Tephrochronologic Constraints on Temporal Distribution of Large Landslides in Northwest Argentina Reginald L. Hermanns, Martin H. Trauth, Samuel Niedermann,1 Michael McWilliams,2 and Manfred R. Strecker Institut fu¨r Geowissenschaften, Universita¨t Potsdam, Postfach 601553, D-14415 Potsdam, Germany (e-mail:
[email protected])
ABSTRACT Two morphologic settings in the northwestern Argentine prone to giant mountain-front collapse—deeply incised narrow valleys and steep range fronts bordering broad piedmonts—were analyzed through detailed investigations of fossil landslides and related fluvio-lacustrine sediments. Nine different rhyodactic tephra layers were defined by geochemical fingerprinting of glass, morphology of pumice, stratigraphic relationships, and mineralogy. The age of three tephra could be determined either directly by 40Ar/39Ar dating or relatively by 14C dating of associated sediments: Paranilla Ash (723 5 89 ka), Quebrada del Tonco Ash (∼30 ka), and Alemanı´a Ash (∼3.7 ka). These units permit correlation of several spatially separate landslide deposits. Landslide deposits in narrow valleys were generated in the late Pleistocene between 40 and 25 ka and in the Holocene since ca. 5 ka and correspond to periods characterized by increased humidity in subtropical South America. In contrast, the age of large landslides in piedmont regions is significantly greater but more difficult to define by tephrochronology. However, selected deposits from this second environment have cosmogenic nuclide exposure ages of 140–400 ka. Because of the large distance of the collapsed mountain fronts from eroding streams and because of important Quaternary displacement along the mountainbounding faults, we suggest that strong, low-frequency seismic activity is the most likely trigger mechanism for most of the landslides in this environment.
Introduction 1979; Adams 1981; Schuster et al. 1992; Crozier et al. 1995). Establishing a temporal framework for large landslide events is often hampered by a lack of datable material, especially in arid and semiarid regions. This dilemma, however, is eased when landslide deposits are associated with volcanic tephra layers and lacustrine sediments that may serve as marker horizons. Situated east of the Central Andean volcanic zone (CVZ), the Argentinian Andes and the adjacent foreland between 247309and 277309S represent an ideal region to develop a tephra-based chronostratigraphy of catastrophic mass movements in an active mountain belt (fig. 1). This region is tectonically active and contains more than 50 voluminous landslide deposits. In addition, this part of the Andes is characterized by an arid climate, which guarantees a high preservation potential for landslide deposits. In the CVZ, more than 60 volcanoes have been repeatedly active throughout the Quaternary (Francis and de Silva 1989). Historic tephra
Large mountain-front collapse resulting in deposits of several million cubic meters is a common phenomenon in tectonically active orogens (e.g., Shreve 1966; Plafker and Ericksen 1978; Keefer 1984). Dating such catastrophic mass movements is essential for the evaluation of mechanisms conditioning and triggering landslides, the assessment of their recurrence, and finally, for the appraisal of future landslides as potential natural hazards. The need to understand landslide mechanisms as well as spatial and temporal landslide distribution is particularly important in regions frequently affected by earthquakes, which may ultimately trigger such large mass movements in rocks already preconditioned for failure (e.g., Nikonov and Shebalina Manuscript received November 30, 1998; accepted August 31, 1999. 1 GeoForschungsZentrum Potsdam, Telegrafenberg, D-14473 Potsdam, Germany. 2 Geophysics Department, Stanford University, Stanford, California 94305-2215, U.S.A.
[The Journal of Geology, 2000, volume 108, p. 35–52] q 2000 by The University of Chicago. All rights reserved. 0022-1376/2000/10801-0003$01.00
35
36
R. L. HERMANNS ET AL.
nary activity, thus potentially representing the result of seismic triggering (Fauque and Strecker 1988; Gonza´lez Dı´az and Mon 1996; Hermanns and Strecker 1999). Although the timing of these avalanches is still poorly understood, there is another influence on controlling landsliding. Radiocarbondated deposits of landslide-dammed lakes indicate that some avalanches were generated during a more humid period between 40 and 25 ka (Trauth and Strecker 1999). However, in order to judge the relationship between the geomorphic setting of mountain-front collapse and landslide frequency, climate, and paleoseismic activity in the northwest Argentinian Andes, a more detailed chronology of mountain-front failures is needed. A large data set of the landslides with volumes between 0.03 and 0.375 km3 (Hermanns and Strecker 1999), based on detailed satellite image analysis, air photo interpretation, and fieldwork, permits a statistically representative analysis of the temporal distribution of these voluminous mountain-front failures. In this article, we present new chronostratigraphic data, based on tephra that will be grouped on the basis of geochemical studies of glass shards, pumice morphology, stratigraphic relations, and mineralogy. These tephras will allow us to date several landslide events and describe their temporal clusters and various morphologic settings. Figure 1. Tectonic provinces (different shades of gray) of the central Andes in Argentina, Bolivia, and Chile and location of volcanoes with Quaternary activity (filled circles), modified after Jordan et al. (1983) and de Silva and Francis (1991).
in the eastern Andean ranges and the foreland from vulcanian- and subplinian-type eruptions indicate ash transport toward the east-southeast, with an areal coverage 1112,000 km2 (Glaze et al. 1989; Felpeto and Ortiz 1997). Although recent eruptions produced only thin ash layers that were subsequently eroded (Felpeto and Ortiz 1997), older Quaternary eruptions produced deposits thick enough to be preserved and have stratigraphic relations to large landslide deposits. Most of the landslide deposits with volumes 1106 3 m in the Argentinian Andes result from rock avalanches (sturzstroms; Hermanns and Strecker 1999). Previous studies show that such deposits lie in clusters (fig. 2), always along lithologically, structurally, and topographically preconditioned mountain fronts (Wayne 1994; Hermanns and Strecker 1999; Strecker and Marrett 1999). In addition, the avalanche deposits often exist in the vicinity of faults having signs of multiple Quater-
Geologic Setting The CVZ of the central Andes (fig. 1) is one of three important Andean provinces to experience Cenozoic volcanism (e.g., Thorpe et al. 1982), whereas the Andean sector between 287 and 337S is amagmatic (Jordan et al. 1983). In its southern part, the CVZ has a transverse zonation with a calc-alkaline association in the west and shoshonitic volcanics in the east (De´ruelle 1978). In both segments, rhyodacitic to rhyolitic volcanics also exist (e.g., Francis et al. 1989; de Silva 1991; Coira and Kay 1993). Silicic eruptions were dominated by explosive ignimbrite volcanism, resulting in several major resurgent caldera complexes (de Silva 1989). To the east of the volcanic region is the intraAndean Puna plateau and the fault-controlled valleys of the Cordillera Oriental and the northern Sierras Pampeanas (fig. 1). The ca. 4000-m-high Puna plateau is characterized by intervening ranges of reverse-fault-bounded basement blocks that reach elevations of about 6000 m (fig. 2a, 2b). The Cordillera Oriental is a fold and thrust belt (Mon 1976; Grier et al. 1991) of Precambrian basement and overlying unmetamorphosed Cambrian to Tertiary sediments (Reyes and Salfity 1973; Omarini
Journal of Geology
TEMPORAL DISTRIBUTION OF LARGE LANDSLIDES
