Petrographic, mineralogical, and climatic analyses, and risk maps for conservation strategies

Wojciech Bartz, Jacek Kościuk, Maria Gąsior, Teresa Dziedzic

doi:10.37190/arc200212

Abstract

El Fuerte de Samaipata is a pre-Hispanic archaeological site in Bolivia that has been on the UNESCO World Heritage List. Its main part – the rock – is densely covered with a complex arrangement of terraces, platforms, water reservoirs, channels, and petroglyphs. The rapidly progressing erosion of the rock is making the petroglyphs less and less clear, and some are no longer recognisable. The main topic of this study is to indicate all risk factors conducive to erosion and to create risk maps identifying the most vulnerable areas that require immediate conservation intervention. Parallel mineralogical and petrographic studies show that the Samaipata rock is a quartz-rich, porous, well-sorted sandstone, classified as quartz arenite or subarcosic arenite. The cement of the rock is composed of quartz overgrowth and ubiquitous, pore-filling hematite-clay aggregates containing non-expanding kaolinite, illite, and expanding smectite. The rock exhibits different stages of weathering, from relatively fresh to strongly altered and heavily cracked. In comparison to fresh rock, the latter has cement enriched in clay minerals and is depleted in hematite due to weathering and the dissolution of the iron-bearing phase.

Full article view is only available on bigger screens.

2020

2(62)

DOI: 10.37190/arc200212

Introduction

El Fuerte de Samaipata is a pre-Hispanic archaeologi-

cal site in Bolivia, in the province of Florida and the de-

partment of Santa Cruz, on the eastern slopes of the An-

des at an altitude ca. 1890–1925 m. Due to its historical

and cultural value, the site was inscribed on the UNESCO

World Heritage List in 1998.

The entire archaeological site covers about 40 hectares

and consists of two main parts:

– An administrative and ceremonial complex of ca. 50

buildings typical of provincial Incan architecture located

in the southern part of the site;

– A natural rock (ca. 80 × 250 m) towering over the Pi-

ray valley in the northern part of the site.

The natural rock was the main subject of the project

“Architectural examination and complex documentation

of Samaipata (Fuerte de Samaipata/Bolivia) site from the

World Heritage List”, of which this study is a small excerpt.

Wojciech Bartz*, Jacek Kościuk**, Maria Gąsior***,

Teresa Dziedzic****

Petrographic, mineralogical, and climatic analyses, and risk maps

for conservation strategies

Analizy petrograficzne, mineralogiczne i klimatyczne oraz mapy ryzyka

dla strategii konserwatorskiej

The whole rock is densely covered with a complex

arrangement of terraces, platforms, water reservoirs, and

channels. Between them are numerous petroglyphs de-

picting animals or geometric forms. The present form of

the rock is the result of the activities of various local cul-

tures over almost 1200 years, who treated the rock as a sa-

cred place (wak’a). However, Incan characteristics clearly

dominated this rock after it was incorporated into the Inca

Empire (Tawantinsuyu) in the 2

half of the 15

century.

Apart from the 16

century mentions by Spanish chro-

niclers [1] and a short description from the 1

quarter of

the 19

century by Alcide Dessaline d’Orbigny [2], the

site only became a subject of systematic scientic interest

in the 20

century [3], [4]. The range of studies signi-

cantly increased in 1970 due to the establishment of the

Archaeological Research Centre in Samaipata [5] and re-

search by German scholars [6]–[11].

Outline and main aim of the study

In the last quarter of the 20

century, intensive research

combined with the removal of vegetation growing on the

rock and the accumulated organic layers led to the expo-

sure of the entire ridge of the rock (Figs. 1A, B). This

not only enabled detailed research to be conducted on

previously unavailable parts of the whole complex, but

also contributed to the popularisation of El Fuerte as an

important tourist attraction.

The acceleration of rock erosion was an unforeseen

side eect of this exposure. The rapidly progressing ero-

* ORCID: 0000-0002-7267-2776. Institute of Geological

Sciences, University of Wrocław, e-mail: wojciech.bartz@uwr.edu.pl

** ORCID: 0000-0003-0623-8071. Faculty of Architecture,

Wro c ław University of Science and Technology.

*** ORCID: 0000-0002-9152-3792. Ceramics and Glass Con-

servation and Restauration Department, Eugeniusz Geppert Academy of

Art and Design in Wroclaw.

**** ORCID: 0000-0003-2855-0074. Faculty of Architecture,

Wrocław University of Science and Technology.

136 Wojciech Bartz, Jacek Kościuk, Maria Gąsior, Teresa Dziedzic

sion of the rock makes petroglyphs less and less clear, and

some are no longer recognisable.

The rst attempts to address the problem of erosion go

back to the beginning of the 21

century, but the prob-

lem is complicated due to the nature and the size of the

rock itself, and it has not yet been satisfactorily resolved.

Research has focused mainly on mineralogical and pe-

trographic analyses of the rock and nding suitable con-

solidators, recipes for putties to seal natural cracks, and

methods to stop the spread of mosses, lichens, and algae

[12]. Much less attention has been paid to a holistic view

of the problem, and in particular to determining all risk

factors conducive to erosion and drawing up risk maps

identifying the most vulnerable areas requiring immedi-

ate conservation intervention. This subject, in addition to

complementary mineralogical and petrographic research,

is the main topic of this study.

Methodology and methods

Apart from anthropogenic factors, the most signicant

reasons for the progressing erosion of the rock are climat-

ic and biological hazards. Depending on the rock proper-

ties and susceptibility to individual threats, each of these

factors will interact with the rock dierently. Therefore, it

is important to thoroughly understand all the properties of

the rock – physical, chemical, and mineralogical, and at

the same time to identify all risk factors.

