Introduction

Misti volcano has a history of explosive eruptions that emplaced numerous tephra-fall deposits (TFDs; Legros 2001; Mariño et al. 2016; Harpel et al. 2021). The city of Arequipa (Fig. 1), built upon many of these deposits, has a growing population of over 1,100,000 (Instituto Nacional de Estadística e Informática 2018) and critical infrastructure in high volcanic hazard zones (Mariño et al. 2007). Arequipa is nestled among the massive Plio-Pleistocene Pichu Pichu (Guevara 1968; Bernahola Portugal 2018) and Pleistocene Chachani (Aguilar et al. 2022) volcanic complexes, and Misti, a voluminous, youthful cone with documented Holocene eruptions (Thouret et al. 2001). Misti is currently fumarolic (Birnie and Hall 1974; Moussallam et al. 2017; Vlastelic et al. 2022) and has experienced historical periods of unrest and enhanced degassing, some of which were possibly accompanied by small-volume phreatic or phreatomagmatic eruptions and the most recent of which occurred in 1985 (Hantke and Parodi 1966; Masías Nuñez del Prado 1997). A mid-fifteenth century eruption, which built the summit cinder cone and distributed ash locally, was incorporated into the indigenous population’s oral history and possibly provoked human sacrifices (Chávez Chávez 1992; Socha et al. 2021). Misti’s most recent Plinian eruption occurred ~ 2 ka, emplacing voluminous deposits, which now underly parts of Arequipa (Harpel et al. 2011; Cobeñas et al. 2012). Thouret et al. (2001) broadly delineated Misti’s history. Yet to date, only the Autopista, Sandwich Inferior, Sandwich Superior, and 2-ka deposits have been individually investigated (e.g., Cacya et al. 2007; Harpel et al. 2011; Escobar 2021), leaving deposits from most of Misti’s eruptions uncharacterized and their hazards implications unknown.

Fig. 1
figure 1

a Map of Peru with Arequipa and Misti indicated by the black hexagon. b Isopach map for the Sacarosa tephra-fall deposit. Thickness in centimeters is indicated at each point. The locations of Arequipa, the airport, Misti, Chachani, and Quebrada San Lazaro are indicated. Black stars indicate the Aguada Blanca hydroelectric installation. c Pumice isopleth map. Average maximum pumice size in centimeters is at each point. Dark gray shading in a and c represents Arequipa’s urbanized area. Isopachs and isopleths indicated with solid lines are constrained by field data and dashed lines are inferred

Full size image

Within the Andean Central Volcanic Zone (CVZ), five Holocene or historical eruptions with Volcanic Explosivity Index (VEI; Newhall and Self 1982) 5 or greater are documented (Siebert et al. 2010) with only Misti’s ~ 2-ka (VEI 5), Huaynaputina’s 1600-CE (VEI 6), and Cerro Blanco’s 4.2-ka-cal-BP (VEI 7) eruptions exhaustively volcanologically characterized (e.g., Adams et al. 2001; Harpel et al. 2011; Fernandez-Turiel et al. 2019). Huaynaputina’s paroxysm remains the most powerful historical Andean eruption, which in addition to catastrophic economic and social consequences for Peru, caused global impacts such as unseasonably cold weather, crop failures, and epidemics in Asia during 1601 CE (e.g., de Silva et al. 2000; Fei et al. 2016; Mariño et al. 2021). Whereas such events are rare, occurring once a decade or more worldwide (Siebert et al. 2010), they are commonly under-represented in eruption catalogs (Kiyosugi et al. 2015). Recent work at other volcanoes in the Peruvian CVZ, such as Ubinas (Wright et al. 2018; Samaniego et al. 2020), Sara Sara (Cueva 2016; Rivera et al. 2020a), and Yucamane-Calientes (Rivera et al. 2020b) suggests that more VEI 5 or greater eruptions probably have occurred but remain to be characterized in detail.

A distinct group of at least three biotite-bearing dacite and rhyolite deposits emplaced sometime between about 50 ka and 20 ka punctuates Misti’s history, which is dominated by pyroxene- and amphibole-bearing andesites (e.g., Legros 1998; Thouret et al. 2001; Mariño et al. 2016). Available data from these felsic deposits, emplaced during an eruptive period which we informally name the Cayma stage, imply that they are products of powerful explosive eruptions (e.g., Legros 1998; Mariño et al. 2016; Cuno 2018). The Sacarosa TFD was emplaced by one of these eruptions sometime between 33.7 ka and 20 ka (Legros 2001; Mariño et al. 2016; Cuno 2018).

Characterizing the magmatic conditions leading to the Sacarosa eruption is the first step in understanding the origin and evolution of the broader Cayma stage eruptions (Topham et al. 2021), which is key to understanding if future similarly powerful explosive silicic events could occur. Additionally, multiple subsequent eruptions, including Misti’s most recent Plinian event at 2 ka, produced mingled pumice of dominant andesite and minor felsic components (Tepley et al. 2013; Takach et al. 2021, in review). Such felsic components in products from recent eruptions highlight the continued impact of silicic magmas on Misti’s eruptive behavior and importance in understanding them.

Future powerful explosive eruptions would severely disrupt Arequipa. By characterizing the Sacarosa TFD, its distribution, and its eruption’s dynamics, we provide the first comprehensive description of one of Misti’s Cayma stage eruptions and its implications for future hazards. We further determine the Sacarosa magma’s pre-eruption conditions and provide the first insights into Misti’s late Pleistocene silicic magmatic system.

