Intended to write about recently active volcanoes in Japan and ran across Usu in southern Hokkaido, the northernmost main island of Japan. Turns out Usu is the most recent activity associated with the neighboring Toya Caldera created some 110,000 years ago. The caldera is filled with a lake and four islands in the center named Nakajima Island. This looks like an interesting place to explore.
Lake Toya is a volcanic caldera in the Shikotsu – Toya National Park which is also part of the Toya Caldera and Usu Volcano Global Geopark. The lake is nearly circular measuring 10 x 9 km and fills most of the caldera. Four islands made up of domes and a pyroclastic cone are in the center of the lake. Collectively, the islands are referred to as Nakajima Island.
Mount Usu (aka Usuzan and Usu-yama) is a small stratovolcano and dome complex that sits on the southern rim of the Toya caldera. Its top is cut by a somma caused by a flank collapse some 7,000 – 8,000 years ago. It is topped with multiple domes and cryptodomes. Three of these domes O-Usu, Ko-Usu and Showa – Shinzan and seven cryptodomes were formed since 1663. The growth of the Showa – Shinzan dome during the 1943 – 1945 eruption was documented by the local postmaster, the first detailed documentation of a growing lava dome.
The Nakajima (aka Naka-jima) Islands are a tuff cone, multiple dacitic to andesitic lava domes and a dacitic cryptodome in the center of Lake Toya. These formed some 50,000 years ago when a resurgent dome in the center of the caldera breached the surface of the lake. They are heavily forested.
The town of Toyako, population just over 9,200, surrounds most of the southern portion of the lake with Sobetsu, population nearly 2,700, located on the eastern side. As with most parks, recreation is the primary economic engine, with hiking, fishing, camping, and other outdoor pursuits available to visitors. The Smithsonian GVP lists nearly 215,000 living within 30 km of the lake with 3.2 million within 100 km.
The picturesque lake is popular in manga and anime and is used to model lakes in popular series.
Lake Toya is the northernmost Japanese lake that never ices over, though there is a competing claim with another nearby lake.
Online searches for information about Lake Toya and the surrounding parks and geoparks turn an incredible number of hits for potential visitors and tourists. Trails, museums, ropeway rides, gardens, temples, and hot springs are all featured attractions.
You can find the Usu Volcano Observatory here.
The dedicated Usu webcam no longer seems to exist. However, there are multiple webcams at Lake Toya that do capture the volcano. Here is one.
Toya, Usu and Nakajima started out as primarily andesitic / basaltic volcanoes. Domes formed at Usu since 1663 are rhyolites to dacites. There are extensive dacites and rhyolites from the climatic eruptions that formed the Toya caldera. Petrology suggests there are a pair of magma chambers under the volcano. The deepest at 10 km and the shallow one at 4 – 6 km. The upper chamber has a chemical gradient which would be expected from an evolving magma chamber.
The ancestral Toya dates as far back as perhaps 12 – 13 Ma. Oldest eruptive products are pyroclastics, lava flows, and hyaoclastics in the western coastal areas, indicating some portion of early activity was below sea level. Other andesitic volcanoes were active in the region at the time. There are welded tuffs in the SE coast of the lake, though origins of these are unknown. The largest sulphur mine in Hokkaido was located on the eastern flanks of the pre-caldera system.
In the late Pleistocene there was a large eruption near the center of the present-day Lake Toya, forming the Toya caldera. The ejecta was the Toya Pyroclastic Flow Deposit. Total volume of the eruption was over 20 km3 dense rock equivalent (DRE) along with widespread tephras and ash fall estimated at over 150 km3. This eruption is dated around 110,000 years ago. The pyroclastic flows from this eruption created a flat plain surrounding the caldera. Lake Toya is the third largest caldera lake in Japan.
Subsequent eruptions from the caldera lake started some 50,000 years ago. These built a pyroclastic cone and lava domes. There are an arc of maar remains around Higashiyama, one of the islands that make up the greater Nakajima Island. The island is made of multiple lava domes, at least one of which did not break the surface of the lake. Pottery and stone tools dating some 3,000 years have been found on one the islands, making it a unique archeological site in Japan.
