Introduction

The Trans-Mexican Volcanic Belt (TMVB) is an E–W trending zone located between 19° and 20°N latitude, extending ca. 1,000 km from the Pacific to the Gulf of Mexico (Fig. 1). Its origin is related to the subduction of the Cocos Plate beneath the North American Plate. Pelado, Guespalapa, and Chichinautzin monogenetic scoria-and-cinder cones are located within the Sierra del Chichinautzin Volcanic Field (SCVF) in the central part of the TMVB. The SCVF is a volcanic highland elongated in an E–W direction, extending from the flanks of Popocatépetl (Popo) stratovolcano in the east to the edge of the Toluca valley in the west (Fig. 1). This volcanic field represents the volcanic front of the TMVB in this area and is on the continental drainage divide that separates the Basin of Mexico City from the valleys of Cuernavaca and Cuautla to the south. According to Fries (1962), the Basin of Mexico drained to the south before the Pleistocene. Since then, formation of the SCVF sealed the basin to the south (Mooser 1963).

Fig. 1.
figure 1

Sketch map showing the SCVF and location of Pelado, Guespalapa, Chichinautzin, and other monogenetic cones mentioned in the text. Major stratovolcanoes, archaeological sites, and cities are also shown

Full size image

The SCVF has one of the highest concentrations of monogenetic volcanoes in the entire TMVB, and includes visibly at least 220 overlapping Quaternary cinder cones, associated lava flows, tephra sequences, and lava shields intercalated with alluvial sediments that cover an area of approximately 2,500 km2 (Bloomfield 1975; Martin 1982; Lugo-Hubp 1984). Volcanic rocks in this area are mostly andesites with subordinate basalts and dacites that define a calc–alkaline series (Gunn and Mooser 1971; Swinamer 1989; Wallace and Carmichael 1999).

Individually, the monogenetic volcanoes forming the SCVF have small volumes, but collectively, they cover a wide area and are more voluminous than most composite volcanoes within the TMVB. Bloomfield (1975) reported a cinder cone density of 0.1/km2 and Swinamer (1989) estimated a total volume of 1,600 km3 for the entire SCVF. This high volume might be in error because it implies a thickness of >600 m for the entire Quaternary volcanic pile. Unfortunately, borehole measurements revealing the thickness of the pile at different locations are not yet available.

Lava flows within the SCVF vary considerably in their morphology. Most are compound aa-flows, less commonly blocky, and rarely pahoehoe flows. Paleomagnetic measurements on volcanic rocks from the SCVF (Mooser et al. 1974; Herrero and Pal 1978; Urrutia and Martin 1993) indicate that most exposed rocks were produced during the normal Brunhes Chron and are therefore younger than 0.73–0.79 Ma. This is not surprising, since most of the cinder cones and lavas display very young morphological features and are covered by and intercalated with poorly developed soils.

Most research in the SCVF has concentrated on Xitle volcano and associated flows which underlie the southern part of Mexico City (e.g., Siebe 2000). Xitle has been studied in detail because its lavas destroyed and buried the ancient pre-Hispanic town of Cuicuilco, located 7 km NE of the cone (e.g., Cummings 1923; 1933). Libby (1955) dated charcoal beneath flows from Xitle for the first time at 2,422±250 years b.p. Since then, organic materials below the products from Xitle's eruption have been dated by the radiocarbon method by several authors. Although most published radiocarbon dates cluster around 2,000 years b.p. (e.g., Heizer and Bennyhoff 1958; Crane and Griffin 1958), more recent investigations (González et al. 2000; Siebe 2000) indicate that Xitle erupted around 1,670±35 years b.p. and that Cuicuilco was probably abandoned as a direct consequence of this eruption.

Xitle has been traditionally regarded by most scientists as the youngest volcano within the SCVF that had an impact on the basin of Mexico and was witnessed by humans (e.g., Scandone 1979). Exceptions from this view include Delgado et al. (1997), Rodríguez-Lara (1997), and Arana-Salinas (1998), who provided radiocarbon dates indicating that Pelado, Guespalapa, and Chichinautzin cones are younger than Xitle. Indeed, Landsat Thematic mapper images (Fig. 2) show that the morphology of lava flows and other features of these volcanoes could be younger than Xitle.

Fig. 2.
figure 2

Landsat Thematic Mapper satellite image of the central part of the SCVF showingP Pelado , G Guespalapa, and Ch Chichinautzin monogenetic volcanoes. T Texcal lava flow, Kc Cretaceous limestones, Tf Tertiary Tepoztlán Formation, and C city of Cuernavaca are also shown. Arrows indicate a break in slope produced by an inferred E–W trending fault that forms the southern limits of the Sierra Chichinautzin horst

Full size image

The main purpose of the present study was to determine with greater confidence the radiocarbon ages of these three volcanoes. In addition, we estimated volumes and areas covered by their lava flows, determined the possible impact of their eruptions on pre-Hispanic populations, and prepared a preliminary risk map that considers information gathered during this study. The present study is part of a broader project aimed at establishing a detailed tephrochronology of the area and determining the recurrence intervals and probable magnitude of future monogenetic eruptions that could affect Mexico City and its infrastructure.