Methods

We document the Sacarosa TFD at 40 sites with deposit thickness recorded at 36 of these locations. At 28 sites, the five largest pumice were measured and averaged. Data were collected from sites on flat or gently sloping surfaces and where the deposit was not obviously over-thickened by dry ravel. At locations where post-emplacement erosion of the deposit is evident, thicknesses are considered minima. The scarcity of lithics precluded their measurement. Samples from eight sites, including sub-samples from the deposit’s layers, were hand sieved to determine their grain-size distribution. Componentry was determined for bulk samples by point-counting 1000–1756 clasts per sample for the 125–500 μm fractions. For grain sizes greater than 2–4 mm, components were segregated, weighed, and their masses converted to volume percentages using average clast densities. Density was measured on 89 pumice from six bulk tephra samples.

Whole-rock geochemical analysis includes X-ray fluorescence on nine samples at the Washington State University GeoAnalytical Laboratory (WSU) and Hamilton Analytical Laboratory (HAL) following the methods of Johnson et al. (1999). Trace-element concentrations were measured for one sample at WSU using inductively coupled plasma-mass spectrometry (ICP-MS) and three samples at HAL using laser-ablation ICP-MS following the methods of Conrey et al. (2023).

The Sacarosa pumice were petrographically characterized in six thin sections and their glass and phenocryst compositions subsequently determined at Oregon State University using a CAMECA SX-100 electron microprobe with five wavelength-dispersive spectrometers and high-intensity dispersive crystals for high-sensitivity analysis. Glass and minerals were analyzed using a 15 keV accelerating voltage and 10 and 30 nA sample currents, respectively. Beam sizes were 1 μm for Fe-Ti oxides, 5 μm for plagioclase, amphibole, and biotite, and 10 μm for glass. Count times ranged from 10 to 120 s depending on the element and desired detection limit. Alkali migration was mitigated by applying zero-time intercept functions. Typical one standard deviation precision for each phase is listed in Table 1. Back-scattered electron images were obtained on this instrument using the CAMECA PeakSight software.

Table 1 Representative compositions of glass, plagioclase, amphibole, biotite, and Fe-Ti oxides
Full size table

Charcoal and organic material were collected for 14C dating at three sites within 10 cm of the Sacarosa TFD’s basal contact. At one of the sites, two additional samples of organic material were collected from a layer of reworked sediment ~ 3–4 m above the Sacarosa deposit. The samples were dated by Beta Analytic Inc., PaleoResearch Institute, and the U.S. Geological Survey Radiocarbon Laboratory in Reston, VA, using the accelerator mass spectrometry method. Ages were calibrated in OxCal 4.4.4 (Bronk Ramsey 2009; Bronk Ramsey and Lee 2013) using a mixed calibration curve model allowing for any proportion of the IntCal20 (Reimer et al. 2020) and SHCal20 (Hogg et al. 2020) curves to impact the modeled ages (Ogburn 2012; Marsh et al. 2018). Published ages were calibrated using the same method. For a detailed discussion of our 14C age calibration and mixed curve model, see Electronic Supplementary Material (ESM) 1.

Cayma stage stratigraphic context

The Sacarosa TFD is one of the voluminous Cayma stage deposits given a variety of names by previous workers. Legros (1998, 2001) characterized and designated the units from lowest to highest as R1, R2, D1, and D2, with the R2 TFD intercalated between coeval pyroclastic-density–current (PDC) deposits, including the underlying R1 PDC deposit (Fig. 2). Mariño et al. (2016) provided additional descriptions and informally designated the same TFDs Fibroso I (R2), Sacaroso (D1), and Fibroso II (D2). The TFD dated by Ayala-Arenas et al. (2019) is ostensibly correlated to Mariño et al.’s (2016) Fibroso I unit. Cacya et al. (2007) and Cuno (2018) subsequently applied the names Sacaroide and Sacarosa to the Sacaroso, respectively.

Fig. 2
figure 2

a Stratigraphic columns showing the Cayma stage deposits south and southwest of Misti with correlations to deposits in previous investigations. The stratigraphic columns are not to scale and do not represent the units’ relative thicknesses or grain sizes. The stratigraphic positions of the dated paleosols and their ages are indicated. TFD is tephra-fall deposit and PDCD is pyroclastic-density–current deposit. b Typical outcrop southwest of Misti showing the Sacarosa TFD, associated Cayma stage units, and their stratigraphic positions. The shovel’s handle is 50 cm. Charcoal from the paleosol underlying the Sacarosa TFD at this site was dated. c Outcrop south of Misti showing the Sacarosa TFD, Chuma TFD, intervening layers, and their stratigraphic positions. Contacts of individual Cayma stage tephra-fall deposits are highlighted. Charcoal and organic material from the paleosols beneath the Sacarosa and Chuma TFDs at this site were dated. Note the people on the right side of the outcrop for scale

Full size image

We follow Cuno (2018) and informally apply the name Sacarosa to the unit variously referred to as D1, Sacaroide, and Sacaroso (Legros 2001; Cacya et al. 2007; Mariño et al. 2016). The similarity of the Fibroso I and Fibroso II names on unrelated units causes confusion. So, we informally rename the TFDs Cogollo and Conchito, respectively (Fig. 2). We also informally name and describe for the first time the Anchi TFD, which crops out between the Cogollo and Sacarosa TFDs (ESM 2).

Southwest of Misti, our observations generally support Legros’ (2001) stratigraphy, but south and east of Misti, the Sacarosa TFD is locally overlain by perhaps two or more Cayma stage TFDs. At one location, the outcrop is complex (Fig. 2c), not fully characterized, and faulting may have caused repeated sequences of TFDs. Several units at the site, nonetheless, appear primary and in stratigraphic order despite such faulting and minor erosional unconformities. We informally name the uppermost Cayma stage unit at the site the Chuma TFD and provide a preliminary description of it in ESM 2. The Sacarosa TFD is the third from the base of the Cayma stage. Legros (2001) also observed additional Cayma stage deposits southeast of Misti (Fig. 2c), which may correlate to units between the Sacarosa and Chuma deposits. ESM 2 provides additional discussion of the Cayma stage stratigraphy and some of its important units.