The methodology adopted in these studies was to ex-

amine each of these factors separately, and then to analyse

the obtained results to understand how they interact. Cli-

matic factors directly aecting the rock include aeolian

processes (wind erosion), thermal shock eects, and pre-

cipitation. Biological factors include the eects of plant

vegetation and the spreading of lichens and algae. The

shape of individual parts of the rock also needs to be con-

sidered – this can either intensify the eects of risk factors

or minimise them. Particularly important is the inclination

of slopes and their orientation towards cardinal compass

directions, areas of natural and anthropogenic inltration

of rainwater (cracks, holes, etc.), zones of permanent ac-

cumulation of rainwater (articial reservoirs and natural

depressions), zones of water euence, and deposits of

organic material favourable for plant vegetation. Due to

this, accurate data obtained by 3D laser scanning of the

whole rock

and photogrammetric documentation

has

been very important.

Analysis of the current state

of erosion

The mutual interaction of all risk factors and the local

properties of heterogeneous sedimentary rock has resulted

in dierent types of erosion – peeling, cracking, exfolia-

tion, and powdering. During the two seasons of eldwork

(2016, 2017), the main stages of erosion and areas of oc-

currence were identied.

The rst stage of erosive phenomena is micro-cracks

appearing on the rock surface (Fig. 2A). Their presence

can potentially be associated with thermal shock and the

alternation between the soaking and drying of porous

rock. Practically, 50% of the Samaipata rock surface is

covered with such micro-cracks (Fig. 2B).

Cf. J. Kościuk, B. Ćmielewski, M. Telesińska, A. Kubicka, 3D

terrestrial laser scanning of El Fuerte de Samaipata , in the same issue

of “Architectus”.

B. Ćmielewski, I. Wilczyńska, C. Patrzałek, J. Kościuk, Digital

close-range photogrammetry of El Fuerte de Samaipata, in the same

issue of “Architectus”.

A B

Fig. 1. The southern slope

of the rock:

A – during the time of research

by Erland Nordenskiöld

(around 1923)

(photo from archives

of the Samaipata Archaelogical

Research Centre);

B – in April 1974

(photo courtesy of Damielo

Montenegio Fernández)

 Petrographic,mineralogical,climaticanalyses,riskmaps / Analizypetrograficzne,mineralogiczne,klimatyczne,mapyryzyka 137

The second observed erosion stage, especially where

the natural stratication of the rock is oblique in relation

to its surface, is exfoliation – individual layers separate

from the bedrock (Fig. 3A). Exfoliation processes are ac-

celerated by the presence of vegetation whose root sys-

tem penetrates and breaks existing micro-ssures. It is

estimated that about 25% of the entire rock surface is at

risk of exfoliation (Fig. 3B). A strong correlation is visible

between the zones of micro-cracks and the zones where

the surface layers of the rock are exfoliating.

In the third stage of erosion, detached fragments of rock

completely lose contact with the ground and begin to move

(Fig. 4A). The progression of this process causes their fur-

ther fragmentation into ever smaller pieces that are easily

moved by rainwater and wind. Accumulating deposits of

such material, additionally enriched with organic particles

blown by the wind, are an excellent breeding ground for

growing plants whose root system will penetrate the rock

deeper and deeper. Zones in which loose lithic material is

deposited (Fig. 4B) are increasing every year.

Climatic and topographical factors

El Fuerte de Samaipata is located in the southern hem-

isphere at 18° latitude – in the zone between the equator

and the Tropic of Capricorn. This fact is essential for con-

sidering the impact of climatic factors on possible conser-

vation hazards, mainly since the main ridge of the rock

Fig. 2. First stage of erosion:

A – an example of the first phase of erosion (micro-craks) on 0.3 mm-resolution orthoimage (photo by J. Kościuk);

B – areas of numerous micro-cracks in the lithic surface (elaborated by T. Dziedzic)

Fig. 3. Second stage of erosion:

A – an example of the second phase of erosion (exfoliation) on 0.3 mm-resolution orthoimage (photo by J. Kościuk);

B – areas of exfoliation of the lithic surface (elaborated by T. Dziedzic)

Fig. 4. Third stage of erosion:

A – an example of the third phase of erosion (loose fragments of the rock lying on the surface) on 0.3 mm-resolution orthoimage (photo by J. Kościuk);

B – zones of loose lithic material (elaborated by T. Dziedzic)

138 Wojciech Bartz, Jacek Kościuk, Maria Gąsior, Teresa Dziedzic

is latitudinally oriented. Strong rays of sun heat both the

northern and the southern slope of the hill. The combina-

tion of this with the 24-hour cycle of temperature changes

and cold winds creates the risk of thermal shock. Identi-

cation of areas where such risk may occur was, therefore,

a priority for this study.

The rst step in this type of analysis was to determine

which fragments of the complex surface of the rock are

most exposed to sunlight. These types of calculations are

usually carried out to determine the energy balance of

a building, or to assess the possibility of using photovolta-

ic cells. Software for this kind of simulation, for example,

EnergyPlus, uses integrated meteorological data – EPW

les that contain information about the temperature and

intensity of solar radiation (direct and diused). However,

there is no weather station on El Fuerte de Samaipata or

even in the town of Samaipata itself; hence, there is no de-

tailed meteorological data for this place that would allow

cumulated solar energy to be calculated. The problem had

to be solved dierently.

In our case, we were not interested in the absolute val-

ues of accumulated solar energy expressed in W • h/m

, but

only the relative dierences in the insolation of individu-

al parts of the rock. Therefore, instead of calculating the

energy balance, a virtual 3D model of the whole hill was

used. The model was embedded in the virtual 3D space

according to geographical coordinates and correctly ori-

ented towards north. The two most opposing moments of

the solar year were selected – summer solstice (21

June)

and winter solstice (21

December). At equal intervals

corresponding to daylight hours, ten visualisations were

rendered for 21

June (Fig. 5) and twelve for 21

Decem-

ber

(Fig. 6). The rendering tool used was Bentley Micro-

Station V8i software (version 08.11.09.459), which ena-

bles the realistic visualisation of solar lighting according

to geographical location and specic date. The values of

the horizontal and vertical angles of incidence of sunlight

were additionally veried using Stellarium software (ver-

sion 0.16.0). As a rendering algorithm, ray tracing was

selected, which also includes diused light and ground

reection for skylight, as well as cloud cover (cloudless

conditions were selected) and air transparency (due to the

high humidity of air in this region of Bolivia, the option

“rural” was chosen).

