Bull Volcanol (2001) 62 : 493518

DOI 10.1007/s004450000105
RESEARCH ARTICLE
Nancy K. Adams Shanaka L. de Silva
Stephen Self Guido Salas Steven Schubring
Jason L. Permenter Kendra Arbesman
The physical volcanology of the 1600 eruption of Huaynaputina,
southern Peru
Received: 9 July 1999 / Accepted: 15 August 2000 / Published online: 9 February 2001
Springer-Verlag 2001
Abstract Volcn Huaynaputina is a group of four
vents located at 1636'S, 7051'W in southern Peru
that produced one of the largest eruptions of historical
times when ~11 km
3
of magma was erupted during
the period 19 February to 6 March 1600. The main
eruptive vents are located at 4200 m within an ero-
sion-modified amphitheater of a significantly older
stratovolcano. The eruption proceeded in three stages.
Stage I was an ~20-h sustained plinian eruption on
1920 February that produced an extensive dacite
pumice fall deposit (magma volume ~2.6 km
3
).
Throughout medialdistal and distal parts of the dis-
persal area, a fine-grained plinian ashfall unit overlies
the pumice fall deposit. This very widespread ash
(magma volume ~6.2 km
3
) has been recognized in
Antarctic ice cores. A short period of quiescence
allowed local erosion of the uppermost stage-I depos-
its and was followed by renewed but intermittent
explosive activity between 22 and 26 February (stage
II). This activity resulted in intercalated pyroclastic
flow and pumice fall deposits (~1 km
3
). The flow
deposits are valley confined, whereas associated co-ig-
nimbrite ash fall is found overlying the plinian ash
deposit. Following another period of quiescence, vul-
canian-type explosions of stage III commenced on 28
February and produced crudely bedded ash, lapilli,
and bombs of dense dacite (~1 km
3
). Activity ceased
on 6 March. Compositions erupted are predominantly
high-K dacites with a phenocryst assemblage of pla-
gioclase>hornblende>biotite>FeTi oxidesapatite.
Major elements are broadly similar in all three stages,
but there are a few important differences. Stage-I
pumice has less evolved glass compositions (~73%
SiO
2
), lower crystal contents (1720%), lower density
(1.01.3 g/cm
3
), and phase equilibria suggest higher
temperature and volatile contents. Stage-II and stage-
III juvenile clasts have more evolved glass (~76%
SiO
2
) compositions, higher crystal contents (2535%),
higher densities (up to 2.2 g/cm
3
), and lower temper-
ature and volatile contents. All juvenile clasts show
mineralogical evidence for thermal disequilibrium.
Inflections on a plot of log thickness vs area
1/2
for the
fall deposits suggest that the pumice fall and the plin-
ian ash fall were dispersed under different conditions
and may have been derived from different parts of the
eruption column system. The ash appears to have
been dispersed mainly from the uppermost parts of
the umbrella cloud by upper-level winds, whereas the
pumice fall may have been derived from the lower
parts of the umbrella cloud and vertical part of the
eruption column and transported by a lower-altitude
wind field. Thickness half distances and clast half dis-
tances for the pumice fall deposit suggests a column
neutral buoyancy height of 2432 km and a total col-
umn height of 3446 km. The estimated mass dis-
charge rate for the ~20-h-long stage-I eruption is
2.410
8
kg/s and the volumetric discharge rate is
~3.610
5
m
3
/s. The pumice fall deposit has a dispersal
index (Hildreth and Drake 1992) of 4.4, and its index
of fragmentation is at least 89%, reflecting the domi-
nant volume of fines produced. Of the 11 km
3
total
volume of dacite magma erupted in 1600, approx-
imately 85% was evacuated during stage 1. The three
main vents range in size from ~70 to ~400 m. Align-
Editorial responsibility: W. Hildreth
N. K. Adams S. L. de Silva (
)
) S. Schubring
J. L. Permenter K. Arbesman
Department of Geography, Geology, and Anthropology,
Indiana State University, Terre Haute, IN 47809, USA
E-mail: gesilva@scifac.indstate.edu
Phone: +812-2372269
S. Self
Hawaii Center for Volcanology, SOEST, University of Hawaii,
Honolulu, HI
G. Salas
Universidad de San Agustin, Casilla 1023, Arequipa, Peru
494
ment of these vents and a late-stage dyke parallel to
the NNWSSE trend defined by older volcanics sug-
gest that the eruption initiated along a fissure that
developed along pre-existing weaknesses. During stage
I this fissure evolved into a large flared vent, vent 2,
with a diameter of approximately 400 m. This vent
was active throughout stage II, at the end of which a
dome was emplaced within it. During stage III this
dome was eviscerated forming the youngest vent in
the group, vent 3. A minor extra-amphitheater vent
was produced during the final event of the eruptive
sequence. Recharge may have induced magma to rise
away from a deep zone of magma generation and stor-
age. Subsequently, vesiculation in the rising magma
batch, possibly enhanced by interaction with an
ancient hydrothermal system, triggered and fueled the
sustained Plinian eruption of stage I. A lower volatile
content in the stage-II and stage-III magma led to
transitional column behavior and pyroclastic flow gen-
eration in stage II. Continued magma uprise led to
emplacement of a dome which was subsequently
destroyed during stage III. No caldera collapse
occurred because no shallow magma chamber devel-
oped beneath this volcano.
Keywords Huaynaputina Peru Plinian eruption
Stratigraphy Physical volcanology Vent evolution
Introduction
The 1600 eruption of Volcn Huaynaputina, located at
1636'S, 7051'W in southern Peru (Fig. 1), has been
recently recognized as the largest historical eruption
in South America and one of the largest in the past
1000 years (VEI 6; Thouret et al. 1997, 1999; de Silva
and Zielinski 1998). Since the account of the eruption
given by Bullard (1962), it has become the subject of
speculation as the putative source of a major acidity
spike in the South Pole (Antarctica) and GISP
Fig. 1 Location map shows the regional context of Volcn
Huaynaputina. The triangles are the active volcanoes of the Cen-
tral Volcanic Zone arc. 1 Sabancaya; 2 Chachani; 3 Misti; 4 Ubi-
nas; 5 Ticsani. Huaynaputina is located in the Moquegua prov-
ince of Peru. The approximate known extent of ash fall on land
from known exposures (mostly reworked), historic reports, and
ash layers recorded in ice caps of surrounding mountains is also
shown. Note the approximate extent of the stage-I pumice fall
deposit (stippled field) which was dispersed primarily to the
WSW. Key historic recordings are also given
495
(Greenland) ice cores (Hammer et al. 1980, 1981;
Thompson et al. 1986; de Silva and Francis 1990, 1991;
Palais et al. 1990). Historical, parochial, and anecdotal
accounts synthesized by de Silva et al. (2000) indicate
that this major plinian eruption commenced on 19
February 1600, and ash fall continued until 6 March
1600, a period of 16 days. The eruption was locally
and regionally devastating (de Silva et al. 1998, 2000;
Thouret et al. 1999), and aerosols from the eruption
may have been at least partly responsible for the
worst growing conditions for northern hemisphere
trees and the coldest winter in the past 600 years (de
Silva and Zielinski 1998; Briffa et al. 1998).
The 1600 eruption is important for our understand-
ing of volcanically induced climate perturbations in
the historical past, and the utility of the ash as a
regional stratigraphic marker has already been noted
(Thompson et al. 1986; de Silva and Francis 1991;
Thouret et al. 1997; Thompson et al. 1998).
This paper describes the eruption stratigraphy and
products of the 1600 eruption of Huaynaputina based
on detailed field work at and around the source vents
and in distal areas. We then focus on the physical vol-
canology of the main plinian phase. Since the grain-
size character of fall deposits, including stratification
and fluctuations in size, reflects processes of fragmen-
tation and eruption (cf. Walker 1973; Houghton et al.
1999), such features hold vital clues about eruptive
processes. We therefore invert these physical data
from the main plinian fall deposit to yield the key
eruptive parameters. Finally, we describe the vents,
relate them to the deposit stratigraphy, and present a
preliminary model for the evolution of the eruption.
The 1600 deposits have also been studied recently
by another group (Thouret et al. 1997, 1999). Our
interpretations very generally agree with those of this
group, for instance, that the eruption can be divided
into three main stages, but differ from theirs in some
important aspects which are summarized herein and
discussed in detail elsewhere.
Site of the 1600 eruption
Apparently, no individual volcano had been locally
recognized at the site of the 1600 eruption prior to
that date and the pre-1600 topography is described as
a low ridge in the center of a Sierra (Vasquez de
Espinosa 1942; Bullard 1962). The eruption site was
named Huaynaputina after the fact and consists of
four vents that erupted through an unnamed pre-exist-
ing composite volcano (Figs. 2, 3). Three of the four
vents formed in 1600 are spectacularly located at
~4200 m within a 2.5-km-diameter amphitheater
carved into an older composite cone (Fig. 3), which
has since also become known as Huaynaputina. These
are some of the highest plinian vents yet described.
The fourth vent is located on the southern flank of
the host volcano. The 1600 vents, and recent (postgla-
cial) dacite bodies in the immediate area, are aligned
along a NNWSSE trend on which the nearby volca-
noes Ubinas and Ticsani are also aligned (Figs. 1, 2).
Fig. 2 Physiography of the
area surrounding Huaynapu-
tina. Contour interval is
200 m. Note that Huaynapu-
tina is located on the eastern
edge of an elevated plateau.
The steep drainages off this
plateau channeled and focused
primary and secondary pyro-
clastic flows into the main
drainage of the Rio Tambo.
Pyroclastic flow deposits are
concentrated in these drain-
ages, mainly to the east and
south. Historic reports of the
Rio Tambo being dammed
probably correlate with
emplacement of these deposits
at the mouths of the tributary
canyons: Quebradas Huayna-
putina; Tortoral; Aguas Blan-
cas; and del Volcn. The main
dam was probably at the con-
fluence of Quebrada del Vol-
cn and the Rio Tambo. UAB
Upper Aguas Blancas
496
This lineament is slightly oblique to, and approx-
imately 50 km behind, the main arc of the Central
Volcanic Zone of the Andes (de Silva and Francis
1990).
We thus differ with Thouret et al. (1999) in two
important aspects. Firstly, they identify the amphi-
theater as a caldera, but based on morphological sim-
ilarities with nearby extinct volcanoes, we think that
the amphitheater is a glacial cirque which developed
in an old sector collapse of the ancient volcano. The
flanks of the ancestral edifice show clear evidence of
glaciation. The cirque was subsequently reamed out
and modified during the 1600 eruption to produce the
feature we see presently, which we thus prefer to call
an amphitheater. Secondly, we recognize only three
vents inside the amphitheater, whereas Thouret et al.
(1997, 1999) recognized four or five such vents. Our
fourth vent is outside the amphitheater; they make no
mention of this.
The basement to the volcanoes in the area around
Huaynaputina is composed of voluminous, extensive,
Late Tertiary ignimbrite sheets of the Late Miocene
Sennca Formation and volcaniclastics and andesites of
the latest Miocene to earliest Pliocene Capillune For-
mation (Tosdal et al. 1981). The older glaciated host
volcano to Huaynaputina, and others like it, are part
of the Pliocene to Pleistocene Barroso Group (James
et al. 1976; Tosdal et al. 1981). The Late Tertiary vol-
canics are approximately 500 m thick below Huayna-
putina and overlie almost 2 km of Mesozoic and
Paleozoic sediments and intrusions. Young postglacial
volcanic activity is assigned to the Quaternary (Tosdal
et al. 1981) and includes the Arequipa volcanics
(James et al. 1976) and, in the area of Huaynaputina,
the Ubinas volcanics (James 1981); the latter consists
of youthful volcanoes such as Volcn Ubinas, Volcn
Ticsani, and Huaynaputina. Volcn Ubinas is typical
of the composite cones on the volcanic front and is
dominated by effusive eruptions of andesite and sub-
ordinate dacite. Ticsani and Huaynaputina, on the
other hand, are different in that they have been the
site of voluminous eruptions of silicic, mainly dacite,
magma. The youngest activity of Ticsani, approx-
imately 30 km to the southeast, must predate 1600
because there are no historical reports. Ticsani is a
Holocene (non-glaciated) dacite flow-dome complex,
from which the most recent activity was a sub-plinian
eruption that produced an east-directed pumice fall
deposit and a steep proximal cone of non-expanded
dacite along a short fissure aligned NNWSSE. While
superficially similar in color, vesicularity, and mineral-
ogy to that of Huaynaputina, the Ticsani dacite is
chemically distinct (Arbesman and de Silva 1998; S.
