Sundial, calendar and Khmer temples

 

Prasat Sdok Kok Thom at Equinox

notes on astro-archaeology


''At the equinox, both gods and demons see the sun in the horizon''
(Surya Siddhanta, chapter XII, verse 47)

 

Intro:

     Between the solar-lunar events at Prasat Phanom Rung 14 days before and after equinox in spring 2009 I spent the time doing field-research in the eastern provinces of Thailand, Prachinburi and Sra Kaeo. The area is geographically the western part of the Cambodian plain and an amount of some 50 ancient Khmer temples witness the prosperity of the former empire of Kambuja. The temples span over an era from the 6th to 13th century.
     The most magnificent Khmer sanctuary in the area is Prasat Sdok Kok Thom (also transcribed as Prasat Sdok Kak Thom and Prasat Sdok Kuk Thom) build in the 11th century.

     The temple also hosted the famous K. 235 inscription, dated by a description of the location of the sun, the moon and the planets when the main deity was installed.
 

   

Above: Central tower, libraries and surrounding corridors.
Right: Plan of Prasat Sdok Kok Thom

Courtesy to FAD for the plans above.  
   

ASTRO-ARCHAEOLOGY

Basic data for calculations on astro-archaeology are:
1) location of the site, 2) orientation of the structure and 3) date of the temple.

     Further data for interpretation of orientation and lay-out would include history, art-history, ethnography, religion, linguistics, mythology, ancient astronomy/astrology etc.; 
astro-archaeology is an interdisciplinary study.

1.0: LOCATION

The coordinates were determined on site by use of use of handheld GPS (Garmin GPSmap 60CSx) in the central tower: N13.84356 E102.73729 (Datum: WGS 84).
 

(Try and copy/paste the coordinates into Google Map, Google Earth or PointAsia and see the location from above)
   

2.0: ORIENTATION

2.1: Compass

     At a visit to Prasat Sdok Kok Thom in 2004 compass readings indicated the orientation to be nearly equinoctial: 89.5 - only from equinoctial straight east.
     As the error-margin of compass readings is in the range of and magnetic deviation was not know for certain (-2.5 was used tentatively) the site was expected to be orientated 90.0 true east.


Above: Prasat Sdok Kok Thom and its associated eastern barai (courtesy to PointAsia).

2.2: GPS

     GPS-measurements (waypoints and tracks) were conducted in 2004, 2006 and 2009; all confirming the likeliness of equinoctial orientation, 90.0 true east. The error margin of a handheld GPS is app. on structures larger than 100 m.
     If a temple has long ceremonial walk-ways and/or a large baray the error is reduced. The Sdok Kok Thom eastern baray is anyway not included in orientation calculations because the exact location of the corners could not be determined. The western and eastern moats are apparently orientated straight north-south; the eastern corners of the eastern corners were troublesome due to erosion and rumors about land-mines and robbers did not invite to full ground-truthing...


Above: GPS-map based on waypoints and track-points.

.
     Because compass-readings, GPS-measurements, and satellite images all indicated an orientation close to or exact  equinoctial, I decided to let the rising sun at equinox clarify the exact orientation. I therefore spent the early mornings around spring equinox alone in the temple corridors hoping that the eastern horizon would be without clouds giving opportunity to take photos of the rising sun from the centre-line of the structure. The diameter of the sun is app. and solar observations/photos have an error-margin in the range of less than 1/10

2.3: Satellite images

   
  Above: Satellite image (courtesy to PointAsia) edited in a CAD-CAM program.

The yellow stippled line and the red full-lines are orientated 90.0. The dotted red line is orientated 89.5.
The orientation of the ruin is apparently 90.0
 
 
   
 

Above: Satellite image (courtesy to PointAsia) edited in a CAD-CAM program.