Grey-scale visualisations for individual hours of day-

light were then stitched together into one image, separate-

ly for 21

June and 21

December, and the resulting im-

ages were divided into six classes of grey. Each of the grey

classes was assigned one of six basic colours from red for

the most illuminated surfaces throughout the whole day,

to blue for those that remain mainly in the shadows. In this

way, maps were obtained showing to some extent which

fragments of the rock could have warmed up the most at

these two dierent moments of the year (Figs. 7, 8), so

zones in which one could expect thermal shock due to the

rapid temperature changes between day and night.

During our stay on El Fuerte in June–July 2016 and

2017, we repeatedly experienced such temperature jumps.

At the end of the workday, the rock was heated in many

places to almost 40 °C, while the next morning it was cov-

ered with frost and ice.

The eect of thermal shock is intensied by strong

winds blowing over El Fuerte. The specicity of the ter-

rain, with its parallel ranges of mountains and valleys

(Figs. 9A, B), means that NNW winds dominate there

(Fig. 9C).

Precipitation is another essential factor from the point

of view of conservation hazards. The annual rainfall for

the nearby town of Samaipata reaches 1000 mm – for

comparison, this is almost twice as much as for London.

For El Fuerte, located nearly 300 m higher than the town,

annual rainfall is undoubtedly even higher, but detailed

meteorological data are not available.

The virtual 3D model of the rock was used again to

simulate the inuence of wind on the eect of thermal

shock. The model was illuminated from a NNW direction

(335° from the north) and rendered using the same soft-

ware. In this way, the most illuminated parts of the hill

simulated the surfaces most exposed to the wind, and thus

those that cooled down the fastest (Fig. 10). This chart is

also an approximate image of the zones where wind ero-

sion eects may occur.

Fig. 5. Samples of visualisations of the rock insolation on 21

June for the fourth (A), sixth (B), and eight (C) hour of daylight

(elaborated by J. Kościuk)

Fig. 6. Samples of visualisations of the rock insolation on 21

December for the fourth (A), sixth (B), and eight (C) hour of daylight

(elaborated by J. Kościuk)

 Petrographic,mineralogical,climaticanalyses,riskmaps / Analizypetrograficzne,mineralogiczne,klimatyczne,mapyryzyka 139

Threats related to atmospheric precipitation are two-

fold. Rainwater that accumulates on the rock surface for

a long time causes hydrolysis and the subsequent wash-

ing-out of the binder of the sandstone rock. If the temper-

ature drops below 0 °C, water freezing in the rock crevices

will break them. Fortunately, the months with the highest

rainfall (December to March) are also the warmest, when

the temperature usually does not fall below freezing point.

In both cases, however, the decisive factor inuencing the

possible retention of rainwater is the inclination of the

slope and the occurrence of local depressions.

Fortunately, one of the thematic styles of displaying

3D objects in Bentley MicroStation V8i is slope inclina-

tion. With this, as in all previous cases, the results were

divided into six classes using a colour scale in which the

red corresponded to surfaces where the inclination did not

exceed 8°, and the blue meant a slope greater than 75°

(Fig. 11).

Fig. 7. Rock insolation on 21

June (elaborated by J. Kościuk)

Fig. 8. Rock insolation on 21

December (elaborated by J. Kościuk)

Fig. 9. The topographical situation of El Fuerte de Samaipata (A and B) and the wind rose (C)

(sources: http://suncalc.net; https://www.meteoblue.com/en/weather/forecast/modelclimate/samaipata_bolivia_11494517 [accessed: 12.01.2018])

140 Wojciech Bartz, Jacek Kościuk, Maria Gąsior, Teresa Dziedzic

Fig. 10. Exposure to the prevailing wind direction (elaborated by J. Kościuk)

Fig. 11. Slope inclination analysis (elaborated by J. Kościuk)

Fig. 12. Depressions without drainage (elaborated by J. Kościuk)

Local depressions were also searched for in the Bentley

MicroStation V8i environment – contour lines were gen-

erated every 10 cm and those that formed enclosed areas

were selected (Fig. 12). The results were presented this

time in a dierent colour scale. The results of all these

partial analyses were further used to prepare risk maps for

conservation strategies.

Biologically inducted erosion

In addition to the climatic and topographic aspects men-

tioned above, biological factors also promote rock erosion.

The scope of the entire project did not include genetic test-

ing, which would have required the separate consent of the

Bolivian Ministry of the Environment. Biological factors

 Petrographic,mineralogical,climaticanalyses,riskmaps / Analizypetrograficzne,mineralogiczne,klimatyczne,mapyryzyka 141

have also already been the subject of partial studies conduct-

ed by Sonia Avilés [12]– [14]. Nevertheless, numerous plants

growing on the rock, as well as lichens, mosses and algae

(Fig. 13), distract not only for aesthetic reasons, but mainly

due to their impact on erosion. This impact is not only the al-

ready mentioned process of breaking micro ssures by pen-

etrating plant roots, but also the biodegradation of the binder

of the ne sandstone particles, which is even more invasive.

Mineralogical and petrographic analyses

Mineralogical analyses were done for 10 samples taken

from the Samaipata rock and one sample of locally occurring

clay material that is used for restoration and lling cavities

in the Samaipata rock. The rock material was split into two

portions with a stone cutting saw. Afterwards, all samples

were laboratory dried at temperature of ca. 40 °C, to a con-

stant weight. The rst halves were impregnated with a low

viscous epoxy-resin under vacuum conditions and used for

the preparation of a petrographic thin section according to

procedure described by Jan Elsen [15]. Petrographic analyses

were performed using a Zeiss Axiolab polarising microscope.

The second halves were analysed by means of

– powder X-ray diraction (XRD);

– thermal analysis (DSC-TG);

– scanning electron microscopy (SEM-EDS).

In the latter case, analyses were performed for two rep-

resentative samples (polished thin sections, 30 nm carbon

coat) chosen on the basis of petrographic observation.

A Jeol JSM IT-100 scanning electron microscope (SEM)

equipped with an Oxford EDS system operating at an ac-

celeration voltage of 16 kV was used.