Schubring et al., unpublished data). Near Huaynapu-
tina are other young unnamed dacite lava bodies and
a larger, possibly monogenetic dome called Cerro El
Volcn (Fig. 2). The young dacites of the Ubinas vol-
canics belong to a separate magmatic episode from
that which produced the older composite cones of the
Barroso Group. Some of these young dacites have
erupted within the older edifice at Huaynaputina, but
the older edifice is the accidental host, and not ances-
tral to these. We think that intersection of the 1600
fissure with the older edifice and its extinct hydrother-
mal system focused the explosive eruption there, prob-
ably using pre-existing weaknesses to re-establish a
conduit to the surface.
We concur with de Silva and Francis (1990, 1991)
who classified the 1600 Huaynaputina eruption as the
formation of a monogenetic volcano.
The eruption and its effects
Anecdotal and parochial accounts provide a colorful
history of the eruption and its effects (Table 1; e.g.,
Vasquez de Espinosa 1942; Barriga 1951). For a
detailed account and sources, the reader is referred to
de Silva et al. (2000). Precursory seismicity began on
15 February and the eruption began at approximately
5 p.m. on 19 February, lasting approximately 16 days
until 6 March. During this time, constant strong earth-
quakes, intermittent ash, and pumice fall plagued the
population and wreaked widespread destruction (Ta-
Fig. 3 Topography of Volcn Huaynaputina. The three main
vents (13) are nested on the floor of the amphitheater cut into
an extinct stratovolcano. The nested vent cluster is the focus of
fumarolic activity and exposes a zone of dacite dykes which are
aligned NNW. The extra-amphitheater vent (4) is probably
located on an en-echelon extension of the 1600 dyke
497
ble 1). Bullard (1962) makes reference to ash fall as
late as 15 or 16 March 1600. Since there are no
accounts of earthquakes or eruptions coincident with
these later occurrences, they were probably the result
of ash being redistributed by the wind. Simkin and
Siebert (1994) tentatively mention an eruption of
Huaynaputina in 1667, to which Thouret et al. (1997)
ascribe to one of their four vents. We have not been
able to ascertain the source of this information in the
historical and parochial literature, and we are unsure
of its veracity. We note that Volcn Ubinas did erupt
in 1667 (Simkin and Siebert 1994) and it is likely that
the reference is to that eruption . Our observations
indicate that all the events, deposits, and vents we
describe here were formed in 1600.
The 1600 eruption resulted in complete destruction
of local communities and the loss of at least 1000 lives
as a result of accumulation of thick fall deposits and
by the passage of pyroclastic and secondary mass
flows (de Silva et al. 2000). Deposition of mass-flow
material also resulted in the damming of rivers.
Breaches of these dams in the Rio Tambo resulted in
flooding and transport of communities, livestock,
crops, and fish out to the sea. A thriving colonial wine
industry was dealt a death blow by the eruption (Rice
and Smith 1989). The loss of such resources com-
pounded to a significant economic burden on this
region.
Climate proxy data indicate that the summer of
1601 was the coldest of the past 600 years throughout
much of the northern hemisphere, and among the
coldest of the past 1500 years in Fennoscandia (Briffa
et al. 1998). High levels of H
2
SO
4
found in ice core
layers dated to 15991604 (including dating error)
from both polar regions (Hammer et al. 1980; Delmas
et al. 1992), and observed atmospheric phenomena
(e.g. Lamb 1970; Scuderi 1990). strongly implicate the
Huaynaputina eruption in forcing the cooling. De
Silva and Zielinski (1998) argue that the 100 Mt of
H
2
SO
4
estimated from the South Pole ice core (Del-
mas et al. 1992) can reasonably be related entirely to
the Huaynaputina eruption, because both the ice core
and the volcano are located in the southern hemi-
sphere. They estimate that the mean northern hemi-
sphere loading for Huaynaputina would be approx-
imately 4210 Mt. These two hemispheric estimates
yield a global average of approximately 70 Mt H
2
SO
4
for the Huaynaputina eruption. While this is only 35%
of the ~200 Mt (global average) estimated for the
1815 eruption of Tambora, the greatest sulfur-pro-
ducing eruption of at least the past ~750 years, these
data nevertheless indicate that the atmospheric pertur-
bation of the Huaynaputina eruption was not only far
greater than any eruption in the twentieth century,
but it also was one of the greatest of historical time.
Deposit stratigraphy and eruption sequence
The 1600 deposits are well exposed in a few accessible
sections on the amphitheater rim and walls and pro-
vide controls on the proximal stratigraphy (Fig. 4).
Table 1 Summary of Events of the 1600 eruption of Huaynaputina as recorded in historic, parochial, and anecdotal accounts from
Arequipa and vicinity. (Adapted from de Silva et al. 2000)
Date Event Stage
15 February Regular earthquakes begin
18 February 9 p.m. Earthquakes increasing in strength
10 p.m. People in Arequipa woken up by strong earthquakes
19 February Earthquake swarm throughout the day increasing in intensity to three or
four movements every 15 min
11 a.m. to
1 p.m.
Two major quakes of intensity 11 on the Mercalli scale felt in Arequipa
(buildings fall)
5 p.m. Tremors so frequent that they felt continuous Stage I
6 p.m. The whole region was dark and dry, lightning was prominent
Explosions every few seconds
Ash began to fall
20 February Major ash fall, explosions, and earthquakes continue
2 p.m. Was like midnight in Arequipa
2025 cm of ash recorded in Arequipa
Intermittent roars and explosions and continued earthquake activity
21 February Clearer, not much ashfall
22 February Renewed ashfall Stage II
23 February Quiet
24 February More ash fall and intense earthquakes
25 February Very cloudy and dark, minimal fall 40 h of darkness
26 February Renewed ash fall
27 February Quiet
28 February Large earthquake felt in Arequipa, renewed ashfall Stage III
11 March Big earthquake, ashfall
Ashfall and earthquakes continue
16 March Ashfall finally stops
12 April Atmosphere finally cleared
498
Observations from widely dispersed deposits from sec-
tions in medial and distal distances (Figs. 5, 6) reveal
that some layers are not represented in the proximal
area, and these allow a more complete stratigraphic
framework for the eruption to be developed.
Stage I: Main plinian phase (1920 February)
On the WSW wall of the amphitheater, 500 m later-
ally and 300 m above the three-vent cluster, the lower-
most deposit of the 1600 eruption is a coarse, well-
sorted pumice fall deposit 1823 m thick (Fig. 4). We
identify five distinct intervals (AE; Fig. 4) through
this section. The base of the 1600 sequence (A) is
defined at the onset of the coarse pumice fall that
overlies a pre-1600 ash with a weak soil developed on
it in some sections. Thouret et al. (1999) have also
identified this as the base of the sequence. Intervals
BE are lithic-rich intervals that may record periods
of increased vent erosion and vent-wall collapse
through the eruption.
From the base, the deposit is reversely graded up
to the 12 m level. From 12 to 17 m is the coarsest part
of the deposit with abundant smashed pumice frag-
ments (up to 60 cm) and exotic lithics as large as
Fig. 4 Ultraproximal section of 1600 deposits on the WSW rim
of the amphitheater located in Fig. 3). The stage-I deposit is a
rare example of a sequence formed due to ballistic (low foun-
taining) emplacement of clasts from the outer collar of the erup-
tion column. Summary density and crystal content data are also
shown. See also Fig. 6A for field context
h
Fig. 5 Stratigraphic sections through the 1600 deposits. Sections
along the dispersal axis are shown from Omate, 12 km from
source, to a locality located 105 km from source. Also shown is
a section through Upper Aguas Blancas ~4 km SE of the amphi-
theater and a section 40 km from the source off the dispersal
axis. At all the medial and distal localities, the main plinian
deposit displays a zone near the middle enriched in coarse clasts
and lithics. This sequence is consistent and can be linked back
to the amphitheater rim deposits (Fig. 4). The index map shows
in bold the relative locations of all the sketched sections
499
500
3.5 m. Two distinct intervals of coarse lithic concentra-
tions, intervals C and D, occur at the 12- and 15-m
levels. These are dominated by lava fragments from
the host volcano and underlying volcanics and sed-
iments. From 17 m to the top, the deposit is normally
graded for both coarse pumice and lithics until the
uppermost 50 cm, where the lithic rich interval E caps
the sequence.
In the lower part of this stage-I deposit (up to
12 m) the dominant lithics are hydrothermally altered
rhyolite and sandstone, many of which are pervasively
altered and stained yellow. Approximately 5 m from
the base is a 2-m-thick zone (interval B) which is par-
ticularly rich in hydrothermally altered lithics. In the
upper part of the deposit, the lithics are dominantly
andesitic lava clasts and there is a general increase in
the lithic content upward.
Fig. 6AD Ultraproximal section and medial fall deposits of the
1600 eruption of Huaynaputina. A Amphitheater rim sequence
shows the disparate deposits which characterize the three differ-
ent stages of the eruption. This is the sequence that is depicted
in Fig. 4. View to the west. Stage-I pumice fall deposit is ~20 m
thick in this view. B A 2.0-m-thick section of the stage-I pumice
fall deposit near the town of Omate, 12 km west of the vent
along the dispersal axis. Note the presence of a middle zone
enriched in coarser clasts. On top of the millet-seed lapilli layer
is the fine-ash couplet of stage II (stn. 47; Fig. 5). C Detail of
the boundary between the stage-I and stage-II deposits near
Calicanto ~10 km south of the vents. Uppermost part of the
stage-I fall deposits is the millet seed layer, overlain by the
distinct couplet of fine ash intercalated with pumice fall. D Typi-
cal medial section of the stage-I fall deposit at ~60 km from the
vent on the dispersal axis (stn. 23; Fig. 5) . Note the coarser
pumices in the middle and the stage-II couplet which is well
developed here. The upper deposit of pumice and ash above the
couplet is reworked. The bottle is 15 cm tall
501
Away from the vent area, the pumice fall deposit
has a westerly dispersal and forms an exponentially
thinning sheet out to the coast approximately 150 km
to the west. The characteristic of an inversely graded
basal part under a normally graded top, reaching a
maximum clast size near the middle, is maintained
throughout the lateral extent of the deposit. The
coarse middle layer occurs ~4065% up through the
stage-I deposit and accounts for ~1035% of the
deposit at various localities, but generally forms a
smaller fraction with distance. In near-medial areas
(Figs. 2, 5, 6B), such as Omate (~12 km along the dis-
persal axis), Upper Aguas Blancas (~4 km SE of the
vents), and Quinistaquillas (~15 km to the south), the
uppermost part of the stage-I pumice fall deposit is a
well-sorted layer that consists of rounded lapilli of
consistent grain size (1.53 mm). This millet-seed
layer can be recognized throughout proximal and
medial exposures (Fig. 6C,D) and is finer grained at
Aguas Blancas and Quinistaquillas than at Omate,
suggesting that it too has a westerly dispersal. In many
localities, the upper surface of the millet-seed layer
has been reworked or eroded, suggesting that a short
period of quiescence followed the main stage-I plinian
eruption. A couplet of normally graded fine-lapilli and
ash layers which belong to stage II of the eruption lies
unconformably on the erosion surface (Fig. 6B).