Deposit characteristics

The Sacarosa TFD is white but locally weathered to pale yellow, commonly crops out at or within several meters of the modern surface, and mantles paleo-topography beneath it. As Legros (2001) noted, it commonly overlies a dark brown paleosol, which frequently contains charcoal. This paleosol is sufficiently pervasive and distinct among the poorly developed paleosols and organic-poor sandy and silty layers of reworked sediment that more commonly underly Misti’s deposits that it facilitates identification of the Sacarosa TFD. The unit consists of two massive layers, about equally thick, separated by a diffuse contact, and differentiated by the upper layer’s slightly coarser grain size and poorer sorting (Fig. 3; ESM 3). The unit and both of its layers are well sorted (per Cas and Wright 1987). With mean grain sizes (Mdø) from 1.9–0.6 mm (ESM 3), the Sacarosa TFD is commonly finer-grained than other TFDs of similar thickness at the same outcrop. Within the area investigated, the deposit’s 125–500 μm fraction also consists mostly of loose crystals (ESM 4). Sacarosa pumice are subangular to subrounded (Fig. 3b, c). Lithics are aphanitic and porphyritic, dominantly red or orange with a minor proportion of gray clasts (Fig. 3d), and conspicuously sparse, composing ≤ 0.2 vol.% of the deposit (ESM 5) at proximal sites. We do not observe lithic-rich layers, despite Mariño et al. (2016) reporting such concentrations at the unit’s base and middle.

Fig. 3
figure 3

a Typical outcrop of Sacarosa tephra-fall deposit. Dashed line indicates the contact between the unit’s upper and lower layers. Black pencil is 14 cm long. b Typical type 1 pumice. c Typical type 2 pumice. d Typical lithic types present in the deposit. These are exceptionally large examples

Full size image

Petrology and geochemistry

We differentiate two types of Sacarosa pumice based on their textural characteristics (Table 2). Type 1 pumice are volumetrically dominant, have an average density of ~ 500 kg/m3 (ESM 6; Harpel et el. 2023), larger crystals, bimodal vesicle sizes, and are white. Some type 1 pumice have small, uncommon domains of deformed glass and bubbles. Type 2 pumice are light gray, fine grained, microvesicular, have slightly oxidized glass and phenocrysts, and an average density of ~ 800 kg/m3 (ESM 6; Harpel et al. 2023). Rare pumice with mingled type 1 and 2 textures are also present, composing ≤ 4 vol.% of the deposit that is > 2 mm (ESM 6). The proportions of the two pumice types are the same in the deposit’s upper and lower layers. Glass compositions are predominantly rhyolitic from 69 to 75 wt.% SiO2, while 10% of analyses are within an andesitic-dacitic compositional range of 62 to 66 wt.% SiO2 (Table 1). The latter glass compositions are primarily related to type 2 pumice. Glass in both pumice types lacks microlites. Phenocrysts of plagioclase, amphibole, biotite, and Fe-Ti oxides are present in both pumice types, though type 2 contains only trace amounts of biotite. Quartz is not present in our samples, contrary to Mariño et al.’s (2016) observation.

Table 2 Key characteristics for Sacarosa pumice types 1 and 2
Full size table

The Sacarosa pumice have dacitic whole-rock compositions (Table 3; Harpel et al. 2023) with type 1 slightly more silicic (65.0–65.6 wt.% SiO2; normalized to an anhydrous basis) than type 2 (64.0–64.2 wt.% SiO2). Legros’ (1998) D1 whole-rock data coincide with our Sacarosa data, thus supporting our correlation (Figs. 2 and 4a). The deposit’s trace-element concentrations (Fig. 4b) have variable enrichment in large ion lithophile elements (e.g., Rb, Ba, Th, K) and light rare earth elements (e.g., La, Ce), a depletion in heavy rare earth elements (e.g., Yb), and strong negative anomalies in high field strength elements (e.g., Nb). Type 1 pumice are slightly enriched in Rb (52–61 ppm) and Zr (173–179 ppm), and slightly depleted in Sr (535–593 ppm) compared to type 2 (Rb = 42–50 ppm, Zr = 144–156 ppm, Sr = 629–682 ppm).

Table 3 Whole-rock geochemistry for Sacarosa pumice
Full size table
Fig. 4
figure 4

Major and trace element concentrations of the Sacarosa pumice compared with Misti’s other eruptive products (gray field) from the past 120 ky (Rivera et al. 2017). a Total alkali-silica diagram (Le Bas et al. 1986) showing the Sacarosa pumice’s dacitic composition, subtle compositional differences between pumice types, and overlap with Legros’ (1998) D1 deposit. b Trace element concentrations of the Sacarosa pumice normalized to primitive mantle after Sun and McDonough (1989)

Full size image

Plagioclase compositions range from An30 to An64, with one core at An71 (Fig. 5a; Table 1; ESM 7). Crystals range in size from 0.5 to 1.5 mm, have albite twining, and are relatively inclusion free, blocky, and subhedral to euhedral. We define three end-member plagioclase populations which, in order of decreasing relative abundance, are (1) relatively simple plagioclase defined principally by oscillatory zoning, (2) plagioclase with one to two prominent resorption surfaces and calcic overgrowths, and (3) subhedral to anhedral crystals with complex and patchy zoning (Fig. 6). A subset of the first two textures exhibits patchy and or resorbed cores. Amphibole compositions range from magnesiohornblende to edenite and pargasite (Mg# ~ 0.62–0.69) (Fig. 5c, Table 1). Amphibole phenocrysts are 0.25–0.6 mm, generally subhedral to euhedral, exhibit both basal and diamond-shaped crystal forms, and display light green to medium brown pleochroism. Biotite phenocrysts occur as blocky, subhedral to euhedral crystals, range from 0.3 to 1.5 mm and exhibit light brown to medium brown pleochroism. Compositionally, the biotite phenocrysts are relatively Mg-rich (Mg# ~ 0.63–0.69) and plot in the biotite field nearest the siderophyllite end-member (Fig. 5b, Table 1). Amphibole and biotite lack reaction rims, while plagioclase and amphibole exhibit extension-cracked crystals, many of which contain interstitial melt fibers or foam (Fig. 6). Fe-Ti oxides compose ~ 1% of the pumice, with magnetite dominant and ilmenite sparse. Phenocrysts are subhedral, but exhibit both anhedral and euhedral crystal forms, ranging from 0.05 to 0.2 mm, with most ~ 0.1 mm.

