Search USGSSearch

USGS · pp1791

The Novarupta-Katmai eruption of 1912 - largest eruption of the twentieth century; centennial perspectives

Abstract

The explosive outburst at Novarupta (Alaska) in June 1912 was the 20th century's most voluminous volcanic eruption. Marking its centennial, we illustrate and document the complex eruptive sequence, which was long misattributed to nearby Mount Katmai, and how its deposits have provided key insights about volcanic and magmatic processes. It was one of the few historical eruptions to produce a collapsed caldera, voluminous high-silica rhyolite, wide compositional zonation (51-78 percent SiO 2 ), banded pumice, welded tuff, and an aerosol/dust veil that depressed global temperature measurably. It emplaced a series of ash flows that filled what became the Valley of Ten Thousand Smokes, sustaining high-temperature metal-transporting fumaroles for a decade. Three explosive episodes spanned ~60 hours, depositing ~17 km 3 of fallout and 11±2 km 3 of ignimbrite, together representing ~13.5 km 3 of zoned magma. No observers were nearby and no aircraft were in Alaska, and so the eruption narrative was assembled from scattered villages and ship reports. Because volcanology was in its infancy and the early investigations (1915-23) were conducted under arduous expeditionary conditions, many provocative misapprehensions attended reports based on those studies. Fieldwork at Katmai was not resumed until 1953, but, since then, global advances in physical volcanology and chemical petrology have gone hand in hand with studies of the 1912 deposits, clarifying the sequence of events and processes and turning the eruption into one of the best studied in the world. To provide perspective on this century-long evolution, we describe the geologic and geographic setting of the eruption - in a remote, sparsely inhabited wilderness; we review the cultural and scientific contexts at the time of the eruption and early expeditions; and we compile a chronology of the many Katmai investigations since 1912. Products of the eruption are described in detail, including eight layers of regionwide fallout, nine packages of ash flows, and three lava domes that followed the explosive pyroclastic episodes. Changes in the proportions of coerupting rhyolite, dacite, and andesite pumice documented for the fallout and ash-flow successions, which are locally interbedded, permit close correlation of those synchronously emplaced sequences and their varied facies. Petrological correlation of the sequence of deposits near Novarupta with ash layers at Kodiak village, 170 km downwind, where three episodes of ashfall were recorded (to the hour), provides key constraints on timing of the eruptive events. Syneruptive collapse of a kilometer-deep caldera took place atop Mount Katmai, a stratovolcano centered 10 km east of the eruption site at Novarupta, owing to drainage of magma from beneath the Katmai edifice. Correlation of ~50 earthquakes recorded at distant seismic stations (including 14 shocks of magnitude 6.0 to 7.0) to fitful caldera collapse provides further constraints on eruption timing, because layers of nonjuvenile breccia and mud ejected from Mount Katmai during collapse pulses are intercalated with the pumice-fall layers from Novarupta. Structure of the Novarupta vent, a 2-km-wide depression backfilled by welded tuff and inferred to be funnel-shaped at depth, is described in detail, as is the 4-km-wide caldera at Mount Katmai. Discussions are also provided concerning: (1) the impact on global climate of the great mass of sulfur-poor but halogen-rich aerosol ejected into the atmosphere by the rhyolite-dominated eruption; (2) chemical and mineralogical effects of the fumarolic acid gases; and (3) the timing of several syneruptive landslide deposits sandwiched within the pumice-fall sequence. Secondary posteruption phenomena characterized include impounded lakes, ash-rich debris flows, phreatic craters on the ignimbrite sheet, responses of glaciers to the fallout blanket and to beheading by caldera collapse, growth of new glaciers inside the caldera, and gradual filling of the caldera lake. Structure, composition, and ages of the several andesite-dacite stratovolcanoes, closely clustered near Novarupta, all of which remain fumarolically and seismically active, are summarized. But among them only Mount Katmai extends compositionally to include basalt and rhyolite. The petrological affinities of 1912 magmas erupted at Novarupta with pre-1912 Katmai lavas are outlined, and various chemical, mineralogical, isotopic, and experimental data are assembled to construct a model of preeruptive magma storage beneath Mount Katmai. The monograph concludes by comparing the 1912 eruption with several other well-studied large explosive eruptions, 14 of them historical and 9 prehistoric. Finally, we retrospectively review the historical difficulties in understanding what had actually taken place at Katmai in 1912 and the century of progress in volcano science that has allowed most of it to be figured out.

Explore related subjects

90° N90° S · 180° W ← longitude → 180° E
Source-reported bounding extent: 40° to 75° latitude; -170° to -110° longitude. This indicates report coverage, not an exact sampling location. View area on OpenStreetMap.