At medialdistal and distal distances, the pumice
fall deposit is directly overlain by an ash fall deposit
of significant but varied thickness. We think this ash is
the slower-settling fine ash complementary to the
stage-I pumice fall deposits, and we describe it hence-
forth as the plinian ash. Glass shards from this ash
deposit chemically match the glass composition of the
stage-I pumice, suggesting that they erupted concur-
rently. The ash deposit extends regionally throughout
southern and central Peru (Fig. 1), consistent with his-
torical reports of ash fall at the time of the eruption
(de Silva et al. 2000). Furthermore, shards from the
ash chemically match ash particles from ice cores in
Antarctica (de Silva and Zielinski 1998), indicating
significant hemispheric distribution of the plinian ash.
Stage II: Pyroclastic flow phase (2226 February)
A sequence of intercalated pyroclastic flow units and
thin fall layers of limited dispersal overlie the main
plinian deposit of stage I (Figs. 4, 6A). On the amphi-
theater rim the sequence has three thin (<1 m),
brown-colored flow units with lag breccia lenses sep-
arated by lithic-rich lapilli fall deposits (Fig. 7B).
These proximal flow units have a coarse, sugary-tex-
tured matrix of ash, lithics, and crystals in which the
lithic breccias are supported and there are sag struc-
tures under the biggest lithics. The brown color may
derive from fragmented hydrothermally altered mate-
rial in the ash matrix that retains moisture efficiently.
No pumice was found in these proximal flow units.
Overlying this sequence are two more lapilli fall units
containing pumice clasts that are notably denser, more
crystal-rich, and finer-grained than from those of stage
I. These deposits are notably whiter than the stage-I
pumice fall deposit and have vague lenticular layering,
with each layer having subtle inverse grading of the
whole clast assemblage. This suggests that the top of
the deposit was unstable as it accumulated on the
steep slopes and was slumping slightly down slope. It
is thus uncertain what the true thickness of the deposit
should be. At the top of this unit is another zone of
larger lithics and a fine 2- to 3-cm fine ash layer that
accumulated during the hiatus at the end of stage II.
Away from the amphitheater, some of the best-pre-
served sections are found in the upper reaches of
Quebrada Aguas Blancas (Figs. 2, 5). Directly over-
lying the stage-I plinian fall deposit are two additional
thin (510 cm) layers of brown co-ignimbrite ash inter-
calated with fall deposits (Fig. 7A). Above this cou-
plet are three fines-depleted, lithic-rich ignimbrite
flow units separated by thin lapilli fall layers and
brown ash (Figs. 4, 7A). The lowermost unit is proba-
bly secondary because it is dominated by rounded
lithics that we believe are due to milling of the lithics
as dense flows slumped off topographic highs to accu-
mulate in the drainages and valleys. A matrix-support-
ed, inversely graded pumice raft at the top of this flow
unit probably formed during this secondary deposi-
tion. The similarity of the brown ash and the lapilli
layer sequence to the amphitheatre wall proximal flow
units and lapilli layer sequence, and the similar
number of breccia beds to flow units, allows correla-
tion of the Aguas Blancas section with the proximal
sequence, and places the pyroclastic flow deposits as
stage-II products. The intercalation of these fall and
flow deposits suggests that the latter are at least partly
intraplinian, and the ash beds indicate short hiatuses
in activity.
Pyroclastic flow deposits are concentrated in steep
local valleys, particular to the east and southeast of
the volcano up to 20 km from source (Fig. 2). Con-
trary to the assertions of Thouret et al. (1999), we are
not aware of any primary flow deposits more distal
than those indicated in Fig. 2. Major (>2500 m
elevation) topographic barriers bordering the Rio
Tambo and Omate valleys appear to have limited the
distribution of the flows. Historical accounts describe
that the Rio Tambo was dammed during the eruption
and that major flooding resulted from breached dams.
At least three locations along the Rio Tambo show
evidence for breached dams of ignimbrite, the largest
at the confluence with the Quebrada del Volcn near
Quinistaquillas (Fig. 2).
At medial distances, outside the local drainage net-
work, no pyroclastic flow deposits are found, and the
only stage-II deposit that persists is the distinct couplet
of normally graded, fine-lapilli, and ash fall that uncon-
formably overlie the main plinian deposit of stage I.
This couplet is prominent around Quinistaquillas,
502
Omate, and out to ~50 km west and southwest from
source (Fig. 6BD). It ranges in thickness from ~1 to
10 cm, thinning with increasing distance from the vents.
We believe the couplet is equivalent in time to the low-
ermost beds of stage II seen at Upper Aguas Blancas.
In the Quebrada del Volcn and the Pampa de
Calicanto (Fig. 2), near Quinistaquillas, three to six
ignimbrites are exposed in different sections. The
three thickest (up to 3 m) are similar to the three
exposed at Upper Aguas Blancas, but the others are
Fig. 7AD Details of the stage-II and stage-III deposits of the
1600 eruption. A Section through stage-II sequence in Upper
Aguas Blancas (UAB in Fig. 2) shows two of the main ignim-
brite deposits which were deposited during this stage (Fig. 5).
Also shown are the co-ignimbrite ash and lapilli fall couplets. B
Detail of the lower ultraproximal stage-II proximal ignimbrites
with lag breccias; these are intercalated with lithic-rich pumice
fall layers. It is believed that these can be correlated with the
deposits at Upper Aguas Blancas (see A). See text for further
explanation. C Fine-grained pyroclastic flow deposits in Lower
Aguas Blancas; these were deposits from the dilute feather
edges of pyroclastic flows travelling down (toward the viewer)
the Aguas Blancas valley. They contain abundant accretionary
lapilli (e.g., top left) dispersed throughout the deposit. Interbed-
ded layers of fall (observer is pointing at one) and outsized
pumice clasts (arrows) suggest that the flows were intraplinian
in nature. D Partial section (40 cm) through the crudely bedded
coarse tephra fall of stage III. The juvenile material is all dense
and poorly vesicular, crystal-rich dacite very similar to the dikes
and dacite of the ejecta-ring of vent 3
503
typically lithic-poor and more pumice and fines-rich.
The lithic-rich units are massive with no clear grading
or bedding and are again dominated by rounded
lithics, consistent with re-mobilization or sloughing
down from the steep slopes surrounding the valley.
The pumice-rich units are also massive, but show
reverse coarse-tail grading with several pumice rafts
toward the top of the units. To reach these points,
pyroclastic flows had to travel over extremely steep
topography, the vertical drop of almost 1000 m at the
head of the Quebrada del Volcn being the most
extreme. This would have contributed to significant
loss of fines as co-ignimbrite ash. Cascading down the
steep slopes would likely cause ingestion into the
flows of copious amounts of air, resulting in severe
fluidization and elutriation of fines. The flow deposits
in this area are also fines-depleted as a result.
The most distal exposures of the pyroclastic flow
deposits are found in the lower reaches of the drain-
ages surrounding Huaynaputina as they meet up with
the Rio Tambo. At Aguas Blancas, pyroclastic flow
deposits form a thick valley-fill sequence, along with
several reworked layers of fall and older debris flows
from Cerro El Volcn. Rare exposures of fine-grained
pyroclastic flow deposits with abundant accretionary
lapilli distributed throughout most of the deposit are
found on the flanks of the valley (Fig. 7C). Crude
bedding parallel to the slopes and interbedded ephem-
eral fall layers are found. These layers can be traced
into normal pumice-rich and lithic-rich ignimbrite in
the valley. We interpret these as the deposits of the
feather edges of pyroclastic flows traveling down the
Aguas Blancas valley. Steam generated by the heated
waters in the Aguas Blancas and the Rio Tambo
played a role in the formation of the accretionary lap-
illi. We estimate a total of ~2 km
3
of flow deposits
with an unknown volume of co-ignimbrite ash, yield-
ing a DRE volume of >1 km
3
.
We have found no surge deposits associated with
this eruption. Rare exposures of thin, primary flow
deposits with poorly developed, unimodal, low-angle
cross-stratification and minor up-flow scours are the
only evidence of turbulent flow. This is surprising con-
sidering the steep, rough topography that the flows
traveled over. We have not examined flow deposits at
the confluence of Q Huaynaputina and the Rio Tam-
bo. This is the most likely place that any surge-type
flows might be generated as pyroclastic flows inter-
acted with dammed rivaer Tambo waters. In medial to
distal locations (4080 km from source), thin centi-
meter-scale, crystal-rich units with very-low-angle,
large-wavelength cross-stratification are intercalated
with beds of wind-reworked ash; the former never
contain up-flow scours or antidunes, typically pinch
and swell, are commonly ephemeral lenses, are associ-
ated with channels, and contain abundant quartz
(some aeolian) all factors indicative of post-emplace-
ment reworking. These have been interpreted as crys-
tal-rich flow and surge units by Thouret et al. (1999).
At some distal locations, a fine brown ash layer
overlies the plinian ash of stage I. The distinct color
and the high Si-rhyolite (SiO
2
~76%) composition of
the shards confirms that this is distal co-ignimbrite ash
and allows us to differentiate it from the plinian ash
and identify it in distal exposures. This deposit is very
discontinuous and poorly preserved.
Stage III: Vulcanian phase (28 February to 6 March)
A sequence of thin (1 cm to decimeter) but crudely
bedded, well-sorted, fines-poor deposits containing
poorly vesicular, crystal-rich dacite overlies the
stage-II deposits in the proximal amphitheater rim sec-
tion (Figs. 4, 6A, 7D). Outside the amphitheater, the
fall deposits have a more restricted dispersal area than
the stage-I fall deposit and are distributed primarily to
the WNW. In medial locations as far as 40 km west of
the vent cluster, well beyond stage-II pyroclastic flows,
a massive coarse ash layer that overlies the stage-I
deposits can be correlated with these dense juvenile
fall deposits on the basis of continuity, composition,
and mineralogy. In the amphitheater wall sections, the
base of the dense juvenile fall sequence has several
thin ash layers that we think are phreatomagmatic
based on their extremely fine-grained nature and small
volume. The upper part of the sequence and the
present surface of the deposit contains abundant large
ballistic blocks and bombs. One such bomb, approx-
imately 10 m below the rim of the amphitheater
directly south of the vents, measures 622 m and
shows clastogenic, autobrecciated, and flow-banded
textures typical of those commonly found in lava
domes. Many of the smaller bombs are non- to poorly
vesicular, have bread-crusted surfaces, and show pris-
matic jointing and flow banding. Coarse-ash and lap-
illi-covered accumulations of ballistically emplaced
lithic blocks and the stage-III dacite form elongate
mounds that form a pattern concentric to the amphi-
theater on the outer flanks of the pre-1600 edifice out
to 400 m from the amphitheater rim. Alongside these
mounds are abundant impact craters. The association
of thinly bedded fall deposits of dense, juvenile ejecta
(lapilli and ash) with abundant ballistic bombs and
blocks of similar dacite, the close similarity of this
dacite to dome material in the vents, and the rel-
atively limited dispersal of the fall units lead us to the
conclusion that these deposits were produced by a
series of closely spaced vulcanian-type eruptions. We
estimate ~1.5 km
3
of tephra and a DRE of ~1 km
3
for
the stage-III deposits. We discuss stage-III deposits
around the vents in a later section of this paper.