XRD and DSC-TG analyses were done on a sieved frac-

tion of <63 μm that mostly contained rock cement. XRD

measurements were taken with a Siemens D 5005 dif-

fractometer, using Bragg–Brentano geometry and Co Kα

Fig. 13. Examples of various forms

of plants, lichens, fungi, and algae

growing on the rock surface

(photo by J. Kościuk)

142 Wojciech Bartz, Jacek Kościuk, Maria Gąsior, Teresa Dziedzic

radiation, in the 2Theta range 4÷75 deg. DSC-TG analy-

sis was carried out with a PerkinElmer STA 6000 thermal

analyser, from 40 °C up to 999 °C, at a constant heating

rate of 15 °C/min in N

atmosphere.

The sample of the locally occurring clay was powdered

and suspended in ca. 500–700 ml of distilled water. After-

wards, it was allowed to settle gravitationally (12 hours)

to sediment most of the particles not part of the clay frac-

tion (greater than ca. 4 µm in diameter). Subsequently, the

cloudy supernatant was siphoned and then centrifuged at

5000 rpm for 10 min. The obtained centrifuged sediment

was used to prepare three oriented mounts. For each one,

a few drops of the suspended clay sample were placed onto

a slide using a pipette, and the slides were left to dry. Finally,

the rst slide was saturated with ethylene glycol, the second

one was heated for one hour at 550 °C, and the last one was

left untreated. XRD measurements were made using a B

Siemens D 5005 diractometer with Bragg–Brentano geo-

metry and Co Kα radiation in the 2Theta range 4÷35 deg.

Results and discussion

The research was intended not only to identify the

sources of conservation risks, but also to determine their

potential scope. To this end, a number of risk maps were

developed, and physicochemical, mineralogical, and pe-

trographic analyses were used to characterise the local

sandstone and determine its weak points.

Climatic and topographical analyses

The juxtaposition of insolation data (Figs. 7, 8) with

simulations of exposure to prevailing winds (Fig. 10) de-

termined the zones for which there is the greatest risk of

thermal shock. For the June solstice, these are main ly the

northern slopes of the hill and all surfaces of niches, ter-

races, and platforms with northward exposure (Fig. 14).

It should be clearly stated here that the presented

risk scales do not use absolute values. This is particu-

larly important when comparing June data with Decem-

ber data. The risk of negative temperatures on El Fuerte

appears mainly in June, July, and August, and these are

the months in which the lowest maximum temperatures

are recorded. Together with September, this is also the

period of the strongest winds. Temperatures dropping be-

low zero and strong winds are the two most important

factors for the risk of thermal shock during this period

of the year.

Fig. 14. Risk of thermal shock on 21

June (elaborated by J. Kościuk)

Fig. 15. Risk of thermal shock on 21

December (elaborated by J. Kościuk)

 Petrographic,mineralogical,climaticanalyses,riskmaps / Analizypetrograficzne,mineralogiczne,klimatyczne,mapyryzyka 143

Risks of thermal shock for the December solstice are

completely dierent. Mainly all horizontal surfaces are

threatened, and on slopes, even southern ones, the risk is

much lower (Fig. 15).

In turn, mid-November to mid-March is a period when

the minimum temperature does not fall below zero. The

winds are also much weaker, but the maximum temper-

atures are among the highest in the year. An important

factor is also the dierence in the angle of incidence of

sunlight. From November to March, at these latitudes, in

the middle of the day, the sun’s rays fall almost perpen-

dicularly, so all horizontal surfaces heat up much more

strongly. Risk levels dened as “high” for June and De-

cember are therefore hardly comparable.

The risk of thermal shock is not the only threat to El

Fuerte. The previous section also specied the impact of

wind erosion (Fig. 10), slope inclination (Fig. 11), and lo-

cal depressions (Fig. 12). Summing up these four factors

gives us a full picture of the threats caused by climatic

factors and the site topography. The resulting map of ac-

cumulated risk factors (Fig. 16) indicates that, apart from

local depressions that were already discussed separately,

the most negative factors add up on all horizontal areas.

Therefore, from this point of view, conservation priorities

should focus on these areas, and particularly on all petro-

glyphs located there. This, however, does not exhaust the

list of all hazards regarding El Fuerte.

Biologically induced erosion

El Fuerte de Samaipata is a red sandstone rock with

high porosity. It contains iron compounds, which aect

both its colour and resistance to weathering. It is a rock

uplift exposed to destructive atmospheric factors: chemi-

cal, mechanical, and physical. Most processes are initiated

in the presence of water in various forms (rainfall, fog,

and condensation). Wind and associated abrasion process-

es are also highly destructive.

All these factors aect the intensive weathering of the

rock. There are deep cracks in sandstone in many places. The

largest run diagonally from the north-west to the south-east.

The rock is covered with reliefs of varying heights.

Due to the presence of articial (reliefs with symbolic

decoration) and natural depressions, there are numerous

deposits where water, loose rock fragments, and organic

material deposited by the wind collect. The accumulation

of humidity and organic material creates favourable con-

ditions for the development of lichens, and as a conse-

quence of their soil-forming activity, also the growth of

higher vegetation and mosses (Fig. 17). These plants are

Fig. 16. Accumulated risk factors: slope inclination, depression, thermal shock, wind erosion (elaborated by J. Kościuk)

Fig. 17. Green plant zones (elaborated by T. Dziedzic)

144 Wojciech Bartz, Jacek Kościuk, Maria Gąsior, Teresa Dziedzic

strongly rooted. Attempts to remove them can damage the

rock, and root expansion leads to the cracking and break-

ing of the rock.

The growth of numerous types of fungi, algae, and li-

chens including Lecanora, Cladonia, Aspicilia, Xantho-

ria, and Rhizocarpon is observed on most of the rock sur-

face (Fig. 18). Lichens cause moisture in the sandstone to

be retained, create colourful spots on the entire surface of

El Fuerte, and the acids they produce cause rapid progres-

sive chemical erosion of the rock [16].

Lichens have a particularly destructive eect, as they

form spots that are crusty and very sti. Their presence

causes not only drastic colour changes, but also the trans-

formation of the stone structure and, as a consequence,

the powdering and aking of the surface attacked by

them. This type of lichen is observed on the surface of

El Fuerte in the northern part, and is particularly abun-

dant in one of the most valuable parts of the monument

– within the relief that depicts the “Big Snake” (the rattle-

snake) and on the surface of what is known as the “Choir

of Priests”.