Fig. 5
figure 5

Mineral composition and classification diagrams. a Histogram of plagioclase phenocryst and microphenocryst An content. b Biotite classification after Rieder et al. (1998). c Calcic amphibole classification applying Leake et al.’s (1997) definitions based on (Na + K)A contents. Amphibole represented by gray triangles are classified according to the names listed in gray, whereas those represented by black circles correspond to names in black. Biotite and amphibole parameters are given in atoms per formula unit (apfu)

Full size image
Fig. 6
figure 6

Representative back-scattered electron images of the Sacarosa plagioclase’s end-member types (a–c) and an extension-cracked plagioclase crystal (d). a Oscillatory zoned crystals; b subhedral to anhedral crystals with patchy, complex zoning; c crystals with resorption surfaces followed by calcic overgrowth; d an extension-cracked crystal with melt fibers (yellow arrows) that bridge the crack. White box in c indicates the portion of the crystal shown in d. Scale bars are 250 μm

Full size image

The compositions of the Sacarosa pumice are within Misti’s characteristic medium- to high-K calc-alkaline suite (2.0–2.3 wt.% K2O; Fig. 4a) but exemplify one of the volcano’s most differentiated magmas (Thouret et al. 2001; Rivera et al. 2017). The unit’s trace-element concentrations (Fig. 4b) are also within Misti’s normal range with characteristics typical of calc-alkaline continental margins (Grove and Kinzler 1986; Wilson 1989; Pearce and Peate 1995).

Magma conditions and ascent

Pre-eruptive magmatic conditions are determined with Putirka’s (2016) amphibole-only and amphibole-melt thermobarometers and Ghiorso and Evans’ (2008) Fe-Ti oxide thermometer and oxygen barometer. Putirka’s (2016) amphibole-only and amphibole-melt thermometers yield temperatures well within uncertainty of each other. We approximate an amphibole temperature of 815 ± 30 °C and Fe-Ti oxide temperature of 799 ± 30 °C, consistent with oxide inclusions present throughout the amphiboles and the unit’s dacitic composition. Ghiorso and Evans’ (2008) oxygen barometer also indicates a fO2 of Ni-NiO (NNO) + 1.5 (Table 4). Amphibole-melt pairs indicate the Sacarosa magma’s pre-eruptive storage system was located at 310 ± 170–370 ± 160 MPa, or ~ 9–11 km depth (depths hereafter represent kilometers beneath Misti’s summit; Fig. 7). Oxide pairs are in equilibrium according to Bacon and Hirschmann’s (1988) criteria, and Putirka’s (2016) test also shows amphibole-melt pairs are in equilibrium.

Table 4 Key characteristics of the Sacarosa tephra-fall deposit and its eruption
Full size table
Fig. 7
figure 7

Pressure and temperature (P–T) determinations using Putirka’s (2016) amphibole-only and amphibole-melt thermobarometric equations. Open symbols show P–T determinations for individual amphibole-melt pairs using two different pressure equations. The mean P–T from 34 amphibole-melt pairs for the respective equation used are indicated by a solid line, and the corresponding expected standard error of estimate for each method is shown by a shaded region

Full size image

Sacarosa pumice textures indicate a rapid magma ascent rate typical of Plinian eruptions. The lack of reaction rims on amphibole and biotite phenocrysts are consistent with the inferred pressures at which amphibole last equilibrated, indicating that the magma did not linger at shallow depths prior to eruption. Additionally, the magma ascended sufficiently rapidly that microlites did not form and enough overpressure developed in decompressing melt inclusions to fracture their host phenocrysts (Fig. 6e). Melt fibers and foam filling the cracks in the phenocrysts indicate the process was syn-eruptive (Spieler et al. 2004; Kennedy et al. 2005; Miwa and Geshi 2012). The 1980 Mount St. Helens dacite ascended from similar depths in 4 h (Scandone and Malone 1985; Endo et al. 1990) with decompression occurring sufficiently quickly that amphibole breakdown did not occur and the glass is microlite-free (Rutherford and Hill 1993). Experimental replication of such textures further indicates that they cannot exist if the magma takes more than a few hours to days to reach the surface (Rutherford and Hill 1993; Geschwind and Rutherford 1995). While we cannot currently quantitatively constrain the Sacarosa magma’s ascent rate, such qualitative evidence and comparisons suggests that it rose from depth rapidly.

Distribution, volume, and column height

The Sacarosa TFD mantles the paleo-topography over more than 830 km2, thickening and coarsening toward Misti, its inferred source (Fig. 1). The unit’s dispersal axis is southwest, similar to many of Misti’s other TFDs (e.g., Legros 2001; Cacya et al. 2007; Harpel et al. 2011). The thinnest Sacarosa deposit that we document along its dispersal axis is 24 cm, implying that the unit was emplaced over a much wider area than our data indicates. We did not observe the Sacarosa unit east of Quebrada (Spanish for ravine) San Lazaro despite thicknesses over 100 cm just west of the channel (Fig. 1b). East of the quebrada, erosion removed it, or younger deposits buried it.