Usu started erupting some 20,000 years ago beginning with basaltic – to basaltic-andesitic eruptions with lavas / scoria on the south wall of the Toya caldera. Early eruptive products flowed into the caldera, raising the water level of the lake. It also flowed outside as the edifice grew.
7 – 8,000 years ago, the summit of Usu collapsed and debris from this flank collapse flowed down the southern slope of the volcano to the sea. This avalanche engulfed and covered the Toya Pyroclastic Flow Deposits and formed the complex shoreline south of the volcano. The avalanche appears to be driven by an eruption, making it a hot avalanche. This left Usu with a double ring structure on the south wall of the Toya Caldera.
The volcano itself is around 6 km in diameter, 500 m above the surrounding countryside. It has a 1.8 km sommma, a lateral volcano (Donkoro-yama Scoria Cone), three dacite lava domes and multiple cryptoodomes. Lava domes of Usu are typically very viscous dacite magmas. The surface of the domes are red brick colored layers aligned in two zones NW – SE.
No volcanic ejecta was identified for several thousand years after the flank collapse, indicating the volcano was relatively quiet between the collapse some 7,000 years ago until the 1663 eruption. Human archeological sites were identified in the debris flow indicating the region was inhabited before the eruption.
Historical eruptions resumed from 1663, starting with Plinian pumice eruptions. Since then, Usu has erupted in pumice, ash and phreatic eruptions. There have been pyroclastic flows, surges, hot mudflows, lahars and formation of multiple lava domes and cryptodomes.
Usu erupted eight times since 1663. These eruptions were generally silicic magma, pumice, ash eruptions with dome emplacement. 1663 was a pyroclastic surge. 1769, 1822 and 1853 were pyroclastic flows. 1910 and 2000 were volcanic mudflows. General duration of pyroclastic eruptions was 2 – 24 months, depositing pyroclastic flows around the central volcano. Five of the eruptions were summit eruptions accompanied by pumice ejection (Plinian to sub Plinian) in their initial stages. Eruptions in 1910 – 1943 – 1945 and 2000 were flank eruptions from the NE / NW foot. These were initially phreatic or phreatomagmatic. Magmas erupted were thick dacites.
Earthquakes warned of the August 1663 eruption for a few days before the eruption. Massive ashfall buried houses and burned people (pyroclastic surge). The eruption ended the end of August. Total ashfall volume was up to 2.5 km3 of pumice followed by pyroclastic surges. Large blocks were mixed in the surge deposits indicating destruction of a dome. A crater plain on the Usu Gairinzan Trail called Minami Gairinzan was formed by a low-temperature pyroclastic surge immediately after the Plinian eruption phase of the 1663 eruption. Other parts of Gaininzan were formed by the flank collapse 7,000 years ago. This is the largest historic eruption at VEI 5.
Video of Usu 2000 eruption uploaded by Geoff Mackley to YouTube in 2009. Video is 9+23 long.
There was a small eruption in the late 17th century that deposited a pumice fall, volcanic ash and pyroclastic surge. There is no detail of this eruption.
The January 1769 eruption began with a Plinian / sub Plinian plume. Later stages of this eruption saw pyroclastic flows as “fire fell everywhere.” Flows were on SE, SW and N sides of the cone. The Kousu Lava Dome is thought to have formed in the later stages of this VEI 4 eruption.
The VEI 4 1822 eruption produced pyroclastic flows that killed 103, the deadliest eruption known from Usu. This village was previously evacuated the first day of the eruption. They unfortunately returned after three relatively quiet days before the second pyroclastic flow.
Earthquakes for the VEI 4 April 1853 eruption began about 10 days before the eruption began. Eruption was Plinian to sub Plinian with extensive pumice and ash fall 2 – 3 m thick on the eastern flank. There was a pyroclastic flow toward the end of the eruption. Eruption produced the Ousu Lava Dome which not only continued to smoke for two years after the eruption ended in May, but also continued to grow for the next 55 years. This eruption created a cryptodome on the SE side of the new dome.