504
Petrology of the 1600 juvenile products
The juvenile products of the 1600 eruption are pre-
dominantly high-K dacites. In stage I the dacitic
pumice define a range of 64.4 to 65.8% SiO
2
(Table 2)
and contain 1720% crystals (vesicle-free basis) of
plagioclase+hornblende+biotite+FeTi oxides+apatite.
The density of individual fragments of pumice from
~1 to 8 cm in length ranges from 1.0 to 1.3 g/cm
3
,
whereas bulk densities of the stage-I plinian lapilli fall
deposit range from 0.80 to 1.03 g/cm
3
. The glass in the
pumice is rhyolitic (~73% SiO
2
; Table 2). Pumice
from the ignimbrites and fall deposits of stage II gen-
erally overlap the compositions of pumice from the
stage-I plinian fall deposit but extend to higher silica
values (Table 2). However, interstitial glass in the
stage-II juvenile clasts is a high-Si rhyolite with ~76%
SiO
2
and is therefore significantly more evolved than
the 73% SiO
2
of the interstitial glass in stage-I
pumices (Arbesman and de Silva 1998). The pheno-
cryst assemblage in stage-II juveniles is plagioclase+
hornblende+biotite+FeTi oxides+apatite, the same as
stage I but with crystal contents that are much higher
2530% (vesicle-free basis). The stage-II juvenile com-
ponents are also notably more microlite-rich than
stage I. The density of the stage-II pumice is also
higher and ranges from 1.21.8 g/cm
3
, whereas non-ex-
panded juvenile blocks have densities of 2.02.2 g/cm
3
.
The composition of the stage-III dacites overlaps stage
I and stage II but extends to more SiO
2
-rich composi-
tions (Table 2). The stage-III dacite contains the same
mineral assemblage and evolved glass composition as
the stage-II juvenile blocks but has generally higher
crystal contents (up to 40%) and contains abundant
microlites in the groundmass. Densities of individual
blocks range from 1.9 to 2.2 g/cm
3
.
We note that while Thouret et al. (1999) indicate
that stage-I juvenile material is distinct from that of
later stages in their whole-rock major element data,
our data show considerable overlap in compositions.
Physical characteristics of the stage-I plinian deposits
Distribution and volume
Two distinct but complementary deposits were
produced from the same eruption column during the
main plinian stage I of the 1600 eruption: (a) a pumice
fall deposit that forms a continuous, exponentially
thinning sheet; and (b) a deposit of fine ash that
widely overlies the pumice fall in distal locations but
extends far beyond it. De Silva and Zielinski (1998)
estimated a total tephra volume for the plinian stage I
of ~19 km
3
by applying an exponentially thinning law
to the combined pumice and fine-ash deposits. Thou-
ret et al. (1999) appear to have used a similar treat-
ment; however, our new data indicate that the plinian
pumice fall deposit and the plinian ash deposit have
different distributions and thinning characteristics and
hence cannot be considered as one exponentially thin-
ning deposit. We calculate here the separate volumes
represented in each of these deposits to arrive at the
total volume erupted during stage I. This treatment is
consistent with the assertion of Bonnadonna et al.
(1998) that in eruptions where significant volumes of
fine ash are produced (as in this eruption), the expo-
nential thinning law (cf. Pyle 1989; Fierstein and
Nathenson 1992) will result in gross underestimation
of the volume erupted if the volume of fine ash is not
taken into account.
Thickness and volume of the pumice fall deposit
of stage I
Isopachs for the pumice fall deposit (Fig. 8A) define a
westerly dispersal with a large increase in the spacing
of the isopachs between the 20- and 15-cm isopachs.
To the WSW, the 2-m isopach extends 12 km, the
20 cm extends 60 km, and the 5 cm extends ~140 km.
The deposit thins most abruptly to the east, from 1 m
Table 2 Summary petrological characteristics of 1600 Huaynaputina juvenile clasts
Stage I Stage II Stage III
SiO
2
a
64.465.8 n=21 64.065.7 n=12 64.566.2 n=10
MgO
a
1.91.7 2.11.7 1.71.8
Glass
b
72.80.7 n=8 (103) 75.80.4 n=4(52) 75.80.3 n=4(46)
Plagioclase
c
An
27
Ab
69
Or
4
An
59
Ab
40
Or
1
An
41
Ab
57
Or
2
An
55
Ab
44
Or
1
An
30
Ab
66
Or
4
An
57
Ab
41
Or
2
Biotite Mg#
d
58.573.1 60.669.5 52.362.6
Hornblende Mg#
e
57.772.7 57.769.8 56.166.3
Magmatic temperatures (C)
f
83922 77323 and 84722
f
77725
H
2
OAb exchange
g
3.25.3% 3.03.9% 3.14.0%
H
2
OAn exchange
g
2.95.0% 1.72.4% 1.92.6%
a
XRF analyses of bulk juvenile clasts on an anhydrous basis; n
no. of samples
b
Microprobe analyses on an anhydrous basis; n samples (no. of
analyses)
c
Range of microprobe analyses of phenocrysts
d
Range of microprobe analyses; Mg#=ionic(Mg/Mg+Fe+Mn)
e
Range of microprobe analyses; Mg#=ionic(Mg/Mg+Fe+Mn)
f
Average of temperatures obtained from co-existing FeTi
oxides using QUILF of Andersen et al. (1993)
g
Average of two distinct sets of co-existing FeTi oxides
h
Based on rim compositions of plagioclases and residual glass
compositions using method of Housh and Luhr (1991)
505
Fig. 8 A Isopach map of the
Huaynaputina stage-I pumice
fall deposit. B Comparison of
thickness data for 1600 Huay-
naputina plinian tephra fall
of Thouret et al. (1999) with
our data from A. Also shown
are our thickness measure-
ments of the plinian ash
deposit; these represent the
sections where we were con-
fident that the deposit rep-
resented at least a minimum
primary thickness. The pumice
fall deposit directly underlies
the plinian ash deposit at most
of these locations
506
to 1 cm in ~10 km. This rapid thinning reflects the
strong influence of dry-season easterlies that produced
the strong westerly dispersal. The re-entrants in the
distal isopach segments near the coast could reflect
the influence of westerly coastal winds and turbulence
coming off high topography on the finer clasts. It
should be noted that the measurements made are
minimum thicknesses owing to common reworking on
the top of the main fall unit.
Our thickness data are in marked contrast to those
of Thouret et al. (1999), whose data show the 1-cm
isopach extending some 50 km to the east of the vents
and up to 5 cm of 1600 Huaynaputina deposits around
and to the east of Ticsani (Fig. 8B). On the basis of
chemistry, we have established that the pumice fall
around and to the east of Ticsani is not the 1600
Huaynaputina pumice fall deposit, but that from Tic-
sani. We have not found the 1600 pumice fall this far
east (Fig. 8A). Furthermore, the thickness data of
Thouret et al. (1999) are consistently much greater
than ours throughout medial and distal locations. We
note that their data are for plinian tephra and we
surmise that they must have used total tephra (pumice
and ash?) thicknesses in their isopach map. We
believe this approach to be untenable because each
subunit has different distribution and thinning charac-
teristics and therefore results using such an approach
should be treated with caution.
The square root of the area enclosed by each iso-
pach is plotted vs thickness on a log scale (Fig. 9) fol-
lowing Pyle (1989). The exponential decay of the data
can be represented by three discrete straight-line seg-
ments, with the distal segment extrapolated to infinity.
Using the method of Fierstein and Nathenson (1992),
integration under the three line segments, with an
adjustment for the volume calculated from the final
mapped isopach (1 cm) to infinity following Pyle
(1995), yields a minimum mapped volume of ~7 km
3
for the stage-I pumice fall deposit.
The straight-line segments in Fig. 9 are similar to
those seen in other eruptions like the 1932 eruption of
Quizapu, Chile (Hildreth and Drake 1992), and the
1912 eruption at Novarupta, Alaska (Fierstein and
Hildreth 1992). At Quizapu, the first segment is attrib-
uted in part to a ballistic component, the second to
deposition of pumice from the convecting column, and
the third to deposition of fine ash and lapilli from the
umbrella cloud. Hildreth and Drake (1992) also note
that, whereas it is not the case at Quizapu, the third
segment is often attributed to contribution of co-ig-
nimbrite ash to distal fallout. At Novarupta, it has
been suggested by Fierstein et al. (1997) that the first
segment results from a collar of low-fountaining ejecta
that sheaths the high-velocity core of an eruption col-
umn. Like Novarupta and Quizapu, at Huaynaputina
the first segment is the proximal ejecta ring that
resulted from ejecta of diminished velocities sloughing
off the sides of a high-velocity, stable-cored column.
This ejecta ring has been preserved at Huaynaputina
because there was no caldera collapse subsequent to
the eruption. The second segment for the Huaynapu-
tina data represents fallout of coarse pumice from the
convecting column, and the third segment results from
deposition of fine lapilli and ash from the umbrella
cloud. Note that the overlying plinian ash layer is not
included above. We later suggest that its transport and
deposition was decoupled from that of the pumice fall.
Bulk densities of the pumice fall deposit measured
in the laboratory range from 0.8 to 1.03 g/cm
3
(Ta-
ble 3). Because the pumice is dacitic, a density of
2.4 g/cm
3
was used to calculate the volume DRE of
2.6 km
3
.
Thickness and volume of the plinian ash
Directly on top of the pumice fall deposit in far-me-
dial through distal locations, and extending well bey-
ond the pumice fall deposit, is the fine-ash deposit
from the stage-I plinian eruption. We think this
deposit was produced by the slow-settling fines that
remained aloft longer than the pumice. Implied in this
interpretation is that the stage-I umbrella cloud was
strongly stratified according to particle Reynolds
number (cf. Bonnadonna et al. 1998) with fine ash of
high Reynolds number being elutriated into the upper-
most portions. These fines probably remained sus-
pended for at least several hours after the end of the
stage-I eruption but had to have deposited regionally
prior to the deposition of co-ignimbrite ash from stage
II a few days later. The more widely distributed ash
may have remained aloft for even longer in an exten-
sive umbrella cloud that was subsequently dispersed
by stratospheric winds as far south as Antarctica and
at least 1000 km to the west. This does not imply
abnormal behavior of the fines in this eruption, as
Fig. 9 Plot of thickness (log scale) vs square root of isopach
area (see Fig. 8A) for the stage-I pumice fall deposit of the 1600
eruption. Pumice fall volume of 7 km
3
is calculated by integra-
tion under the three line segments with an adjustment from the
last mapped isopach to zero thickness (Fierstein and Nathenson
1992; Pyle 1995). The two inflection points are thought to rep-
resent (a) the outer limit of low-fountaining ejecta preserved on
the amphitheater rim, and (b) onset of deposition of fine lapilli
and ash from the umbrella cloud
507
decoupling of fines (and the resultant deposit) from
the coarser components is consistent with the decoup-
led fallout paths and independent dispersal expected
of fine particles in plinian eruption clouds (e.g., Hild-
reth and Drake 1992; Fierstein and Hildreth 1992;
Bonnadonna et al. 1998). Instead, the excellent preser-
vation of the deposits from the 1600 eruption may be
revealing rarely preserved evidence of typical plinian
behavior (cf. Fierstein and Hildreth 1992).
Historical accounts indicate significant ash fall as
far north as Lima (~1000 km) and as far south as
Arica in Chile (Fig. 1). Thus, ash fell over an area of
more than 300,000 km
2
. The presence of ash particles
in the South Pole ice core in Antarctica (de Silva and
Zielinski 1998) and reports of ash fall on a ship
1000 km off the Peru coast (de Silva et al. 2000) indi-
cate very widely distributed ash fall. Constraining the
volume of such ash is fraught with problems, however.