1983; Salfity and Marquillas 1994); it is cut by deeply incised valleys (fig. 2a, 2b). In contrast, the Sierras Pampeanas are late Cenozoic Laramide-type uplifts composed of late Proterozoic metamorphic basement rocks that contrast with highly erodible Tertiary clastic sediments in the adjacent intramontane basins (Caminos 1979; Mon 1979; Jordan et al. 1983). The basins are further characterized by alluvial-fan deposits and coarse gravel associated with multiple, gently inclined pediments that abut the steep mountain fronts (Strecker et al. 1989). Along tectonically active mountain fronts and in areas where antecedent rivers cross the uplifting ranges of the Cordillera Oriental, voluminous landslide deposits often associated with lacustrine and terrace deposits exist (fig. 2a, 2b) (Hermanns and Strecker 1999; Strecker and Marrett 1999). Common to all Cenozoic sedimentary deposits in the intramontane basins is their association with intercalated volcanic ash layers related to multiple eruptions in the CVZ (Marshall and Patterson 1981; Strecker et al. 1989; Grier and Dallmeyer 1990; Malamud et al. 1996). Methods In the study area, 26 tephra samples were taken from nine different Quaternary sedimentary sequences associated with landslide deposits. At the outset, we did not know the stratigraphic relations among these tephra. Our task was therefore to characterize the samples geochemically and use statistical analyses of these data supplemented by scanning electron microscope (SEM) observations of the glass morphology, stratigraphic relations, and mineralogy of the tephra (table 1) to identify groups of samples that were deposits of single eruptions. Once their identity as time-synchronous marker horizons was known, these tephras were named after type localities. Finally, we could apply limited chronological data from 40Ar/39Ar and 14C dating in a more efficient manner. The chemical composition of volcanic glass contained in such pyroclastics is one of the most distinctive and consistent characteristics by which tephra from individual eruptions can be identified (e.g., Sarna-Wojcicki 1976; Westgate and Gorton 1981; Bogaard and Schmincke 1985). Our tephrochronological study is therefore based mainly on electron-microprobe analysis (EMA) of volcanic glass shards (table 2). Tephra samples were disaggregated in water and wet sieved at intervals of 63, 200, and 500 mm. The 200–500-mm-size fraction was placed in an ultrasonic bath to remove adhering material and sieved
37
again with water and acetone. Mafic minerals were removed using a magnetic separator. Glass shards were picked from the felsic concentrate, and biotite from the mafic concentrate, avoiding samples with vesicles, crystalline intergrowth, or alteration. EMA was carried out according to the principles outlined in Frogatt (1992), using a 15-kV excitation potential, a 20-nA beam current, and a 10-mm beam diameter. Up to 22 glass shards from each of the 26 tephra beds were analyzed for Na, Si, Mg, Al, K, Ca, Ti, Mn, and Fe. The beam was defocused to minimize volatilization of water or dispersion of sodium. Up to nine single measurements per glass shard were performed, using commercially available silicate and oxide minerals as standards. Oxide concentrations for each sample generally ranged between 90% and 99%. In order to eliminate measurements of disproportionately strongly altered glass shards, CO2 and H2O contents were measured by infrared absorption. The resulting percentages of volatiles were subtracted from the ideal volatile free value of 199% of the EMA measurements. All totals below this value were not included in further analysis. All EMA data were normalized to 100% to facilitate proper comparison (table 2). The large number of analyses per sample allows a statistical analysis of the data, involving an examination for outliers, alteration trends, and multiple populations—a standard procedure in tephrochronologic investigations with enormous data sets (e.g., Smith and Nash 1976; Sarna-Wojcicki et al. 1987). The distribution of sodium in a few samples tended to have a weak negative skew, suggesting a loss during glass hydration (fig. 3; Cerling et al. 1985). After eliminating outliers, the median of the measurements of all elements was used as a measure of central tendency, which is less sensitive to outliers and skewness than arithmetic means (Swan and Sandilands 1995). Before statistical analysis, the data were pretreated by log-ratio transformation (e.g., Aitchison 1984, 1986). This ensures data independence (on an oxide-to-oxide basis) and avoids the constant sum normalization constraints imposed by EMA of volcanic glass with variable composition: yi = log (x i /xd), where yi denotes the transformed score (i = 1, 2, ) , d 2 1) of some raw oxide score xi. The procedure is invariant under the group of permutations of the components, and any component (oxide) can be used as divisor xd. In this study, we adopted the log-transformation approach and selected Al2O3 as the divisor. The oxide Al2O3 has
Journal of Geology
TEMPORAL DISTRIBUTION OF LARGE LANDSLIDES
only a small variance within the data set and is generally little affected by alteration processes; hence, it is the most appropriate divisor. Using the mobile K2O as divisor (e.g., Stokes et al. 1992) did not lead to significantly different results. Principal component analysis (PCA) and hierarchical cluster analysis (CA) have been used to identify groups of ash layers based on the chemical composition of glass shards (e.g., Sarna-Wojcicki 1976; Smith and Nash 1976; Sarna-Wojcicki et al. 1987; for a review of the principles of PCA and CA see Swan and Sandilands 1995). The advantage of CA over other methods such as discriminant function analysis is that an a priori set of known samples is not required. Each of the groups identified using CA is considered to represent a single volcanic event and can therefore be used as a stratigraphic marker horizon. The CA and PCA were carried out using Matlab routines contained in the PLS Toolbox provided by Eigenvector Research, Manson, Washington. The routine pca was used to perform PCA on autoscaled data. Autoscaling (or variance scaling) essentially puts all variables on an equal basis in the analysis. This is important for variables showing large differences in the absolute values of mean and variance. The PCA redefines the nineelement coordinate system by ranking each new perpendicular axis based on the maximum separation of the samples. It is generally found that the data can be adequately described using far fewer factors than the original variables. The variance that each principal component (PC) captures can determine the number of PCs to remain in the model; the leftover variation is considered noise (table 3). Unfortunately, there is no simple method for automating the determination of the number of PCs to retain in the model. However, a rule of thumb is to start from the smallest eigenvalue and to go back to the larger ones, looking for a sudden jump in the values. It is then appropriate to choose those PCs for the model that include this jump. The routine cluster performs k-means clustering on autoscaled data using Mahalanobis distances based on raw principal component scores. The Mahalanobis distance is a measure of the differences between the means of k multivariate groups.
39
Applying the PCA on the log-ratio-transformed and autoscaled data listed in table 1, it is appropriate to choose the first two PCs for the model (fig. 4). The scores for PC 1 and PC 2 provide a firstorder relation between the different ash layers (fig. 5), defining distinct and well-separated tephra clusters. However, there are a number of randomly distributed data points without any relationship to a single tephra group. Next, we employed cluster analysis on autoscaled data to define the groups of ash layers. The graphical output of cluster analyses, the dendrogram choosing the first two PCs, displays information regarding the Mahalanobis distance between samples and groups of samples (fig. 6). For example, g63 is the sample with the greatest distance from the other samples in the dendrogram. Other samples, such as g43, g37, g47, g45, g35, g65, and g21, are in close proximity to each other and are very similar, suggesting that they represent a single volcanic event. Finally, in order to develop a chronostratigraphy for the tephra layers and avalanche deposits, selected biotites from two ash layers were dated by the incremental-heating 40Ar/39Ar method. Calculated ages are based on multiple crystal analyses. In addition, AMS-14C dates from peat, mollusks, and charcoal sampled from fluvio-lacustrine sediments associated with landslide deposits were used for further age constraints. All 14C ages cited here were converted to calendar years using the principles outlined in Stuiver and Reimer (1993) to compare ages obtained by different dating methods. Tephra Classification Tephra in northwestern Argentina is made of pumiceous, fine- to coarse-grained glass particles with an admixture of lapilli and lithic fragments. The thickness of tephra layers varies between a few centimeters and 2 m; generally, tephra layers in the study area thicken from north to south. All tephras are dominantly white and rhyodacitic in composition, with SiO2 contents between 66.8% and 72.8% and total alkali contents ranging from 8.4% to 8.9%. The tephras contain significant amounts of quartz, potassium feldspar, and plagioclase; bi-
Figure 2. a, Topographic map showing distribution of mountain ranges, intramontane basins, and narrow valleys and indicating type localities of the defined tephra layers (table 1). b, Generalized geologic map of the study area and distribution of large landslide deposits (after unpublished map by J. Sosa Gomez, pers. comm.; Allmendinger et al. 1983; Strecker et al. 1989; Marrett et al. 1994; Hermanns and Strecker 1999). S.L.B. = Sierra Laguna Blanca, Q = Quebrada. Open circles denote rock-avalanche deposits in narrow valleys; filled circles correspond to rock-avalanche deposits in piedmont environments. Selected rock-fall deposits are marked by a circle with a cross. Letters indicate positions of profiles illustrated in figure 8.
