The 946 eruption of Paektu Mountain in Korea and China, also known as the Millennium Eruption or Tianchi eruption, was one of the most powerful volcanic eruptions in recorded history and is classified as a VEI 7 event. The eruption resulted in a brief period of significant climate change in Manchuria. The year of the eruption has not been precisely determined, but a possible year is A.D. 946. The eruption ejected about of tephra and collapsed the mountain into a caldera, which now contains the crater lakeHeaven Lake. The eruption began with a strong Plinian column, and ended with voluminous pyroclastic flows. An average of of Plinian ashfall and coignimbrite ashfall covered about of the Sea of Japan and northern Japan. This ash layer has been named the "Baegdusan-Tomakomai ash". It probably occurred in winter in late A.D. 946. This was one of the largest and most violent eruptions in the last 5000 years, along with the Hatepe eruption of Lake Taupo, the 1257 eruption of Mount Samalas near Mount Rinjani, and the 1815 eruption of Mount Tambora.
In 1996, Dunlap reported a high-precision wiggle-matching age determined at the University of Arizona as 1039 ± 18 AD. However, in 1998, Liu reported measurements from the center to the edge of the wood, followed by fitting with a high-accuracy tree ring calibrating curve, the obtained age of the Millennium eruption was determined to be 1215 ± 15 AD. In 2000, Horn reported another wiggle-matched radiocarbon dating with an AMS-mass spectrometer, and the interval of highest probability is 969 +24/-15 AD, which is widely used. In the 2000s, at least 5 high-precision wiggle-matching ages had been reported: 930–943 AD, 926 ± 10 AD, 945–960 AD, 931 ± 10 AD, and 946 ± 6 AD.
2013 wiggle-matching dating
Xu et al., reported 27 best wiggle-match datings from a single partially charred 264-year-old tree, which is Yin et al., also reported 82 best wiggle-matched AMS ages of samples from four carbonized logs, which is 938/939. However, the result of Xu et al. used a "regional offset" in their ages to decrease the error, and their new date was obtained from the longer tree-ring sequence with the higher analytical precision of ±25 years, on a 260-year tree-ring sequence that covers three consecutive wiggles around A.D. 910, A.D. 785, and A.D. 730. Since longer dated tree-ring sequence, finer sample resolution, and higher analytical precision all facilitate more and tighter tie-points for better WM dating. The new date is believed to represent yet the best high-accuracy and high-precision WM chronology for the Millennium eruption. Xu's wood samples were cut from a tree growing in the area about 24 km from the vent of Changbaishan volcano, it is not clear if volcanic CO2 emission before the eruption could affect the samples and produce ages that are slightly too old. The best WM dates for the Millennium eruption use the outliers-removed subset of the original measurements and also account for the effect of possible regional offset, and yielded two nearly identical WM ages of and, where overall and combined agreement indices of the models reach their highest values. Therefore, the average of these two WM ages represents the best modeled WM age for the Millennium eruption.
Historical records
The book Goryeosa describes "是歳天鼓鳴赦" and "定宗元年天鼓鳴" in 946 A.D. Also, the book «興福寺年代記» records "十月七日夜白灰散如雪" on 3 Nov 946. The thunders may relate to the Millennium eruption, and white ash rain may relate to B-Tm ash. Three months later, on 7 February, "十四日，空中有聲如雷鳴" and "正月十四日庚子，此日空中有聲，如雷" were recorded in «貞信公記» and «日本紀略». Another similar record is on 19 Feb 944, in «日本紀略» "廿三日丙申，子刻，振動，聲在上". Based on the historical record, the eruption may have started on February 944 or November 946, reaching a climax in February 947.
Ice-core
Sun et al., found volcanic glass in Greenland, which could well have originated in the Millennium eruption magma. The age of the volcanic glass layer is 939/940 A.D. However, Sigl et al., found out that ice-core chronologies are 7 years offset, and the Millennium eruption glass layer should be in 946/947 A.D. This conclusion is consistent with wiggle-matching dating and history records.
Multi-proxy dating
Oppenheimer et al., A radiocarbon signal of 775 CE in a subfossil larch engulfed and killed during the initial explosive eruption, combined with glacial evidence from Greenland, dates the eruption to late 946 CE. This date rules out the Millennium Eruption contribution to the collapse of the Balhae in 926 CE. They also did not see a consequent cooling signal in tree-ring-based reconstructions of Northern Hemisphere summer temperatures. The new date focuses attention on the chronicle from a temple in Japan that reports "white ash falling like snow" on 3 November 946 AD.
