SN 1006: the brightest supernova (2024)

Despite the rapid advances in astronomy in recent years, centuries-old observations continue to play an important role today. This is especially true in the study of supernovae, violent stellar outbursts that may lead to either the complete dissipation of a star or the collapse of its core to an extremely compact neutron star. Modern telescopic searches have detected many supernovae occurring in external galaxies, and as recently as 1987 a supernova was observed in the Large Magellanic Cloud, a satellite of our own galaxy. Yet the last supernova to be observed in the Milky Way itself occurred as long ago as 1604, a few years before the advent of telescopic astronomy. Hence for observations of galactic supernovae explosions we still have to rely on sightings made with the unaided eye. Fortunately, an impressive array of early records survives, from a variety of cultures.

Remnant

At its height, a supernova can shine as brightly as several hundred million Suns, reaching an absolute magnitude of around —19. Hence even without optical aid such a spectacular explosion may be detected at a distance of a few kiloparsec. After the original flare-up, the decline in brightness is very gradual, over a period of several years. Furthermore, the supernova remnant (SNR) — a rapidly expanding cloud of gas and dust — continues to act as a powerful source of electromagetic radiation (especially in X-rays and radio waves) over several thousand years (figure 1). Successful identification of the original outburst enables the precise age of the associated SNR to be fixed — itself an important result in the study of its evolution. Furthermore, even archaic observations often yield important information about the peak brightness and light variation. Hence the search for historical records of supernovae is much more than a purely scholarly pursuit.

Over the past two millennia, reports of numerous sightings of “temporary stars” of various kinds can be found scattered throughout the pages of history, notably from East Asia (China, Japan and Korea), Europe and the Arab world. These objects probably represent a mixture of comets, variable stars and novae — as well as the occasional supernova. In our joint researches on historical supernovae and their remnants, David Green and I (e.g. Stephenson and Green 2002) have mainly focused on those new stars that were said to remain fixed for at least six months. This restriction effectively eliminates most spurious sightings of supernovae. During such a long interval, comets move across an extensive arc of sky, while even bright novae usually fade to far below the unaided eye limit.

The earliest known new stars that rate as potential supernovae occurred in AD 185 and 393. Both periods of visibility were long, respectively 20 months and eight months. However, the surviving records are brief and each star appeared in an area of sky where there are many SNRs. As a result, the remnants of these two new stars have yet to be confidently identified. No further supernova candidates were reported until after 1000. However, since then, five new stars — sighted in 1006, 1054, 1181, 1572 and 1604 — were visible for periods from between six months and around three years. All five stars appeared to have remained stationary, and in each case several independent records are preserved, which provide a good fix on the position. Correcting for precession, each location lies close to that of an established SNR.

By far the most brilliant supernova recorded in history was seen in 1006. This spectacular celestial outburst attracted widespread attention and around 30 separate reports are preserved — from China, Japan, various parts of the Arab world and Europe. The supernova was only lost to view at least three years after it first appeared. Fortunately, the various positional records enable the remnant of SN 1006 to be established confidently. Thorough investigations of many of the occidental and oriental sources were first made by Goldstein (1965) and Goldstein and Ho Peng Yoke (1965) and since then several further records of the supernova have come to light, from both China and the Middle East (Stephenson and Green 2002). Despite the extreme brilliance of the star, it has attracted significantly less attention than the supernova of 1054 (which produced the Crab Nebula). The aim of this article is to attempt to redress the balance.

1006 East Asian records of SN

It comes as no surprise to find that the most careful observations of the new star of 1006 originate from East Asia. Around this time, oriental astronomers were the only observers to maintain a systematic watch of the night sky. Their main motive was astrological; celestial changes were regarded as presaging important terrestrial events. In China, the new star was recorded in as many as nine separate sources, notably in various sections of the Songshi (“History of the Song Dynasty”: 960—1279). Other valuable Chinese sources include the Song Huiyao (“Important Documents of the Song Dynasty”), covering the period 960—1220, and the Xu Zhixi Tongjian Changbian (“Long Draft of the Continuation of the Comprehensive Mirror as an Aid to Government”), a detailed chronicle covering 960—1126.