For instance, present-day thicknesses are considerably
less than those reported at the time of eruption, the
fine-ash deposit is extensively reworked, and primary
thicknesses in most of the area where ash is known to
have fallen are difficult to find. Moreover, the pres-
ence of co-ignimbrite ash in some distal locations
requires that care be taken to identify the plinian ash;
however, we believe the distinctive brown color and
the evolved glass shard composition (~76% SiO
2
) of
the co-ignimbrite ash compared with that of the glass
from plinian pumice (73% SiO
2
) allow us to distin-
guish the two. Another potential pitfall is the presence
of aeolian sand in reworked exposures of the plinian
ash deposit along the coast and as far inland as Are-
quipa. The aeolian component was relatively easily
identified by the presence of frosted and rounded
quartz. Despite these pervasive difficulties, 25 loca-
tions with primary deposits of plinian ash were identi-
fied and thicknesses were measured (Fig. 8B). Note
that, whereas there is a range of thicknesses, the ash is
considerably thicker than the pumice fall at most loca-
tions. These data suggest that the plinian ash has a
thickness maximum ~100150 km from the vents. This
thickness maximum may represent premature flushing
out of ash shards that flocculated (e.g., Sarna-Wojicki
et al. 1981); otherwise, the shards might have been
transported further afield and distributed more evenly.
There are many localities where over 1 m of ash is
found. Although these thicknesses are clearly not pri-
mary, they are consistent with historical accounts of
over 1 m of ash in the areas surrounding Arequipa
soon after the eruption. For instance, in Camana, on
the coast directly west of the volcano, a palm tree is
said to have collapsed under the weight of ash (cf. de
Table 3 Representative physical data for the 1600 Huaynaputina plinian fall deposit
Sample Approximate
distance from
vent (km)
Azimuth
direction
from vent
Md
(f)
Sorting
(s
f
)
Wt.%
<1 mm
Wt.%
<63 m
Bulk
density
(g/cm
3
)
Wt.%
lithics
Wt.%
crystals
HY-044A 117 255 0.25 1.75 144.47 10.30 2.53 11.49
HY-042A 124 260 10.50 1.35 161.15 10.20 0.89 6.54 44.41
HY-005 125 320 10.60 0.90 173.15 10.05 2.53 11.49
HY-041A 126 258 10.85 1.33 166.24 10.50 1.03 7.49 49.74
HY-101A 128 188 0.90 1.58 137.38 10.31 8.14 22.50
HY-040 132 260 10.90 1.00 177.60 10.29 0.85 3.56 58.91
HY-034 133 257 10.90 1.05 179.03 10.07 2.03 31.42
HY-097 134 200 10.15 1.20 153.35 10.04 5.23 37.31
HY-007 135 310 10.90 0.90 181.93 10.51 0.89 3.20 41.71
HY-098 137 209 10.30 1.35 155.51 10.49 3.61 35.41
HY-031 137 249 10.80 1.00 175.77 10.03 2.37 50.47
HY-029 143 265 11.20 0.90 189.43 10.25 0.92 2.90 63.16
HY-019 146 244 11.10 0.85 188.32 10.12 3.36 46.65
HY-027 147 258 11.10 0.90 189.46 10.07 3.06 59.41
HY-012A 153 269 11.68 0.75 197.65 11.09 3.51 50.98
HY-015 155 259 11.45 0.70 197.44 10.38 0.84 2.96 62.67
HY-011 163 275 11.70 0.75 198.56 10.30 3.04 62.04
HY-081A 195 276 11.50 0.83 199.84 10.25 1.47 39.43
HY-088 195 245 11.80 0.53 199.75 10.59 0.57 74.94
HY-116 196 245 11.70 0.70 199.73 10.03 0.93 1.56 56.78
HY-114A 197 250 11.70 0.70 199.83 10.22 0.18 46.08
HY-120 198 277 12.00 0.50 199.43 10.22 2.37 56.29
HY-113 199 248 11.70 0.55 199.80 10.19 0.80 0.88 61.41
HY-082A 100 275 12.52 0.81 199.91 19.15 0.45 44.20
HY-111 107 250 12.10 0.68 199.79 13.61 0.85 76.51
HY-083 110 269 12.50 0.60 100.00 13.93 0.82 0.19 48.63
HY-117 111 250 11.85 0.50 199.77 11.21 1.50 49.75
HY-108 115 300 12.70 0.90 199.87 14.04 0.14 68.31
The 28 samples represent the whole thickness of the fall deposit
at that location. The samples were hand sieved and separated
into different components to establish weight percents. Bulk
density was determined in the laboratory after drying by tapping
compaction of tephra in a container of known volume
508
Silva et al. 2000, and references therein). Given the
strong westerly dispersal of the plinian deposits, a sig-
nificant volume of fine ash was also transported off-
shore and deposited in the ocean consistent with the
report from a ship 1000 km off the coast (de Silva et
al. 2000). Data for ash thicknesses preserved in drill
cores from the Pacific Ocean are a matter of ongoing
investigation.
Recent work on ice caps in the Central Andes also
constrains the thickness of plinian ash in far-field loca-
tions. At Quelcaya, Peru (~300 km to the north), and
at Cerro Sajama, Bolivia (~300 km to the southeast),
drill cores through permanent ice caps contain primary
thicknesses of 8 and 11 cm of 1600 ash, respectively
(Thompson et al. 1986; Thompson et al. 1998; M.E.
Davis, pers. commun.). The composition of the ash in
the cores is identical to the interstitial glass in stage-I
pumice and plinian glass shards with ~73% SiO
2
.
These disparate data sources help us to constrain at
least a minimum volume of plinian ash. We suggest
that at least 10 cm of plinian ash was deposited in a
roughly semi-circular area with a radius of 300 km
west of a line extending from Quelcaya in the NW to
Sajama in the SE. We then assume a 1-cm isopach
area ten times as large as the 10-cm isopach. This is
reasonable in light of the fact that the 1-cm isopach
for the plinian lapilli fall has an area ~15 times that of
the 10-cm isopach (Fig. 8A). Finally, expecting that
the deposit thins exponentially, the volume of the ash
is estimated using the single-line segment equation of
Pyle (1995) and adjusting for the extrapolation to
infinity. This treatment yields ~19 km
3
of ash. Aver-
age bulk density of the ash deposit we have measured
is 0.8 g/cm
3
, which yields a minimum DRE volume of
~6.2 km
3
of plinian ash.
Total volume erupted during the stage-I plinian
eruption
Combining the estimate of 19 km
3
for the fine ash
with the 7-km
3
estimate for the pumice fall gives a
total tephra volume of 26 km
3
of stage-I deposits and
a DRE of 8.8 km
3
. Note that de Silva and Zielinski
(1998) and Thouret et al. (1999) used a density of
1.2 g/cm
3
to calculate the DRE in their work. In the
former case, this density was obtained from indurated
and compacted ash and therefore is not fully repre-
sentative. We note that 1.2 g/cm
3
is the density of
individual pumice blocks and therefore unlikely to be
representative of the bulk fall deposits. Our average
densities of 0.89 g/cm
3
for the pumice fall and 0.8 g/
cm
3
for the ash were determined in the laboratory
(Table 3), and we use these in our DRE estimates.
The new minimum estimate of 8.8 km
3
DRE is there-
fore much better constrained and confirms that this
eruption is the largest historical volcanic eruption in
South America and ranks as one of the world's largest
historical eruptions.
Granulometry and components
Clast size distribution
The maximum clast size (maximum pumice, MP; max-
imum lithic, ML) found at each location was deter-
mined in the field by measuring and averaging the
long axes of three to five clasts found in the deposit.
The deposits preserved along the amphitheater rim
contain pumice and lithic blocks >1 m across (Fig. 4).
These are best seen in the WSW sector along the
amphitheater walls. Pumice blocks as big as 15 cm fell
11 km along the dispersal axis to the WSW but only
as far as 2 km NE of the amphitheater (Fig. 10A).
Likewise, 1 cm pumice lapilli are found 40 km along
the dispersal axis and only 13 km NE. Similarly, lithic
clasts as big as 1 cm are found ~18 km WSW from the
amphitheater, ~12 km NE, and 25 km S (Fig. 10B).
The isopleths defining the distribution of ML appear
more concentric than isopachs or MP isopleths; how-
ever, although the ML isopleth data is more evenly
distributed, there is still a slight elongation south and
west of the volcano.
Following Hildreth and Drake (1992), areas
enclosed by the isopleths and isopachs are compared
with the data from similar plinian eruptions (Fig. 11).
The dispersive power of the different historical erup-
tions can be compared in each panel. No clear rank
for each eruption is established by evaluating these
three criteria of dispersive power, but the diagram
does illustrate the significance of Huaynaputina com-
pared with well-known historical eruptions. For Quiza-
pu, Hildreth and Drake (1992) observed that inflec-
tions between straight line segments on their plot of
the square root of area vs thickness on a log scale
(e.g., Fig. 9) could also be recognized on a loglog
plot of area vs thickness like in Fig. 11. At Huaynapu-
tina, the inflection between the 10- and 20-cm iso-
pachs seen in Fig. 9 is also found on the loglog plot
of area vs thickness (Fig. 11). This signals the change
from deposition from the convecting column to depo-
sition from the umbrella cloud.
Componentry and sorting
Pumice of the stage-I pumice fall deposit is typically
white and occurs as subangular to subrounded clasts
with medium vesicularity. Rare dark-gray pumice with
poor vesicularity and subrounded clasts constitutes
<<1 wt.% of the deposit. Plagioclase makes up most
of the crystal population, with accessory amphibole,
biotite, and opaques. Lava and sandstones consistently
constitute the majority of accidental clasts in the
deposit with the latter dominating the lower half and
the former the upper half of the deposit.
Lithic content of the deposit varies, but in general,
the weight percent of lithics decreases with distance
509
Fig. 10 Proximal isopleths of
the main plinian fall deposit at
A Huaynaputina for average
maximum pumice (MP), and
B for average maximum lithics
(ML)
510
from the vents. This decrease is steepest NW of the
amphitheater. Based on visual estimates, lithics exceed
10 wt.% of the fallout within 110 km of the vents.
Componentry indicates that as far as 28 km from the
vents at an azimuth of 188, there is up to 8 wt.%
lithics in the deposit, and at 34 km at an azimuth of
200, lithics constitute ~5 wt.% (Table 3). Beyond
34 km in all directions, the weight percent of lithics
continuously decreases. By ~95 km downwind, the
lithic content is ^1 wt.%. It is important to note,
however, that sparse lithics are present at very distal
localities. The quantity of such clasts is difficult to
determine because lithics smaller than 2f (0.25 mm)
are extremely difficult to distinguish from other dense
constituents in this deposit. Based on the fallout vol-
umes calculated above and our componentry, lithics
comprise ~0.177 km
3
or ~2.5% of the pumice fall.
This is ~0.7% of the total volume of stage I. Lithics
are very sparse in the ash fall.
Grain-size data, componentry, and bulk density
were determined for 28 samples chosen to provide
complete geographic coverage of the pumice fall
deposit (Table 3). Histograms showing grainsize dis-
tribution, componentry, and sorting for 6 of the 28
sieved samples (Fig. 12) exhibit a unimodal grainsize
distribution, characteristic of a fall deposit. The hig-
hest proportion of crystals appears to have a size
range limited between 1 and 0.125 mm (0 and 3f).
When the Md
f
for the Huaynaputina data is plotted
against s
f
, the clustering of points is characteristic of
a well sorted, pyroclastic fall deposit. The sorting (s
f
)
ranges between 0 to 2f (10.25 mm) with the majority
falling close to 1f (0.5 mm), and the median diameter
(Md
f
) ranges between 1 and 3f (20.125 mm).