Table 1.
Compilation of Location, Glass Morphology, Biotite Contamination, and Age of Sampled Tephra
Sample
Location (latitude/longitude)
Quebrada La Yesera Tuff: g09 g67
267009S/657459W 257579S/657449W
Glass morphology Y–bubble wall junction, double concave plates Y–bubble wall junction, double concave plates
Biotite contaminationa
Age (ka)
um
)
um
)
El Pen˜on Tuff: g71 g61
267299S/677499W 277109S/667099W
Lapilli tuff Elongated, tubular vesicles, thick vesicles walls
um pm
1200
Cerro Paranilla Ash: g05
267059S/657449W
Elongated, tubular vesicles, thick vesicles walls Elongated, tubular vesicles, thick vesicles walls Elongated, tubular vesicles, thick vesicles walls Elongated, tubular vesicles, thick vesicles walls
um
)
um
723 5 89
um
)
um
)
)
g07
257349S/657589W
g25
267059S/657449W
g31
257349S/657589W
)
Ruinas del Rinco´n Ash: g63
267389S/667259W
Irregularly shaped, oval vesicles, thin walls
pm
)
El Paso Ash: g04
257599S/657459W
Elongated, tubular vesicles, thick vesicles walls Elongated, tubular vesicles, thick vesicles walls
um
54 5 221
pm
)
277049S/667489W 257599S/657459W 257349S/657589W
Elongated, tubular vesicles Elongated, tubular vesicles Elongated, tubular vesicles
pm pm pm
133
277049S/667499W
Irregularly shaped, oval vesicles, thin walls, strong intergrowth with biotite
pm
11.31 5 .13
277109S/667089W
Elongated, tubular vesicles, thin vesicles walls Elongated, tubular vesicles, thin vesicles walls Elongated, tubular vesicles, thin vesicles walls Elongated, tubular vesicles, thin vesicles walls
pm
)
pm
)
pm
)
pm
13.63 5 .07
pm
)
pm
)
um
!3.63 5 .07
um
)
um
)
pm
)
um
!5.93 5 .05
g17 Quebrada del Tonco Ash: g15 g29 g39 Villa Vil Ash: g33 Buey Muerto Ash: g27 g49
277049S/667489W
g69
257599S/657469W
g73
267319S/657449W
Alemanı´a Ash: g21
a
277109S/667079W
267059S/657449W
g35
277049S/667499W
g37
267319S/657449W
g43
267059S/657449W
g45
267389S/667249W
g47
277109S/667089W
g65
257429S/657429W
um = unimodal; pm = polymodal.
Elongated, tubular vesicles, thick vesicles walls Elongated, tubular vesicles, thick vesicles walls Elongated, tubular vesicles, thick vesicles walls Elongated, tubular vesicles, thick vesicles walls Elongated, tubular vesicles, thick vesicles walls Elongated, tubular vesicles, thick vesicles walls Elongated, tubular vesicles, thick vesicles walls
) )
Journal of Geology
TEMPORAL DISTRIBUTION OF LARGE LANDSLIDES
otite is a minor or accessory constituent. In some cases, accessory minerals such as zircon and apatite were identified as intergrowths in biotite. Tephra classification is based on chemical glassshard fingerprinting because of the unambiguity of homogenous glass composition within a tephra layer (e.g., Sarna-Wojcicki 1976; Westgate and Gorton 1981). Although these glass shards generally have rhyolitic compositions, precise EMA revealed significant differences, used to group several tephra layers (table 1). Further discrimination is based on EMA compositions of biotites; unimodal biotite compositions were found for nearly half of the samples and are taken as evidence for primary air-fall deposition. Conversely, redeposition of tephra layers is indicated by biotites with polymodal compositions (table 1). Of the 26 tephra samples, seven related tephra groups and two individual tephra could be identified (tables 1, 4). Among the different groups, one tephra unit occurs only locally, whereas the others are widely distributed throughout the entire area. Since their exact eruption source is unknown, the tephra are named after type localities. Terminology is used after Fisher and Schmincke (1984). The term “ash” is used for unconsolidated tephra with a grain size smaller than 2 mm; tuff is the consolidated equivalent of ash. Table 2.
41
Figure 3. Histogram of Na2O concentration of EMA of glass in ash layer g67 (60 samples). Quebrada La Yesera Tuff. This tuff (g09, g67) is a fine-grained and slightly lithified tephra deposit; mafic minerals only occur as accessories. The glass shards typically show Y-shaped bubble wall junctions or double concave plates (fig. 7a). This tuff occurs as a continuous layer around 20 m below the top of a terrace at the confluence of the Calchaquı´es and Santa Marı´a rivers in the southern
Median Glass Shard Major Element Composition of Tephra Layers in Northwest Argentina
Sample
n
Na2O
SiO2
MgO
Al2O3
K2O
CaO
TiO2
MnO
FeO
g09 g67 g71 g05 g07 g25 g31 g63 g61 g04 g17 g15 g29 g39 g33 g27 g49 g69 g73 g21 g35 g37 g43 g45 g47 g65
14 49 3 17 25 14 21 31 7 6 30 11 8 14 4 8 20 7 13 26 45 24 58 38 23 12
3.58 3.61 2.41 2.37 2.30 2.51 3.12 3.29 3.02 2.39 2.26 3.83 3.61 3.63 2.36 3.28 2.84 1.87 2.08 3.57 2.59 3.54 2.81 2.96 2.99 3.42
77.93 77.61 77.48 78.09 77.96 77.66 77.12 75.55 77.82 79.01 78.99 77.74 77.94 77.80 77.43 78.26 78.63 79.67 79.54 78.07 78.74 77.90 78.59 78.49 78.46 77.82
.10 .11 .06 .10 .09 .10 .09 .13 .05 .05 .04 .04 .04 .05 .14 .04 .05 .05 .05 .05 .04 .04 .05 .04 .04 .04
13.22 13.34 13.41 13.04 13.06 13.04 12.97 14.17 12.70 12.90 13.08 12.84 13.06 13.12 13.18 12.81 13.01 13.23 13.21 12.91 13.28 13.02 13.22 13.15 13.21 13.02
3.89 4.06 4.96 4.84 4.94 4.90 4.95 4.62 4.91 4.16 4.35 4.39 4.36 4.31 4.74 4.45 4.25 4.01 4.06 4.30 4.26 4.36 4.21 4.24 4.17 4.39
.51 .50 .87 .73 .73 .73 .73 1.08 .81 .83 .69 .48 .50 .50 1.03 .52 .52 .51 .51 .51 .52 .49 .50 .50 .50 .51
.10 .11 .07 .14 .13 .13 .14 .12 .08 .08 .07 .09 .08 .08 .17 .06 .09 .07 .07 .07 .07 .07 .07 .07 .08 .07
.06 .08 .04 .05 .06 .05 .06 .09 .04 .05 .05 .12 .10 .09 .04 .08 .05 .07 .09 .09 .09 .08 .09 .10 .08 .10
.62 .61 .60 .76 .76 .80 .78 .89 .55 .54 .47 .44 .42 .49 .90 .47 .49 .49 .44 .47 .47 .47 .46 .45 .46 .50
Note.
n = number of analyses. Values are in weight-percent oxide, recalculated to 100% on a fluid-free basis.