Although the usual oriental term for a strange starlike object was kexing (“guest star”), in China, the supernova was consistently described as a Zhoubo (“Earl of Zhou”) instead. Choice of this term seems to be an allusion to the colour of the star, described in various Chinese works as either “yellow”, “golden” or “golden yellow”. An early history of China, dating from the 7th century, asserts that a Zhoubo is a “large (star) of a brilliant yellow colour”. However, in practice, this term was rarely used to describe actual observations of new stars.

In China, the astrological significance of the supernova was particularly emphasized. At the time, the Song Dynasty Emperor Zhenzong (reigned 998—1022), who was greatly influenced by divination, was on the throne. From the start of his reign, sacrifices were offered to the bright star Canopus, known in China as Shouxing (“Longevity Star”). A systematic watch for this star at heliacal rising was maintained annually around the time of the autumnal equinox.

In 1004, two years before the supernova appeared, an invasion by the Khitans from the north had compelled the Song court to temporarily abandon the capital of Bian (modern Kaifeng) and escape south. To establish peace, it was necessary for the Song to pay a huge annual tribute of silk and silver to the Khitans. Hence, when the supernova appeared, it is perhaps not surprising that the Song astrologers were at pains to reassure the Emperor that it was a propitious omen. So favourable was the prognostication submitted by the astrologer Zhou Keming that the Emperor promoted him to high office!

Of the many Chinese reports of the supernova, two — in the Song Huiyao and the Songshi Astronomical Treatise — are especially comprehensive (see extracts i and ii below). Both accounts are rather long and in the case of the Song Huiyao record I have extracted only the more significant details. For fuller information, see Stephenson and Green (2002).

(i) “Jingde reign period of Emperor Zhenzong, fifth lunar month, first day [30 May 1006]. The Director of the Astronomical Bureau reported that previously, on the second day of the fourth lunar month [1 May 1006], during the initial watch of the night, a large star had been sighted. It was yellow in colour and it appeared to the east of Kulou and to the west of Qiguan. Its brightness had steadily increased. (Its position) was measured as 3 du (deg) in (the lunar lodge) Di. Subsequently the star (further?) increased in brightness…” (Song Huiyao).

Throughout East Asia, the night was divided into five equal geng (watches). The initial watch, when the star was first sighted, would begin after dusk. Kulou and Qiguan were adjacent star groups in the constellations of Centaurus and Lupus, considerably to the south of the celestial equator. Di, the third of the 28 xiu (lunar lodges), defined the region of right ascension (some 15° wide) in which the star appeared. East Asian astronomy was equatorial, whereas its occidental counterpart was based on the ecliptic.

(ii) “Jingde reign period of Emperor Zhenzong, fourth lunar month, (day) wuyin [6 May 1006]. A Zhoubo star was seen; It appeared to the south of Di and 1 du (deg) to the west of Qiguan. Its form was like the half Moon and it had pointed rays. It was so brilliant that one could scrutinize things (jian). It passed through (li) the east of Kulou. During the eighth lunar month [27 August — 24 September 1006], following the wheel of Heaven (Tianlun), it entered the horizon. During the 11th month [24 November — 22 December 1006] it reappeared at Di. From this time onwards it was regularly seen in the 11th month in the early morning at the eastern direction, while in the eighth month at the southwest it entered the horizon” (Songshi, chapter 56).

The latter text — from the Songshi— is illustrated in figure 2. The expression jian means to examine closely or scrutinize. Just what was meant here by the term “passed through” (li) is not clear; the expression may merely mean “spent its time in”. However, the subsequent record makes it clear that the star remained fixed (“at Di”) for a period of several years.

In Japan, the supernova was also reported in a variety of sources: diaries of courtiers, chronicles, privately compiled histories, etc. Fortunately, early last century the Japanese scholar Kanda Shigeru (1935) made an extensive compilation of the diverse astronomical records of his country from AD 600 to 1600, thus rendering them readily accessible. In this work he included seven Japanese accounts of the 1006 star, which was consistently described using the traditional term “guest star”.