Along the dispersal axis, the weight percent of crys-
tals increases relative to the weight percent of pumice
as the distance from the vents increases, e.g., at
~20-km the crystals in HY-044A constitute ~11 wt.%
of the deposit, whereas at ~100 km in HY-113 they
compose ~60 wt.%. As expected, clast size decreases
and the deposit becomes better sorted away from the
vent cluster along the dispersal axis. Within 30 km of
the vents, s
f
>1, whereas >100 km from the vent
cluster, s
f
~0.65.
Off the dispersal axis, crystal content and sorting
varies greatly. At 25 km north of the vents, HY-005
consists of only ~11.5 wt.% crystals, whereas only
35 km south, HY-097 consists of ~40 wt.%. Due west,
at 24 km, HY-042A is characterized by ~44 wt.% crys-
tals and at 110 km HY-082 has ~49 wt.% crystals.
Thus, along the WSW dispersal axis an increase of
~50 wt.% in crystals occurs from 20 to ~100 km from
the vent cluster, whereas off the dispersal axis due
west from 24 to 100 km from the vent cluster the dif-
ference in crystal content is only 5 wt.%. While sort-
ing off the dispersal axis is generally poorer and shows
greater variability compared with along the axis, the
same trend of improved sorting with distance from the
vent cluster is seen. The graphical standard deviation
(s
f
) for most samples off-axis varies between 0.5 and
1.6. For instance, HY-120, ~100 km west of the vents,
has s
f
=0.5, indicating a very well-sorted deposit,
whereas HY-101 A, ~30 km south of the vents, has
s
f
=1.58, one of the least sorted samples of this
deposit.
Approximately 89 wt.% of the pumice fall deposit
at Huaynaputina is finer than 1 mm. This translates to
97% of the total volume of stage I. Unfortunately, the
estimation of the portion finer than 63 m in the
pumice fall deposit was difficult as only a very small
amount of material finer than 63 m was found during
sieving. Our best estimate is that ~2% of the pumice
fall deposit is finer than 63 m. Given that the fine
plinian ash deposit is almost completely finer than
63 m, this computes out that 67% of the material
erupted during stage I was finer than 63 m. The large
Fig. 11 A comparison of the
dispersive power of several
plinian eruptions. Areas in
square kilometers enclosed by
thickness isopachs and iso-
pleths of maximum pumice
and maximum lithic sizes, all
in millimeters; diagram mod-
ified from Hildreth and Drake
(1992) to include Huaynapu-
tina data (bold)
511
percentage of fines, as well as the increased weight
percent of crystals in medial and distal samples, sug-
gest a high degree of fragmentation and efficient sort-
ing within the umbrella cloud. Most of the material
finer than 63 m appears to have been elutriated into
the upper portions of the umbrella cloud and even-
tually deposited to produce the voluminous plinian
ash. Componentry in the finer-than-63-m fraction
was determined using a petrographic microscope
because the binocular microscope proved unsatisfac-
tory.
Eruptive parameters of the 1600 stage-I plinian
eruption
Column dynamics, duration, and rates of thinning
To describe the rate at which a fall deposit thins, Pyle
(1989) proposed the thickness half-distance parameter
b
t
, the distance over which the thickness halves. Based
on the isopach map (Fig. 8A) and the plot of area vs
thickness on a log scale (Fig. 9), b
t
for the Huaynapu-
tina stage-I pumice fall deposit was estimated for the
Fig. 12 Histograms illustrate
grain size and componentry
(wt.%) for selected samples
(Table 3) on and off the dis-
persal axis. Azimuth, distance
from the crater, deposit thick-
ness, and the Inman param-
eters median diameter (Md
f
)
and sorting (s
f
) are shown for
each sample. Lithics are
shown in black, free crystals
stippled, and pumice white.
Grain size is measured in phi
f; each size class is denoted
by the mesh size that retains
it. Index map shows the rel-
ative location of the samples.
Md
f
vs s
f
plot for the samples
in Table 1. Ninety-nine per-
cent of sieve analyses for fall
deposits occur within the solid
line 1 (after G.P.L. Walker
1971)
512
proximal deposit as 1 km, for the medial deposit as
6.5 km, and for distal portions of the deposit as
50.5 km (Table 4).
Similar to the parameter b
t
, Pyle also proposed a
parameter to describe the rate at which clast size
decreases away from a source. The maximum clast
half distance, b
c
, is defined as the average distance
across which the maximum clast-size halves. Based on
the isopleth maps for ML and MP (Fig. 11) and a plot
of the area enclosed by the isopleth vs clast size on a
log scale (Fig. 13), b
c
for MP is 8.3 km and b
c
for ML
is 4.0 km (Table 4).
According to Pyle (1989), the rate at which clast
sizes decrease (i.e., the slopes of the straight line seg-
ments that best fit the data in Fig. 13) can be used to
estimate H
B
, or the neutral buoyancy height of the
eruption column. The relationship between H
T
, or the
total column height, and H
B
can then be approx-
imated by H
B
/H
T
~0.7. Based on these relationships
and using the slopes derived from our MP- and
ML-h A plots, H
B
is 32 and 24 km, and H
T
is ~46 km
and ~34 km, respectively.
Mass discharge and volumetric discharge rates can
be calculated if the duration of the eruption is known.
Historical accounts indicate that the stage-I eruption
lasted anywhere from ~20 to 24 h (see above and
Table 1). Most of these accounts are from people in
Arequipa, 70 km away from the volcano, and are
descriptions of felt earthquakes and ash fall. Given
the difficulty of pinpointing the cessation of activity
using such accounts, we also tried to constrain the
duration using known accumulation rates of plinian
fall deposits. Wilson and Hildreth (1997) compiled a
database of accumulation rates for several plinian fall
deposits based on column heights. Using their data
and our estimation for H
T
of between 3446 km, the
accumulation rate for 1600 stage-I pumice fallout at
Omate (~12 km along the dispersal axis) was
0.030.035 mm/s. This accumulation rate yields a cal-
culated duration of the eruption of 1619 h. Given
that thicknesses we have measured are minimum
thicknesses, this calculated duration must be consid-
ered a minimum. Using the estimated volumes of
26 km
3
of tephra and 8.8 km
3
DRE, a 20-h duration
yields a mass discharge rate (MDR) of 2.810
8
kg/s
and a volumetric discharge rate (VDR) of
3.610
5
m
3
/s for stage-I plinian eruption. Alternatively,
if the temperature of the erupting mixture can be esti-
mated, this can be combined with H
T
to estimate
MDR and VDR using the method of Sparks (1986).
Magmatic temperature estimates based on FeTi
oxide equilibria define a range of 780815 C for the
stage-I pumice (Table 2) and these are a reasonable
proxy for the temperature of the erupting mixture.
The temperatures combined with a column height
(H
T
) of 3446 km yield MDR and VDR of
2310
8
kg/s and 810
4
110
5
m
3
/s, respectively
(Fig. 14) concordant with the other methods (Table 4).
These parameters firmly place the plinian eruption of
1600 Huaynaputina among the largest and most
intense historical eruptions (Table 5).
Table 4 Volcanological data for the 1600 Huaynaputina stage-I
plinian eruption
Vent radius
a
220 m
Volume of ejecta
b
26 km
3
; 7 km
3
(lapilli fall)+
19 km
3
(ash fall)
Bulk density (g/cm
3
) 0.9 (lapilli fall); 0.8 (ash fall)
Volume DRE
c
8.8 km
3
; 2.6 km
3
(lapilli fall)+
6.2 km
3
(ash fall)
Duration
d
20 h
Total column height (H
T
) 3446 km
Neutral buoyancy height (H
B
) 2432 km
b
t
(distal) 50.5 km
b
t
(medial) 6.5 km
b
t
(proximal) 1.0 km
b
c
(MP) 8.3 km
b
c
(ML) 4.0 km
MDR (plot)
e
2310
8
kg/s
MDR (20 h)
f
2.810
8
kg/s
VDR (plot)
g
810
4
110
5
m
3
/s
VDR (20 h)
h
3.610
5
m
3
/s
Dispersal index
i
4.4
Fragmentation index
j
89 wt.% (pumice fall);
96 wt.% (total)
a
Average radius of vents 1 and 2
b
Minimum volume
c
Based on a dacite magma density of 2.4 g/cm
3
d
Minimum duration based on historic accounts and estimate cal-
culated according to accumulation rates by column height given
by Wilson and Hildreth (1997). See text for discussion
e
Mass discharge rate; refers to plot by Sparks (1986)
f
Calculated using the volume of the DRE for the total tephra
deposit of ~8.8 km
3
and the duration of 20 h
g
Volumetric discharge rate; refers to plot by Sparks (1986)
h
Calculated using total volume of 26 km
3
and duration of 20 h.
i
Following Hildreth and Drake (1992)
j
Following Walker (1971)
Fig. 13 Plots of clast-size (maximum lithic and maximum
pumice) on a log scale vs square root of isopleth area (from
Fig. 10) for the pumice fall deposit. The slope of the best fit line
can be used to estimate H
B
, or the neutral buoyancy height of
the eruption column; H
B
can be used likewise to estimate b
c
, or
the max clast half distance, and H
T
, or the total column height
(Pyle 1989)
513
Dispersal and fragmentation indices
In common with most fall deposits, Walker's indices
of dispersal (D) and fragmentation (F; Walker 1973,
1981) are difficult to establish at Huaynaputina
because of the arbitrariness of the definition of T
max
.
If we use Walker's (1973) definition of F as the per-
centage of material finer than 1 mm in the deposit
where the 0.1 T
max
isopach crosses the dispersal axis,
and T
max
for Huaynaputina as 20 m (proximal deposits
preserved at the amphitheater rim), D is 3100 km
2
and F is 44 wt.%. When plotted on a DF plot, Huay-
naputina would lie in the phreatoplinian field. These
estimates suggest a high degree of fragmentation of
the magma and a high-intensity eruption consistent
with the grain-size data described previously. Never-
theless, we feel the classification of the 1600 plinian
eruption as phreatoplinian is unreasonable and reflects
the difficulty in choosing T
max
(cf. Hildreth and Drake
1992). Fierstein et al. (1997) have argued that the
ultraproximal deposits from the 1912 eruption of
Novarupta are overthickened due to the accumulation
of localized deposits that do not extend into the
coeval sheet-forming pumice fall. In a similar way, the
ultraproximal deposits at Huaynaputina are also over-
thickened and T
max
based on these is not appropriate
for Walker's (1973) method.