Morphology, area covered, and volume of erupted products

Pelado, Guespalapa, and Chichinautzin were targeted in the present study because of their youthful morphologies. They are closely spaced and located in the topographically highest central part of the SCVF near the pass between Mexico City and Cuernavaca (Figs. 3 and 4). Their vents are situated on the continental drainage divide at altitudes above 3,400 m. For this reason some of their lavas drained into the Basin of Mexico (Atlantic) while others flowed towards the valley of Cuernavaca (Pacific). Chemical analyses (Siebe et al. 2003) revealed that Pelado is quite homogeneous and composed of high-Si andesite (ca. 60 wt% SiO2) while Guespalapa and Chichinautzin show a much wider compositional variation ranging from basalt to basaltic andesite (48–56 wt% SiO2).

Fig. 3.
figure 3

Schematic geologic map of the central portion of the SCVF where the monogenetic cones Pelado, Guespalapa, and Chichinautzin are located. Points where radiometrically dated charcoal samples were obtained are also shown

Full size image
Fig. 4.
figure 4

Sketch map showing the Basin of Mexico and location of pre-Hispanic settlements. The approximate extent of the Texcoco lake system is also shown

Full size image

Rodríguez-Lara (1997) produced a geologic map of the area (Fig. 3), calculated the areas covered by lava flows, and estimated volumes of erupted materials (Table 1). In addition, he determined different morphological parameters that characterize the cones following the methods described by Porter (1972) and Wood (1980a; 1980b). The areas covered by the scoria cones and lava flows were measured using a compensating polar planimeter together with topographic maps and aerial photographs. For simplicity, the volumes of the cinder cones were crudely calculated using the formula of a right circular cone where π r2 equals the area of the cone base (neither the apex of the cones, nor the volume occupied by vesicles or interstices between clasts were considered). The volumes of the lava flows were calculated using the aerial measurement determined with the planimeter multiplied by an average of thickness measurements at flow margins, as well as at a few available road cuts.

Table 1. Morphometric parameters for Pelado, Guespalapa, and Chichinautzin monogenetic volcanoes
Full size table

Porter (1972) established the following relations between basal cone width (Wco), crater diameter (Wcr), and cone height (Hco) for fresh scoria cones on Mauna Kea: Hco=0.18 Wco and Wcr=0.40 Wco. These relations, originally derived from cones in a single location are apparently valid for scoria cones in diverse tectonic settings and with a range of compositions and particle sizes (Wood 1980a). Normally, fresh cones composed of 2–10 cm diameter scoria have slopes of 32–33°, reflecting the equilibrium angle of repose of such materials (Porter 1972). Redistribution of material by weathering and mass wasting preferentially removes the top of the cone and deposits this material around the flanks. This results in a continual decrease in flank slope. Cone heights, height/diameter ratios, and slopes systematically decrease as cones degrade, offering the possibility of roughly dating cones by simple measurements of their morphology (Wood 1980b). Because of their great differences in eruptive rate, chemical composition, and related differences in magma viscosities, Pelado, Guespalapa, and Chichinautzin lavas vary greatly in their morphology. With this background in mind, Pelado, Guespalapa, and Chichinautzin will be discussed individually below.

Pelado

Pelado (3,620 m a.s.l.) is one of the most perfectly shaped cones in the entire SCVF, with a height (Hco) of 180 m and a basal diameter (Wco) of 1,050 m (Fig. 5). This yields a Hco/Wco ratio of 0.17, which deviates only slightly from the ideal ratio of 0.18 for fresh cones. In comparison, Martin (1982) and Swinamer (1989) reported Hco values of 150 and 200 m for Pelado, respectively. As a result, Martin (1982) calculated Hco/Wco=0.14 and Swinamer (1989) calculated Hco/Wco=0.18. This suggests that variations in the method of measurement can have important consequences when comparing cinder cones solely on the basis of geomorphometry.

Fig. 5.
figure 5

Aerial view of Holocene high-silica andesite Pelado volcano (3,620 m a.s.l.) toward the SW. T Tuxtepec scoria cone and H Cerro Huilote (3,685 m a.s.l.) are indicated for orientation. The lack of deeply incised gullies at its outer slopes and excellent preservation of its bowl-shaped summit crater attest to its young age. Photograph taken by C.S. during a helicopter flight on January 9, 1996

Full size image

Pelado's crater is well formed with a diameter (Wcr) of 450 m and a depth of 45 m. Its rim can be accessed by a dirt road serving a telecommunications tower constructed on the summit of its southern rim. The scarcity of pine trees growing on its flanks seems to be related to its high altitude in close proximity to the timberline, rather than to its youthfulness. The cone displays a small asymmetry, with the southwestern quadrant of the rim standing higher than adjacent quadrants. The opposing quadrant also stands higher than the intervening quadrants, producing a bipeaked rim. This asymmetry might have resulted from prevailing wind action during the eruption, since the highest rim of such cones invariably lies on the same side as the tephra plume, which was deposited downwind (Porter 1972). A prevailing southward direction of paleowind direction during Pelado's eruption is also suggested by greater thickness of ash fallout measured at distal sections to the south. Pelado's cone consists predominantly of stratified, loose, angular, dark-gray to reddish cinders of Strombolian type, the estimated mean particle size falling in the range of 1–10 cm. These features indicate that scoria fragments were relatively cool and emplaced from a higher volcanic column. Although graded bedding is common, sorting is generally poor with fine lapilli and ash forming a loose matrix between scoria fragments and bombs. The slopes of the cone are steep (ca. 30°) with few gullies and a general lack of large-scale erosion features.