By far the most detailed of the Japanese accounts of the 1006 supernova is to be found in the Meigetsuki (“Diary of the Full Moon”): the diary of the Japanese poet-courtier Fujiwara Sadaie (1163—1241). There is an intriguing story behind Fujiwara's inclusion of this account in his diary. On a date corresponding to 4 December 1230, a “guest star” was sighted in Japan. As it happens, a parallel record of this same object from Northern China, detailing the motion of the star, clearly demonstrates that it was a comet. However, the event inspired Fujiwara to compile a list of eight further guest stars recorded in earlier Japanese history. These observations — made in the years 642, 877, 891, 930, 1006, 1054, 1166 and 1181 — were inserted en bloc in his diary on a date corresponding to 13 December 1230. Amazingly, three of these last four guest stars (appearing in 1006, 1054 and 1181) were supernovae; the remaining object (in 1166) was a comet.

The relevant part of the record of the 1006 supernova, as cited (along with reports of the supernovae of 1054 and 1181) by Fujiwara in his Meigetsuki, may be translated as follows:

“Kanko reign period of Emperor Ichijo In, third year, fourth lunar month, second day, guiyou [1 May 1006]. After night had fallen, there was a large guest star in Qiguan. It was like Mars, and it was bright and scintillating. For successive nights it was seen clearly at the southern direction. Some (people) suggested that it was a change in Qichen Jiangjun (κ Lup) itself” (Meigetsuki, volume 12).

Japanese astrography was based on that of China and we can be fairly confident that at least the bright stars of Qiguan were identical in both countries.

Arab and European records

Arab astronomers in general tended to show little interest in sporadic (and hence unpredictable) phenomena such as comets and new stars; they were much more concerned with cyclic (and hence predictable) events involving the Moon and planets. According to Aristotle, whose views were widely accepted in the medieval Arab world, the heavens were immutable; change could only occur in the sub-lunar regions. Hence comets and other intermittent celestial events were regarded as atmospheric phenomena — and thus of minimal astronomical importance. Reporting of comets and new stars was usually left to chroniclers, who — in general — had little astronomical knowledge. Hence in Arab records of SN 1006, technical information is usually rare. However, at al-Fustat in Egypt, a physician named Ali ibn Ridwan gave a careful description of the star:

“I shall now describe a spectacle which I saw at the beginning of my studies. This spectacle appeared in the zodiacal sign of Scorpio, in opposition to the Sun. The Sun on that day was 15 deg in Taurus and the spectacle in the 15th deg of Scorpio. This spectacle was a large circular body, 2½ to 3 times as large as Venus. The sky was shining because of its light. The intensity of its light was as bright as that of the Moon a little more than one quarter illuminated. It remained fixed, moving daily with its zodiacal sign until the Sun was in sextile with it in Virgo, when it disappeared suddenly…” (Ali ibn Ridwan: Commentary on the Tetrabiblos of Ptolemy).

Following the ancient Greek practice, the sign Taurus covered a range in celestial longitude from 30° to 60°, while Scorpio extended from 210° to 240°. Later in his account, Ali ibn Ridwan cites — for the time when the star first appeared — the longitudes of the Sun, Moon and five bright planets, as well as additional data. He presumably calculated these various figures from tables. On the basis of this information we may compute the date of first appearance of the new star as 30 April 1006. The expression “sextile” is an astrological term, indicating a difference in longitude by 60° (or two full zodiacal signs). When the star disappeared, the Sun would thus be in the 15th deg of Virgo, implying that the star was lost to view around 2 September.

In Europe, where there was little astronomical interest at this period, reports of the supernova are confined to chronicles. The only detailed European account of the supernova is contained in a chronicle of the Swiss monastery of St Gallen. Clearly the star left a vivid impression on eyewitnesses; its description is the only entry in the chronicle for the entire year:

“AD 1006. A new star of unusual size appeared; it was glittering in aspect and dazzling the eyes, causing alarm. In a wonderful manner it was sometimes contracted, sometimes spread out, and moreover sometimes extinguished. Nevertheless, it was seen for three months in the extreme limits of the south, beyond all the constellations which are visible in the sky” (Annales Sangallenses Maiores).

Although the supernova was much better placed for visibility in the southern hemisphere, no observation of the star can be traced from places further south than San'a', the Yemeni capital (latitude 15°N). It seems that in more southerly latitudes the star — although without doubt seen by very many people — failed to be recorded. In particular, a search of Indian sources (Narlikar and Bhate 2001) has proved negative.