On the surface of the El Fuerte rock, there are also rel-

atively numerous traces of vandalism. These are mainly

inscriptions carved in rock of various sizes and locations.

Fortunately, since the introduction of restrictive regula-

tions that prohibit climbing on the rock, such activities are

no longer recorded.

The greatest damage is observed in the eastern and

southern parts of the rock. Damage in the eastern part

is mainly caused by exposure to wind and abrasion and

rainwater and its associated washing and rinsing of the

sandstone surface. In the southern part, the vertical sur-

face of the rock is destroyed, which is probably related to

the action of water accumulating in the rock structure and

just leaking on the southern slope of the hill.

On the northern slope of the rock, in a place invisible

to visitors, next to similar samples already made by Sonia

Avilés, an attempt was made to mechanically clean the

surface of the rock from lichens, fungi, and algae using

soft brushes and distilled water. The cleaned surface was

impregnated with Biotin biocide 40% (solution in a mix-

ture of ethyl and isopropyl alcohol).

The next test was carried out on a crude, lichen-cov-

ered piece of rock, which was impregnated with Biotin R

biocide alcohol solution (manufactured by CTS, Italy) in

various concentrations: 10%, 20%, and 40%.

It was observed that the application of Biotin R caused

the decay of lichens at the site (Figs. 19, 20). The eect

Fig. 18. Zones of intensive occurrence of lichens, fungi, and algae (elaborated by T. Dziedzic)

Fig. 19. Prior to application of 40% solution of Biotin R

(photo by M. Gąsior)

Fig. 20. Seven days after application

(photo by M. Gąsior)

 Petrographic,mineralogical,climaticanalyses,riskmaps / Analizypetrograficzne,mineralogiczne,klimatyczne,mapyryzyka 145

was dependent on the concentration applied – the higher

the concentration, the more eective the decay. Unfortu-

nately, within a short period of time, it was found that the

disinfected area of the rock was quickly inhabited by other

lichen specimens. This is a natural process – the special

agent used did not restrain the growth for much longer.

It should also be mentioned that disinfection of the

whole rock would require the use of a signicant amount of

chemicals, which would not be without impact on the natu-

ral environment. These types of actions can possibly be rec-

ommended only to protect the most valuable petroglyphs.

Petrographic and mineralogical analyses

Mineralogical analysis showed the investigated sam-

ples are clastic sedimentary rocks sandstones (Fig. 21).

Despite a few minor dierences, they share a number of

common petrographic features. They consist mainly of

quartz grains and a lower amount of feldspar and lithic

grains (rock fragments) (Table 1). The only exception is

sample 4, where the volume of quartz, as well as other

aforementioned constituents, is much lower (Table 1).

Accessory minerals, including tourmaline and zircon, are

very scarce (below 0.1 vol. %).

Rock porosity varies widely, from 3.5 vol. % up to

16.6 vol. % (Table 1). Moreover, there is a clear correla-

tion between porosity and cement volume (R = –0.67) as

samples with a larger cement volume usually exhibited

lower porosity. On the other hand, there is no clear corre-

lation between the total volume of the framework grains

and rock porosity (R = –0.15). Samples strongly enriched

in the framework are depleted in the cement (R = –0.63).

Matrix grains (<∼0.06 mm) occur in very small quanti-

ties. Thus, according to the most popular classication of

sandstones [17], the rock is quartz arenite (samples 4, 8,

11; Fig. 21) or subarcosic arenite (samples 5, 6, 7, 9, 10,

12, 13; Fig. 21).

The cement in the sandstone is in the form of quartz

overgrowth and the black/brown mass seen in plane-po-

larized light (Fig. 22) is composed of iron-oxide miner-

als (hematite, goethite) mixed with very ne, plate-like

minerals (sericite and/or clay minerals). Their proportion

is variable. In many cases, iron-oxide minerals dominate

(samples 5, 6, 7, 13), whereas most samples (4, 8, 9, 10,

11, 12) have cement enriched in clay minerals.

Sample 4 is slightly dierent in comparison to most

of the other samples. It exhibited unusually high porosity

and, at the same time, a high content of cement (Table 1).

An exact microscopic analysis of the sample showed that

it is heavily cracked (Fig. 22) and the framework grains

are disseminated in an abundant ne-crystalline mass. The

latter shows a yellowish tint, and it is more translucent

than typical cement that is seen as black/brown in plane-

polarized light. The observed dierences presumably re-

sult from a signicant degree of weathering of the rock.

Scanning electron microscopy

Mapping analyses of representative areas of rock sam-

ples, as well as point analyses, show that some elements

are abundant and relatively evenly distributed in the rock,

while others occur in minor amounts and only in places.

Moreover, there are some correlations between certain

elements. Thus, silicon and oxygen are ubiquitous, while

iron, potassium, and aluminium occur less frequently. On

this basis, it can be concluded that the vast majority of the

framework grains are quartz. The locally observed con-

centrations of potassium, aluminium, silicon, and oxygen

correspond to sparse alkali feldspar grains. The occurrence

of aluminium is, as stated above, associated with alkali

feldspars, but this element is also located in the intergran-

ular spaces. Its occurrence is associated with the pres-

ence of silicon and oxygen bound together in platy clay

minerals composed of close kaolinite or montmorillonite

Constituent

Sample number

4 5 6 7 8 9 10 11 12 13

Quartz 29.4 66.4 60.1 57.8 68.2 58.8 56.7 62.7 65.4 64.0

Feldspar 0.8 6.3 6.2 6.7 3.7 5.7 5.6 2.8 9.8 5.4

Lithic grains 1.0 4.7 3.0 3.2 3.5 1.7 4.8 1.4 1.9 1.6

Pore 25.0 7.3 5.9 3.5 10.1 16.6 14.5 6.1 9.1 15.7

Cement 43.8 15.3 24.8 28.8 14.5 17.2 18.4 27.0 13.8 13.3

Table 1. Results of modal analysis in vol. % (elaborated by W. Bartz)

Fig. 21. Classification of investigated Samaipata sandstone samples

(elaborated by W. Bartz)

146 Wojciech Bartz, Jacek Kościuk, Maria Gąsior, Teresa Dziedzic

(smectite). Sporadically, this kind of platy clay additional-

ly contains small amounts of potassium or iron and mag-

nesium, representing illite, sericite, or chlorite crystals,

respectively. The iron is evenly distributed in the inter-

granular spaces and is associated with oxygen. It indi-

cates that the crystalline mass (the cement) lling the pore

spaces between the framework grains contains iron oxide

minerals – hematite, limonite, or goethite. Sporadically

observed and weak calcium bands found on a few EDS

spectra imply the presence of small amounts of calcite.