Pelado's cone grew on relatively flat terrain and is surrounded by a broad shield of compound aa lava flows. The shield is regular in shape with an average diameter of 7–8 km and slopes from 5–10°. Oxidation and weathering has sufficiently decomposed cinders to form a loamy soil above the lava that supports a pine forest that includes abundant underbrush. The lava flows associated with Pelado's cone issued from flank fissures. The two largest flows followed pre-existing valleys and flowed towards the S–SE and N–NE. The S–SE flow is the longest at 8.25 km, and was constrained by the eastern margins of the pre-existing shields of Tesoyo and Tres Cumbres cinder cones (Fig. 3). The N–NE flow was channeled through a valley between the pre-existing shields of Oyameyo and Mezontepec cinder cones and reached a maximum distance of 6.75 km. The length of a lava flow is influenced by eruptive (flow) rate, eruptive volume, initial viscosity and marginal cooling. Walker (1973) observed that high effusive rates produced the longest flows. Specifically, the length of a channel fed flow is a function of the highest rates of effusion averaged over the first few days of eruption (Walker 1973). A useful parameter for characterizing the extent of a lava flow is its aspect ratio. Walker (1973) defined this ratio as the length of the flow divided by the thickness. Assuming an average thickness of 15 m for both flows, aspect ratios of 550 and 450 were computed for Pelado's S–SE and N–NE flows, respectively. The 5–25-m-steep front margins of Pelado's aa flows, as well as the stepped surface of the shield, both indicate that Pelado was formed by viscous evolved lavas. Approximately 1.5 km W of Pelado's cone, the otherwise regular shape of the lava shield is disturbed by a small topographic protuberance named Tzotzocol (tzotzocolli = earthen jar in Náhuatl, the language spoken by the Aztecs and other related ethnic groups). Closer inspection in the field and chemical analyses (Siebe et al. 2003) showed that Tzotzocol represents the remnants of an older cinder cone that was almost completely covered by Pelado's products.

Pelado and its lavas cover an area of 79 km2 which is ca. 25% larger than the area of 63–65 km2 computed earlier by Swinamer (1989), Arana-Salinas (1998), and Márquez et al. (1999). The reason for this discrepancy is because these earlier workers did not recognize that the distal part of the S–SE flow belonged to Pelado. The minimum volume of erupted products was estimated by multiplying the area covered by lava flows with an average thickness determined by observations at lava flow margins. An accurate average thickness is very difficult to determine because the pre-existing topography is not known. Using average thicknesses of 15–25 m, a total volume of 1–2 km3 is obtained. The volume of the cone was calculated at 0.052 km3 using the formula of a circular cone (the real volume comprising the scoria cone is certainly smaller since we neither substracted the volume of the apex of the cone, nor the volume occupied by vesicles and interstices between clasts). Its relatively small volume confirms the general observation that scoria cones are the most visible result of many eruptions, but commonly represent the least significant fraction of the entire output of an eruption (Wood 1980a).

Recent field research by Arana-Salinas (1998) has revealed the occurrence of pyroclastic flow deposits from Pelado. In addition, Márquez et al. (1999) mention the occurrence of hydromagmatic base-surge deposits belonging to the sand-wave and planar facies with well-developed cross-bedding structures. Since this issue is important for volcanic hazard assessments, we visited the localities described by these workers and concluded that in both cases the deposits under discussion are composed of reworked ash fallout. The outcrop described by Márquez et al. (1999) shows a sequence of distal silty–sandy ash fallout, which is finely laminated and intercalated with layers that display cross bedding. The ash in each layer is well sorted and normally graded. The cross-bedding can be attributed to re-deposition by moderate to strong winds. Typical features diagnostic for hydromagmatic activity such as accretionary lapilli are missing. The area mentioned by Arana-Salinas (1998) is a small alluvial fan that overlies Pelado lavas and consists of a sequence of unconsolidated fluviatile layers. These decimeter-thick layers are composed mostly of dark-gray ash intercalated with discrete layers of subrounded pebbles. When wet, the layering is not clearly visible and the entire sequence is massive in appearance.

Guespalapa

Guespalapa is a dense cluster of four relatively small (80–150 m high) cones that were produced along an E–W trending monogenetic fissure (Fig. 6). They are named by local shepherds as El Hoyo, Manteca, Caballito, and El Palomito (from E to W) and are not included in the official 1:50,000 topographic map published by the Instituto Nacional de Estadística, Geografía e Informática (INEGI). The Guespalapa cinder cones consist principally of interbedded deposits of fragmental pyroclastic material produced by moderately explosive eruptions. The E–W alignment and spatial distribution of the Guespalapa cones is consistent with the general crustal structure and fracture system of the region (Siebe et al. 2003).