     The landscape around Prasat Sdok Kok Thom is searched for contours indicating an associated urban layout with similar orientation as the temple, but without results (Inscription K.235 mentions that the land belonging to the temple was marked towards north, south, east and west). Red lines are orientated cardinal. The red rectangle is drawn symmetrical around the centre-line of the sanctuary and its processional path, showing that the baray has another centre-line.
     The green stippled lines deviate 2.0 from 90.0; the yellow 3.6. No ground-truthing was conducted.
     The red arrow marked '400 m north' shows the location where Aymonier 100 years ago reported about an ancient road from Angkor.

 

 

2.4: Solar observations

Expectations / calculations:

Spring equinox 2009 was 20 March, 18:44 local time.

At sunrise the 20th the sun rose at an azimuth a little more than 90.0 and rising towards south (see picture) the sun would not be observable straight east.

The 21st the sun rose at 8941' and was calculated to be orientated straight east at 06.19 (see picture).

The horizon east of Prasat Sdok Kok Thom is obscured by vegetation, which was reduced by cutting the day before; but the local horizon was still more than altitude zero (sea-level), so the best day to observe the sun was the 21 of March hoping than the sun had ascended above the trees when passing straight east.



 

Right: 21 March 2009

Below: 20 March 2009

 
Above: First visibility of the sun was at 06:20   Above: 06:21 the 21st of March 2009

     The orientation of the sun was calculated to be straight east at 06.19 - soon after which the sun became visible; but it was not possible to take photos from the centre-line of the door-ways! My experiences from solar observations through the door-ways of Prasat Phanom Rung immediately told me that the doors were not orientated straight true east and that the sun would be aligned with the centre-line of the door-ways the next day.
 

   
Above: 06:36 the 21st   Above: 06:57 the 21st   Above: 07:16 the 21st

     The rays of the rising sun penetrating the door-ways of Prasat Sdok Kok Thom do not reach and illuminate the linga in the central tower; but this would not bother a God residing in the sacred central room, the garbha-griha. Shiva would know when the sun rises straight east marking the moment when the sun has completed half of his travel from his southernmost to his northernmost sunrise point - from winter solstice to summer solstice.
     The garbha-griha has one open door and three doors blinded by stones, but Shiva's 3rd eye would anyway be able to watch the sunset through the western stone-door and watch the Northern Star through the northern blind-door as well. The priests and other humans need an open door to serve the God.

     The best sunrise observation point for photographers and other mortals is at the lowest step of the staircase leading up to the God. From this point there is app. 60 m to the easternmost door of the eastern gopura.
     There is only space for one observer, so if the locals want to make the solar event at Sdok Kok Thom into a tourist show as at Prasat Phanom Rung the only way to do so is by placing a video-recorder at the lowest steps and use large monitors for the audience for watch the sunrise.
NB: Actually the prasats are 'palaces of the gods'; not showrooms for commoners.

 
  Above: 07:16 the 21st

 

 

 
Above: First visibility of the sun the 22nd was at 06:18   Above: 06:20 the 22nd of March 2009
 
Above: 06:21 the 22nd of March 2009   Above: 06:23 the 22nd of March 2009

     At the first visibility of the sun its orientation was 89.6 (above) and was located slightly south (right) of the centre-line of the doorways, indicating that the orientation of the doors is app. 89.5.
     The diameter of the sun is and considering how easy it is to orientate a construction without modern sophisticated tools as theodolites etc the error margin is unacceptable; the intention of the ancient master-builders must have been to orientate the temple straight east towards the equinoctial point.

     As building construction is my professional training I decided to take a look at the door-ways from another angle: From the entrance, the point of the rising sun:

 
Above: The door-ways and the processional pathway seen from 200 m east of the easternmost door.
The alignment follows the processional walk-way.
  Above: The door-ways and the processional pathway seen from 200 m east of the easternmost door.
The alignment follows the doors.


     Comparing the two photos above it is evident that the centre-line of the doors is not identical to the centre-line of the processional walk-way!
     If the door-ways are orientated 89.5 as indicated by the solar observations above then the processional pathway and the total structure of the temple is most probably orientated 90 true east.