42
R. L. HERMANNS ET AL.
Cordillera Oriental (fig. 2a). In the Quebrada La Yesera (Cafayate section) and Casa de Los Loros sections (fig. 8g, 8h), the terrace is overlain by two generations of rock-avalanche deposits. In the Quebrada La Yesera, these rock avalanches are in turn overlain by the deposits of a landslide-dammed lake; stratigraphic relations show that this lake was most likely caused by the Casa de Los Loros rock avalanche farther downstream. The age of the Quebrada La Yesera Tuff is not well constrained; the AMS-14C age of mollusk shells from lake deposits define a minimum age of 32,480 5 150 yr (Trauth and Strecker 1999). The high degree of consolidation and deformation features suggests, however, that the terrace deposits and the intercalated Quebrada La Yesera Tuff may be older than the unconsolidated Cerro Paranilla Ash (see next section) sampled in a similar setting. Cerro Paranilla Ash. The ash (g05, g07, g25, g31) is relatively coarse grained and contains abundant biotite crystals up to 1 mm in diameter. The average TiO2 content of the biotites typically ranges from 4.8% to 4.9%, compared with values between 2.8% and 4.2% in biotites from other tephras. Glass shards are pumiceous with elongated, tubular vesicles and thick vesicle walls (fig. 7b). The ash also occurs in the Quebrada del Tonco and 90 km to the south (figs. 8f, 8j, 9a). In the Quebrada del Tonco (fig. 2a), it covers eroded Cretaceous strata and is in turn overlain in direct contact by rock-avalanche debris. Stratigraphically, this tephra layer is below the Tonco Ash. Immediately west of Cerro Paranilla (fig. 8f), the Cerro Paranilla Ash occurs within terrace gravel overlain by landslide debris. It is also found as a lens in reworked landslide deposits, which in turn are overlain by two younger landslide deposits. Biotites from the Cerro Paranilla Ash bed were dated with the 40Ar/39Ar method, resulting in a plateau age of 723 5 89 ka and a concordant isochron age of 763 5 136 ka. El Pen˜on Tuff. This tuff (g71) typically contains Table 3.
Figure 4. Eigenvalues and cumulative proportion of total variance explained by the first eight principal components (cf. table 3).
biotite-bearing pumice lapilli up to 2–3 cm in diameter (fig. 7c). It is poorly lithified and appears to be reworked as indicated by the large number of lithic fragments. The tuff forms a continuous layer, 2–3 cm thick, about 80 m below the top of an uplifted pediment remnant on the western piedmont of Sierra Laguna Blanca (fig. 8d). The 247 eastdipping pediment remnant contains a sequence of at least three rock-avalanche deposits. A 21Ne-exposure dating of the avalanche deposits results in minimum ages of ca. 140–200 ka for these deposits (Hermanns 1999; Niedermann and Hermanns 1999). However, since the rock-avalanche deposits do not show any evidence of deformation, the tilted pediment and intercalated El Pen˜on Lapilli Tuff are probably significantly older. Of similar geochemical composition is an unlithified ash lens (g61) with tubular pumiceous shards and elongated, tubular vesicles and thick vesicle walls. The ash occurs in a basin on top of the youngest of five landslide deposits (figs. 8a, 9b)
Tabular Output from the Principal Component Analysis
Principal component number
Eigenvalue of Cov (X)
1 2 3 4 5 6 7 8
4.36e 1.50e 9.97e 6.05e 3.41e 1.17e 6.57e 1.88e
1 1 2 2 2 2 2 2
00 00 01 01 01 01 02 02
Percentage variance captured this PC
Percentage variance captured total
54.45 18.73 12.46 7.57 4.26 1.47 .82 .24
54.45 73.19 85.65 93.22 97.48 98.94 99.76 100.00
Note. For each principal component, the table gives the value of the associated eigenvalue of the covariance matrix Cov (X) (i.e., correlation for autoscaled data), the percent variance captured by the PC, and the cumulative variance captured.
Journal of Geology
TEMPORAL DISTRIBUTION OF LARGE LANDSLIDES
Figure 5. Data points for the first versus second principal component (PC) calculated for individual ash layers from table 2.
in the western Aconquija piedmont (Hermanns and Strecker 1999). It lies 2 m below the top of the basin fill. In this area, the age is relatively defined by the few-decimeters-thick carbonate cementation. Pediment surfaces several tens of kilometers away with similarly advanced carbonatization have 40Ar/ 39 Ar and fission-track ages of 0.6–1.2 Ma (Strecker 1987). This ash is significantly older than the Buey Muerto and the Alemanı´a ashes because of its stratigraphic position. In contrast to the tephra sample from the Sierra Laguna Blanca pediment, this tephra is unlithified and does not contain lapilli fragments; it is believed to belong to the same eruptive event because of its geochemical composition and similar age at both localities. The state of lithification may be because of different depositional histories, and the difference in grain size because of the distance to eruptive centers. While Sierra Laguna Blanca lies close to active Quaternary volcanoes (some tens of kilometers), the distance from these centers to the Sierra Aconquija is 1100 km. Ruinas del Rinco´n Ash. The ash (g63) is exposed in an erosional cut at the base of a landslidedammed basin in the piedmont of Sierra Chango Real at the southeastern Puna border (fig. 2). The ash is therefore older than the Villa Vil Ash sampled from a surficial layer on the landslide debris (fig. 8c). It is unlithified and contains irregularly shaped pumiceous glass characterized by spherical to oval vesicles with thin walls (fig. 7d). The ash is 5 cm thick and has been reworked. El Paso Ash. The ash (g04, g17) occurs at two
43
localities (fig. 9c). It is unlithified and contains tubular pumiceous shards with elongated vesicles and thick vesicle walls. It overlies the older of two landslide deposits (fig. 8g). Biotites provide a 40Ar/39Ar plateau age of 54 5 221 ka, which is not significantly different from a zero age. The young age is corroborated by the age of mollusks from lake sediments, 32,480 5 150 yr, that also overlie the El Paso landslide. Because of the duration of this lake (roughly 3000 yr, estimated by varve counts; Kleinert et al. 1997), this ash has a maximum age of about 36 ka. The El Paso ash correlates geochemically with a tephra layer sampled 135 km to the south in the footwall area of the Sierra Aconquija reverse fault. At this locality, it documents a minimum offset of this fault of 10 m since deposition. Quebrada del Tonco Ash. The ash (g15, g29, g39) occurs as a 15-cm-thick continuous layer in a small basin on the landslide deposits in the Quebrada del Tonco (fig. 8j). It belongs to a widespread tephra layer of polymodal biotite composition, indicating retransportation (fig. 9d). The glass is pumiceous, tubular shaped, and contains elongated vesicles. About 200 km south at Villa Vil, the same unit exists as a lens in reworked landslide debris (fig. 8b). The chronostratigraphic position of the Que-
Figure 6. Dendrogram showing relationships between individual samples and sample groups, using Mahalanobis distances based on raw principal component scores of autoscaled EMA data. Data were log-ratio transformed using Al2O3 as divisor, after removal of outliers, in order to ensure independence (on an oxide-to-oxide basis). Lower values indicate more similar geochemical composition; larger values indicate more different composition.
44
Table 4.