The Sacarosa TFD’s thickness and distribution permit only four isopachs at 40 cm or thicker (Fig. 1b; ESM 8). Plotting deposit thickness versus square root of the area enclosed within each isopach (A1/2) yields a single line segment, from which Pyle (1989) and Fierstein and Nathenson’s (1992) method provides a minimum bulk volume of 0.4 km3 (Table 4). Bonadonna and Costa’s (2012, 2013) Weibull method using Daggitt et al.’s (2014) application yields a similar volume of 0.3 km3.

The crystal enrichment and scant fine vitric material in our Sacarosa samples indicate that its eruption generated a significant amount of fine ash that was deposited downwind of the mapped area. Its minimum volume excludes such fine ash, which can account for a significant proportion of a deposit’s volume (Walker 1980; Rose 1993; Bonadonna et al. 1998). Estimating volumes for deposits like the Sacarosa TFD, which lack distal isopachs is challenging. Bonadonna and Houghton’s (2005) method accounts for such distal ash, yielding a bulk volume of 1.7 km3 using proximal (B) and distal (C) integration limits of 1 and 1000 km, respectively. Their method, however, is sensitive to integration limits and problematic for deposits like the Sacarosa TFD, which lack proximal and distal isopachs, leading to uncertainty in the resulting volume (Sulpizio 2005; Bonadonna and Costa 2013; Biass et al. 2019).

Well-preserved TFDs yield multiple line segments on log10 of the deposit thickness versus A1/2 plots (Rose 1993; Bonadonna et al. 1998). The Sacarosa data, however, only define a single line segment on such plot (ESM 8), but the distal segment representing the fine ash distributed far from the vent is notably absent. We apply Sulpizio’s (2005) method to estimate that the inflection point between the deposit’s proximal line segment, which is defined by our data, and its distal segment, for which we lack data, is located at A1/2 = 20 km and 23 cm thickness. The distal line segment’s slope is further estimated to be k = 0.0749–0.0142 using Sulpizio’s (2005) Eqs. 5–7. Using such slopes with the inflection point, Fierstein and Nathenson’s (1992) method for calculating deposit volume from multiple line segments yields a volume of 0.6–3.2 km3. The lower value’s similarity to the unit’s minimum volumes derived from the other methods indicates that it does not represent the deposit’s complete volume. Such lower values also underestimate the volume of deposits with less than 70% of their bulk volume in proximal areas (Sulpizio 2005). Considering the evidence for abundant fine ash produced during the Sacarosa eruption, it is reasonable to expect that more than 30% of the deposit’s volume was emplaced distally. Additionally, the Sacarosa TFD thickness and distribution are similar to those of other eruptions with bulk volumes from 1 to 10 km3 (e.g., Sarna-Wojcicki et al. 1981; Scasso et al. 1994; Fontijn et al. 2011), and the unit is significantly thicker and more widely dispersed than Misti’s 2-ka TFD (ESM 8), which has a minimum volume of 0.2–0.6 km3 and an estimated total bulk volume of 1.4 km3 (Harpel et al. 2011; Cobeñas et al. 2012). As a result, the higher value of about 3 km3 better represents the Sacarosa TFD’s bulk volume.

Sacarosa PDC deposits would also increase the deposit’s bulk volume, but such deposits associated with the Sacarosa eruption are yet to be identified. While large-magnitude eruptions that do not produce PDCs are documented (Williams and Self 1983; Fontijn et al. 2011; Harpel et al. 2019), they are rare, and at least 70% of Plinian eruptions do generate such phenomena (Newhall and Hoblitt 2002). Multiple Cayma stage PDC deposits crop out (Legros 2001; Thouret et al. 2001; Mariño et al. 2016), but only the R1 deposit is correlated to its eruption. Due to the presence of such non-correlated Cayma stage PDC deposits, future investigation may reveal that one or more are co-genetic with the Sacarosa TFD and would increase the unit’s bulk volume.

We convert bulk to dense rock equivalent (DRE) volume by assuming typical TFD bulk densities of 800–1000 kg/m3 (Walker 1981; Sparks et al. 1997) to calculate the unit’s mass. We then use a rock density of 2400 kg/m3 to convert the unit’s mass to a DRE volume of 1.1–1.3 km3 (Table 4). The dearth of lithics precludes needing to remove accidental material from the DRE volume.

Applying Carey and Sparks’ (1986) method to the pumice isopleth data (ESM 8) yields an eruption column neutral buoyancy height (HU) of 14–21 km above the crater and winds of ~ 10–20 m/s. Sulpizio’s (2005) empirical relation of k to maximum column height (HT) yields a similar value of 18 km above the crater. Pumice, especially large clasts, often break upon impact skewing resulting column heights (Bonadonna et al. 2013), but the Sacarosa TFD is relatively fine grained and few of its pumice appear broken. Sparks’ (1986) empirical relationship between HU and HT and adjustment for Misti’s elevation (~ 6 km asl) indicates that the maximum column height reached 24–36 km asl (Table 4) during the Sacarosa eruption.

Eruption dynamics

The textural homogeneity of each of the Sacarosa TFD’s layers indicates steady eruption and wind conditions during emplacement of each, but the increase in grain size from the Sacarosa deposit’s lower to upper layer implies a shift to either more vigorous eruption conditions or a lesser degree of fragmentation. Changing wind direction is not the cause of the up-section coarsening since the layers are of about equal thickness throughout the unit’s distribution. The diffuse contact between layers and lack of evidence for a time break indicate emplacement by a single continuous eruption.