Activity started again in July 1910 with earthquakes and then eruption from the northern foot of the volcano. These continued intermittently forming multiple explosion craters. The number of these grew from 15 in August to over 45 in November. Eruptions were all phreatic producing volcanic blocks. A high temperature, fast-moving volcanic mud flowed out of six of the craters. There was uplift in the area of the eruption and the Mt. Yosomi-yama cryptodomes formed. Genta Ana Crater is one of the 45 craters created during this eruption. Magma from this eruption powers current hot springs and fumaroles. The Lake Toya hot springs at the foot of Usu on the southern shore of the lake are referred to as the “volcano’s blessing” and are a symbolic theme of the geopark. This sequence was classified as a VEI 2.
Earthquakes began in Dec. 1943 near the eastern peak of Usu. Phreatic explosion took place in late June. There were dozens of explosions through October that formed seven craters and a cryptodome some 250 m above sea level. By December triangular domes grew from the cryptodome. This grew through Sept. 1945 and was named Mt. Showa – Shinzan topping out at 407 m. The local postmaster documented its creation and growth in a series of sketches, one of the first documentation of a growing dome. The dome is still active with some parts of it reaching temperatures over 300 C. Inflation dammed part of a local river forming the Shinzan Numa swamp. It also destroyed a portion of a railway near the eruption. This is classified as a VEI 2.
The first earthquakes for the 1977 eruption began in August. The Plinian plume began the next day reaching at least 12 km. The plume dusted the area east of the volcano with a significant amount of ash and pumice. The first stage of this eruption stopped around a week later. Magma was dacitic. Explosions formed three craters at the east foot of the Kousu dome. Ashfall during the first week was thick enough to collapse nearby houses. Rain mixed with it blanketed nearby agriculture and forestry. Uplift continued after the first week of the eruption expanding the fumarole field, uplifting the crater floor and growing a new cryptodome (Usu-shinzan).
The second stage of the eruption began in November with a small phreatic eruption. The phreatic explosions continued through January with frequent phreatomagmatic explosions July – October. A low temperature pyroclastic flow was emplaced. By late October 1978, heavy rains dislodged the ashfall causing widespread lahar flows at the foot of Usu, damaging many homes and killing three. Ginnuma was a swampy pasture which was the penetrated by the biggest crater created by the 1978 explosions. The 1977 – 1982 eruption sequence created a new cryptodome at the summit of Usu – Shinzan. The northern side of the summit is known as Northern Somma. It was pushed and deformed toward the lake during this activity. The Usu – Shinzan dome grew some 170 m taller during these eruptions. It continued to grow through 1982. This series of eruptions was classified as a VEI 3.
The Usu Volcano Observatory was established on Usu just before the start of the 1977 – 1982 series of eruptions. It detected precursory movements and seismic events leading to a successful evacuation of residents before the 2000 eruption.
Increasing seismic activity in March 2000 got to the point where an evacuation was ordered March 30. The first phereatomagatic explosion took place on the western foot of Usu, near the Mt. Kompira Yama dome, on a western extension of the 1910 crater chain. What is described as fierce activity, probably multiple phreatomagmatic explosions, opened over 65 craters during the course of the eruption. Lahars and hot mud flows during this eruption did not harm anyone due to the evacuation, though there was extensive property damage with at least 850 homed damaged, 119 destroyed, along with damage to local roads, water and sewer lines. This sequence ended Sept. 2001 and was a VEI 2.
The Nakajima Islands grew through the top of a resurgent dome in the middle of Lake Toya. Although no hyoclastics are found, it appears the islands grew through a resurgent dome emplaced following the caldera formation. There are eight lava domes, a cryptodome, and a tuff cone making up the islands.
The only volcanic cone on the islands is the Higashiyama pyroclastic cone. It is the source of the only large volume pyroclastic fall deposit from the islands. Activity appeared to begin with phreatomagmatic explosions that shifted to magmatic explosions and ended with magmatic effusions that created the Higashiyama dome. The tuff cone and neighboring dome texture suggest that the eruptions took place in a water-rich environment or a shallow water environment. Lake fill timing is unknown but assumed to be shortly after the caldera was formed.
There is mudstone and sandstones around the Hokutou-miasaki and Kannon-jima domes. These suggest that the caldera floor was uplifted prior to dome emplacement. The domes are thought to have bene formed by dacitic to andesitic magma.
The resurgent dome of the Nakajima islands is much smaller than that of Valles caldera. This may be due to relatively soft caldera floor deposits of non-welded pyroclastic tuffs. The Toya pyroclastic deposits are not welded for the most part.