R. L. HERMANNS ET AL.
Averaged Chemical Composition (and 1j Errors) of Tephra Groups Na2O
Quebrada La Yesera Tuff Cerro Paranilla Ash El Pen˜on Tuff El Paso Ash Quebrada del Tonco Ash Buey Muerto Ash Alemanı´a Ash
3.60 2.64 2.71 2.33
(.02) (.36) (.43) (.10)
SiO2 77.77 77.75 77.65 79.00
(.23) (.39) (.24) (.02)
MgO .11 .10 .05 .05
(.00) (.01) (.01) (.01)
Al2O3 13.28 12.95 13.06 12.99
(.08) (.18) (.50) (.12)
K2O 3.98 4.89 4.93 4.26
(.12) (.06) (.04) (.13)
CaO .50 .67 .84 .76
(.01) (.13) (.04) (.10)
TiO2 .11 .15 .08 .08
(.00) (.04) (.01) .01)
MnO .07 .05 .04 .05
(.01) (.01) (.00) (.00)
FeO .61 .80 .57 .50
(.00) (.07) (.04) (.05)
3.69 (.12) 77.83 (.10) .04 (.01) 13.01 (.15) 4.35 (.04) .49 (.01) .08 (.01) .10 (.01) .45 (.03) 2.52 (.66) 79.03 (.69) .05 (.00) 13.06 (.20) 4.19 (.20) .52 (.01) .07 (.01) .07 (.01) .47 (.02) 3.15 (.41) 78.26 (.38) .04 (.00) 13.11 (.15) 4.28 (.08) .51 (.01) .07 (.00) .09 (.01) .47 (.02)
brada del Tonco Ash is best defined in the Cafayate section (fig. 8g), where it is overlain by lake deposits containing mollusk shells dated at 32,480 5 150 yr. Villa Vil Ash. This redeposited ash (g33) occurs within sediments of a landslide-dammed lake at the southeastern Puna border (figs. 2a, 8b). The ash typically shows strong intergrowths of tiny biotites with irregularly shaped pumice. The pumiceous glass has spherical to oval vesicles and thin vesicle walls. AMS-14C-dated organic material from lake deposits 5 m above the Villa Vil Ash yields a minimum age of 1311 5 132 yr for both the tephra and the landslide (Fauque 2000). Buey Muerto Ash. This ash (g27, g49, g69, g73) belongs to an extensive deposit (fig. 9e) characterized by tubular pumiceous glass with elongated, tubular vesicles with thin walls. It has polymodal biotite compositions at all sites and occurs in the western piedmont of Sierra Aconquija (fig. 8a) within a terrace deposit that covers a rock-avalanche deposit. In the vicinity of Villa Vil (fig. 8b), an equivalent ash occurs within reworked landslide debris, whereas at the Cafayate section, it was found within reworked lake deposits (fig. 8g). The ash is constrained by peat deposits with an age of 3632 5 70 yr (Beta-122114) overlying it in direct contact, which are in turn overlain by the Alemanı´a Ash (fig. 8e). This age correlates with an ash sampled in the Lerma basin south of Salta (fig. 2a), dated at 4400 5 190 to 6300 5 250 yr by over- and underlying organic horizons, respectively (Malamud et al. 1996). Alemanı´a Ash. This ash (g21, g35, g37, g43, g45, g47, g65) is widespread in Holocene deposits (fig. 9f). It contains small amounts of tiny and thin biotites, which are difficult to analyze by EMA. Similar to other tephras, the pumiceous glass typically has elongated, tubular vesicles and thick vesicle walls. East of Cerro Paranilla (fig. 8f), it occurs about 1 m below the top of an alluvial-fan deposit. At Alemanı´a (fig. 2), it is 1 m below the top of a terrace gravel overlying lacustrine sediments related to a rockfall (fig. 8i). The maximum age of 3632 5 70 yr (confined by the dated underlying
peat deposit in the Cumbres Calchaquı´ basin; fig. 8e) is consistent with the age of charcoal of 4540 5 40 yr (Beta-129760) sampled about 1.20 m below the ash-bearing terrace gravel at Alemanı´a and with mollusk shells sampled from an unknown position of underlying lake sediments having an age of 5926 5 50 yr (fig. 8i; Wayne 1999).
Discussion In most cases, the occurrence of large landslides in the study area is restricted to two different environments: (1) narrow valleys with steep walls characterized by rock anisotropies oriented toward the valley and/or young thrust faults and (2) broad piedmont regions with alluvial fan–covered pediments in front of steep mountain fronts characterized by a variety of rock-strength anisotropies and active range-bounding faults (Hermanns and Strecker 1999; Strecker and Marrett 1999). The oldest landslides documented exist in unrestricted piedmont regions, whereas late Pleistocene and Holocene landslide clusters prevail in narrow and deeply incised valleys. Ages of Landslides. All landslide deposits in piedmont regions rest on top of alluvial fans or fancovered pediments. Possible tephra layers predating these landslides are most likely covered or eroded and were not found in shallow depth exposures. Tephra postdating the landslide deposits lie on the deposit surfaces or are, rarely, preserved together with other detritic and air-transported sediments in small basins behind the frontal rims of lobate deposits. The morphology of landslide source areas indicates that the piedmont landslide deposits are significantly older than the landslides clustered in narrow valleys (fig. 10). Typically, breakaway scarps and sliding surfaces are eroded or poorly preserved. The inference of an old age for these deposits is also supported by the degree of soil development on the avalanche deposits. For example, the frontal rims and lateral levees of the Aconquija slides are cemented by several decimeters of carbonate hori-
Journal of Geology
TEMPORAL DISTRIBUTION OF LARGE LANDSLIDES
45
Figure 7. Scanning electron micrographs of glass shards of ash units g67, g07, and g63, and field photograph of the lapilli tuff g71. a, Mostly bubble wall Y-junction shards and a few bubble wall shards with large radius or curvature. b, Tubular pumiceous shard with elongated, tubular vesicles or “capillaries” and thick vesicle walls. c, Lapilli tuff with 2–3-cm large lapilli (arrows indicate thickness of tephra layer). d, Irregularly shaped pumiceous shard with spherical to oval vesicles and thin vesicle walls.
zons, related to eolian carbonate input and indicating an age of several hundred thousand years (Fauque and Strecker 1988). Similar horizons associated with tephra incorporated in old pediment surfaces in the neighboring Calchaquı´ Valley were dated by the 40Ar/39Ar and fission-track methods and are on the order of 0.6–1.2 Ma (Strecker 1987). Cosmogenic-nuclide dating of a series of landslides on the western Sierra Laguna Blanca piedmont has resulted in exposure ages of 389 1 17/ 2 24 to 139 1 16/ 2 22 ka, also documenting greater ages for landslide deposits in piedmont regions (fig. 10; Hermanns 1999; Niedermann and Hermanns 1999). Multiple landslides in narrow valleys perpendicular or oblique to structural trends commonly occur together with terrace and lake deposits. Pre-
served erosional remnants of tephra-bearing deposits are as old as 723 5 89 ka (Cerro Paranilla), but such deposits are rare exceptions. Most avalanche deposits in these valleys cluster into two distinct time periods (fig. 10). The earlier and quantitatively more important cluster occurs at about 33,000 yr. In the Cafayate section, the older of two landslide deposits is overlain in up- and downstream directions by lake sediments. The lake sediments upstream have an age of 32,480 5 150 yr (Trauth and Strecker 1999). These sediments are in turn covered by a younger landslide mass that contains injection dikes of lake sediment, indicating that the sediments at that time were still water saturated (Hermanns and Strecker 1999). Varve counts in an undeformed lacustrine section suggest that the lake existed for about 3000 yr before the
Figure 8. Composite profiles of 10 localities (indicated in fig. 2), showing the stratigraphic relations of different deposits and tephra layers. All sections are drawn in the same scale, except e. Note that tephra lenses sampled from redeposited landslide debris (e.g., b) postdate stratigraphically lower landslides but do not necessarily predate stratigraphically higher landslides.
Journal of Geology
TEMPORAL DISTRIBUTION OF LARGE LANDSLIDES
47
Figure 9. Sample sites and minimum spatial distribution (hachured areas) of six tephra layers (a–f). Thickness of tephra layer is given in brackets in all cases where the tephra exists as a continuous layer for several meters. Shaded areas correspond to regions at 3000 m elevation. Crosses denote sampling localities.
second landslide covered the unconsolidated sediment (Kleinert et al. 1997). Landslide deposits from the adjacent Casa de Los Loros section must be relatively coeval because the lake sediments in the Cafayate section from the downstream direction resulted from a barrier located at the Casa de Los Loros section (Hermanns and Strecker 1999). Landslides at Villa Vil and in the Quebrada del Tonco also seem to correlate with this time of enhanced landsliding activity. The deposits are overlain by
the Quebrada del Tonco Ash, which postdates the older landslide deposits in the Cafayate section, giving a minimum age for these deposits. The interpretation that these minimum ages are close to the true age is supported by morphologic observations of the breakaway zones of the landslides. All of these landslides and two landslides from the Cerro Paranilla area were generated in coarse clastic sedimentary rocks, which dip toward the valley. Breakaway scarps are well defined, and sliding sur-
48
R. L. HERMANNS ET AL.
Figure 10. Volcanic ash, radiocarbon, and cosmogenic-nuclide stratigraphy for the studied landslides. Sequences are based on the chemical similarities of individual ashes, which displayed high degrees of correlation throughout all the statistical analyses. Ca. = Casa, Q. = Quebrada, S. = Sierra, C. = Cumbres.