Hildreth and Drake (1992) suggested other criteria
to describe the dispersal and fragmentation of a plin-
ian deposit. For example, if the ratio of areas enclosed
by the 20-cm and 1-m isopach is used as a dispersal
index, the value at Huaynaputina is 4.4. Values for
other deposits include ~4 for Quizapu, ~7.5 for Vesu-
vius AD 79, and ~11.5 for the Taupo plinian (Hildreth
and Drake 1992). The weight percent of the whole
deposit beyond the 10-cm or 1-cm isopach has also
Fig. 14 Graph of the total column height (H
T
) vs the volumetric
and mass discharge rates from Sparks (1986), modified to
include Huaynaputina data. The light gray area is that enclosed
by the curves for 1000 and 600C (temperatures of eruptive mix-
ture) for tropical and temperate atmospheres for the plume
model developed by Sparks (1986). The dashed curves are for
the same limits but were calculated by Sparks (1986) using his
model equation. If H
T
is assumed to be 3446 km and magmatic
temperatures are estimated between 780 and 815 C, the mass
discharge rate and volumetric discharge rate can be approx-
imated at 2310
8
kg/s and 810
4
110
5
m
3
/s (see Table 4)
Table 5 A comparison of eruptive parameters of significant his-
toric plinian eruptions. MDR inferred magma discharge rate;
DRE volume of plinian fall deposit (dense rock equivalent,
density of 2400 kg/m
3
); b
c
clast half distance; b
t
thickness half
distance; VEI volcanic explosivity index; p proximal; m medial;
d distal. (Data sources are Carey and Sigurdsson 1989; Pyle
1989; Hildreth and Drake 1992; Fierstein and Hildreth 1992)
Eruption Column MDR DRE Time Total mass b
c
b
t
height (km)
a
(kg/s) (km
3
) (h)
b
(kg)
c
(km) (km)
El Chichon A 1982 27 8.010
7
0.30 12.6 7.510
11
2.50 12.5
El Chichon B 1982 32 1.510
8
0.40 11.8 1.310
12
El Chichon C 1982 29 8.510
7
0.40 13.3 1.010
12
Santa Maria 1902 34 1.810
8
8.60 35.1 2.210
13
4.20 21.6
Mount St. Helens May 18, 1980 19 1.910
7
0.30 19.1 7.110
11
~3.30 12.834.4
Novarupta IA1912 2326 0.7110
8
2.10 <11 4.810
12
Novarupta II1912 2225 0.6210
8
1.99 26 4.810
12
Novarupta III 1912 1723 2410
7
1.66 10 4.010
12
Vesuvius 79 AD 32 1.510
8
2.10 19.5 6.1010
12
4.20 17.6
Tambora 1815 (upper) 44 3.810
8
1.20 12.2 1.2010
14
Tambora 1815 (lower) 33 1.510
8
0.50 12.1 1.2010
12
Quizapu 1932 2532 1.510
8
4.05 1825
Huaynaputina 1600 stage I 3446 2.810
8
8.8 20 2.110
13
8.34(MP) 11.0(p)
3.97(ML) 16.5(m)
50.5(d)
a
Inferred eruption column height in kilometers above the volcano
c
Mass of fall, surge, and pyroclastic flow deposits
b
Duration of plinian fall
514
been suggested as a measure for the dispersal of finer
ash. For Huaynaputina, these values are 84 and 17%,
respectively. At Quizapu, these values are 75 and
17%. If the weight fraction of the whole deposit finer
than 1 mm is used as an index of fragmentation, F
rises to 96% at Huaynaputina when the fine ash is
included. These data are consistent with a dominant
volume of fine ash being produced during the stage-I
plinian eruption. Two factors that may have contrib-
uted to the efficient fragmentation of the magma are
(a) the interaction of the magma with a hydrothermal
system, and (b) the relatively high crystal content of
the magma. The former would have resulted in a
phreatomagmatic interaction and attendant fine frag-
mentation. The second factor is largely unexplored in
the literature, but the presence of large amounts of
crystals may provide more interfaces for fracture and
therefore generate abundant fines that can be distrib-
uted distally.
Vents
Of the four Huaynaputina vents, three are nested in
the amphitheater of the host edifice. The remnant of a
dacitic dome in the northern amphitheater wall
(Fig. 15A) and a dike of similar lithology in the south-
ern amphitheater wall (Fig. 15B) are expected to date
from considerably before 1600. The four vents, the
amphitheater dome and dike, and the pre-1600 Cerro
El Volcn dome are all aligned along a broad NNW
SSE trend. We conclude that the 1600 vents were
located within a pre-existing structural element or
consistent stress field.
Vent 1, the northernmost of the four vents, has a
maximum dimension of approximately 70 m. It is an
elongate trough which appears to have an extension in
the northern amphitheater wall adjacent to the
pre-1600 dome remnant there. We interpret this as the
remnant of an early fissure-like vent, most of which
was obliterated by subsequent vents (Fig. 15A,C).
Most of Vent 2 is also obscured by the latest of the
three vents, vent 3, and its host structure, but the
exposed remnant suggests a crater of maximum
dimension of ~400 m that would have occupied almost
the entire area of the vent cluster. The vent walls dip
inward at ~40, defining a flared vent and exposing a
pervasively altered red-brown crystal-rich dacite that
contains orthopyroxene. This is distinct from any of
the other dacite bodies, wall rock, or ejecta found in
the vent area, leading us to conclude that it is either
part of the host volcanic edifice or another pre-1600
dome. The rim and outer flanks of vent 2 consist of
layers of poorly to non-vesicular dacite lapilli and
coarse ash that is part of the ejecta ring from vent 3.
Vent 3 is the latest and best preserved of the vents
in the amphitheater. It is funnel shaped, with a diam-
eter of ~200 m and depth of ~80 m. The steep (~60)
inner walls reveal ~75 m of massive, crystal-rich dacite
mineralogically distinct (no orthopyroxene) from that
seen in vent 2 but similar to the other pre-1600 and
1600 dacites. Alteration imparts a yellowish hue to the
walls. Overlying the dacite are the deposits of an
ejecta ring. These are ~3 m thick on the rim and con-
sist of welded dacitic spatter and bedded poorly vesic-
ular dacitic clasts of lapilli to block size. The upper-
most 1 m contains a significant coarse ash matrix.
These late-stage deposits are the uppermost in the
vent area. Away from the rim, flanks of a low mound
surround vent 3 on all sectors except the north. On
the eastern side it is clear that this mound is nested
within vent 2, whereas on the western side the mound
is revealed to consist of blocky dacite talus as it
emerges from beneath the ejecta-ring deposits. We
interpret these features to suggest that vent 3 was
excavated in a 1600 dacite dome that was emplaced or
being emplaced within vent 2. The ejecta ring drapes
over the dome to define the low mound. The lithology
and composition of the juvenile blocks in the ejecta
ring are the same as those from the stage-III deposits
on the amphitheater rim and beyond.
Three areas of active fumaroles are found around
the rim of vent 3, two at the rim-edge exposures of a
dike, and one of the western flank where abundant
native sulfur is exposed. The dike in the northern sec-
tor of vent 3 is ~20 m across and extends into the
elongate trough of vent 1 (Fig. 15C). This dike is vis-
ible again in the southern wall of vent 3 and its rim-
edge exposure is marked by active fumaroles and a
mound of dacite. Away from the rim, no direct expo-
sure of the dike is seen, but a series of small dacite
bodies on the amphitheater floor extend SSE along
the floor (Fig. 15A) of the amphitheater and suggest
its extension in that direction. The lithology and com-
position of the massive dacites of vent 3 (dome and
dike) are the same as the ejecta ring and the stage-III
tephra outside the amphitheater and indicate that they
are coeval.
A set of minor but well-defined concentric normal
faults have developed in a narrow (~20 m) band
around the eastern and southern sides of vent 3
(Fig. 15A). Their trace is continuous with, and concen-
tric to, the arc of the vent 2 remnant. Although the
throw on these faults is minor at the surface
(~510 cm), they clearly accommodate a series of
small steps down or back into vent 2. We interpret
these as settling faults that accommodated deflation
after the eruption of vent 3 when uprise of magma
ceased.
Vent 4 is a small crater ~70 m in diameter and
~30 m in depth. Its walls reveal ~15 m of stage-I plin-
ian fall overlain by ~2 m of stage-II ignimbrite and
3 m of stage-III bedded dacitic deposits (Fig. 16D).
Abundant flow-banded, breadcrusted bombs of crys-
tal-rich dacite are strewn around the rim and overlie a
thin white ash exposed on the eastern and western
rims. We interpret this vent as the product of a minor
explosive (phreatomagmatic) outburst from an en-
515
echelon segment of the main feeder dike and the final
gasp of the 1600 eruption. The insignificant volume of
juvenile material (ash and bomb) ejected suggests a a
very small magmatic component.
Discussion
Timing of events in the period 19 February
to 6 March 1600
The 1600 eruption began abruptly on 5 p.m. on 19
February along an en-echelon fissure system that
trended NNWSSE across the old edifice. Precursory
earthquakes several days before the eruption (Ta-
ble 1) may have been signaling magma movement to
relatively shallow levels as magma rose along a series
of dykes to erupt at fissures. Coarse pumice directly at
the base of the stage-I fall sequence indicates that the
plinian column established very rapidly. The intensity
of the eruption increased with time, as evidenced by
general reverse grading in the pumice fall deposit until
approximately halfway through the eruption sequence.
A large central vent, vent 2, was established and its
development may be recorded by the extremely lithic-
rich coarse layer C. Mining of an old hydrothermal
system is suggested by the abundance of hydrother-
Fig. 15AD Key features of the 1600 vents and host amphi-
theater. A View from amphitheater rim northeastwards shows
the main vent cluster. Vents are numbered sequentially 13
according to the perceived progression. Black arrows point to
exposures of the same dike that traverses the vent cluster and
amphitheater floor. Remnant of the pre-1600 dacite dome
reamed out during the eruption lies on the northern extension
of dike trend. Large white arrows indicate the trace of several
small faults that define a narrow (20 m) discontinuous zone con-
centric to the arc of vent 2. Small white arrows point to slumps
in ejecta from vent 2. The ejecta-ring of vent 3 includes the
deposits on the rim of vent 2. These and the uppermost beds on
the amphitheater rim are all poorly vesicular dacite of stage III.
B Southern amphitheater wall shows the internal stratigraphy of
the ancestral stratovolcano. These are dominantly andesitic lavas
and pyroclastics, but two large silicic pyroclastic layers occur
toward the top of the 250- to 300-m-tall sequence. White depos-
its of the 1600 eruption form the rim. Note the pre-1600 dike
(arrow) that lies on an extension of the main NNWSSE trend.
C Closer view of the exposure of the dike between vents 1 and
3 (left black arrow in A). The exposure is approximately 70 m
high. This massive crystal-rich dacite is poorly vesicular and dis-
plays strong vertical jointing. One of the active fumarolic zones
is on the western margin of the dike on the rim of vent 3. The
view is to the east. D The small crater that forms the fourth
(and final) vent of the 1600 eruption. This crater is ~30 m deep
and exposes 15 m of the stage-I plinian fall, approximately 5 m
layer of pink ignimbrite, and 5 m of stage-III tephra fall. A thin
ash deposit with abundant bombs forms the rim
516
mally altered accidental lithics in the pumice fall
deposit. As a consequence of vent widening, the erup-
tion intensity progressively decreased and this is
recorded by the normal grading of pumice in the
upper part of the plinian deposit. At its peak, around
the time of layer C, the eruption cloud consisted of a
convecting column up to 2433 km and a giant
umbrella cloud that extended up from 3446 km. Dur-
ing the final stages a less powerful but consistent erup-
tion produced the millet-seed layer. This general
sequence was punctuated by violent bursts and vent
wall collapse throughout the course of the eruption to
produce the coarse lithic-rich zones within layers B, C,
D, and E. The stage-I eruption had a duration of
~20 h and ceased on 20 February at approximately
2 p.m. (Table 1). The main amphitheater was
enlarged, due to reaming out and mass wasting, to
almost its present dimensions by this time.
Fig. 16 Evolution of the vent system during the 1600 eruption.
See text and legends for further explanation
517
Quiescence from 20 to 22 February was terminated
by the onset of intermittent explosive activity of stage
II. Vent 2 was reactivated and produced the sequence
of intraplinian pyroclastic flows and minor fall depos-
its (Fig. 16). A set of closely spaced plinian or subplin-
ian eruptions is implicated. The lithic-rich proximal
lag breccias attest to vigorous vent erosion during the
ignimbrite-generating stages. The transition between
collapsing and stable columns was probably due to
physical changes in the vents. This stage lasted
~4 days and ceased on 26 February.
Toward the end of stage II, a dike-fed dome was
being emplaced into vent 2 and the final stages of this
may be recorded by the relative quiescence described
in historical records between 26 and 28 February
(Fig. 16). Subsequent disruption of this dome by
closely spaced vulcanian-type eruptions resulted in the
formation of vent 3, its ejecta ring, and coarse tephra
fall. Historical accounts of large earthquakes preced-
ing ashfalls are consistent with vulcanian explosions
and may also indicate that magma movement beneath
Huaynaputina continued throughout stage III, fueling
the eruptions. Exhaustion of the magma supply
beneath vent 3 resulted in deflation and settling along
the concentric faults. A short, discrete blast to form
vent 4 was the last event of this eruption (Fig. 16). All
activity ceased on 6 March. Fumarolic activity contin-
ues presently.