There is a terrace on the volcanic domes and cone of Nakajima around 56 m above the current level. This means that the water level of the lake rose after dome emplacement. There are at least four terraces on Toya’s inner caldera wall, all below the 56 – 57 m level. Water level rise is thought to be due to geological uplift or injection of volcanic debris from the newly formed Usu.
The evolution of the Nakajima Islands is now thought to start with the caldera formation 110,000 years ago with a violent silicic eruption and associated pyroclastic flows. The caldera shortly filled with water forming a caldera lake. The caldera floor uplifted 45,000 – 40,000 years ago with another silicic magma intrusion which formed a small resurgent dome. Volcanism began through the top of the dome forming a number of lava domes, a cryptodome, and a tuff (pyroclastic) cone. The formation of Usu some 19,000 years ago raised the lake level some 56 m producing a terrace. Since then, the lake level decreased to the present level.
Hokkaido sits near the southern end of the Okhotsk Plate. Basic tectonics in this region are defined by subduction of the Pacific Plate under the Okhotsk Plate. As such, volcanoes in Hokkaido and north into the Kuril Islands are volcanic arc volcanoes. North of Hokkaido, the subduction line is the Kuril Trench. South of Hokkaido, it becomes the Japan Trench. Midway down the Japanese mainland we also have the multiple junction between the Philippine Sea, Pacific, Okhotsk, and Eurasian Plates. This junction is by definition, complex. The entire region sees massive megathrust earthquakes as the subduction proceeds in fits and starts.
The southern tip of the Okhotsk Plate is depicted from time to time as is own microplate, with that dividing line very close to underlying the Toya region. The Toya region is also abeam the transition of the Kuril Trench to the Japan Trench.
Final piece of the puzzle is extension in the Japan Sea and Okhotsk Sea to the west of Japan. To the north, this has formed the Kuril Basin between Sakhalin Island and the Kuril Islands. East of the Japanese mainland, this is called the Japan Basin. Both basins were formed by similar processes, rotation and oceanward retreat of an arc-trench system. And it is the formation of the Kuril Basin that makes things interesting.
The Pacific Plate has been subducting under the Eurasian and Okhotsk Plates since at least 42 Ma. This subduction generated volcanism along the NE Japan and Kuril arcs. Back arc spreading of these arcs separated them from the Eurasian and Okhotsk blocks, forming the Japan, Yamato and Kuril Basins. An oblique collision between the Eurasian and Okhotsk Plates took place at Hokkaido and Sakhalin at least 15 Ma. This joined the Japan and Kuril volcanic arcs. The collision also interrupted volcanic activity in the Kuril arc for a time.
The Miocene back arc volcanism some 15 Ma was 200 km wide, with volcanic activity spreading southward. In the last 2 Ma, the volcanic arc is around 60 km wide, volcanism fixed at the volcanic front. The suture line for the two volcanic arcs is NE of the Toya complex. This suture line also intersects with the line of tectonic weakness separating the Eurasian Plate from the Okhotsk Plate running north through the long axis of Sahkalin Island, south through Hokkaido and forming yet another triple junction east of central Hokkaido. Fortunately, there is not significant tectonic movement along this line.
Hokkaido itself is divided into at least six tectonic belts with the Toya complex on the western belt.
Like the tectonics of Hokkaido, Toya is a complex volcanic system that started as an andesitic / basaltic – andesitic system and has been trending more dacitic and rhyolitic in recent centuries. In historic times, it has erupted perhaps every 20 – 50 years, leading it to be grouped among those volcanic systems that are “due to erupt.” Of course, no volcanic system is ever due to erupt.
What we do know is that it has very active hydrothermal system, a shallow magma chamber that is periodically destabilized with new andesites / andesitic – basalts, capable of at least 4 VEI 4+ eruptions in the last 500 years, and a VEI 7+ caldera forming eruption 110,000 years ago. The region is now a parkland, though there are people living on Usu itself. From time to time, they relearn the same lesson that Hawaiian Big Island residents recently relearned about Kilauea, which is not to make the lifestyle choice to live on active volcanoes. Unfortunately, unlike Kilauea, recent eruptions from Usu are intensely phreatic, with explosions, ashfall, lahars and hot mudflows.