faces are only weakly eroded, attesting to a relatively youthful age. In addition, multiple breakaway surfaces and two separate lacustrine units caused by landslides in the Quebrada del Toro dated at 33,550 5 190 and 29,580 5 130 yr also suggest pronounced landsliding activity in northwestern Argentina during this period (Trauth and Strecker 1999). Ages of these landslide dams are probably within an error range of several hundred to a few thousand years, similar to the age of the related lacustrine deposits. This interpretation is based on observations from the Cafayate sections and empirical data of landslide-dammed lakes showing that 85% of landslide-dammed lakes last !1 yr (Costa and Schuster 1988). The younger, less pronounced cluster of landslides is Holocene in age and is made up of the youngest rock avalanche near Villa Vil, the large rock fall at Alemanı´a, and possibly a rock avalanche
deposit at Brealito (fig. 10). Near Villa Vil, the landslide was dated with a minimum age of 1311 5 132 yr on organic material from a small landslidedammed lake (Fauque 2000). The lake deposit and the Alemanı´a rockfall are overlain by the ∼3700-yr Alemanı´a Ash. The landslide at Brealito is also inferred to be of Holocene age because of the pristine morphology of the breakaway scarps, intact sliding surfaces, and depositional morphology; 1-cm-deep and sharp-edged grooves, produced by the movement of the rock mass on the smooth sliding surface, are perfectly preserved. This surface provides a 21Ne-exposure age of !24,000 yr. However, this age is considered to be too old, as no cosmogenic Ne was detected unequivocally. All excess 21Ne could be characterized as radiogenic Ne; hence, this age represents the maximum analytical error (2j). Coeval with this age group may also be a giant granitic rockfall deposit with pristine surface mor-
Journal of Geology
TEMPORAL DISTRIBUTION OF LARGE LANDSLIDES
phology which was generated 14 km northwest of Brealito (fig. 2a). Significance of Temporal Landslide Distribution.
The pronounced differences in the temporal distribution of landslides in narrow valleys versus broad piedmont regions result from distinct conditioning parameters in both environments. Tectonically and lithologically, both settings are similar and prone to generating large landslides. Evidence for repeated Quaternary fault movements along steep mountain fronts in combination with a variety of rockstrength anisotropies suggests that paleoseismicity played an important role in generating large rock avalanches in these environments (Hermanns and Strecker 1999). However, the late Pleistocene and Holocene clusters of landslides in northwestern Argentina are coeval with increased humidity and runoff in this part of the Andes (e.g., van der Hammen and Absy 1994; Grosjean et al. 1997; Turcq et al. 1997; Trauth and Strecker 1999). With respect to the total number, the most important landslide clusters in valleys lie in the Quebrada de Cafayate and in the Quebrada del Toro. Both valleys are characterized by large catchment areas of up to 19,800 km2 and important runoff (EVARSA 1994; Malamud et al. 1996) originating in regions with elevations 14000 m that were also affected by multiple late Pleistocene glaciations (Fox and Strecker 1991). Glacial advances could be dated in the Cordillera Oriental at ca. 27,970 5 190 and after 5280 5 200 yr (Zipprich et al. 1998). Higher runoff in the course of climate change would have resulted in enhanced scouring, undercutting, and landsliding along the structurally preconditioned mountain fronts and valley walls. Similar relations have been demonstrated for enhanced landsliding activity in other regions of the world (e.g., Reneau and Dethier 1996). Furthermore, greater humidity and enhanced seasonality may have increased groundwater flow, which increased seepage body forces and lowered critical thresholds in rocks susceptible to failure by changing the effective stress (Trauth and Strecker 1999). Under such circumstances, in addition to scouring, landsliding could have been aseismic or triggered by even smaller earthquakes with lower levels of ground motion. The multiple occurrences of young landslide deposits in these valleys is remarkable, but the quantity of preserved landslide deposits in the stratigraphic record may not represent the total number of Quaternary landslide deposits in these environments. Because of the highly dynamic environments in these narrow valleys, older deposits may have been removed. For example, cut-and-fill terrace deposits upstream from the approximately
49
29,500-yr landslide deposits in the Quebrada del Toro attest to repeated temporary base-level changes in this area that may be caused by a repeated already-obliterated landslide barrier. Although there are no unambiguous relations between these deposits and landslide remnants, the spatially limited occurrence of multiple-terraced deposits immediately upstream from the landslide area suggests a causative relationship. The dominant occurrence of late Pleistocene and Holocene landslides is therefore a combined result of increased mass movements in the deeply incised valleys of the Cordillera Oriental at these times and the low preservation potential of older deposits of similar origin. Although the ages of piedmont-region landslides are more difficult to define, these landslide deposits have a higher preservation potential. It is likely that these environments are less sensitive to climatic change, as no clear correlation of landslide occurrence with humid climates exists. The landslide ages also do not correlate with any other climatedriven morphologic changes in the Argentinian Andes (Siame et al. 1997). Because of the large distance between the avalanche sites and trunk streams, the higher discharge of these rivers during wetter climatic conditions did not affect mountain fronts in this setting. In addition, mountain-front failure in this environment is not correlated with glacial processes, such as in the Alps, for example (e.g., Porter and Orombelli 1981), because mountain-front failure occurred on slopes below the glaciated parts of the ranges or because avalanche deposits are situated along mountain fronts without any sign of glacial overprint. Instrumentally recorded large earthquakes are rare in this region (Chinn and Isacks 1983). However, this region is characterized by the occurrence of frequent low-magnitude events (Cahill et al. 1992; Assumpc¸a˜o and Araujo 1993). Nevertheless, large prehistoric earthquakes must have occurred, as documented by the formation of Quaternary fault scarps along reverse-fault-bounded mountain fronts and within late Pleistocene fan deposits (Strecker et al. 1989; Cahill et al. 1992; Marrett et al. 1994; Hermanns 1999). For example, the maximal 36,000-yr-old El Paso Ash, sampled from the southern Sierra Aconquija reverse fault, indicates a minimum displacement of 10 m, suggesting that several strong, shallow-seated earthquakes have occurred along this mountain front. Because of the low frequency of large landslides at mountain fronts with broad piedmonts, it is plausible that cliff collapses in this environment were most likely triggered by large earthquakes with long recurrence
50
R. L. HERMANNS ET AL.
intervals. We cannot unambiguously prove the seismic origin of these landslides since no other independent dateable paleoseismic features are preserved. This is in line with observations by Densmore and Hovius (1999), who suggested that seismic events focus landsliding to ridge crests and hilltops where effects of fluvial erosion are negligible, while climatic events trigger landslides near hillslope toes. Conclusions Two settings with distinctive temporal distribution for generating large landslides exist in northwestern Argentina: narrow valleys and broad piedmont regions. While landslides in narrow valleys can often be accurately dated by tephrochronology, occur in temporal clusters, and are relatively young, the age of landslides in piedmont regions is more difficult to constrain because of the different depositional environment. In general, however, large landslides in piedmont regions are significantly older, and a tendency toward temporal clustering in the late Pleistocene and Holocene is not observed. We suggest that the young landslide clusters and higher landsliding frequency in narrow valleys represent effects of climatic change affecting the distant high-catchment areas. In contrast, large landslides along tectonically active mountain fronts paralleling broad piedmont regions cannot be correlated with climate change; for these regions, we suggest that the more likely trigger mechanism was low-frequency/high-magnitude seismic events. Landslides in these settings may thus represent effects of important paleoseismic events. At a larger regional scale, the tephra associated
with landslide-related deposits in northwestern Argentina reveals an explosive Quaternary eruptive history of the CVZ. Although the distal tephra record analyzed here is not complete and mainly involves late Pleistocene deposits, at least nine important rhyodacitic eruptive events are recorded. The identification and correlation of tephra thus provides a chronostratigraphic framework for Quaternary deposits in northwestern Argentina and defines three important temporal marker beds at 723 5 89, ∼30, and ∼3.7 ka.