Within 15–20 km from the vent, the Sacarosa TFD’s < 1 mm size fraction composes 50–80 wt.% of the deposit, which is a degree of fragmentation commonly associated with phreatomagmatism. Such an eruption mechanism, however, is excluded by the unit’s sorting (< 1.75; ESM 3), scarcity of lithics, and paucity of fine vitric ash in proximal areas (Walker 1973; Gonnermann 2015; Houghton and Carey 2015). High overpressure, induced by the rapid ascent of the Sacarosa magma, correlates to fragmentation efficiency and can produce fine-grained deposits (Kueppers et al. 2006). The abundance of loose crystals lacking adhering glass and sub-angular to sub-rounded pumice also indicates secondary fragmentation that would further reduce grain-size (Jones et al. 2016). The inferred vigor of the Sacarosa eruption, its column height, and its bulk volume all indicate a VEI 5 Plinian event (Newhall and Self 1982), confirming Mariño et al.’s (2016) initial assessment.

Mass discharge rates correlate to column height (Sparks 1986; Sparks et al. 1997; Mastin et al. 2009). Applying Mastin et al.’s (2009) empirical relation to our column height yields mass eruption rates (MER) of 7.7 × 106–4.1 × 107 kg/s, assuming a magma density of 2400 kg/m3. For the Sacarosa TFD’s DRE volume, a MER of this magnitude implies an eruption duration between 17 h and five days. A comparison with other eruptions of similar volume (e.g., Rose 1972; Fierstein and Hildreth 1992; Hildreth and Drake 1992), however, indicates that a duration of several days is unlikely. Our minimum duration assumes peak discharge conditions throughout the eruption because our MER is based upon maximum column height, which is inferred from isopleth data. The unit’s isopleth data and its derivative MER, however, more likely represent the eruption conditions during emplacement of the unit’s coarse upper layer, while the finer grain size of the lower layer suggests emplacement by a lower eruption column height and MER. Consequently, the eruption likely took longer than our minimum duration and as a result, we infer that it lasted for tens of hours (Table 4).

Age

We constrain the Sacarosa TFD’s emplacement with ages from overlying and underlying units within our Bayesian model developed in OxCal (Bronk Ramsey 2009). By incorporating stratigraphic relations as independent information, the model can refine calibrations and statistically determine the age of the Sacarosa TFD and associated Cayma stage TFDs with 95.4% confidence based on their stratigraphic positions among each other and the available ages. Within the model, queries can also be performed to statistically determine the length of time between events or determine the probable sequence of events where stratigraphic relations are not known (Bronk Ramsey 2009). For our model and a detailed explanation of its components, see ESM 1.

Charcoal samples collected in this study from the paleosol underlying the Sacarosa TFD yielded ages of 39.3–37.7 ka cal BP (all calibrated ages reported at 95.4% confidence) and 39.4–37.7 ka cal BP. In contrast, humic acids from organic material provided an age of 32.9–31.8 ka cal BP (Table 5; Harpel et el. 2023). Legros (1998, 2001) and Thouret et al. (1996, 1999, 2001) reported two additional ages of 41.1–35.7 ka cal BP and 41.0–36.0 ka cal BP, respectively, from the paleosol underlying the Sacarosa TFD. All the ages from the paleosol underlying the Sacarosa TFD are statistically indistinguishable except our humic acids age (ESM 1), which typically yield minimum ages representing the component’s average residence time rather than a deposit’s emplacement age (Trumbore 2000).

Table 5 Radiocarbon and modeled ages for the Cayma stage deposits
Full size table

Organic material and microcharcoal from the reworked layer beneath the Chuma TFD, which overlies the Sacarosa TFD, yielded ages of 30.4–29.3 ka cal BP and 12.8–12.2 ka cal BP, respectively, providing minimum age constraint. A third minimum age constraint is 29.5–27.8 ka cal BP (Legros 2001; Thouret et al. 2001) from a deposit stratigraphically above the Sacarosa TFD and other Cayma stage deposits. An additional age of 39.5–33.3 ka cal BP ostensibly provides a minimum age for the Cayma stage (Legros 2001; Thouret et al. 2001), but the unit it was derived from only has a direct overlapping relation with the Cogollo TFD (Legros 1998), while its relations with other Cayma stage units, including the Sacarosa TFD, are not known. As a result, we use it in our model to constrain the Cogollo TFD’s minimum age rather than the entire Cayma stage. Using such constraints within our Bayesian age model, we infer that the Sacarosa TFD was emplaced between 38.5 ka cal BP and 32.4 ka cal BP.

Implications of the Sacarosa TFD

Timing, quantity, and magnitude of explosive Cayma stage eruptions

By applying the Sacarosa TFD as a marker bed, it becomes apparent that multiple Cayma stage TFDs delineate a more complex period of explosive volcanism than just the deposits from the three eruptions described by Legros (2001) and Mariño et al. (2016). Our preliminary investigation reveals deposits from at least five and perhaps six or more explosive eruptions during the stage (ESM 2). The first documented Cayma stage eruption emplaced the Cogollo tephra-fall and PDC deposits, while the Sacarosa TFD was produced by the third event (Fig. 2). Overlapping relations suggest at least one more unnamed TFD likely crops out between the Conchito and Chuma deposits. The TFD dated by Ayala-Arenas et al. (2019) crops out individually and its stratigraphic relation with the other Cayma stage deposits is not directly observed. Nonetheless, its age indicates that it is not the Cogollo TFD and was likely emplaced between the Anchi and Chuma TFDs (ESM 1).

Our modeling of new and published 14C ages indicates that the Cayma stage deposits were emplaced during a period of explosive volcanism beginning perhaps as early as 44.9 ka cal BP and lasting about 8.9–15.5 ky (ESM 1). The Cogollo TFD was emplaced sometime from 44.9 to 38.7 ka cal BP by the stage’s first eruption, which was followed by the Anchi (43.2–38.3 ka cal BP), Sacarosa (38.5–32.4 ka cal BP), and Conchito (37.1–30.5 ka cal BP) eruptions. Subsequent Cayma stage eruptions possibly emplaced multiple unnamed TFDs and culminated with the Chuma eruption about 30.3–28.8 ka cal BP.