Thermal analysis

A strong endothermic eect was visible at 570 °C with

no associated weight loss. This was the result of the pol-

ymorphic transformation of quartz [18, p. 55], [19]. At

temperatures below 100 °C, a strongly marked weight

loss was visible on the DTG curve (rst derivative of the

weight curve), and this is associated with an endothermic

eect. It should be interpreted as the process of releasing

unbound (hygroscopic) water. Less common (samples 9,

8, and 10) was a poorly marked deection of DTG and

DDSC (rst derivative of the heat-ow curve), which was

visible slightly above the temperature of 100 °C. This can

be interpreted as the process of dehydration of clay mine-

rals (illite and/or smectite) and/or gypsum [18, pp. 83,

76–80, 125]. Another strong weight loss was observed in

the temperature range of 200–400 °C. This was accom-

panied by endothermic eects well marked on the DDSC

curve. Within this temperature range, iron oxyhydroxides

(goethite, lepidocrocite) undergo dehydroxylation [18,

pp. 64, 65], [19, p. 38]. Some of the samples (7, 8, 9, 10)

showed additional minor endothermic and weight loss ef-

fects within this temperature region. Presumably, this was

due to the occurrence of both forms of FeOOH – goethite

and lepidocrocite. The dehydroxylation temperature of

these minerals diers by several dozen degrees Celsius

[18, pp. 64–85]. A very strong weight-loss reaction was

also observed at a temperature close to 500 °C, and was

accompanied by an endothermic reaction. Clay minerals

undergo dehydroxylation at similar temperatures. A very

strong weight loss suggests their abundance. Depending

on the type of clay mineral, the peak temperature of the

weight-loss eect shifts, and for smectite is 680 °C, illite

550 °C, and kaolinite 520–590 °C [18, pp. 68–84], [19,

pp. 28–34]. Due to the fact that some of the samples clear-

ly showed reactions below 500 °C, it can be assumed that

kaolinite was present in the sample. The split-peak behav-

iour observed in a few samples suggests an overlapped

dissociation of two dierent phyllosilicates – kaolinite and

Fig. 22. Samaipata sandstone sample 4 (upper row) and sample 7 (bottom row)

viewed in plane-polarized light (A) and cross-polarized light (B)

(elaborated by W. Bartz)

 Petrographic,mineralogical,climaticanalyses,riskmaps / Analizypetrograficzne,mineralogiczne,klimatyczne,mapyryzyka 147

illite. Some samples showed a relatively weak weight-loss

reaction with the associated endothermic reaction in the

temperature range of 630–660 °C. This could be inter-

preted as the result of the thermal dissociation of calcium

carbonates (calcite) with a low degree of crystallinity [18,

p. 108].

X-ray powder diffraction

The results showed the prevailing presence of quartz

due to the peaks at 3.34 Å, 4.25 Å, and 1.81 Å (Fig. 24).

Compared to quartz, less intense peaks were attrib-

uted to feldspar minerals – microcline and orthoclase,

evidenced by stronger peaks: 4.22 Å, 3.26 Å, 3.25 Å

and weaker peaks: 3.31 Å, 3.77 Å, 4.22 Å, respectively

(Fig. 23). Although the grains of minerals discussed be-

long to the frameworks and are relatively larger, a small

number of the nest fraction (the matrix) might have

passed through the sieve.

The next component identied by XRD was iron oxide

– hematite, representing the cement of the rock. Its pres-

ence is shown by the peaks observed at 2.70 Å, 2.52 Å,

and 1.69 Å (Fig. 23).

All of the obtained diraction patterns are marked with

peaks located at low 2Theta values: 10.10 Å, 7.18 Å, and

17.50 Å. They illustrate the presence of clay minerals: illite

(or structurally similar muscovite), kaolinite, and smectite,

respectively. Illite peaks were also found at 3.35 Å and

5.03 Å, while kaolinite peaks were found at 3.58 Å and

4.33 Å. Peaks located at 4.49 Å and 2.57 Å were attribut-

ed to smectite. Aluminium hydroxide (gibbsite) was also

identied on the basis of peaks 4.85 Å and 2.38 Å (Fig. 23).

The occurrence of crystalline phases within the ana-

lysed samples, identied by means of XRD, was strong-

ly varied. Quartz, alkali feldspar, hematite, illite (and/or

muscovite) were found in all samples. Unlike them, gibb-

site was very rare and was only found in sample 10. A fre-

quent constituent was clay minerals – kaolinite, identied

in samples 5, 7, 10, and 13 and smectite, found in samples

5, 7, 8, 11, and 13 (Fig. 23).

Properties of local clay

The clay fraction of the local clay material that could

potentially be used for restoration is a mixed-layer min-

eral – chlorite-smectite, kaolinite, and illite (or structur-

ally similar muscovite; Fig. 24). Illite was identied on

the basis of peaks at 10.06 Å, 3.33 Å, and 4.96 Å that

remained unaltered by ethylene glycol saturation and

heating to 550 °C. Kaolinite peaks were found at 7.21 Å,

3.57 Å, and 4.47 Å in the untreated sample and disap-

peared when the material was heated to 550 °C. The

mixed-layer smectite-chlorite was recognized on the basis

of peaks at 14.43 Å, 3.33 Å (overlapping with the most

intense quartz peak), and 7.21 Å. The small shift of peak

001 (d = 14.43 Å) after ethylene glycol saturation, visible

in the form of a wide peak at 15.5–17.5 Å, conrmed the

presence of the smectite layer. This particular peak shifted

to 8–10 Å when the sample was heated (Fig. 24).