Dates of first and last sighting

Chinese, Japanese and Arab sources agree closely on the date of first appearance of the new star in 1006. As noted above, a date of 30 April may be deduced from the rather cryptic information given by Ali ibn Ridwan. A Chinese work, the Song Huiyao, indicates a discovery date of 1 May, as does the Japanese Meigetsuki. However, although a further Chinese account — in the Songshi Astronomical Treatise — cites 6 May instead, three separate Arabic texts imply discovery on 2 or 3 May.

The supernova was said by Ali ibn Ridwan to be in exact opposition with the Sun at the time of discovery. Hence conjunction with the Sun would occur six months later, and thus close to 1 November. Various sources suggest that the supernova was lost to view in the evening twilight around September. For instance, a Chinese history — the Songshi— asserts that the star “entered the horizon” during the eighth lunar month (some time between 27 August and 24 September). In Japan, offerings were made to the star on 3 September and again on 14 September at the various shrines, while on 21 September a request for a general amnesty was made to Emperor Ichigo In “because of the appearance of the large star”. Presumably the star still remained visible on this last date. In Egypt, the date of disappearance, as specified indirectly by Ali ibn Ridwan, was 2 September. Several other Arab sources also imply a duration of visibility of around four months.

Once the star had been lost to view, only Chinese astronomers seem to have watched for its reappearance later in the year. A brief note in the Imperial Annals of the Songshi (chapter 7), announces that “on the day renyin in the 11th month the Zhoubo star was again seen”. This date corresponds to 26 November. Although the Songshi Astronomical Treatise (chapter 56), only notes that the supernova was recovered “during the 11th month” (24 November — 22 December), it asserts that the star “was again seen at Di”. Here we have direct evidence that the position of the star had not changed significantly.

However, the subsequent description in the Songshi Astronomical Treatise is remarkable: “From this time onwards it was regularly (changyi) seen in the 11th month in the morning at the eastern direction while in the eighth lunar month at the southwest it entered the horizon.” Evidently the supernova continued to be visible — rising in November/December and setting in August/September — for several years! Unfortunately, the exact number of years of visibility is not specified. However, a further source, the Xu Zizhi Tongjian Changbian, implies a duration of at least three years and affirms that the star remained at Di:

“Dazhong Xiangfu reign period, second year, ninth lunar month, (day) gengwu [9 October 1009]. (The Emperor) issued a command that from this day the ‘Suburban Sacrifice’ should be offered to both the Zhoubo star at Di lunar lodge and Shouxing (Canopus), and that this arrangement should be maintained for ever. This met with the approval of the Hanlin academicians and the astronomers” (Xu Zizhi Tongjian Changbian, chapter 72).

This intriguing report is illustrated in figure 3. The date 9 October, which would fall during the annual period of invisibility (between heliacal setting and rising) of the supernova, is nearly 3½ years after discovery. Unfortunately, no further records of the supernova can be definitely traced after 9 October 1009.

Brightness of the supernova

Despite the extreme southerly location of the supernova, Chinese, Arab and European records separately testify to its superlative brilliance around maximum phase, often comparing its brightness with that of the Moon! In China, one source reported that the star was “huge”, adding that “its bright rays were like a golden disc”. The Songshi affirmed that: “Its form was like the half Moon and it had pointed rays. It was so brilliant that one could scrutinize things.”

In Egypt, Ali ibn Ridwan described the supernova as “a large circular body, 2½ to 3 times as large as Venus”. He added: “The sky was shining because of its light. The intensity of its light was as bright as that of the Moon a little more than one quarter illuminated.” In Baghdad, it was said that: “It had dazzling rays … and brilliant light like that of the Moon.” Two European sources also emphasize the brilliance of the new star. At St Gallen in Switzerland it was “glittering in aspect and dazzling the eyes, causing alarm”, while at Benevento in Italy it was reported that “a very brilliant star gleamed forth”. In marked contrast, descriptions of the brightness of SN 1054 are unimpressive, the only comparison object being Venus. Clearly SN 1006 was an extraordinary object!

A Yemeni chronicle, citing an observation made at San'', went so far as to assert that the star's “light rays were similar to the rays of the Sun”. Although generous allowance must surely be made here for hyperbole, it is evident that the supernova was indeed extremely brilliant. At this latitude (15°N), the star would be much higher in the sky than at the other places of observation — most of which were between about 30° and 35° north of the equator. Further, the thin air at San'' (2500 m above sea-level) would scarcely dim the star's light.