The Sacarosa TFD is the first of the Cayma stage’s units to be described in detail and its eruption characterized. It is both the third most voluminous eruption in the CVZ and Misti’s most voluminous eruption yet documented. Other Cayma stage TFDs erupted during Misti’s phase of explosive silicic eruptions are of comparable thicknesses and distributions to the Sacarosa TFD, implying eruptions of similar volume and explosivity (Legros 2001; Mariño et al. 2016). Additionally, several Cayma stage deposits, including the Cogollo and Sacarosa TFDs are thicker and more widely distributed than many of Misti’s younger, andesitic units (Legros 2001; Harpel et al. 2021). Up to six or more deposits emplaced in perhaps 8.9–15.5 ky indicate a period of particularly intense explosive volcanism at Misti.

Post-Cayma stage eruption magnitudes

Of Misti’s more recent late Pleistocene and Holocene andesitic deposits, the Autopista, Pampa de los Huesos, and possibly Sandwich Inferior TFDs (Cacya et al. 2007; Escobar 2021; Harpel et al. 2021) are of sufficiently comparable thicknesses and distributions to the Sacarosa TFD to infer emplacement by similarly powerful and voluminous eruptions.

The Sacarosa TFD clearly represents a more voluminous and powerful eruption than many of Misti’s other post-Cayma stage TFDs, including the 2-ka TFD (ESM 8), indicating that VEI 5 eruptions do not occur at the frequency they did during emplacement of the Cayma stage deposits. Younger analogous TFDs, nonetheless, imply that despite its lack of characterized eruptions and historical slumber, Misti can still produce powerful paroxysmal explosive events, and the volcano’s eruption history should reflect this. Recent investigations at Ubinas, Sara Sara, and Yucamane-Calientes volcanoes, Peru (e.g., Wright et al. 2018; Rivera et al. 2020a, b), and Cerro Blanco Caldera, Argentina (Fernandez-Turiel et al. 2019), further illustrate that VEI 5 or greater eruptions are also probably more common at volcanoes throughout Peru and the CVZ than previously documented.

Magma genesis and ascent

Deposits from Misti that are older than the Cayma stage are compositionally similar to deposits subsequent to the Cayma stage (Thouret et al. 2001; Rivera et al. 2017). Such deposits indicate that Misti’s magmatic evolution is more complex than Legros’ (2001) rhyolite-dacite-andesite progression. While unlikely, were such a compositional shift to biotite-bearing silicic magmas to recur at Misti, it could herald the onset of thousands of years of intense, explosive activity.

The Sacarosa eruption’s parent magma likely followed Misti’s normal magmatic path, experiencing assimilation-fractional crystallization processes between its mantle source and ~ 10 km-depth magma chamber (Mamani et al. 2010; Rivera et al. 2017). This depth is similar to those inferred for other eruptions at Misti (Ruprecht and Wörner 2007; Tepley et al. 2013; Takach et al. in review), indicating that it is a typical magma storage level. Once there, it continued evolving for an unknown period into one of Misti’s more silicic magmas. While resorption surfaces and complex zoning in some Sacarosa plagioclase record convection and intrusions into the magma chamber, the event that triggered ascent of the Sacarosa magma remains unidentified. Preliminary petrologic evidence additionally indicates some differences in magmatic conditions may exist among the adjacent Cayma stage units (Topham et al. 2021), indicating that they likely experienced slightly different evolutionary paths leading to eruption.

The Sacarosa magma rose from its ~ 10-km-deep chamber rapidly, culminating in the paroxysmal eruption. Misti’s 2-ka eruption also resulted from magma ascending from a similar depth in ≤ 5 days. The 2-ka magma, however, interacted with a shallower silicic magma body and was heralded by several shallow intrusions during the preceding 50–60 days (Tepley et al. 2013), some of which possibly triggered small-volume phreatic/phreatomagmatic explosion(s) (Harpel et al. 2011; Cobeñas et al. 2012).

Monitoring

The Sacarosa TFD is the product of a voluminous explosive eruption of a dormant volcano, one initiated from a rapidly ascending magma that necessitates considering its monitoring ramifications. The paleosol beneath the Sacarosa TFD, lack of tephra deposits between it and the Anchi TFD, and both units’ modeled ages indicate Misti was dormant for centuries to millennia before the Sacarosa eruption. Volcanoes reawakening from such slumber to produce VEI 5 or greater eruptions usually experience weeks to months of anticipatory unrest (e.g., Zen and Hadikusumo 1964; Endo et al. 1981; Jiménez et al. 1999).

The Sacarosa magma was staged at ~ 10 km depth, providing opportunities to detect unrest during emplacement and recharge of the parent magma, and upon its ascent from this level to erupt. Magma rising to erupt typically triggers earthquakes (McNutt and Roman 2015; White and McCausland 2016), deforms the crust (Poland and de Zeeuw-van Dalfsen 2021), and exsolves gas (Kern et al. 2022). While rare eruptions with very little precursory activity are known to occur (Roman and Cashman 2018), Misti’s 2-ka eruption provides evidence that the volcano’s next eruption would likely be preceded by several weeks or months of forewarning (Harpel et al. 2011; Cobeñas et al. 2012; Tepley et al. 2013). Nonetheless, during both the 2-ka and possibly Sacarosa eruptions, the ultimate eruption-triggering intrusion ascended rapidly, implying that even with detectable precursory unrest, events can evolve quickly. The potential for a rapid-onset explosive eruption, even with precursory activity, highlights the value of a robust, multidisciplinary monitoring network. It additionally emphasizes that monitoring evidence for magma ascending within Misti’s upper 10 km could be concerning, especially in consideration of the magma inferred to currently reside beneath the volcano (Moussallam et al. 2017; Vlastelic et al. 2022). Continuous engagement with stakeholders could further facilitate a rapid and appropriate response at the onset of unrest.