ACKNOWLEDGMENTS
This work is part of the Collaborative Research Center 276 “Deformation Processes in the Andes” supported by the German Research Council in a grant to M. Strecker. The success of this project was made possible through the support of many friends and colleagues, namely B. Aban, F. Aban, R. Alonso, A. Bourdin, I. Capdevila, L. Fauque, R. Gonzales, F. Hongn, J. J. Marcuzzi, J. A. Salfity, A. Villanueva, and J. Viramonte. EMA and SEM were carried out at the GFZ Potsdam and at the Naturhistorisches Museum Berlin. We appreciate help by O. Appelt, P. Claeys, D. Rhede, and E. Wa¨sch. C. Fischer prepared most samples for analysis. R. Naumann (GFZ Potsdam) performed XRF and infrared absorption analyses. The AMS radiocarbon date was performed by Beta Analytic, Miami, Florida. We thank N. B. Gallagher and B. M. Wise from Eigenvector Research, Manson, Wash., for support while using the Matlab PLS Toolbox. We gratefully acknowledge thorough reviews by T. Jordan and three anonymous reviewers.
REFERENCES CITED
Adams, J. 1981. Earthquake-dammed lakes in New Zealand. Geology 9:215–219. Aitchison, J. 1984. The statistical analysis of geochemical composition. Math. Geol. 16:531–564. ———. 1986. The statistical analysis of compositional data. New York, Wiley. Allmendinger, R. W.; Ramos, V. A.; Jordan, T. E.; Palma, M.; and Isacks, B. L. 1983. Paleogeography and Andean structural geometry, northwest Argentina. Tectonics 2:1–16. Assumpc¸a˜o, M., and Araujo, M. 1993. Effect of the Altiplano-Puna plateau, South America, on the regional intraplate stresses. Tectonophysics 221:475–496. Bogaard, P., and Schmincke, H. U. 1985. Laacher See Tephra: a widespread isochronous late Quaternary te-
phra layer in central and northern Europe. Geol. Soc. Am. Bull. 96:1554–1571. Cahill, T.; Isacks, B. L.; Whitman, D.; Chatelain, J.-L.; Perez, A.; and Ming Chiu, J. 1992. Seismicity and tectonics in Jujuy Province, northwestern Argentina. Tectonics 115:944–959. Caminos, R. 1979. Sierras Pampeanas noroccidentales, Salta, Tucuman, Catamarca, La Rioja y San Juan (second). Simp. Geol. Rep. Arg. Acad. Nac. Ciencias, p. 225–291. Cerling, T. E.; Brown, F. H.; and Bowman, J. R. 1985. Lowtemperature alteration of volcanic glass: hydration, Na, K, 18O and Ar mobility. Chem. Geol. 52:281–293. Chinn, D. S., and Isacks, B. L. 1983. Accurate source depths and focal mechanisms of shallow earthquakes
Journal of Geology
TEMPORAL DISTRIBUTION OF LARGE LANDSLIDES
in western South America and in the New Hebrides island arc. Tectonics 2:529–563. Coira, B., and Kay, S. M. 1993. Implications of Quaternary volcanism at Cerro Tuzgle for crustal and mantle evolution of the Puna Plateau, Central Andes, Argentina. Contrib. Mineral. Petrol. 113:40–58. Costa, J. E., and Schuster, R. L. 1988. The formation and failure of natural dams. Geol. Soc. Am. Bull. 100: 1054–1068. Crozier, M. J.; Deimel, M. S.; and Simon, J. S. 1995. Investigations of earthquake triggering for deep-seated landslides, Taranaki, New Zealand. Quat. Int. 25: 65–73. Densmore, A., and Hovius, N. 1999. The topographic signature of bedrock landsliding. Conference of the European Union of Geosciences, Strasboug. J. Conf. Abstr. 4:445. De´ruelle, B. 1978. Calc-Alkaline and shoshonitic lavas from five Andean volcanoes (between latitudes 217459 and 247309S) and the distribution of the Plio-Quaternary volcanism of the south-central and southern Andes. J. Volcanol. Geotherm. Res. 3:281–298. de Silva, S. L. 1989. Altiplano-Puna volcanic complex of the central Andes. Geology 17:1102–1106. ———. 1991. Styles of zoning in Central Andean ignimbrites: insights into magma chamber processes. In Harmon, R. S., and Rapela, C. W., eds. Andean magmatism and its tectonic setting, Boulder, Colorado. Geol. Soc. Am. Spec. Pap. 265:217–232. de Silva, S. L., and Francis, P. W. 1991. Volcanoes of the Central Andes. Berlin, Springer, p. 1–216. EVARSA (Evaluacio´n de Recursos Sociedad Anonima). 1994. Estadı´stica Hidrolo´gica (Vols. 1, 2). Buenos Aires, Secretarı´a de Energı´a, p. 1–651. Fauque, L. 2000. Villavil rockslides, Catamarca Province, Argentina. In Evans, S. G., and DeGraff, J. V., eds. Catastrophic landslides. Boulder, Colorado. Geol. Soc. Am. Rev. Eng. Geol., in press. Fauque, L., and Strecker, M. R. 1988. Large rock avalanche deposits (Sturzstro¨me, sturzstroms) at Sierra Aconquija, northern Sierras Pampeanas, Argentina. Eclogae Geol. Helv. 81:579–592. Felpeto, A., and Ortiz, R. 1997. Ash dispersion model of the 1993 Lascar eruption. International Association of Volcanology and Chemistry of the Earth’s Interior (IAVCEI) Volcanic Activity and the Environment, abstr. Puerto Vallarta, Mexico, p. 157. Fisher, R. V., and Schmincke, H.-U. 1984. Pyroclastic rocks. Springer, Berlin, 472 p. Fox, A. N., and Strecker, M. R. 1991. Pleistocene and modern snowlines in the Central Andes (24–287S). Bamberg. Geogr. Schriften 11:169–182. Francis, P. W., and de Silva, S. L. 1989. Application of the Landsat Thematic Mapper to the identifiation of potentially active volcanoes in the central Andes. Remote Sens. Environ. 28:245–255. Francis, P. W.; Sparks, R. S. J.; Hawkesworth, C. J.; Thorpe, R. S.; Pyle, D. M.; Tait, S. R.; Mantovani, M. S.; and McDermott, F. 1989. Petrology and geochem-
51
istry of volcanic rocks of the Cerro Galan caldera, northwest Argentina. Geol. Mag. 126:515–547. Frogatt, P. C. 1992. Standardization of the chemical analysis of tephra deposits: report of the ICCT working group. Quat. Intern. 13/14:93–96. Glaze, L. S.; Francis, P. W.; Self, S.; and Rothery, D. A. 1989. The 16 September 1986 eruption of Lascar volcano, North Chile: satellite investigations. Bull. Volcanol. 51:149–160. Gonza´lez Dı´az, E. F., and Mon, R. 1996. El Origin de las Lagunas de Yala, Provincia de Jujuy (247059 de latitud y 657289 de longitud oeste). 13th Congr. Geol. Argentino y 3d Congr. Explor. Hidrocarb. Actas 4:209–217. Grier, M. E., and Dallmeyer, R. D. 1990. Age of the Payogastilla Group: implications for foreland basin development, NW Argentina. J. S. Am. Earth Sci. 3: 269–278. Grier, M. E.; Salfity, J. A.; and Allmendinger, R. W. 1991. Andean reactivation of the Cretaceous Salta rift, northwestern Argentina. J. S. Am. Earth Sci. 4: 351–372. Grosjean, M.; Valero-Garce´s, B. L.; Geyh, M. A.; Messerli, B.; Schotterer, U.; Schreier, H.; and Kerry, K. 1997. Mid- and late-Holocene limnology of Laguna del Negro Francisco, northern Chile, and its palaeoclimatic implications. Holocene 7:151–159. Hermanns, R. L. 1999. Spatial-temporal distribution of mountain-front collapse and formation of giant landslides in the arid andes of northwest Argentina (24–287 S, 65–687 W). Unpub. Ph.D. thesis, Universita¨t Potsdam, Potsdam, 123 p. Hermanns, R. L., and Strecker, M. R. 1999. Structural and lithological controls on large Quaternary rock avalanches (sturzstroms) in arid northwest Argentina. Geol. Soc. Am. Bull. 111:934–948. Jordan, T. E.; Isacks, B. L.; Allmendinger, R. W.; Brewer, J. A.; Ramos, V. A.; and Ando, C. J. 1983. Andean tectonics related to geometry of subducted Nazca plate. Geol. Soc. Am. Bull. 94:341–361. Keefer, D. K. 1984. Landslides caused by earthquakes. Geol. Soc. Am. Bull. 95:406–421. Kleinert, K.; Trauth, M. H.; and Strecker, M. R. 1997. A Pleistocene lacustrine phase in the Quebrada de Cafayate, NW Argentina: evidence for variations in paleohydrology from multi-tracer analysis. Terra Nova 9:631. Malamud, B. D.; Jordan, T. E.; Alonso, R. A.; Gallardo, E. F.; Gonzalez, R. E.; and Kelley, S. A. 1996. Pleistocene lake Lerma, Salta Province, NW Argentina. 13th Congr. Geol. Argentino y 3d Congr. Expl. Hidrocarb. Actas 4:103–114. Marrett, R. A.; Allmendinger, R. W.; Alonso, R. N.; and Drake, R. E. 1994. Late Cenozoic tectonic evolution of the Puna Plateau and adjacent foreland, northwestern Argentine Andes. J. S. Am. Earth Sci. 7: 179–207. Marshall, L. G., and Patterson, B. 1981. Geology and geochronology of the mammal-bearing Tertiary of the Valle de Santa Marı´a and Rı´o Corral Quemado, Catamarca Province, Argentina. Fieldiana Geol. 9:80 p.