Above:
The door-ways and the processional walk-way
before renovation and apparently aligned (!)
Photo: Courtesy to
Fine Arts Department

 

2.5: Photo analysis:

     Photo analysis is based on perspective drawing where the extended contours of a construction will meet in the perspective point.
     If two or more door-ways have identical size, level and centre-line, then the perspective lines will meet in the perspective point on the local horizon of the construction (red lines on photos below). If the construction is build at sea-level and orientated straight east the perspective point will be identical to the equinoctial point: 90 true east and altitude zero.
     A photo of the rising sun through a set of door-ways (with identical size, level and centre-line) can give very exact information of the path of the rising sun when analyzed in a CAD-CAM program. The diameter of the sun is close to 0.5 and can be used as a 'measure-stick'. If the latitude of the location is known we also know the path of the rising sun (90 minus the latitude) which can be drawn on the photo as done below (yellow lines). Following the path of the sun to the horizon we know where the sun rose and simple astronomical calculation will tell the path of the rising sun the days before and after the photo was taken.
     This method has been used with success at Prasat Phanom Rung and other temples; but was not successful at Prasat Sdok Kok Thom. The main reason is insufficient knowledge about the levels of the door-ways. A second obstacle was that the door-ways are of different sizes and apparently now sharing the same centre-line. A string or a lazer-light will easily tell the centre-line at the next visit, but a full measurement of level and orientation (as done at Prasat Phu Phek) will not be conducted.

 
Above: The perspective lines connecting the corners of the first and the last door-way do not meet in the same point and pass the corners of the door-ways between.   Above: Sunrise through the eastern gopura of Prasat Sdok Kok Thom the 21st of March 2009.
 

 

     The perspective lines drawn on the photos do not meet in the same point, the perspective point, which is marked by a green circle and tentatively placed on an approximate horizon (blue line).

     The yellow line with an arrow passing the centre of the sun depicts the path of the rising sun and it is obvious that the sun rose very close to the perspective point. The perspective point marks the orientation of the door-ways and as the azimuth of the sun at sunrise the 21st of Marsh was 89.67 the orientation must be a little more: Measured on the photo in CAD-CAM and using the sun as 0.5 the addition should be 0.06 indicating an orientation on app. 89.72 ~ 89.7

     The two yellow stippled lines on each side of the yellow arrow shows the paths of the sun the day before (right) and the day after (left) the 22nd.

 

Right and below: Sunrise through the gates of Sdok Kok Thom the morning after spring equinox 2009.

 
 

 

Summing up on the various ways of determining the orientation of the door-ways:
Compass (on centre-line): App. 89.5
Solar observation: App. 89.5
Photo analysis: App. 89.7
Conclusion: The photo-analysis is the most accurate method, so the orientation on the door-ways is app. 89.7

 

The orientation of the super-structure of Prasat Sdok Thom is equinoctial: 90.0 true east

 

3.0: DATE of Inscription K. 235, stanza 123:

Prasat Sdok Kok Thom was consecrated the 8th of February 1053; when Bhava was installed (Billard).

     Bhava is one of Siva's eight  forms (Śarva, Bhava, Paśupati, Iśna, Bhma, Rudra, Mahdeva, and Ugra). The eight-petalled lotus flower motifs frequently used in Khmer art and architecture probably represent the 'total Shiva', including all his aspects. At Sdok Kok Thom eight-petalled gold leafs were excavated in the eastern gopura and at several levels between the sand-stone blocks in the central tower, where they were deposited in the cardinal directions.
     The linga, symbolizing the Bhava aspect of Shiva, was installed on the above date by performing a Brahmin ritual. The consecration was the final of a series of Brahmin rituals. One of the first rituals would have been when the master-builder or lord-of-the-cord would determined the orientation of the temple by means of the Vedic Circle.


 
   

     Most dated Khmer inscriptions give a Saka year, a lunar month, the number of ascending or waning day, and the weekday - or some of these data.
     The two other dates mentioned in Inscription K.235 mention year, lunar month, and ascending or waning day, from which the Gregorian dates easily can be calculated:
D, 86: Saka 894, 3rd ascending day in the lunar month pausa = Wednesday 11 December 972
D, 90: Saka 901, 3rd ascending day in the lunar month pausa = Wednesday 24 December 979
     The third date mentioned in K.235 gives the Saka year and the horoscope of the planetary positions when the deity was consecrated (right), which makes the calculation more complicated.