Hazards

The Sacarosa TFD represents a powerful VEI 5 eruption; while such an event is not the most likely eruption scenario, it provides a robust example of a paroxysmal event at Misti. Occasionally, VEI 4 events are considered paroxysmal eruptions for Misti, with the 2-ka eruption often used as an example of such an event (Sandri et al. 2014; Thouret et al. 2022). Nonetheless, the Sacarosa eruption demonstrates that VEI 5 events occur at Misti and confirms Mariño et al.’s (2016) assessment that such eruptions are possible. Late Pleistocene and Holocene TFDs, such as the Autopista and Pampa de los Huesos deposits, are analogous to the Sacarosa TFD and provide evidence of the volcano’s continuing capacity to produce similarly powerful Plinian events. Mariño et al. (2007) mapped volcano hazards at Misti for eruptions up to VEI 6, with the zones for large-magnitude eruptions based on deposits from the volcano’s 2-ka and Autopista eruptions and those from the 1600 CE Huaynaputina eruption. Nonetheless, both the 2-ka and Autopista eruptions are applied as examples of VEI 4 events, while the Cayma stage TFDs and subsequent PDC deposits are used as evidence for VEI 5 eruptions (Mariño et al. 2016). Using the Sacarosa TFD for comparison, we suggest that multiple VEI 5 events, including the Autopista eruption, have occurred after the Cayma stage. However, deposits at Misti with distributions and thicknesses sufficiently wider and larger than the Sacarosa TFD to suggest VEI 6 eruptions have yet to be recognized. If present, identification and characterization of deposits from VEI 6 eruptions at Misti could facilitate hazards mapping efforts.

Were a future eruption similar to the Sacarosa event to occur, it could cover Arequipa and the surrounding area with tens of centimeters of tephra (Fig. 1b), causing power failures, polluting water resources, and closing the airport (Blong 1984; Wilson et al. 2012; Thouret et al. 2022). While few fatalities generally occur as a direct result of tephra fall, respiratory irritation, and deaths from roof collapses and other accidents are likely (Blong 1984; Horwell and Baxter 2006). Explosive eruptions also impact wide areas downwind, causing regional disruptions and destroying crops and pastures (e.g., Blong 1984; Jenkins et al. 2015; Elissondo et al. 2016). Fine, distal ash from such an eruption could additionally disrupt regional aviation hundreds to thousands of kilometers downwind (Guffanti et al. 2010; Jenkins et al. 2015; Prata and Rose 2015).

The R1 deposit (Fig. 2a; Legros 2001), while not coeval with the Sacarosa TFD, demonstrates that some of Misti’s Cayma stage eruptions generated PDCs, as is common during VEI 5 eruptions (Newhall and Hoblitt 2002). Modeling results indicate that dense PDCs and dilute surges from Misti can flow 10 km or more from the vent and enter developed areas (Sandri et al. 2014; Charbonnier et al. 2020). Several of Arequipa’s neighborhoods are also built upon and rapidly expanding farther into areas with PDC deposits from Misti’s most recent Plinian eruption at 2 ka (Harpel et al. 2011; Cobeñas et al. 2012; Charbonnier et al. 2020), indicating that people are living within reach of these deadly phenomena. The 2-ka eruption was smaller in volume than the Sacarosa eruption, indicating that the PDC hazard would be exacerbated in the unlikely event of a future paroxysmal eruption like the Sacarosa event. Additionally, despite the arid environment and Misti’s low available snow volume (Delaite et al. 2005; Harpel et al. 2011), lahars could be triggered in the volcano’s drainages due to rain-induced remobilization of loose debris emplaced by such an eruption (Mazer et al. 2020; Thouret et al. 2022).

Conclusions

By characterizing the Sacarosa TFD and its eruption, we document the most voluminous VEI 5 eruption from Misti yet known and provide an example of one of the volcano’s paroxysmal events. The Sacarosa TFD’s 1.1–1.3 km3 (DRE) of magma had temperatures of 799–815 °C, fO2 of NNO + 1.5, and trace-element concentrations typical for Misti. It remains unknown whether the Sacarosa eruption was preceded by antecedent unrest or intrusions into the edifice, but qualitative evidence suggests that the dacitic intrusion that eventually erupted rose rapidly from about 10 km to the surface. Upon the eruption’s initiation, 3 km3 of tephra was emplaced over tens of hours and dispersed to the southwest by winds of ~ 10–20 m/s. The eruption deposited as much as 40 cm of tephra at 20 km downwind from the vent and voluminous fine ash was likely distributed much farther afield. The eruption maintained relatively steady conditions until about half of the deposit was emplaced and it either slightly increased in vigor or its degree of fragmentation decreased. The eruption reached a peak MER of 7.7 × 106–4.1 × 107 kg/s and column height of 24–36 km asl. A similar future eruption, while unlikely, could distribute tens of centimeters of tephra over much of Arequipa, including its critical infrastructure. It could also be accompanied by PDCs and lahars.

The Sacarosa TFD was emplaced between 38.5 ka cal BP and 32.4 ka cal BP by the third eruption of the compositionally distinct Cayma stage. The Cayma stage eruptions began with deposition of the Cogollo TFD and PDC deposits, as early as 44.9 ka cal BP, and emplaced up to eight or more deposits over about 8.9–15.5 ky, including the newly named Anchi and Chuma TFDs (ESM 1). Several of Misti’s Cayma stage and subsequent andesitic TFDs have distributions and thicknesses sufficiently comparable to the Sacarosa TFD, that they represent similarly voluminous and powerful eruptions. Voluminous, VEI 5 explosive eruptions appear more common at Misti than previously documented and likely are also more common in the CVZ than the scant published record suggests.