Eruption volume
The eruption volume has not been well constrained, from 70 to 160 km3. Machida et al. roughly estimated the proximal volume no more than 20 km3, and the volume of distal B-Tm ashfall attains more than 50 km3. A low total bulk volume estimate is 70 km3. Horn and Schmincke used an exponential method for minimum area/thickness and maximum area/thickness to obtain the volume of the Plinian ashfall as 82 ± 17 km3, and used the area-thickness method for ignimbrite to obtain 14.9 ± 2.6 km3. The total bulk volume estimate was 96 ± 19 km3. Liu et al., also used same method with Horn and Schimincke to calculate the volume of Plinian ashfall, and obtained the similar value of 83 km3. However, Liu used a different area-thickness value for ignimbrite. Liu assumed a distribution of ignimbrite within 40 km of the caldera, and an average ignimbrite thickness of 7.47m, yielding a volume of ignimbrite of 37.5 km3. The total bulk volume from this is 120 km3. Guo et al., used the exponential method estimate that volume of ashfall is 135.2 ± 7.8 km3. But Guo assumed the geometry of ignimbrite is a cone, and the volume of ignimbrite could be 20.1 km3. Guo also calculated the volume of valley-ignimbrite, because in a valley the thickness of ignimbrite could be 80 m. Then, the total bulk volume is 161.6 ± 7.8 km3. However, 100–120 km3 has been widely used.
Eruption dynamics
Base on the sequence of pyroclastic, the eruption began with pumice and ash falls, and then an eruption column collapse formed ignimbrite. The column collapse probably was a pulsing collapse, because the ignimbrite and pumice fall deposits are interleaved. Machida et al., divided the Millennium eruption into 4 stages: Baegdu Plinian pumice fall, Changbai pyroclastic flow, Yuanchi tephra falls, and Baishan pyroclastic flow. But Baishan pyroclastic flow may related to post-caldera activity. More recent study indicate that eruption include 2 stages: Plinian pumice fall and unwelded ignimbrite.
Plinian stage
This stage formed a large area of white comenditic pumice and ash. The Plinian eruption column reached about 36 km in height. B-Tm ash and "white ash rain" may related to this stage. Based on variations of grain-size and thickness of pumice, the Plinian stage can divided into 3 parts: early period, climax, and later period.
Early period
In a Plinian pumice-fall section, the grain-size of pumice is reversely graded. The variation of pumice size shows a major fluctuation in eruption column height during this Plinian event. Based on distribution of maximum lithics clasts in the early eruption, the eruption column probably reached 28 km, and mass discharge rate attained 108 kg/s. The Early period may have released 1.88–5.63 × 1019 joule, and the eruption might have lasted for 33.5–115.5 hours.
Climax period
Based on the distribution of crosswind of maximum lithics clasts, the top of eruption column might have reached 36 km, with a mass discharge rate of around 3.6 × 108 kg/s. The distribution of downwind of maximum lithics clasts showed the wind direction at the time to be SE120°, and the wind speed to be 30 m/s. The height of eruption column, the water content of magma, and the temperature of magma indicate that radius of the eruption vent was 200 m. The climax of the eruption may have released 4.18–12.43 × 1019 joule, and the eruption might have lasted for 35–104 hours.
Later period
This period eruption formed the upper part of the Plinian pumice fall, which is the fine pumice. The later Plinian pumice fall and pyroclastic flow occurred simultaneously, because some sections show that pumice fall and ignimbrite are interleaved. Combining the grain-size of pumice and thickness of pumice fall, the height of eruption column of later period was no higher than 14 km, and the mass discharge rate was 5 × 106 kg/s. The later period may released 8.76–26.16 × 1017 joule for Plinian eruption and keeping eruption column.
Ignimbrite stages
In many sections, a large grey ground-surge under an ignimbrite sheet, which might be from the front part of pyroclastic flow, and the unwelded ignimbrite always underlie a large ash-cloud surge. Ignimbrite was deposited within a 40 km radius of the caldera, at an average thickness of 7.47 m. In many valleys, the thickness of ignimbrite may be 70–80 m. The Changbaishan ignimbrite has a low aspect ratio of 1.87 × 10−4. The speed of the initial pyroclastic flow might have been 170 m/s, and 50 m/s at 50 km away from the caldera.
Duration
The vent radius and water content of magma indicate that the average volume discharge rate of the Plinian eruption and ignimbrite was 1–3 × 105 m3/s. A total bulk ejecta of 120 km3 was derived from bulk volumes of pumice fall and ignimbrite of 83 km3 and 37.5 km3, respectively. The ignimbrite-forming eruption may have lasted one and a half to four days, while the Plinian eruption may have lasted three to nine and a half days. The total duration of the eruption may have been four and a half to fourteen days.
Volatiles
Plinian volcanic eruptions can inject a large amount of volatiles and aerosols into the atmosphere, leading to climate and environment changes. Chlorine concentrations in the peralkaline from the Millennium eruption were postulated to have reached up to 2% and an average of 0.44%. The Millennium eruption was thus thought to have emitted an enormous mass of volatiles into the stratosphere, potentially resulting in a major climatic impact.