As smectite is known to increase in volume many times

when watered and shrink when it dries, local clay is com-

pletely unsuitable for conservation purposes. Other clay

deposits need to be found, and their properties need to be

examined before using.

General conclusions and limitations

The formulation of strategies and detailed conservation

programs is not the purpose of this study. These must be

arranged by local specialists familiar with all local admin-

istrative and nancial conditions, as well as the availa-

bility of specic materials and conservation technologies.

Fig. 23. Results of XRD analysis

of sieved fraction <0.063 mm.

Mineral abbreviations:

sme – smectite, ilt – illite,

qtz – quartz, gbs – gibbsite,

kln – kaolinite,

or – alkali feldspar (orthoclase),

mc – alkali feldspar (microcline),

hem – hematite

(elaborated by W. Bartz)

148 Wojciech Bartz, Jacek Kościuk, Maria Gąsior, Teresa Dziedzic

References /Bibliografia

However, the analyses and data gathered above can be

used as a starting point for preparing detailed plans for the

long-term preservation of El Fuerte de Samaipata.

One recommendation can already be applied at this

stage. In light of the described climatic and topographic

factors, as well as the results of petrographic and mineral-

ogical analyses of local sandstone, it is necessary to consid-

er the problem of water retained in local depressions. Two

contradictory recommendations can be formulated for this:

– As smectite swells and shrinks, it is not advisable to

drain rainwater from all the depressions as soon as possi-

ble. Fast and frequent changes (watering and drying) will

promote the transformation of a smectite binder, which

would increase the volume of smectite and lead to faster

erosion of the rock. When considering only smectite, it is

advisable to leave the water in the depressions and allow

it to slowly evaporate.

Moreover, the contrary:

– The prolonged presence of water promotes the trans-

formation of the hematite binder into a ferric hydroxide

gel in the form of soluble oxyhydroxide, which causes

faster binder removal. From this point of view, it is essen-

tial to drain water from the depressions as soon as possi-

ble. The main part of the binder of Samaipata rock is he-

matite, and its dissolution seems to be the main reason for

the erosion of the rock.

Considering both recommendations and the reasons

behind them, it seems more important to drain water from

all depressions as soon as possible.

The size of the rock, its construction, chemical com-

position, position/location, and the multitude of destruc-

tive agents to which it is exposed create a combination

of factors that do not allow eective conservation and

maintenance procedures to be conducted that would ad-

equately protect this natural monument from destruction.

Only small-scale activities are possible for selected parts

of El Fuerte. One of the actions that could be taken and

which would delay the erosion of this valuable and unique

monument is to ll the main cracks in the rock that intensi-

fy erosion by allowing water to inltrate the rock. For this

purpose, it is possible to use mineral llers and binders rec-

ommended for the maintenance and conservation of stone

monuments [20]. This work should be carried out as a sep-

arate project under the guidance of an experienced stone-

art conservator and preceded by careful research and tests

on the use and application of suitable llers and binders.

Fig. 24. Results of XRD analysis

of oriented clay fraction mounts.

Mineral abbreviations:

sme-chl – smectite-chlorite,

ilt – illite, qtz – quartz,

kln – kaolinite, afs – alkali feldspar,

pl – Ca-Na feldspar

(elaborated by W. Bartz)

[1] De Alcaya D., Relación cierta [1636], [in:] I. Combès, V. Tyuleneva

(eds.), Paititi. Ensayos y documentos, Editorial Itinerarios, Cocha-

bamba 2011, 240–251.

[2] Ponce Sanginés C., Alcidesd’OrbignyysuviajeaSamaipataen1832,

[in:] R.D. Arze Aguirre (dir.), Elnaturalistafrancés,AlcideDessaline

d’Orbignyenlavisióndelosbolivianos, Institut Français d’Études

Andines, Plural editores, Lima–La Paz 2002 [1975], 307–315.

[3] Nordenskiöld E., ForschungenundAbenteuerinSüdamerika, Stre-

cker und Schroder, Stuttgart 1924.

[4] Pucher L., EnsayosobreelarteprehistóricodeSamaypata, “Revista

de la Universidad de San Francisco Xavier”, Universidad Mayor de

San Francisco Xavier Sucre, Sucre 1945.

[5] Alcázar de La Fuente R., El SitioCeremonialyAdministrativode

El Fuerte de Samaipata, “Tesape Arandu” 2015, Año 4, Nº 21, 1–9.

[6] Trimborn H., DerskulptierteBergvonSamaipata, “Archäologische

Studien in den Kordilleren Boliviens”, Bd. 3, “Baessler Archiv, Bei-

träge zur Völkerkunde” 1967, NF 5, 130–169.

[7] Meyers A., TrabajosArqueológicosenSamaipata,Departamentode

Santa Cruz, Bolivia,Primera Temporada 1992, “Boletín SIARB”

1993, Nº 7, 48–58, http://siarb-bolivia.org/wp-content/uploads/2019/

08/bol7a.pdf.

[8] Meyers A., Las Campañas Arqueológicas en Samaipata 1994–

1996.SegundoInformedeTrabajo, “Boletín SIARB” 1998, Nº 12,

59–86.

 Petrographic,mineralogical,climaticanalyses,riskmaps / Analizypetrograficzne,mineralogiczne,klimatyczne,mapyryzyka 149

Abstract

El Fuerte de Samaipata is a pre-Hispanic archaeological site in Bolivia that has been on the UNESCO World Heritage List. Its main part – the rock

– is densely covered with a complex arrangement of terraces, platforms, water reservoirs, channels, and petroglyphs. The rapidly progressing erosion

of the rock is making the petroglyphs less and less clear, and some are no longer recognisable. The main topic of this study is to indicate all risk

factors conducive to erosion and to create risk maps identifying the most vulnerable areas that require immediate conservation intervention. Parallel

mineralogical and petrographic studies show that the Samaipata rock is a quartz-rich, porous, well-sorted sandstone, classied as quartz arenite

or subarcosic arenite. The cement of the rock is composed of quartz overgrowth and ubiquitous, pore-lling hematite-clay aggregates containing

non-expanding kaolinite, illite, and expanding smectite. The rock exhibits dierent stages of weathering, from relatively fresh to strongly altered and

heavily cracked. In comparison to fresh rock, the latter has cement enriched in clay minerals and is depleted in hematite due to weathering and the

dissolution of the iron-bearing phase.