Fig. 6.
figure 6

Digitally shaded terrain model of the Guespalapa cone complex prepared using the Integrated Land and Water Information System (ILWIS) software package and Instituto Nacional de Estadística, Geografía e Informática (INEGI) topographic data (maps 1:50,000)

Full size image

In strong contrast with Chichinautzin and Pelado, the Guespalapa cones are not surrounded by a broad lava shield. This is mainly because they formed in a topographically more constrained location, slightly south of the Continental Divide. For this reason all the lavas were channeled southwards and reached extraordinary lengths. In fact, our attention was focused on these cones only after searching for the source of the prominent Texcal basalt lava flow, first mentioned by Ordóñez (1938). Ordóñez noted its young appearance and that its viscosity during flowage must have been quite low, as evidenced by its small thickness and extraordinary length. This lava flow is 24 km long (aspect ratio = ca. 2,400) and is one of the longest of the entire SCVF (Figs. 2 and 3). It reaches far into the plain of Cuernavaca, where it stands out due to its relative lack of vegetation. (Texcal means "oven" or "badland" in Náhuatl). According to Walker (1971; 1973) the most extensive and far-reaching flows (such as Texcal) are simple aa (not divisible into flow units) and are primarily believed to form when the rate of extrusion of lava is particularly high. Of course, they are also far-reaching because of their great volume and low viscosity. Furthermore, Rowland and Walker (1990) determined that the rough-surfaced variety of basalt lava called aa, forms when lava flows at a high volumetric rate (>5–10 m3/s).

At Texcal's most distal margins multiple springs with discharges of several m3/s are located. Due to its rugged surface and difficult access, the Texcal remained uninhabited until the recent rapid expansion of the city of Cuernavaca started to chaotically spread over the surface of this flow, endangering the quality of the springwater.

The total area covered by Guespalapa lavas is 57.6 km2. Assuming an average thickness of 10–20 m, a total minimum volume of 0.58–1.15 km3 is estimated. The volume of each cone was also estimated (Table 1) and yields a sum of 0.05 km3 similar to the volume of Pelado's cone. Hco/Wco ratios vary greatly. While El Hoyo and El Palomito have values ranging between 0.17 and 0.19, typical for fresh cones, Manteca and El Caballito have anomalously high values and therefore also anomalously steep slopes. The reason for this is not well understood but might be related to the order of appearance, sequence of breaching events, etc. of the cones. Closer analysis of the digitally shaded terrain model (Fig. 6) indicates that El Caballito's activity probably lasted longer than Palomito's and Manteca's. Perhaps, Manteca and El Caballito formed earlier and were additionally mantled with ash from the neighboring two cones. The Guespalapa cones, especially El Caballito and El Hoyo, display azimuthal asymmetries with higher W–NW rims. This either provides a measure of low-altitude paleowind directions or indicates consistent breaching of the cones towards the S–SE.

A modern analogue in regard to eruptive style and duration for Guespalapa's activity might be the Tolbachik basaltic fissure eruption in Kamchatka, Russia, which started on 6 July 1975 and resulted in the formation of four N–S aligned cinder cones (Fedotov and Markhinin 1983; Tokarev 1983). The eruption ended after almost 1.5 years and the magma supply was controlled by a deep-seated fault. Ash columns rose to an altitude of more than 13 km and basalt flows covered an area of around 45 km2. Initially, the lava flows emanated from boccas that opened along a N–S fissure and functioned intermittently for not more than several days, although any three or four were at times simultaneously active. Traces of fine ash reached over 800 km, ballistically ejected bombs up to 30 cm in size reached distances of 2 km, and vegetation was totally destroyed over 400 km2. One of the most remarkable features of the eruption was the seismic precursors detected only 19 days preceding the eruption with epicenters distributed over 100 km2. In addition, the eruption induced substantial hydrogeologic and topographic changes in the environment.

Chichinautzin

This cone gave its name to the entire surrounding mountain range, and was used for the first time in the geologic literature by Fries (1962) who grouped all the Quaternary volcanic rocks in this area into the "Formación Chichinautzin". Chichinautzin means "burning lord" in Náhuatl. Such a name strongly suggests that humans might have witnessed its eruption, which in turn implies a very young age.

The above is in accordance with the extremely young appearance of Chichinautzin's dark gray lavas which lack persistent soil cover and vegetation. Their surface is complicated by numerous structural elements characteristic of mobile basalts: gently sloping domes (tumuli), longitudinal levees, pit craters, empty lava channels and tubes, hornitos and blisters. The main shield has a diameter of ca. 5 km and much like Pelado formed on relatively flat terrain on the Continental Divide (Fig. 7). The largest lava flows advanced 14.5 km to the N between the Oyameyo and Cuauhtzin shields toward the Valley of Mexico as well as 9 km to the south toward the valley of Cuernavaca. The fronts of Chichinautzin's aa lava flows are approximately 6 m high and lavas are mostly compound. Best outcrops are along the federal freeway to Cuernavaca which cuts through the most distal part of the southern flows. Assuming an average thickness of 10 m, aspect ratios of 900–1,450 are estimated.

Fig. 7.
figure 7

Aerial view of Chichinautzin volcano (3,490 m a.s.l.) toward the SW. The broad shield has a diameter of ca. 7 km and is crowned by a small scoria cone with a basal diameter of ca. 1 km. Photograph taken by C.S. during a helicopter flight on Oct. 29, 1996

Full size image

Chichinautzin's summit cone (3,490 m a.s.l.) is less prominent than Pelado's, but morphologically more complex. The cone is 160 m high with an Hco/Wco of 0.21 and surrounded by a shield that displays an irregular outline where it formed between and around older shields and cinder cones (Figs. 3 and 7). The area covered by Chichinautzin and its lavas was estimated at 54.9 km2. Assuming an average thickness of 10–20 m, a total minimum volume of 0.55–1.1 km3 is obtained. These values are very similar to those estimated for Guespalapa (Table 1).

