Foundations of Temples of Nepal – a post-earthquake learning!

Sudarshan Raj Tiwari1

The built heritage of Kathmandu Valley, particularly the monuments of the seven zones listed in
UNESCO’s Kathmandu Valley World Heritage Site (KVWHS), suffered massively in the 2015 Gorkha
Earthquake. The collapse of several multi-tiered temples with multiple plinths led some self-styled
experts to speculate that these buildings were inherently unsafe, without foundations and stood on
brick platforms without anchor, and had but a ‘zero capacity for taking lateral loads’! Such tragic
unfounded and untrue devaluation of traditional materials and technology of construction not only
made our heritage buildings as dodders but also portrayed our ancestor builders as without even an iota
of earthquake engineering knowledge and idea as though they lived in thick-headed oblivion of
thousands of years of recurrent cycles of massive earthquakes and damages. However, serious close
study of pattern of debris of many fallen temples showed as if the roofs had come falling straight down
and only a few, like the upper roofs of Basantapur tower of Hanumandhoka palace, Radhakrishna
temple of Swotha, Patan and Nritya Batsala temple of Bhaktapur had been thrown off plane to some
distance. As critical observations and analysis over time would support later, a general understanding
had formed in conservation circles that the heritage disaster was not so much a making of the
earthquake as it was a consequence of material aging, a general state of a decayed buildings plagued by
periods of neglect and absence of structural maintenance, and asymmetrical incompatible interventions
made on to them in the name of repair, restoration and conservation. None of the damaged or
collapsed structures appeared to have suffered damage or failure of foundations or due to problems
associated with foundation.

UNESCO and it’s World Heritage Committee (WHC), weary of our irresponsible assessments and
insensitive reconstruction interventions likely to be detrimental to the outstanding values of KVWHS,
took immediate note of ‘ascertained and potential loss of integrity and authenticity’ to deliberate
whether KVWHS should not be put in the List of Heritage in Danger in June, 2015. Since then, two years
have gone past and we have been so slow and demoralizingly self-deprecating in approaching their
rescue and reconstruction, the monuments zones continue to rue like forlorn landscapes failing to
humor the visitor to the lost civilization but not to chill the Nepali psyche day in and day out. This paper
looks at the state of architecture and engineering of the foundations of the temples to critically
understand and evaluate the interventions made unto them in the hope that they would not be
destroyed at the hands of the callous and the uncaring.

While it was traditional in Kathmandu Valley to reconstruct temples and other heritage buildings
damaged or lost to earthquake or fire disasters upwards from plinth level only and continued use of
previous foundations untouched, the exploration of temple foundations had also been rarely done in
recent times. The archeological excavations at Hadigaun Satyanarayan had revealed that brick strip
foundations with one or two steps were already in use for dyochem-like rectangular religious buildings
since as early as second century BC. For square temples also, similar strip foundation was in use for the
1
Page

1
Professor, Institute of Engineering, Tribuvan University, e-mail – [email protected].
walls forming the sanctum room. The sanctum floor space itself was provided with a navakunda
foundation – the square plan being divided into nine small and equal sized squares using short walls
crossing each other. The ceremonial offerings, such as grains, were deposited here for the pada devata,
the spirits of the subdivided squares or to navagraha, nine ‘planets’ of Hindu/Buddhist astrological sky.
They seemed to form a site of nine plots known as pitha vastu mandala. Similar foundation structures
were also observed in another Siva linga sactum of Deupatan dated to late fifth century AD. This pattern
of foundation appeared to have been used for mandap typology of temples also as corroborated by
classical literature such as Saradatilaka. Masons experienced in temple construction also reported that
such pattern of wall/pit formation
under sanctum was in use until
recent times in all kinds of temples.
Since the divider short walls were
butt-jointed with each other as well
as with the main walls, they
appeared to serve a ceremonial
purpose rather than structural.
They could however have helped
strengthen the base platform of the
temples if the plinth were raised or
made up of a number of terraces,
as became traditional from the mid-
Malla period on.