3.1: Dating the Inscription K. 235, stanza 123:

3.1.1: Aymonier (1901, p. 52) was the first to translate the famous K. 235 inscription of which stanza 123 was only partially translated and the date given as Saka Era 974.

3.1.2: Briggs (1951, p. 170): AD 1052.

3.1.3: Coedes and Dupont (1943, p. 156-154.): ... when ''the sun and the other planets (respectively in) Aquarius, Virgin, Libra, Aquarius, Aquarius, Pisces, Aquarius, Aquarius, and Pisces. The horoscope in Sagittarius. Bhava was erected in the year marked by the (9) orifices, the (7) mountains and the (4) Vedas.'' (my translation) => Saka Era 974.

3.1.4: Billard: 8th of February 1053.

3.1.5: Golzio (2006, p. 153): 1st of February 1051 AD at 06.00... - earlier dated to 8th of February 1053.

3.1.6: Eade  (2006, p. 25 - 28): 8th of February 1053.

 

3.2: My attempt on calculating the date:

3.2.1. Most translators of the Sanskrit text in stanza CXXIII agree on translating the Saka year as 974.

3.2.2. The Saka Era started in 78 AD, so a (too) simple calculation is: Saka 974 = 1052 AD (974 + 78).

3.2.3. As the Saka year actually starts at new-moon before equinox then Saka 974 is the period from the new-moon before equinox 1052 to new-moon before equinox 1053 AD (sometimes written as March 1052 to March 1053).

Note: If we don't know whether Saka 974 is treated as an elapsed or a current year the time-span of Saka 974 is March 1051 to March 1053 AD. But because the two other dates mentioned in inscription K.235 are treated as a elapsed years, we have good reason to assume that Saka 974 is treaded as a elapsed year. (=> Saka 974: March 1052 to March 1053).

3.2.4. New moon before equinox:
Using Yano's Pancanga:
1052  March   4 Wednesday, Saka  974, lunar month: Caitra, Suklapaksa 1 (1st ascending day).
1053  March 23 Tuesday, Saka  975, lunar month: Caitra, Suklapaksa 1 (1st ascending day).
    => Saka 974 spanned from Wednesday the 4th of March 1052 to Tuesday the 23rd of February 1053 and the date of consecration is to be found between these two dates!

Note: If I use modern astronomical parameters, I get the following incorrect dates:
True equinox in 1052 was 14 March and the previous new moon was 4 March. => 1st ascending day = 5 March 1052.
True equinox in 1053 was 15 March and the previous new moon was 21 February => 1st ascending day = 23 February 1052.

3.2.5. ''The sun ...''
     The inscription informs us that the time of consecration was when ''the sun and the other planets were (respectively in) Aquarius, ... '' and as the sun passes Aquarius every year in January-February the date in question must be somewhere in January-February 1053 AD. - so the year is 1053 AD.

3.3.6. ''... and the other planets''
     The chronology of 'the other planets' would give us possibility for a more exact dating; but is not explicitly given. If we assume that the chronology follows the order of week-days ...
 

Weekday

Sunday

Monday

Tuesday

Wednesday

Thursday

Friday

Saturday

-

-

Deity

Surya

Chandra

Yama

Varuna

Indra

Kubera

Agni

Rahu

Ketu

Planet

Sun

Moon

Mars

Mercury

Jupiter

Venus

Saturn

Rahu

Ketu

Zodiacal sign - Coedes

Aquarius

Virgo

Libra

Aquarius

Aquarius

Pisces

Aquarius

Aquarius

Pisces

... we get the following relations: Sun in Aquarius, Moon in Virgo, Mars in Libra, Mercury in Aquarius, Jupiter in Aquarius, Venus in Pisces, Saturn in Aquarius, Rahu in Aquarius, Ketu in Pisces - and can start searching for when such a horoscope happened.