Analysis of a digitally shaded terrain model of Chichinautzin´s summit area (Fig. 8) reveals several crater structures roughly oriented WSW–ENE. To the W–SW of the main cone, a prominent escarpment open to the E is interpreted as the remnant of a cone that formed at an initial vent. Apparently this cone became breached before the activity shifted to the E–SE in the direction of the principal cone. In addition, several other vents aligned WSW–ENE are recognized (Fig. 8). The initial eruptions must have occurred along a fissure and did not rapidly localize to a single vent. As the eruption progressed, effusion of lava became concentrated along a short length of the fissure. In contrast to Pelado and Guespalapa, Chichinautzin cone is constituted by a larger percentage of spatter and may account for its high Hco/Wco ratio of 0.21. Furthermore, most Chichinautzin lava flows are compound and were fed by tubes. According to Rowland and Walker (1990), this means that Chichinautzin lavas were probably emplaced over longer periods of sustained flow at low discharge rates, with the separation of gases from the magma happening before eruption. In contrast, Guespalapa lavas likely resulted from a more rapid emptying of a subsidiary magma chamber with more rapid near-surface vesiculation of gas-rich magma leading to eruptions with high discharge rates and consequently fast-moving channeled flows that traveled much longer distances.

Fig. 8.
figure 8

Digitally shaded terrain model of the Chichinautzin summit cone area prepared with ILWIS using INEGI topographic data (maps 1:50,000)

Full size image

Stratigraphy and radiocarbon dating

As mentioned earlier, most radiocarbon dating efforts in the SCVF have been focused on the determination of the age of the Xitle volcano and the associated Cuicuilco archaeological site. A table containing all related published dates and a thorough discussion of Xitle's age can be found in Siebe (2000).

Studies covering larger areas of the SCVF include Bloomfield (1973; 1975), who concentrated on the western portion of the SCVF which extends into the Lerma basin near Toluca. The age of 41 cones and related lava flows was estimated on geomorphological grounds and calibrated with the aid of radiocarbon dates ranging between 8,390±100 and 40,000 years b.p. Several of these cones are probably older than 40,000 years. Martin (1982) conducted a similar study in the central portion of the SCVF and catalogued volcanoes using degradation stages based on the intensity of erosional modifications. Kirianov et al. (1990) reported ages of soils intercalated with lavas and scoria-fall from the central part of the SCVF. More recently, Arana-Salinas (1998) and Wallace and Carmichael (1999) provided a few scattered additional dates on Pelado and Oyameyo volcanoes. This means that in addition to the more than 30 radiocarbon dates related to Xitle and Cuicuilco, another 25 dates were obtained at different localities within the SCVF by the above authors and are listed in Table 2. Many of these dates were obtained from paleosols intercalated between tephras and do not exactly pinpoint eruptions. Nonetheless, several samples were taken from a reliable stratigraphic context. As a result only 10 from more than 220 volcanic cones in the SCVF have been dated with acceptable precision. The ages of the following cones are best constrained (see also Fig. 1 and Table 2): Tres Cumbres cone is older than 23,200±300 years b.p. (IVAN-504), Tezontle is 21,860±540 years b.p. (Tx-1610 and Tx-1725), Manal cone is older than 18,900±600 years b.p. (IVAN-508), and Cuautl cone is probably slightly younger than 18,560±210 years b.p. (Tx-1590). Oyameyo is between 13,755±95 and 16,145±115 years b.p. (AA-8945 and AA-8944), Raíces cone is younger than 15,570±300 years b.p. (GIN-4059), and La Cima cone is older than 10,410±80 years b.p. (IVAN-509). Tenango is between 8,390±130 and 8,700±180 years b.p. (Tx-1665, Tx-1666, Tx-1667), Tres Cruces cone is slightly younger than 8,390±100 years b.p. (Tx-1606), and Xitle ca. 1,670±35 years b.p. (A-9587 and A-8985).

Table 2. Available (24 new and 27 published) radiocarbon dates for the SCVF monogenetic cones and associated products. Published 14C dates for the Cuicuilco archaeological site were not considered for this table but are listed and discussed in Siebe (2000). All conventional radiocarbon dates have been corrected by reference to the dendrochronological time scales reviewed recently by Stuiver and Becker (1993), Pearson and Stuiver (1993), and Stuiver and Pearson (1993), and graphically illustrated in Fig. 11. Cal 1 sigma and 2 sigma refer to the ranges of confidence level yielded by the Seattle-Groningen methods of probability distribution analyses
Full size table

These dates are not sufficient for determining reliable recurrence intervals for monogenetic eruptions in the area. By choosing three of the morphologically youngest volcanoes for the present study, we attempted to close existing chronological gaps and establish a more accurate recurrence interval for the past 25,000 years. In this context two excellent stratigraphic markers for this region have been identified over the past years. These are the ca. 14,000 years b.p. "Tutti Frutti" pumice from Popocatépetl (Siebe et al. 1995, 1999) and the ca. 10,500 years b.p. "Upper Toluca Pumice" from Nevado de Toluca volcano (Macías et al. 1997, Arce et al. 2003). Both pumice fallout layers originated from large Plinian eruptions and blanket extensive areas around these volcanoes including large parts of the SCVF. Neither of these fallout layers were found overlying deposits of Pelado, Guespalapa, or Chichinautzin.

