18, 946-949. 1103-1106.
 Trace and rare-earth element characteristics of acidic tutfs from Southern Peru and Northern Bolivia 109

APPENDIX cleaved mica muscovite sheets maintained in contact with

NIST reference wafer glass 962a during irradiation. n
Fission- Track Experimental Procedures
glass wafers were used, each of which sandwiched
We collected three obsidian samples with a diameter of between 2 muscovite sheets. After irradiation, tracks were
-2 cm from the upper unit of the SA tutf, in the qda Añas etched in HF 40% for 70 mn at 20ºC. Track counts were
Huayco near Arequipa for determining fission track ages. made with a dry x190 objective and xlO oculars.
We have adopted for the fission-track dating of obsidian
Apparent ages were determined with the "ditference"
samples one of the approaches recommanded by the
technique (Gleadow, 1981), into which no laboratory
lnternational Union of Geological Sciences, subcommis-
heating treatment to erase fossil tracks is applied to the
sion on Geochronology (Hurford, 1990) for the "popula-
samples befare irradiation. After irradiation, two frag-
tion" method. Our procedures, justified in Bigazzi et al.
ments of each sample, respectively irradiatcd and non-
(1988), are numerically equivalent to the onc using the
irradiated, are mounted together in araldite and polished.
zeta calibration (Hurford and Green, 1983).
Tracks were etched with HF 20% at 20ºC for 90 seconds.
One of several mm-sized fragments of each sample Track counts were made either with an Orthoplan or a
was irradiated in the well thermalized position P l of the Jenavert-Zeiss Jena microscope with a xlOO dry objective
Orphée nuclear reactor of the Centre d'Etudes Nucléaires and xlO oculars. Track size was characterized by the
of Saclay. The neutron fluence was monitored from thc measurements of track diameter, defined as the largest
induced fission track densities measured in freshly linear dimension of a track.

SAES 9, 1/2-H