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Wes Hildreth, Judy Fierstein. 2012. The Novarupta-Katmai eruption of 1912 - largest eruption of the twentieth century; centennial perspectives. https://doi.org/10.3133/pp1791

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related USGS reports

Resurvey of the Marble Canyon and Bridge Canyon dam sites in Grand Canyon National Park—Changes in sediment storage and evidence supporting the occurrence of bedrock incision through the mid-20th century

The Bureau of Reclamation developed an extensive plan for a network of dams, water tunnels, and hydropower plants in and around Grand Canyon, Arizona, in the 1940s through 1960s. The two largest of these planned dams were the Marble Canyon and Bridge Canyon Dams on the Colorado River. Though these dams were ultimately never built, Reclamation conducted extensive topographic, bathymetric, and subsurface exploration work at the sites proposed for these dams in the 1940s and 1950s. Resurveys of these dam sites were conducted between 1998 and 2021 to determine the changes in sediment storage at these dam sites caused by the upstream construction and operation of Glen Canyon Dam and by the recession of Lake Mead, the reservoir impounded by Hoover Dam. The resurveys of the Marble Canyon dam sites indicate that the post-1950s changes in sediment storage at these dam sites are broadly consistent with flux-based estimates of voluminous sand erosion from Marble Canyon since the 1963 closure of Glen Canyon Dam. These resurveys also suggest that the pre-dam longitudinal variation in sediment thickness over bedrock played a key role in determining the locations of the sand erosion induced by Glen Canyon Dam; more sand eroded from locations where more sand was present in the 1950s. The resurvey of the Bridge Canyon dam sites indicates that the Colorado River’s incision of the Lake Mead delta is regulated both by bed-sediment grain size and downstream hydraulic controls. Finally, analyses of bed-sediment thickness and sedimentological data at the dam sites, and observations of bed scour and gravel transport, suggest that sufficient bedrock was exposed to allow bedrock incision during commonly recurring pre-dam snowmelt floods that entrained small boulders into transport.

Arizona

The eruptive behavior of distributed volcanism forming low shield edifices—A case study of Sentinel-Arlington volcanic field, U.S.A.

Distributed volcanic fields are present in various tectonic settings worldwide, and their characteristics reflect differing influences from magmatic and tectonic processes. In the southwestern United States alone, there are 37 Quaternary distributed volcanic fields. After the primary period of extensional tectonics in the southern Basin and Range 15–5 million years ago, the Sentinel-Arlington volcanic field developed in southwestern Arizona between 4 and 1 million years ago. The Sentinel-Arlington volcanic field consists primarily of low relief shield volcanoes, a type of distributed volcanism with poorly understood temporal evolution. The Sentinel-Arlington volcanic field is less than 200 kilometers (km) from the Colorado Plateau, Gulf of California, and southern San Andreas Fault system. This work identifies and examines controls on the emplacement of the Sentinel-Arlington volcanic field by documenting shallow and surficial structures as well as eruption characteristics and style through time. The Sentinel-Arlington volcanic field consists of 21 volcanoes with a total of 33 vents over an area of about 770 square kilometers (km 2 ). The prominence of low relief shield volcanoes may be explained by ascent of basaltic magmas through thin Basin and Range crust, without much crustal contamination, and low viscosities common to mafic compositions. Typical eruption characteristics involve the construction of low relief shield volcanoes followed by Strombolian fissure eruptions at the summits or near-summit medial areas that produce scoria lapilli, which may weld to form agglomerate. The total lifetime erupted volume of about 4.3 cubic kilometers (km 3 ) represents an average eruptive flux of approximately 2x10 -3 km 3 per thousand years (k.y.). This erupted volume is low relative to Neogene basaltic intraplate distributed volcanic fields worldwide, which typically range from 10 -3 to 1 km 3 k.y. -1 . Sentinel-Arlington volcanic field eruptions were likely triggered by intermittent rejuvenation of transient magmatic zones that exist in thinned crust. Instantaneous flux from point sources feeding the lava flows is estimated to be on the order of 10 -1 to 10 cubic meters per second.

Arizona

Capitalization of positional (Lower/Middle/Upper) and temporal (Early/Middle/Late) adjectives in the names of formal chronostratigraphic and geochronologic units of the Phanerozoic

Many authors are understandably confused about the capitalization of the words “lower,” “middle,” “upper,” “early,” and “late.” Where these words are used simply as descriptive adjectives, they should be in lowercase; where they form the first word of a formal chronostratigraphic or geochronologic unit name, they should be in uppercase.

Professional Paper