Fig.1: The navakunda pitha mandala of a temple sanctum, Hadigaun Satyanarayan (sixth century CE)

UNESCO funded a team from Durham University (DU), the University Stirling (US) and DOA (shortened
hereafter as DUSA) to undertake post-disaster archeological assessments and evaluations of sites and
monuments within KVWHS in October-November 2015. The major focus of DUSA investigation was in
the ‘ruins making where the Kashthamandap had stood. It revealed that its ‘brick foundation walls
reached depths of two meters and had been set within mud mortar, on a surface prepared prior to
construction and with an organized sediment fill material, which we suggested gave the foundations of
the monument resilience and flexibility during seismic shock’. The uniformity and designed nature of mix
of clay, silt and sand in the filled material suggest a very careful and studied construction practice for
such an early building. This construction of foundations, dated to seventh century CE, had not been
damaged by this or prior seismic activity. The investigation also found a formation of one-brick-thick
cross-walls patterned akin to the navakunda pitha mandala discussed above and seen in earlier smaller
temples. Interestingly, the depth of foundation in all cases has been reported as two meters. This is
equal to 36 courses of brickwork or one byoma of ancient measure. While widths of cross-walls vary
from one brick to one and a half brick breadths, the thickness of main wall appears related to the size of
the temple and ranged from one and a half brick breadths to half a byoma (I meter) in the case of
Kashthamandap.
2
Page
Therefore, it is difficult to agree with DUSA inference that foundation design ‘included cross-walls for
stability, bracing the massive one meter thick and two meter deep outer walls against a central free-
standing pier’. As a matter of fact, such thin and tall divider wall would itself have problems of its own
stability during construction, while freestanding and also while back
filling it. It is also obvious that larger the temple, the pits would be
proportionally larger and free standing wall segment also longer
compounding self-stability problem.

The nine pit formation has a ceremonial and ritual role and its
structural action purely incidental and minor. The finding of another
nine-pit formation with cross-walls of eight courses depth within the
central pit of the outer nine-pit mandala during the 2016 season of
archeological works also substantiates the form’s ritual nature (See
Fig. 3). It is likely that the inner pitha mandala (8 courses in height) is
dated to first consecration (7th century CE) and the outer pitha
mandala (36 courses in total) belongs to its re-consecration dated to
9th century CE. It should be understood that for positioning the four
piers, the outer mandala should be geometrically known and so the
foundation of the outer wall (24 courses with step of 8 courses each)
must have been constructed at the very beginning. Because of the
cuts made in second phase re-consecration, all the eight outer pits of
the outer mandala have lost their geometric purity.

Fig.2: The pitha mandala of Bimalasangha Siva linga, Deupatan (484 CE)

However, the structural intent is clear in the choice of width and depth of strip foundations for main
walls, both inner and outer one, and the designed infill could also add to the structural performance of
foundations though mat action (?). The use of 36-brick course standard foundation depth through
centuries long periods does tell of a serious and stringently practiced building culture of ancient Nepal.
Similarly the standard for outermost wall appears set at 24 courses. DUSA team also believed “that the
use of brick in mud mortar enhances the piers’ resilience to seismic shock”.

DUSA investigation (2015) at other Durbar Squares also exposed the outline of foundation of the outer
walls of Charnarayan in Patan and Nritya Batsala in Bhaktapur, both belonging to the late Malla period.
While the depth of foundation in both cases was made up of thirty six courses of brick laid in mud
mortar, the standard brick foundation was found built over a pad made of river rounded stone boulders.
This appears as a major technological development in the design and construction of foundations in the
Malla period over Lichchhavi standard practice of laying brick straight on excavated earth surface. It may
also be noted here that the original foundation of Chysilin Mandap of Bhaktapur Durbar Square,
destroyed and replaced by a RCC mat by German ‘restorers’ in 1978, was a mat pack of similar river
rounded stone boulders of medium grade. It is speculated that such boulder packs acted as ‘earthquake
shock absorber of some kind’ and empirical micro-tremor studies aimed at understanding the behavior
3
Page

of such boulder packs encased in earth have been initiated at NAST.