Chlorine
McCurry used an electron microprobe to analyze the volatile in glass inclusion of feldspar. McCurry concluded that Millennium eruption may have released 2000 Mt Cl. Liu used chromatography to analyze the average of volatile of 5 whole-rock samples, and the contents of halogen is 0.08%–0.11%. A more recent and more detailed study by Horn and Schmincke used an ion probe to analyze the average of volatile in 6 of matrix glass and 19melt inclusions, and the average of content of Cl in melt inclusions and matrix glass were found to be 0.4762% and 0.3853%, respectively. Horn and Schmincke concluded that the Millennium eruption may have released 45 ± 10 Mt of Cl. Another author, Guo, who studies petrology and geochemistry, shows the average of contents of Cl in melt inclusions and matrix glass to be 0.45% and 0.33%, respectively. They concluded that Millennium eruption may have released 109.88 Mt of Cl, and 15.82 into the stratosphere. The Cl contents in the melt inclusions are similar to those of Mayor Island, and higher than those of Tambora, Krakatau and Pinatubo. The large difference of results between Guo and Horn is because Guo used higher volume and density of magma.
Sulfur dioxide
Liu used chromatography to analyze the average of volatile content of five pumice and obsidian samples, finding the contents of sulfur to be 0.0415%, and Liu assumed the degassing efficiency factor of sulfur is 0.3. Liu estimated that the Millennium eruption may released 40 Mt of sulfur dioxide. However, Horn and Schimincke calculated that only 20% of the sulfur in the magma had been degassed, because 80% of all analyses of inclusions and matrix fall below the detection limit of an ion probe. The results of average contents of sulfur in 19 of inclusions are 0.0455%, Horn assumed the contents of sulfur in matrix glass are 0.025% because 250 ppm is detection limit of the ion probe. They concluded that the total sulfur dioxide released from eruption was only 4 ± 1.2 Mt, but Horn suggests that may be excess sulfur accumulated in the vapor phase. Guo calculated the average contents of sulfur in nine glass inclusions and one matrix glass are 0.03% and 0.017%, respectively. The results of Guo are 23.14 Mt of sulfur dioxide released from eruption, and 3.33 Mt of sulfur dioxide input to stratosphere. The sulfur contents in glass inclusions show the reverse correlation with SiO2 concentrations, indicating that sulfur solubility in magma is controlled by magma differentiation process because of the occurrence of the S-rich fluid inclusions.
Fluorine
Liu used chromatography to analyze the average of volatiles of five pumice and obsidian, finding the fluorine content to be 0.0158–0.0481%. Horn and Schimincke used an ion probe to find an average fluorine content in inclusions of 0.4294%, but fluorine concentrations in matrix glass show a significant bimodal distribution into F-rich and F-poor. In order not to over-estimate syn-eruptive fluorine loss, they considered this bimodal distribution of fluorine for calculating the volatile difference between matrix glass and melt inclusions. The volatile loss is approximately 300 ppm F for melt inclusion and F-rich matrix glass, whereas it is 1900 ppm F for melt inclusion and F-poor matrix glass. Horn concluded that 42 ± 11 Mt of F was released by the eruption. Guo, based on less samples, calculated that F contents in inclusions and matrix glass are 0.42% and 0.21%, respectively. Guo concluded that 196.8 Mt of fluorine was released from eruption, with 28.34 Mt of F injected into the stratosphere. With magma evolving, halogen contents increase irregularly, parallel to the increase of SiO2 concentrations in glass inclusions The large difference of results between Guo and Horn is because Guo used a higher volume and density of magma, and higher difference contents between matrix glass and inclusions.
Vapor phase
Sulfur is not strongly enriched during differentiation, in contrast to water, chlorine, and fluorine. The reason could be pre- or syn-eruptive degassing of a separate vapor phase, such as that postulated for the Pinatubo and Redoubt eruptions. The ultimate source for the excess volatile observed during the 1991 Pinatubo eruption is assumed to be sulfur-rich basaltic magmas underlying, and syn-eruptively intruded, into the overlying felsic magmas. The sulfur-rich trachytic and trachyandesitic magmas which underlay the rhyolitic magma at Changbaishan may have been a possible source for excess sulfur accumulation. If this scenario is realistic, clearcut proxies for the environmental impact of the eruption would be expected. Millennium eruption magmas are predominantly phenocryst-poor comendites plus a volumetrically minor late-stage, more phenocryst-rich trachyte. Sizable glassy but bubble-bearing melt inclusions are widespread in anorthoclase and hedenbergite phenocrysts, as well as in rarer quartz and fayalite phenocrysts. Comparing the relative enrichments of incompatible volatile and nonvolatile elements in melt inclusions along a liquid line of descent shows decreasing volatile/Zr ratios suggesting the partitioning of volatiles into a fluid phase. This suggests that current gas-yield estimates for the Millennium eruption, based on the petrologic method, could be severe underestimates.
Climate effects
The Millennium eruption is thought to have emitted an enormous mass of volatiles into the stratosphere, likely resulting in a major worldwide climatic impact, though more recent studies indicate that the Millennium eruption of Changbaishan volcano may have been limited to regional climatic effects. However, there are some meteorological anomalies in A.D. 945–948 which may relate to the Millennium eruption.