From field observations and morphological analyses carried out during the present study, it is apparent that Chichinautzin is by far the youngest of the three volcanoes studied. The age relationships between the younger Guespalapa and the older Pelado became clear after inspection of the area where lavas from both volcanoes are in direct contact with each other. In this context it is worth mentioning that the Hco/Wco ratios determined for the cones analyzed in this study yielded ambiguous results in respect to their usefulness for establishing relative ages. This conclusion was also reached by other authors working in young volcanic fields (e.g., Conway et al. 1998).

Most outcrops displaying the stratigraphy of these volcanoes are located along the federal freeway and the railroad connecting Mexico City and Cuernavaca. Both cut through lava flows of all three volcanoes. Unfortunately, we were unable to find charcoal samples for radiocarbon dating at these locations. To obtain datable materials associated with these flows several m-deep trenches were dug mostly at lava flow margins. As a result, a total of 24 new samples were collected and dated at the Radiocarbon Laboratory in Tucson, Arizona (Table 2 and Fig. 9).

Fig. 9.
figure 9figure 9figure 9

Stratigraphy at localities where charcoal samples for radiocarbon dating were obtained. Locations related to A Pelado, B Guespalapa, C Chichinautzin, D Oyameyo

Full size image

Age of Pelado

Two previously published C-14 dates relate to Pelado. The first was obtained by Kirianov et al. (1990) who dated the soil on top of one of Pelado's lava flows at 4,070±150 years b.p. (IVAN-506). Because of its stratigraphic position, this age should be considered as a minimum age. More recently, Delgado et al. (1997) published a sensational age of 870 years b.p. obtained by Arana-Salinas (1998) from a "pyroclastic flow deposit." This deposit also bears pre-Hispanic pottery shards and was believed to have originated from Pelado. This extremely young age contradicts the "minimum" age estimated from Kirianov's date. In order to clear up this question, we visited the sampling site situated a few hundred m south of site A2-04 with Lilia Arana-Salinas (Fig. 3). This revealed that the sample was taken from a layer of reworked ash within an alluvial plain covering part of Pelado's S–SE lava flow, rather than from a pyroclastic flow deposit.

As part of the present study, a total of ten charcoal samples were obtained at different locations around Pelado (Fig. 9A and Table 2). Samples A2-14 (190±40 years b.p.), A2-20 (700±75 years b.p.), A2-19 (1,065±50 years b.p.), A2-15 (1,100±80 years b.p.), P1-1 (1,565±60 years b.p.), A2-06 (1,685±45 years b.p.) and A1-01 (2,985±80 years b.p.) were found in layers of reworked dark-gray, sandy–silty ash ponding against distal margins of Pelado lava flows, and hence all represent minimum ages. Some of the charcoal in these layers of reworked ash might have originated from forest fires like those which still occur today during the dry season (Oct.–May) or by fires ignited by younger nearby eruptions.

In contrast samples A2-21 (9,620±160 years b.p.) and A2-25 (10,270±190 years b.p.) were obtained from layers directly underneath Pelado lava flows (Fig. 9A). In addition, sample A2-27 (10,900±280 years b.p.) was obtained from a sandy lahar consisting of Pelado ash and lying directly on top of Pelado fallout tephra. These three samples were taken from unambiguous stratigraphic positions and are therefore believed to closely date Pelado's eruption, which accordingly must have occurred between 9,620±160 and 10,900±280 years b.p. near the Pleistocene/Holocene boundary. Since the Upper Toluca Pumice dated at ca. 10,500 years b.p. (Macías et al. 1997; Arce et al. 2003) does not cover Pelado's products, it is concluded that Pelado's eruption is slightly younger than the Upper Toluca Pumice and might have occurred around 10,000 years b.p.

Age of Guespalapa

This volcano has not been dated previously. Again, several charcoal samples were obtained from reworked ash ponding against lava flow margins (Fig. 9B): samples Gp-II (1,465±190 years b.p.), Gp-III (1,165±55 years b.p.), and Gp-I (690±40 years b.p.) represent minimum ages. The oldest minimum age for Guespalapa was obtained from sample A2-05-II-A (2,835±75 years b.p.) on charcoal in a paleosol directly overlying Guespalapa scoria layers. Only one maximum age was obtained from a paleosol directly underlying the Texcal lava flow at locality C3-04, which yielded an age of 4,690±90 years b.p.. Guespalapa's true age must therefore lie somewhere between 2,835±75 and 4,690±90 years b.p.

Age of Chichinautzin

This volcano has also never been dated before. Samples Ch-IV (850±45), Ch-III (900±95), and A3-01 (1,360±90) are charcoal in reworked sandy–silty ash ponding against Chichinautzin lava flows and represent minimum ages. The best samples with excellent stratigraphic control were found at A2-05-II, a small scoria quarry on the SE slopes of El Caballito cone (Fig. 6). Here, 20 cm of Chichinautzin sandy ash fallout lie directly above a paleosol developed into El Caballito scoria (Fig. 9C). Chichinautzin's ash fallout contains small fragments of charcoal (A2-05-C) dated at 1,835±55 years b.p. These charcoal fragments must have been produced by forest fires ignited by Chichinautzin's incandescent products. The Chichinautzin ash layer is overlain by a charcoal bearing horizon of reworked material that was dated at 2,115±115 years b.p. (sample A2-05-II-B). This slightly older date on top of Chichinautzin ash can be explained by the erosion and subsequent mixing of charcoal of different ages from the ash layer and the paleosol below. It is concluded that sample A2-05-II-C dated at 1,835±55 years b.p. is closest to the age of Chichinautzin's eruption.