More foundations have since been opened and sadly, also continue to be wantonly destroyed in the
name of reconstruction. DOA opened the plinth and foundations of main and perimeter walls of Rato-
Matsyendranath (Karunamaya) temple, Bungamati. In this granthakuta temple too, we find the
foundation of the sanctum core wall built with the standard thirty six courses of fine glazed-on-the-
outside bricks laid on clay mortar. This platform like foundation laid straight on excavated earth surface
and without the river rounded stone boulder pad, portends to be of Lichchhavi period origin, possibly
sixth century CE as the associated popular legend is dated. The core sanctum platform was surrounded
by a two bricks wide strip foundation constructed of fine glazed-on-the-outside bricks aligned under the
colonnade tie band – the space between the square solid platform and the sides of perimeter
foundation was filled solid with coursed bricks laid on sandy dry clay mortar. A further perimeter wall
foundation built to form
the last of the three plinth
steps was built with fair
quality bricks of smaller
size and the space between
the inner and outer
foundation walls was
packed solid with coursed
brickwork on loose sandy
clay mortar. In the name of
engineered earthquake
resistant construction, all
of this elaborate
foundation except the core
platform was cut to below
ground level and thrown
away undocumented. The
historicity, knowledge and
earthquake resilience of
the ancient foundation
system of buttressing a
solid brick sanctum
platform by a number of
tubular brick squares
separated by dry laid solid
brick infill has been lost at
the hands of a callous,
uncaring and sacrilegious
mindset!
4

Fig.3: The double nine-pit pitha mandala of Kashthamandap (DUSA report, 2016)
Page
A disgraceful mass foundation of brick on lime mortar has been built all around the core on a broken
stone boulder pad.

Fig. 4: Destroyed Lichchhavi foundations (left and center) and the disgraceful replacement (right)

In January 2016, Kathmandu Valley Preservation Trust (KVPT) opened up the plinth and foundations of
Bhaidega temple in Patan monuments zone also to discover the use of core platform foundation
retained by a tubular brick square perimeter foundation with the in-between space filled with dry laid
solid brick infill. The use of river rounded stone boulders to form the lower sections of the perimeter
wall was also observed in southern
side of the temple plinth. Realizing
its technical worth and heritage
value, KVPT is planning to ‘retain
this foundation morphology to the
extent possible’. But even here,
the resilient dry laying of coursed
brick infill is being replaced by
semi-monolith forming lime
mortar jointed brickwork!

Fig 5: Foundations of Bhaidegah temple at Patan Durbar Square (KVPT drawing)

Archaeological exploration of foundations and plinths of a number of temples in Kathmandu city core
such as Jaisidewal, Majudegah, Trilokya Mohan Narayan, Jagannath and Gopinath, was also made in
November-December, 2016 by a team of archeologists and students from Durham University, University
of Stirling, University of Sydney and Lumbini Buddhist University led by Robin Coningham (DUSSL).
Although the report of this study have not been made available as yet, from observation of the
excavation works and discussion with team members working on the sites, this author can infer that all
5

the temples explored used the foundation morphology of core sanctum platform ringed by perimeter
Page

wall square with the in-between space packed with dry laid brickwork or soil infill for resilience. While
using this type of foundation in multiple plinth temples with significantly raised sanctum, such as
Jaisidewal, the core platform appears to have been constructed first to its required height as a
freestanding brick tower to support the ritual consecration with mandala of the nine-pit configuration
(or a symbolic stone tablet as seen at Bungamati). This was retained by construction of a perimeter wall
square of commensurate height separated by a gap of space from the core later filled to gain resilience.
The construction of the stepped plinth was done later using short wall perimeter foundations to support
a number of such plinths at a time and was not intended to act as a structural buttress.