Several of the above records enable a plausible estimate of the peak magnitude of the supernova to be made. The fact that Chinese observers were able to “scutinize things” (presumably nearby objects) by the light of the star gives a useful indication of its brightness. Some years ago, I discussed this question with a sea captain who had wide experience of observing in clear and very dark skies. He informed me that only when the Moon is about five days old (approximate magnitude —8.5 or —9), can one start to discern objects by its light, a result which my own observations tend to confirm.

By comparison, the estimate of the star's brightness by Ali ibn Ridwan (“as bright as the Moon a little more than one-quarter illuminated”) would imply a magnitude around —8.5. Summarizing, a peak magnitude for the supernova of around —8.5 seems a fair estimate, especially since I have made no allowance for the appreciable atmospheric absorption at the low altitudes at which most observers viewed the star.

Rather surprisingly, despite the brilliance of the supernova, only a single source (from either NW Africa or Spain) mentions its daylight visibility. However, when the star was first sighted it was almost in opposition to the Sun, and in a much more southerly declination (around —40° compared with +17°). Consequently, from the time the star first appeared it would be below the horizon during the hours of daylight for several weeks, rising after sunset and setting before sunrise.

Location of the supernova

The most careful East Asian observations of the star's position are from China and are contained in the Song Huiyao and Songshi. These are as follows. (i) “It appeared to the east of Kulou and to the west of Qiguan… Its position was measured as 3° in Di (lunar lodge).” (ii) “It appeared to the south of Di and 1° west of Qiguan.” Japanese observers separately described the position as “within Qiguan or “to the south of Qiguan”.

Di (“Root”), the third of the 28 xiu or lunar lodges, consisted of a group of four stars in Libra (α,ι, γ and β Lib). Di also defined a range of right ascension, at the epoch 1006 extending from the meridian of α Lib (13 h 57 m) to that of π Sco (15 h 00 m). In particular, the measured location of the supernova as 3° in Di, and hence 3° east of the meridian of α Lib, would correspond to a RA of approximately 14 h 09 m.

Both Kulou (“Depot Tower”) and Qiguan (“Cavalry Officers”) were fairly extensive star groups, lying to the south of Di. Figure 4a shows a section of a Chinese star chart covering a range in RA from 12 h to 24 h. This chart is a late copy of a now lost star map produced by the great Song Dynasty astronomer Su Song, which was originally printed around 1090. On this figure I have highlighted the star groups Di, Kulou and Qiguan— in each case transliterating their names. In the chart, Kulou is depicted as a single unified asterism, whereas Qiguan consisted of nine separate groups of three stars. Other medieval Chinese star charts depict similar representations of Kulou and Qiguan.

Figure 4b shows an enlargement of the southwest section of figure 4a, focusing on the region in which the supernova was seen. On this latter chart I have marked the range of RA covered by Di lunar lodge (denoted by parallel lines) and identified the celestial equator and ecliptic.

Detailed studies of Song Dynasty star charts and catalogues by Pan Nai (1989) enable most of the stars in Kulou and Qiguan to be identified with fair confidence. Figure 5 is largely based on the research of Pan Nai. However, in producing this diagram I have precessed all star positions back to the epoch 1006. As in figures 4a and 4b, Qiguan is depicted as containing nine individual groups of stars. As noted above, the single star κ Lup represents Qichen Jiangjun; this was regarded as a separate asterism.

The brightest stars represented in figure 5 are between magnitude 2 and 3. These are denoted by large circles; stars between magnitude 3 and 4 are indicated by circles of medium size. For all these brighter stars, the identification — in terms of Bayer Greek letters — seems secure. However, the locations of some of the fainter stars are questionable.

Although the supernova was described in separate Chinese sources as “to the east of Kulou and to the west of Qiguan” and as “1° west of Qiguan”, both are vague descriptions. In particular, precisely what was meant by the latter expression is by no means clear. The most accurate observation would appear to be the measurement of the RA of the star — quoted as 3° in Di lunar lodge. This is depicted in figure 5 by a vertical broken line (RA 14 h 09 m). The star should be expected to be located fairly close to this particular RA, although the real accuracy with which this measurement was made is doubtful.