Key words: Samaipata, conservation, climatic and topographical risk factors, sandstone, mineralogy

Streszczenie

El Fuerte de Samaipata to wpisane na Listę Światowego Dziedzictwa UNESCO prehiszpańskie stanowisko archeologiczne w Boliwii. Jego główna

część to skała ze złożonym układem tarasów, platform, zbiorników wodnych, kanałów i petroglifów. Szybko postępująca erozja sprawia, że petroglify

stają się coraz mniej wyraźne, a niektóre nie są już rozpoznawalne. Głównym tematem badań jest wskazanie wszystkich czynników ryzyka sprzyjają-

cych erozji oraz stworzenie map ryzyka identykujących najbardziej wrażliwe obszary wymagające natychmiastowej interwencji konserwatorskiej.

Badania mineralogiczne i petrograczne wskazują, że Samaipata to bogaty w kwarc, porowaty, dobrze posortowany piaskowiec, sklasykowany jako

arenit kwarcowy lub arenit subarkozowy. Spoiwo składa się z przerostu kwarcu i wszechobecnych, wypełniających pory agregatów hematytowo-

-gliniastych zawierających nierozprężający się kaolinit, illit i rozszerzający się smektyt. Skała wykazuje różne etapy wietrzenia, od stosunkowo świe-

żego do mocno zmienionego i mocno spękanego. W porównaniu ze świeżą skałą ta ostatnia ma cement wzbogacony w minerały ilaste i jest zubożona

w hematyt z powodu wietrzenia i rozpuszczenia fazy żelazonośnej.

Słowa kluczowe: Samaipata, konserwacja, klimatyczne i topograczne czynniki ryzyka, piaskowiec, mineralogia

Acknowledgements /Podziękowania

Thepresentedworkisa partof theresearch projectsponsoredbythe

grant given to the Wrocław University of Science and Technology by

thePolish NationalScience Centre(grant No.2014/15/B/HS2/01108).

Additionally, the municipality of Samaipata, represented by Mayor

FalvioLópesEscalera,contributedtothisresearchbyprovidingtheac-

commodationduringtheeldworkinJuneandJuly2016,aswellasin

July2017.TheMinistryofCultureandTourismofBoliviakindlygrant-

edallnecessarypermits(UDAMNo.014/2016;UDAMNo.060/2017).

The research was conducted in close cooperation with the Centre for

Pre-Columbian Studies of the University of Warsaw in Cusco. Spe-

cialists from many other universities and research centresalso joined

theproject.

[9] Meyers A., Reexiones acerca de la periodización de la Cultu-

ra Inka: perspectivas desdeSamaipata, Orientede Bolivia, [in:]

C. Diez Marín (ed.), ActasdelXIICongresoNacionaldeArqueo-

logíaArgentina, Editorial de la Universdad de la Plata, La Plata

1999, t. 1, 239–251.

[10] El Fuerte de Samaipata: Estudios arqueológicos, A. Meyers,

I. Combès (comp.), Biblioteca del Museo de Historia Universidad

Autónoma Gabriel René Moreno, Santa Cruz de la Sierra 2015.

[11] Combès I., Meyers A., El Fuerte de Samaipata en contexto: Es-

tudios históricos, Biblioteca del Museo de Historia Universidad

Autónoma Gabriel René Moreno, Santa Cruz de la Sierra 2018.

[12] Avilés S., ConservazionedeltempiodellaroccascolpitadiSamai-

pata–SantaCruz,Bolivia(Sudamerica), Tesi di Master: Universi-

tà di Bologna, Sede di Ravenna, Facoltà di Conservazione dei Beni

Culturali, Dipartimento di Storie e Metodi per la Conservazione dei

Beni Culturali, 2002, www.stonewatch.de/media/download/sc%20

04.pdf [accessed: 27.09.2017].

[13] Avilés S., IntroduzioneallaconservazionedellaRocciaScolpitadi

Samaipata, Bolivia, 2011, http://www.rupestreweb.info/samaipata.

html [accessed: 27.09.2017].

[14]

Avilés S., LaconservacióndelaRocaSagradadeSamaipata, [in:]

A. Meyers, I. Combès (comp.), El Fuerte de Samaipata: Estudios

arqueológicos, Biblioteca del Museo de Historia Universidad Autó-

noma Gabriel René Moreno, Santa Cruz de la Sierra 2015, 161–170.

[15] Elsen J., Microscopyofhistoricmortars–areview, “Cement and

Concrete Research” 2006, Vol. 36, Iss. 8, 1416–1424.

[16] Strzelczyk A.B., Karbowska-Berent J., Drobnoustroje i owady

niszczącezabytkiiichzwalczanie, Wydawnictwo Naukowe UMK,

Toruń 2004.

[17] Pettijohn F.J., Potter P.E., Siever R., Sand and sandstone, Springer-

Verlag, Berlin–Heidelberg–New York 1972.

[18] Földvári M., Handbook of thermogravimetricsystem of minerals

and its use in geological practice, Magyar Állami Földtani Intézet

(Geological Institute of Hungary), Budapest 2011.

[19] Wyrwicki R., Analiza derywatograczna skał ilastych, Wydaw-

nictwo Uniwersytetu Warszawskiego, Warszawa 1988.

[20]

Łukaszewicz J.W., Rodzajekamienisztucznych,właściwościorazod-

pornośćnadziałanieczynnikówniszczących, [in:] W. Do ma słowski

(red.), Zabytki kamienne i metalowe ich niszczenie i konserwacja

prolaktyczna, Wydawnictwo Naukowe UMK, Toruń 2011, 47–87.

Documenting petroglyphs

(photo by M. Ziółkowski)

Dokumentowanie petroglifów

(fot. M. Ziółkowski)