52
R. L. HERMANNS ET AL.
Mon, R. 1976. The structure of the eastern border of the Andes in northwestern Argentina. Geol. Rundsch. 65: 211–222. ———. 1979. Esque´ma tectonico de los Andes del norte Argentino. Rev. Asoc. Geol. Argent. 34:53–60. Niedermann, S., and Hermanns, R. L. 1999. The chronology of giant landslides at Sierra Laguna Blanca, Argentina, as deduced from cosmogenic 21Ne. Geophys. Res. Abstr. 1:44. Nikonov, A. A., and Shebalina, T. Y. 1979. Lichenometry and earthquake age determination in central Asia. Nature 280:675–677. Omarini, R. H. 1983. Caracterizacio´n litolo´gica, diferenciacio´n y ge´nesis de la Formacio´n Puncoviscana entre el Valle de Lerma y la Faja Eruptiva de la Puna. Ph.D. thesis, Universidad de Salta, Salta, 220 p. Plafker, G., and Ericksen, G. E. 1978. Nevados Huascara´n avalanches, Peru. In Voight, B., ed. Rockslides and avalanches (Vol. 1). Amsterdam, Elsevier, p. 277–314. Porter, S. C., and Orombelli, G. 1981. Alpine rockfall hazards. Am. Sci. 69:67–75. Reneau, S. L., and Dethier, D. P. 1996. Late Pleistocene landslide-dammed lakes along the Rio Grande, White Rock Canyon, New Mexico. Geol. Soc. Am. Bull. 108: 1492–1507. Reyes, F. C., and Salfity, J. A. 1973. Consideraciones sobre la estratigrafı´a del Cretacico (Subgrupo Pirgua) del noroeste Argentino. 5th Congr. Geol. Argent. Actas, p. 355–385. Salfity, J. A., and Marquillas, R. A. 1994. Tectonic and sedimentary evolution of the Cretaceous-Eocene Salta Group Basin, Argentina. In Salfity, J. A., ed. Cretaceous tectonics of the Andes: earth evolution sciences. Braunschweig-Wiesbaden, Vieweg, p. 266–315. Sarna-Wojcicki, A. M. 1976. Correlation of Late Cenozoic tuffs in the Central Coast Ranges of California by means of trace- and minor-element composition. U.S. Geol. Surv. Prof. Pap. 972. Sarna-Wojcicki, A. M.; Morrison, S. D.; Meyer, C. E.; and Hillhouse, J. W. 1987. Correlation of upper Cenozoic tephra layers between sediments of the western United States and eastern Pacific Ocean and comparison with biostratigraphic and magnetostratigraphic age data. Geol. Soc. Am. Bull. 98:207–223. Schuster, R. L.; Logan, R. L.; and Pringle, P. T. 1992. Prehistoric rock avalanches in the Olympic Mountains, Washington. Science 258:1620–1621. Shreve, R. L. 1966. Sherman landslide, Alaska. Science 154:1639–1643. Siame, L. L.; Bourle`s, D. L.; Se´brier, M.; Bellier, O.; Castano, J. C.; Araujo, M.; Perez, M.; Raisbeck, G. M.; and Yiou, F. 1997. Cosmogenic dating ranging from 20 to 700 ka of a series of alluvial fan surfaces affected by the El Tigre fault, Argentina. Geology 25:975–978. Smith, R. P., and Nash, W. P. 1976. Chemical correlation of volcanic ash deposits in the Salt Lake Group, Utah, Idaho and Nevada. J. Sediment. Petrol. 40:930–939.
Stokes, S.; Lowe, D. J.; and Frogatt, P. C. 1992. Discriminant function analysis and correlation of Late Quaternary rhyolithic tephra deposits from Taupo and Okataina volcanoes, New Zealand, using glass shard major element composition. Quat. Intern. 13/14: 103–117. Strecker, M. R. 1987. Late Cenozoic landscape development, the Santa Maria valley, northwestern Argentina. Unpub. Ph.D. thesis, Cornell University, Ithaca, N.Y. 262 p. Strecker, M. R.; Cerveny, P.; Bloom, A. L.; and Malizzia, D. 1989. Late Cenozoic tectonism and landscape development in the foreland of the Andes: Northern Sierras Pampeanas (267–287S), Argentina. Tectonics 8: 517–534. Strecker, M. R., and Marrett, R. A. 1999. Kinematic evolution of fault ramps and role in development of landslides and lakes in the northwestern Argentine Andes. Geology 27:307–310. Stuiver, M., and Reimer, P. J. 1993. Extended 14C data base and revised CALIB 3.0 14C ages calibration program. Radiocarbon 35:215–230. Swan, A. R. H., and Sandilands, M. 1995. Introduction to geological data analysis. Cambridge, Mass., Blackwell, p. 1–446. Thorpe, R. S.; Francis, P. W.; Hammill, M.; and Baker, M. C. 1982. The Andes. In Thorpe, R. S., ed. Andesites. New York, Wiley, p. 187–205. Trauth, M. H., and Strecker, M. R. 1999. Formation of landslide-dammed lakes during a wet period between 40,000–25,000 yr B.P. in northwestern Argentina. Palaeogeogr. Palaeoclimatol. Palaeoecol. 153:277–287. Turcq, B.; Pressinotti, M. M. N.; and Martin, L. 1997. Paleohydrology and paleoclimate of the past 33,000 years at the Tamandua´ River, Central Brazil. Quat. Res. 47:284–294. van der Hammen, T., and Absy, M. L. 1994. Amazonia during the last glacial. Palaeogeogr. Palaeoclimatol. Palaeoecol. 109:247–261. Wayne, W. J. 1994. Large Andean landslides: indicators of prehistoric seismic epicentres. Geol. Soc. Am. Abstr. Prog. 26:A-216. ———. 1999. The Alemanı´a rockfall dam: a record of Mid-Holocene earthquake and catastrophic flood in northwestern Argentina. Geomorphology 27:295–306. Westgate, J. A., and Gorton, M. P. 1981. Correlation techniques in tephra studies. In Self, S., and Sparks, R. S. J., eds. Tephra studies. Dordrecht, Reidel, p. 73–94. Zipprich, M.; Reizner, B.; Veit, H.; Zech, W.; and Stingl, H. 1998. Upper Quaternary climate and landscape evolution in the Sierra de Santa Victoria (Cordillera Oriental, northwestern Argentina) deduced from geomorphologic and pedologic studies. 16. Geowissenschaftliches Lateinamerika Kolloquium. Terra Nostra 98:180–181.