Note: Of the above nine 'planets' we nowadays only count the five in the middle; which are visible by the naked eye.

3.3.7. The moon in Virgo
     The moon is the 'planet' which as seen from the Earth moves the fastest and is the first to be checked in order to restrict our search:
     A easy-to-use desktop astronomy program (a desktop planetarium with modern parameters) as for example Cybersky gives the moon to enter Libra at app. 16.00 the 6th of February and exit the 8th at app. 21.00 => further search is restricted to the 6th to 8th of February 1053. With this time limit we have to check the Pancanga and HIC to see the results of contemporary Siddhantas (see 3.3.9.).

3.3.8. Mercury, Venus, Mars, Jupiter and Saturn
     The desktop planetarium gives the five visible planets to be in the same zodiacal sign the 6th, 7th and 8th of February 1053:

Planet

Sun

Moon

Mars

Mercury

Jupiter

Venus

Saturn

Rahu

Ketu

Zodiacal sign - Coedes
 

Aquarius
 

Virgo
 

Libra
 

Aquarius
 

Aquarius
(kuca = Gemini)

Pisces
 

Aquarius
 

Aquarius
 

Pisces
 

Zodiacal sign - Modern

Aquarius

Virgo

Scorpio

Pisces

Gemini

Aries

Pisces

? ?

Note: None of the planets are in the zodiacal signs as given by Coedes (!). All are in the following zodiacal sign, except Jupiter. The latter is four signs ahead, which is due to the translation of the Sanskrit text, where the sign not specifically is given as Aquarius, but poetically is referred to as kuca ('female breasts'). Following Dr. Eade's interpretation (2006, p. 26) of kuca as Gemini (twins) then Jupiter was located as written in the inscription.

     The four other planets - except Mercury - are in the first quarter of the following sign, indicating that our assumption that the chronology of the planets follows the order of week-days (3.3.6.) is correct - and that our parameters for calculating are incorrect. We therefore skip modern parameters and turn to Siddhantic parameters on astronomy.
     We know of Siddhantas (canonical text-books) of which the Surya Siddhanta was widely used in SE Asia (Eade, p.  ) and to some extend still is in modern India. The parameters given in the Surya Siddhanta have been computerized and are available as Siddhantic programs like  HIC and Pancanga

3.3.9. Siddhantic programs on ''the sun and the other planets ...''

3.3.9.1. Pancanga
     The Pancanga gives the moon to be in Libra on the 8th and 9th of February 1053. As Mercury is given to be in Aquarius the 8th and be in Pisces the 9th. the latter date is excluded. All other planets do not change sign from the 5th to the 11th.
     The Pancanga locations of the planets are added to the table below and there is a clear accordance between Coedes translation and the Pancanga: The only discrepancy is the location of Rahu. The date of the horoscope is 8 February 1053.

3.3.9.2. HIC
   HIC gives the moon to be in Libra on the 8th and 9th February 1053. Mercury is given to be in Aquarius the 7th and be in Pisces the 8th. All other planets do not change sign from the 7th to the 10th.
     The HIC locations of the planets are added to the table below. Again there is a significant accordance between translation and the computer out-put - except for Mercury. The date of the horoscope is 8 - 9 February 1053.

Planet

Sun

Moon

Mars

Mercury

Jupiter

Venus

Saturn

Rahu

Ketu

Zodiacal sign - Coedes
 

Aquarius
 

Virgo
 

Libra
 

Aquarius
 

Aquarius
=> Gemini

Pisces
 

Aquarius
 

Aquarius
 

Pisces
 

Zodiacal sign - Pancanga

Aquarius Virgo Libra Aquarius Gemini Pisces Aquarius

Gemini

?

Zodiacal sign - HIC

Aquarius Virgo Libra Pisces Gemini Pisces Aquarius ? ?