Compositional aspects
More detailed petrological and geochemical data are
to be presented elsewhere, but available data already
offer some valuable insight into the 1600 eruption.
While the mineralogy of the dacites erupted in 1600 is
the same, subtle differences suggest that these dacites
were not from a single homogeneous magma. During
stage I a less-evolved, hotter dacite with lower crystal
content (1720%) was erupted, whereas more-evolved,
cooler, crystal-richer (2540%) dacite was tapped dur-
ing stages II and III. The increase in density of the
juvenile material from ~1.2 g/cm
3
in stage I to 2.2 g/
cm
3
in stage III indicates differences in volatile con-
tents that may account for style of eruption. Prelimi-
nary estimates of pre-eruptive volatile (H
2
O) contents
using the method of Housh and Luhr (1991; Table 2)
imply that the stage-I magma was a more volatile-rich
magma batch with water contents as high as 5% that
could have sustained a plinian eruption. Lower vol-
atile contents of <4% are suggested for stage-II and
stage-III magmas, consistent with less explosive activi-
ty. These differences may suggest that two different
magma batches were probably tapped (Arbesman and
de Silva 1998; Thouret et al. 1999) or that these
dacites all come from the same zoned column of mag-
ma. The latter interpretation seems more likely
because the short time span between stage I and stage
II is problematic for a model with two separate mag-
mas. In this context, the differences in glass composi-
tion can be reconciled with a zoned magma column if
degassing-induced microlite crystallization resulted in
more evolved residual glass in stages II and III. This
is consistent with the more microlite-rich character of
the groundmass in juvenile clasts stages II and III.
Nevertheless, the hotter equilibration temperatures
(Table 2) of stage I are problematic for this model. In
the context of the zoned magma column model, we
interpret the stage-II and stage-III magma as partially
degassed dacitic magma left after the stage-I eruption,
so the progression from a sustained plinian eruption
to a final vulcanian stage may have been the result of
decreasing volatile content as the eruption progressed.
We will deal with these possibilities elsewhere.
Petrographic observations and mineral chemistry
reveal an abundance of thermal disequilibrium tex-
tures in the form of complex and varied zoning and
sieved textures in plagioclase and amphiboles, and
multiple populations of plagioclase. These suggest an
important role for recharge in the evolution and erup-
tion of the 1600 dacites.
Lack of caldera collapse and nature of the eruption
The lack of caldera collapse despite the significant
volume of magma (~10 km
3
DRE) erupted is intri-
guing. Thouret et al. (1999) suggest that this is a func-
tion of the regional stress field; however, calderas
have developed in other nearby volcanoes, and our
preferred interpretation is that no shallow magma
chamber existed below the site of the 1600 eruption.
Magma probably rose in a plexus of dykes that even-
tually focused into a narrow zone near the surface. In
1600, vesiculation in a rising magma batch may have
been enhanced by volatile addition from recharging
magma and interaction with an ancient hydrothermal
system to trigger and drive the explosive eruption.
We suggest that the young dacite volcanoes, includ-
ing Huaynaputina, are part of a developing silicic vol-
canic field . These silicic eruptions are seen as discrete
(monogenetic) events fed from a deep-seated zone of
magma genesis nourished by the thermal input from
mafic magma. The broad family resemblance of the
dacite magmas in and around Huaynaputina, accom-
panied by subtle variations between each, is easily rec-
onciled with evolution of small batches of magma that
separate from a much larger deep zone of storage
where the general family character is set. We tenta-
tively suggest that the uprise of these magmas is likely
to be fueled by recharge.
Acknowledgements This paper is dedicated to the memory of
the late Peter Francis who encouraged S. de Silva to take on
this project and with whom S. de Silva first visited Huaynapu-
tina in 1989. His support and inspiration will be greatly missed.
The authors are indebted for logistical support to the Instituto
Geofsico del Peru and the Universidad de San Agustin, Are-
quipa. In particular, M. Lazo and F. Ocola are thanked for con-
518
tinued collaboration. The expert driving of J. Chiroque and the
help of the Llameros of Quinistaquillas headed by D. Pedro
were invaluable. Thorough, constructive reviews by C. Wilson
and J. Fierstein, and the comments of a third anonymous
reviewer, are gratefully acknowledged. These, and the insightful
and thorough editorship of W. Hildreth, improved this article
immeasurably. This work was supported by National Science
Foundation grant EAR9614405 to S. de Silva.
References
Andersen DJ, Lindsley DH, Davidson PM (1993) QUILF: a
Pascal program to assess equilibria among FeMgMnTi
oxides, pyroxenes, olivine, and quartz. Comp Geosci
19:13331350
Arbesman K, de Silva SL (1998) Petrology and geochemistry of
the 1600AD Huaynaputina magma. Int Volcanol Congr
Abstr Cape Town, South Africa: 3
Barriga VM (1951) Los Terremotos en Arequipa (15821868).
La Colmena, Arequipa
Bonnadonna C, Ernst GG, Sparks RSJ (1998) Thickness vari-
ations and volume estimates of tephra fall deposits; the
importance of particle Reynolds number. J Volcanol Geo-
therm Res 81:173187
Briffa K, Jones PD, Schweingruber FH, Osborn TJ (1998)
Influence of volcanic eruptions on Northern Hemisphere
summer temperature over the past 600 years. Nature
393:450455
Bullard F (1962) Volcanoes of southern Peru. Bull Volcanol
24:443453
Carey S, Sigurdsson H (1989) The intensity of plinian eruptions.
Bull Volcanol 51:2840
Delmas RJ, Kirchner S, Palais JM, Petit JR (1992) 1000 years of
explosive volcanism recorded at the South Pole. Tellus Series
B, 44:335350
de Silva SL, Francis PW (1990) Potentially active volcanoes of
Peru: observations using Landsat Thematic Mapper and
space shuttle photography. Bull Volcanol 52:286301
de Silva SL, Francis PW (1991) Volcanoes of the Central Andes.
Springer, Berlin Heidelberg New York
de Silva SL, Zielinski G (1998) The global impact of the
1600AD eruption of Huaynaputina, southern Peru. Nature
393:455458
de Silva SL, Adams NK, Permenter JL, Salas G, Self S (1998)
Vent character and eruption stratigraphy of the 1600A.D.
eruption of Huaynaputina Peru. Geol Soc Am Abstr Pro-
gram 30:A45
de Silva SL, Alzueta J, Salas G (2000) The socioeconomic con-
sequences of the 1600AD eruption of Huaynaputina, south-
ern Peru. In: McCoy F, Heiken G (eds) Volcanic disasters in
human antiquity. Geol Soc Am Spec Pap 345
Fierstein J, Hildreth W (1992) The plinian eruptions of 1912 at
Novarupta, Katmai National Park, Alaska. Bull Volcanol
54:646684
Fierstein J, Nathenson M (1992) Another look at the calculation
of fallout tephra volumes. Bull Volcanol 54:156167
Fierstein J, Houghton BF, Wilson CJN, Hildreth W (1997) Com-
plexities of plinian fall deposition at vent: an example from
the 1912 Novarupta eruption (Alaska). J Volcanol Geotherm
Res 76:215227
Hammer CU, Clausen HB, Dansgaard W (1980) Greenland ice
sheet evidence of post-glacial volcanism and its climatic
impact. Nature 288:230235
Hammer CU, Clausen HB, Dansgaard W (1981) Past volcanism
and climate revealed by Greenland ice cores. J Volcanol
Geotherm Res 11:310
Hildreth W, Drake RE (1992) Volcn Quizapu, Chilean Andes.
Bull Volcanol 54:93125
Houghton BF, Wilson CJN, Pyle DM (1999) Pyroclastic fall
deposits. In: Sigurdsson H, Houghton BF, McNutt S, Rymer
H, Stix J (eds) Encyclopedia of volcanoes. Academic Press,
San Diego, pp 555570
Housh TB, Luhr JF (1991) Plagioclase-melt equilibria in
hydrous systems. Am Mineral 76:477492
James DE (1981) Role of subducted continental material in the
genesis of calc-alkaline volcanics in the Central Andes. Geol
Soc Am Mem 154:769790
James DE, Brooks C, Cuyubamba A (1976) Andean Cenozoic
Volcanism: magma genesis in the light of strontium isotopic
composition and trace element geochemistry. Geol Soc Am
Bull 87:592600
Lamb HH (1970) Volcanic dust in the atmosphere; with a chro-
nology and assessment of its meteorological significance. Phil
Trans R Soc Eng 266:425533
Palais JM, Kirchner S, Delmas RJ (1990) Identification of some
global volcanic horizons by major element analysis of fine
ash in Antarctic ice. Ann Glac 14:216220
Pyle DM (1989) The thickness, volume and grainsize of tephra
fall deposits. Bull Volcanol 51:115
Pyle DM (1995) Assessment of the minimum volume of tephra
fall deposits. J Volcanol Geotherm Res 69:379382
Rice PM, Smith GC (1989) The Spanish colonial wineries of
Moquegua, Peru. Hist Archaeol 23:4149
Sarna-Wojicki AM, Shipley S, Waitt RB, Dzurisin D, Wood SH
(1981) Areal distribution, thickness, mass, volume, and grain-
size of airfall ash from the six major eruptions of 1980. US
Geol Surv Prof Pap 1250:577600
Scuderi LA (1990) Oriental sunspot observations and volcanism.
Q J R Astron Soc 31:109120
Simkin T, Siebert L (1994) Volcanoes of the world, 2nd edn.
Geoscience Press, Tucson, Arizona
Sparks RSJ (1986) The dimensions and dynamics of volcanic
eruption columns. Bull Volcanol 48:315
Thompson LG, Moseley-Thompson E, Dansgaard W, Grootes
PM (1986) The Little Ice Age as recorded in the Quelcaya
ice cap. Science 234:361364
Thompson LG, Davis ME, Moseley-Thompson E, Sowers TA,
Henderson KA, Zagorodnov VS, Lin PN, Mikhalenko VM,
Campen RK, Bolzan JF, Cole-Dai J, Francou B (1998) A
25,000 year tropical climate history from Bolivian ice cores.
Science 282:18581864
Thouret J-C, Davila J, Rivera M, Le Pennec JL, Eissen JP,
Gourgaud A, Cotton J, Juvign E (1997) L'eruption explo-
sive de 1600 au Huaynaputina (Prou), la plus volumineuse
de l'histoire dans les Andes centrales. Comp Rend Acad Sci
Paris 325:931938
Thouret J-C, Davila J, Eissen J-P (1999) Largest explosive erup-
tion in historical times in the Andes at Huaynaputina volca-
no, A.D. 1600, southern Peru. Geology 27:435438
Tosdal RM, Farrar E, Clark AH (1981) KAr geochronology of
the late Cenozoic volcanic rocks of the Cordillera Occiden-
tal, southernmost Peru. J Volcanol Geotherm Res
10:157173
Vasquez de Espinosa A (1942) Compendium and description of
the West Indies. Smithsonian Misc Coll 102:1862
Walker GPL (1971) Grain-size characteristics of pyroclastic
deposits. J Geol 79:696714
Walker GPL (1973) Explosive volcanic eruptions: a new classifi-
cation scheme. Geol Rundsch 62:431446
Walker GPL (1981) Plinian eruptions and their products. Bull
Volcanol 44:223240
Wilson CJN, Hildreth W (1997) The Bishop Tuff: new insights
from eruptive stratigraphy. J Geol 105:407439