Age of Oyameyo

We were also able to further constrain the age of Oyameyo, a scoria cone located 4 km NE of Pelado (Fig. 3). Wallace and Carmichael (1999) reported two radiocarbon dates of 13,755±95 years b.p. (AA-8945) and 16,145±115 years b.p. (AA-8944) from layers below Oyameyo ash. These dates therefore represent maximum ages. At locality A2-28 near Parres (Figs. 3 and 9D), we were able to identify Oyameyo ash-and-scoria fallout lying stratigraphically above Popocatépetl's Tutti Frutti pumice (dated at ca. 14,000 years b.p.). At this locality, we dated charcoal from a silty ash-flow directly on top of Oyameyo ash, which yielded an age of 8,315±155 years b.p. (A-10874). We conclude that Oyameyo is younger than 14,000 years but older than 8,315±155 years b.p.

Archaeological implications

According to the present study Chichinautzin volcano (Burning Lord) erupted 1,835±55 years b.p. (60–170 a.d.) which is very close to the Preclassic/Classic boundary of Mesoamerican archaeology (Fig. 10). The Upper Preclassic and Lower Classic periods of Mesoamerican archaeology are characterized by the transition of small tribal villages with subsistence agriculture and a few larger ceremonial centers into more stratified societies with much larger urban centers (e.g., Heydenreich 1975). Toward the end of the Preclassic, Teotihuacán and Cuicuilco were the dominant urban centers in the Basin of Mexico (Fig. 4). By ca. 300–100 b.c. they were very similar in size, character, and regional impact (Sanders et al. 1979). Both were a major demographic focus related to the high productivity of rainfall agriculture and small-scale water-control technologies. At this time Cuicuilco reached its maximum size and architectural complexity. Sanders et al. (1979) estimated a settlement area of at least 400 ha and a minimum population of about 20,000 people. During the period 100 b.c.–100 a.d. Teotihuacán emerged as a center of extraordinary size and population. The great majority (80–90%) of the population of the entire Basin of Mexico had nucleated at Teotihuacán and the population in the southern part of the basin was reduced substantially. At the same time Cuicuilco decreased to a small center (Sanders et al. 1979).

Fig. 10.
figure 10

Graph showing the archaeological time scale for central Mexico (according to Sanders et al. 1979; Porter-Weaver 1993; Millón 1994) and the best available radiocarbon dates for Xitle and Chichinautzin monogenetic cones. The last two major Plinian eruptions of Popocatépetl are also indicated

Full size image

Numerous hypotheses have been offered to explain the northward population shift from Cuicuilco and the southern part of the basin to Teotihuacán near the Preclassic/Classic boundary. One of them suggests that the downfall of Cuicuilco and rise of Teotihuacán was directly attributable to the eruption of Xitle ca. 2,000 years b.p. (e.g., Sarmiento 1994). In a recent paper, Siebe (2000) demonstrated that Xitle's eruption occurred 1,670±35 years b.p. (245–315 a.d.) which is ca. 300 years later than previously established. Therefore Xitle's eruption and Cuicuilco's final destruction apparently occurred ca. 200 years later than the northward population shift. In this context, Popocatépetl also experienced a major eruption dated between ca. 2,200 and 2,000 years b.p. (250–50 b.p.) as revealed by recent studies (Siebe et al. 1996; Plunket and Uruñuela 1998; Panfil et al. 1999). This Plinian eruption had a magnitude (Newhall and Self 1982) of VEI=6, and completely devastated large areas around the volcano, including several Preclassic settlements (e.g., the site of Tetimpa described by Plunket and Uruñuela 1998). Although the eastern slopes of the volcano were most severely affected by Plinian fallout, the NW slopes were also devastated by pyroclastic flows. Survivors in the Valley of Puebla and in the Amecameca-Chalco region located in the SE part of the Basin of Mexico had to migrate to other areas. In short, it seems Popocatépetl played a role in the northward population shift, although it did not affect Cuicuilco directly. A similar but less important effect might be attributable to Chichinautzin, which erupted 1,835±55 years b.p. (60–170 a.d.) shortly after Popocatépetl and also affected the southern Basin of Mexico. This eruption had a magnitude of VEI=2 (similar to Xitle and much smaller than Popocatépetl) and destroyed a considerable area in the immediate vicinity of Cuicuilco.

From the above it can be concluded that a total of three eruptions (Popocatépetl, Chichinautzin, and Xitle) occurred during the period 200 b.p.–315 a.d. at the southern margins of the Basin of Mexico and contributed in different degrees to its depopulation. Popocatépetl's Plinian eruption was the earliest and most destructive of these eruptions and had a much greater environmental and economic impact. During the same period of time, Teotihuacán in the northern part of the basin, experienced continuous growth and expansion.