The positional descriptions in Japanese sources tend to be obscure. Although the Meigetsuki reports the location of the supernova as “within Qiguan”, it also notes that “Some (people) suggested that it was a change in Qichen Jiangjun (“Cavalry General”: κ Lup) itself”. However, an earlier source (the Gyokuyo) indicates that the change in κ Lup occurred at some time between 1011 and 1016. Hence this event was not connected with SN 1006.

The most careful description of the position of the star in Arab records is from Egypt. Here, Ali ibn Ridwan specified the longitude of the star as “the 15th deg of Scorpio”. As in the case of the Chinese measurement, the precision of this estimate is uncertain; in particular, it seems that Ali ibn Ridwan was at pains to stress that the star was in exact opposition to the Sun in longitude.

Fortunately, the fact that the star was visible at all from St Gallen in Switzerland (latitude 47.4°N) sets a useful limit to its southerly declination. As seen from this site, the southern horizon is obscured by distant mountains, with an apparent altitude of up to 5°. A detailed study of the horizon profile at St Gallen indicates that a star with a declination south of about —39° would be invisible (see Clark and Stephenson 1977).

Figure 6 is copied from figure 5, but with several amendments and additions. The zone in RA reported as “3° in Di” has been assigned a width of 1° (assuming the original measurement was carefully estimated to the nearest degree), while the longitude corresponding to the 15th deg of Scorpio (i.e. between longitude 224 and 225) is also marked. In addition, the St Gallen southerly limit of approximately —39° declination is shown. The remnant of the supernova would be expected to lie close to the area of intersection of the two zones corresponding to the Chinese and Arab measurements, and to the north of the St Gallen limit. At this high galactic latitude (around +15°), SNRs are rare and only two lie in the region of sky covered by figure 6: the Lupus Loop (G330.0+15.0) and G327.6+14.6. The former SNR is very large and faint, implying extreme age: several thousand years. However, radio and X-ray observations indicate that an age of around 1000 years for G327.6+14.6 is quite feasible; hence this is the only viable remnant of SN 1006.

As is clear from the SNR positions in figure 6, acceptance of G327.6+14.6 as the remnant of SN 1006 implies errors of around 2° in both the Chinese and Arab positional observations. However, such errors were probably not unusual at this fairly early epoch. Significantly, in 1006 the supernova which produced G327.6+14.6 (declination —37.6°) would lie so close to the horizon at St Gallen that its nightly visibility would be interrupted by the mountains to the south of the monastery. This is in good accord with the report in the Annales Sangallenses Maiores, which asserts that the star was “sometimes extinguished”.

Further south, towards the galactic equator, no further SNRs have been catalogued within about 15° of this site; all other remnants lie no more than 3° from the galactic equator, and hence entirely to the south of the area covered by figures 5 and 6. In addition, not only would the star κ Lup (Qichen Jiangjun) be more than 5° below the St Gallen horizon, no SNR is known to be located within about 8° of this star (in any direction). Hence the early Japanese suggestion — reported in the Meigetsuki— that the new star of 1006 represented a change in Qichen Jiangjun is without foundation.

1006 The remnant of SN

The SNR G327.6+14.6 was first firmly identified as the remnant of SN 1006 by Milne (1971). This remnant has been extensively studied at high resolution in both radio and X-ray wavelengths. The SNR is a limb-brightened shell with no central condensation (see figure 1). Its location at high galactic latitude is more in keeping with an outburst of a star of low mass, resulting in a Type 1a supernova. Various techniques have yielded a result for the distance of the SNR of around 2 kpc. For instance, Winkler (2003) derived a distance of 2.18 ± 0.08 kpc. At this distance, and with a typical absolute magnitude around —19, the expected apparent magnitude would be around —7.5 ± 0.5 (allowing for interstellar absorption of only about 0.3 magnitude so far from the galactic equator).

This result would seem to suggest that the brilliance of the star was significantly overestimated by observers in various parts of the world. In particular, an apparent magnitude of around —7.5 is attained by the Moon at an age of around three days, when it is still barely bright enough to enable objects to be seen by its light. It thus remains to be seen whether future observations of the SNR can reconcile this discrepancy. Whatever the eventual outcome, there can be no doubt that SN 1006 was far and away the brightest supernova on record.

References

Author notes

F Richard Stephenson draws together accounts from China, Japan, Egypt and Europe of a supernova in AD 1006 — the brightest supernova on record.