As part of its restoration campaign, KVPT also opened up the foundations of Charnarayan temple, which
is also patterned after the core platform ringed by square of perimeter wall with the in-between space
packed with soil infill. Some sections of the platform appear deformed with slight bulges. The infill soil
has been removed and packed with brick laid in mud mortar, which gives a better resilience than a lime
mortar jointed brick infill. However, the same agency has opened up and virtually obliterated the
original foundation forms of Manimandap as it cleared ‘the trash under its suspended floor structure.
The proposed ‘reconstruction’ of this pavilion with new designed steel foundations sadly promises to be
as destructive as or worse than the Chyasilin Mandap construction of 1988 both from the perspective of
ethnic heritage value and longer term salvage and restoration.

It can therefore be concluded that the post-earthquake archeological and reconstruction excavations
have revealed use of two distinct types of foundations constructed of brick in mud mortar, both
developing significant resilience against earthquake while transferring the load of temples to ground.
They have confirmed that the mandap-type temple structures used brick strip foundations for the main
walls of the sanctum room while the sanctum floor was ‘supported’ by a set of intersecting cross-walls,
two in each direction, forming a nine-pit ceremonial mandala. In larger structures like Kashthamandap,
the four pillars defining the central square of the mandala were each provided with pier foundations as
deep and wide as the foundations for the main wall. While the brick cross-walls were built thin, the
width of foundation of main walls were made thick, half as much as it was deep. The depth of
foundations was standardized at thirty-six courses of standard bricks laid regularly in well prepared mud
mortar. The mud mortar mix design is very close to present day recommendation of clay 18-22%, silt 40-
45%, and sand 30-40%. The foundations for all other tiered temples or granthakuta temples (i.e.
popularly but wrongly called Shikhara), whether with single or multiple stepped plinth or whether with
sanctum at low or raised level, was formed of a central solid cuboid brick core build up to the required
sanctum level ringed by fairly thick perimeter foundation wall square with the in between space packed
with dry laid bricks or soil infill. For temples with very high sanctums and multiple plinths, a further set
of perimeter wall and tubular infill space is added. The square top of the cuboid core is sized to support
the sanctum and it’s circumambulatory. Both the core and perimeter wall are built by laying regular
bricks in mud mortar. A nine-pit mandala made of cross walls on top section of the core platform
provided ‘ceremonial ground’ for consecration rituals. Because of such requirements, it is believed that
the cuboid core platform is constructed first and rest of the temple construction including the stepped
plinth is then sequenced. This design and method of constructing foundation as number of square
tubular brick walls and in-fill sheaths supporting the core cuboid brick platform appear aimed at
6
Page
developing resilience against earthquakes. In Malla period, these cores and walls were raised on a base
mat or pad of river rounded stone boulders for similar structural behavior.

The above appraisal of reconstruction works also shows that these methods and forms of traditional
foundation of heritage structure are being destroyed recklessly and replaced by new incompatible and
less effective interventions in the name of building earthquake resistance. At Tunaldevi, which has
images and stones dated to fifth century CE and whose intangible festive practices go even further back
in history, DOA has shamelessly put in a RCC paving destroying the ancient archeological layers without
even recording such. This and the destructive rebuilding of Rato-Matsyendranath temple foundations
put to shame even those others we have built back in somewhat acceptable ways. Indeed, we have
taken out and destroyed so many of the fine traditional foundations and replaced with unimaginative
and punitively engineered primitive constructions executed in irreversible technologies and materials
foreign to them, against all norms and much to the chagrin of UNESCO and others who see heritage
value in authentic foundations in local materials and methods, we have put our civilized past and our
ancestors to great ignominy. Such actions continue to threaten to obliterate the values of KVWHS
altogether and forever after.

The so-called reconstruction engineers and architects should inform themselves with available
assessment reports or do serious analysis of their own and work towards saving the heritage value of
the materials, methods and morphologies of the traditional foundations. Destruction of such foundation
structures is destruction of heritage and will eventually destroy our identity. The self-inflicted
destruction of heritage must stop. It has to be the responsibility of the government to dismantle the
wrongs and restore the traditional.

7
Page