Future eruptive scenarios and hazards

The main practical conclusion that can be drawn from the present work is that during the past 10,000 years, at least four monogenetic eruptions (Pelado, Guespalapa, Chichinautzin, and Xitle) occurred within the central part of the SCVF. This means that the average recurrence interval for such a type of eruption in this part of the SCVF is on the order of 2,500 years or less. Assuming an average erupted volume of 1 km3/cone, it is possible to roughly calculate an eruption rate of 4 km3/10,000 years=0.4 km3/1,000 years. If the entire SCVF is considered and Tenango and Tres Cruces cones are included (Fig. 11), a total of at least six cones have formed over the past 10,000 years. This corresponds to a recurrence interval of <1,700 years with an estimated eruption rate of 0.6 km3/1,000 years, a value that differs by more than one order of magnitude from the 11.7 km3/1,000 years estimated by Verma (1999) and Márquez et al. (1999). Indeed, this is happy news of some relief for the nearby population. The reason for this large discrepancy is that Verma (1999) and Márquez et al. (1999) wrongly assumed that the more than 200 cones in the SCVF are younger than 40,000 years b.p. Nonetheless, volcanic activity still poses hazards for the surrounding region, especially for Mexico City. If we consider the entire SCVF and all of the available radiocarbon dates for the past 25,000 years, a more complete and realistic picture starts to emerge (Fig. 11). Accordingly, at least nine monogenetic volcanoes (five basaltic and four high-Si andesitic) were formed during this period. In addition, Popocatépetl experienced at least seven Plinian eruptions (only five are well constrained by the C-14 method), and Nevado de Toluca only two. This means that a minimum of 18 eruptions (nine monogenetic and nine Plinian) occurred in the vicinity of Mexico City during the past 25,000 years. We anticipate that as part of our ongoing radiocarbon dating project, we will find a few (two to –three) additional cones that are also younger than 25,000 years b.p.

Fig. 11.
figure 11

Graph showing the age of radiocarbon-dated eruptions in the vicinity of Mexico City for the past 25,000 years

Full size image

For historical and political reasons, Mexico is a very centralized country. Government services and industrial development have concentrated in Mexico City and more than 30% of the country's industrial activity and of its gross national product are located here. This situation obviously enhances vulnerability.

The freeway from Mexico City to Cuernavaca and the port of Acapulco cuts through lava flows from all four Holocene volcanoes (Figs. 3 and 12). There is a high probability that renewed activity of this type would obstruct this important lifeline. Furthermore, lava flows would most likely follow pre-existing valleys and reach the city limits of Cuernavaca or Mexico City (Figs. 3 and 11). Due to shifting wind directions, ashfall from a monogenetic eruption would reach, with great certainty, both urban areas. Its effects would be more than just a temporary annoyance. Mexico City's sewer system is ca. 10,000 km long and its efficiency could suffer tremendously from ashfall. The same applies to the supply network for electrical power, which is largely subaerial. Ash coatings would likely cause short circuits and subsequent power outages. Airports (Mexico City, Cuernavaca, and Toluca) would probably have to close (at least temporarily). Another issue that requires a more profound investigation is related to the supply of drinking water for Mexico City, whose unusual physical setting in a high, naturally closed basin, uniquely challenges the provision of water for a large urban population. Almost 72% of the city's water supply comes from the Mexico City Aquifer, which underlies the metropolitan area (The Joint Academies Committee, 1995). The SCVF is the most important natural recharge zone for the Mexico City Aquifer due to the high porosity and permeability of its volcanic rocks. It is possible that renewed activity could affect the amount and quality of the water supplies.

Fig. 12.
figure 12

Sketch map showing the central part of the SCVF and major drainages that could be used by lava flows in case of future monogenetic eruptions. Several of these drainages lead directly into the urban areas of Cuernavaca and Mexico City

Full size image

Explosive (phreatomagmatic) activity seems unlikely since characteristic surge deposits were not identified in the field. This conclusion is substantiated by the fact that deep boreholes for wells in the SCVF reached the groundwater table at depths of more than 250 m. This indicates high porosity and permeability of the rocks and rapid groundwater flow.

Volcanoes in the SCVF seem to cover a wide range of possibilities in regard to the duration of monogenetic eruptions. Wood (1980a) compiled data on historic monogenetic eruptions worldwide and concluded that seismic activity started shortly before each of the eruptions, intensified just before the eruption and then died off. The duration of 42 observed historic eruptions (the time for construction of the cone and effusion of associated flows) ranges from one day (Etna, 1792) to 15 years (Jorullo, Mexico, 1759–1774), with a median value of 30 days. Furthermore, 50% of the observed cinder cone eruptions lasted less than 30 days and 95% were over in one year or less. In the SCVF it seems reasonable to assume that basaltic eruptions with high discharge rates (e.g., Guespalapa) would certainly last less than two years while more silicic andesite eruptions (e.g., Pelado) could last longer. For the former, Tolbachik in Kamchatka (e.g., Fedotov and Markhinin 1983) could serve as a modern analogue, while for the latter Paricutín (e.g., Luhr and Simkin 1993) seems to be a modern replay. In both cases 1–2 km3 of lava were erupted, and activity was mostly Strombolian with a VEI somewhere between 2 and 3, and eruptive columns not surpassing 15 km in height.