Zodiacal sign - Modern

Aquarius

Virgo

Scorpio

Pisces

Gemini

Aries

Pisces

? ?
Above: Table comparing computer outputs on 8th February with Coedes' translation of K. 235.
   
Comments:
 
 
I:  

 

There is a significant discrepancy between the computer output based on modern astronomical parameters and the inscription, demonstrating that datings of ancient Khmer inscriptions should be based on contemporary parameters as given in the Indian siddhantas.
II:   There is a significant correspondence between the Pancanga / HIC and the translation.
III:   The Pancanga out-put gives one date: 8th February.
IV:  

The HIC out-put gives two dates, the 8th or 9th February; both with Mercury located in the neighboring sign.


3.3.10. "
The horoscope in Sagittarius"
     This last information
informs us about the time of the day when the consecration ceremony was performed: when Sagittarius rose on the eastern horizon (... with Sagittarius in the ascendant).
     Time = ?
    
I don't  (yet) know how to deal with ascendant in the programs based on Siddhantic parameters. A planetary program would easily tell when Sagittarius rose on the eastern horizon, but the information is irrelevant if we want to know how the ancient calendar makers conceived.

3.3.11. Conclusion on date of consecration: 8th of February 1053.

 

For a more thorough examination of the dating, see 1051, 1052 or 1053?

 

 

4.0.
 
CONCLUSIVE REMARKS ON ASTRO-ARCHAEOLOGY
 
4.1.
 
The temple is like the majority of Khmer Hindu temples equinoctially orientated: The azimuth = 90.0 true east,
implicating that the sun rose and still rises aligned with the structure around true equinox in March and September.
4.2. The orientation of the temple is not related to the consecration date of the linga.
4.3.
 
The consecration date and moment is inscribed as a horoscope and chosen due to considerations about some kind of auspiciousness, the nature of which we do not have any clues.

 

LITERATURE
 
 
Aymonier, E.: Khmer Heritage in the Old Provinces of Cambodia, Paris, 1901.
Briggs, L.P.: The Ancient Khmer Empire, Washington, 1951.
Coedes, G. and Dupont, P.: Les stles de Sdŏk Kăk Thom, Phnom Sandak et Prh Vihăr”, BEFEO XLIII, 1943.
Golzio, K.H.:
 
Chronologie der Inschriften Kambojas. Verifizierung und Umrechnung von Datumsangaben der Śaka-ra. Weisbaden, 2006.
Eade, J.C.:
 
Computers vs Tables, Billard vs Golzio: Two new date-lists of the inscriptions of Kamboja, Canberra, 2006.
Burges, Ebenezer
 
Surya-Siddhanta . A Textbook of Hindu Astronomy. India, 1858. (From: Journal of the American Oriental Society, Vol. 6, 1959-60. Page 141-498)
K.235 online In Thai จารึกสด๊กก๊อกธม ๒ จารึกสด๊กก๊อกธม ๑
K.235 online In French
K.235 online In English: http://instruct1.cit.cornell.edu/courses/hist190/Cambodia.html Cornell University.
   
PROGRAMS:    
Yano, M and Fushimi M. Pancanga:, version 3.13. On-line Siddhantic parameters
Gislen, L. and Eade, J.C. HIC: version 1.0. Free downloadable Siddhantic parameters
Ott, R.C. Astronomic Clock version 10 Modern parameters
Schimpf, M.S. Cybersky Version 4.0.6. Planetarium Modern parameters


Above: 'GPS-reference point'  2002 at Prasat Sdok Kok Thom.
The compass shows magnetic north, the N-arrow depicts true north.
The difference is called magnetic deviation, which depends on location.
At Prasat Sdok Kok Thom the magnetic deviation is approximately 2 (!).

If the compass is not influenced by metal hidden in the concrete
then the modern GPS-reference point is orientated incorrect
and the ancient Khmer master-builders more skilled
than modern high-tech constructers
not using the sun

 

INDEX

24 April 2009 Asger Mollerup

macsida@thai-isan-lao.com

www.thai-isan-lao.com

โดย อ.ทองคำ