23
Arheo 26, 2009, 23-32
Introduction
The quality of spatially oriented studies depends on
many elements, which cannot be easily and uniformly
described. It is also impossible to provide a systematic
taxonometry for them. The data or model accuracy may
affect subsequent processing steps strongly. We can con-
sider the errors as a critical element for determination
something as “appropriate for usage”, according to our
tasks. It means that the error should not be considered
just as potential inconvenience. We can distinguish two
basic groups of errors (Burrough, McDonnell 1998,
241). The first considers data sources (inherent to the
object/entity that is measured or observed) and the sec-
ond considers a processing/operational error
(error/uncertainty propagation considering spatial analy-
ses and modelling). Another view distinguishes between
error in the system and human error (factor). Errors in a
system can be considered as latent design errors. We
should be aware that a low quality of the source data
causes a low quality of the final product. The human
error includes cognitive bias, which can be minimised by
making systems more forgiving or error-tolerant. The
human error is also related with interpretation of final
results of the modelling.
The majority of error sources can be found, described or
evaluated by statistical methods or by other means
(Podobnikar 2009). The measurement errors are split to
random, systematic and gross. The first group of errors
is considered as random variation and would be dimin-
ished with higher quality measurements/observations.
The second group may be eliminated only when the
accurate data are presented. The last group is often con-
sidered as a mistake as it is caused by a fault. Although
the random measurement errors of spatial data are main-
ly considered to be normally distributed, their character
often contains admixtures of poorly explained properties
and other uncertainties that came from difficulties in
separation between particular types of errors. Many stan-
dard methods for measuring and further modelling errors
techniques are known, e.g. known as standard deviation,
root means square error, mean error, rank, …, bench-
marking, flowcharting, control charts,…, inspection, etc.
Error distributions that can be explained with stochastic
rules can be even modelled (simulated), for example by
Monte Carlo simulations.
Analyses and modelling of the cultural landscapes could
be a high level uncertain process if we don’t understand
the nature of data well, and especially if there is not
enough metadata available. Different kind of spatial
analyses require various aspects of knowledge of the
data sets and their properties. 
This paper provides an empirical explanation of the
effect of data uncertainty that influence particular mod-
elling tasks (visibility, path modelling, catchment, analy-
sis, etc.). Three selected approaches to uncertainty analy-
ses and modelling of cultural landscapes in landscape
Consequences of Data and Information Uncertainties for
Spatial Modelling of the Cultural Landscape 
Posledice negotovih podatkov in informacij pri prostorskem 
modeliranju kulturne krajine
© Tomaž Podobnikar
UNG, Fakulteta za znanosti o okolju
ZRC SAZU, Inštitut za za antropološke in prostorske študije
tp@zrc-sazu.si
Abstract: The quality of data and information depends on many ele-
ments that may considerably affect to the results of subsequent spatial
modelling and interpretations in archaeology. This complex needs
careful treatments that enable recognition the nature of the entirely
studied problem. Three applied problems to the uncertainty analyses
and modelling of cultural landscapes in landscape archaeology are
exposed: (1) influence of the sea level changes on the interpretation of
various ancient human activities, (2) analyses of the optimal paths
algorithms and reliability of the data sources for interpretation of the
ancient paths between settlements, and (3) analyses of the visibility
from the significant points and visual communication interpretation.
The case study areas are from Croatia (Central Dalmatia), Slovenia
(Dolenjska) and Mexico (Yucatan, Campeche).
Keywords: uncertainty, quality of data, model and interpretation,
paths, visibility, simulations
Izvleček: Kakovost podatkov in informacij je odvisna od velikega
števila elementov, ki lahko pomembno vplivajo na rezultate nadaljnje-
ga prostorskega modeliranja in interpretacije v arheologiji. Vse to
potrebuje skrbno obravnavo, kar omogoča celovito spoznavanje nara-
ve obravnavanega problema. V okviru arheologije krajine smo opisali
tri probleme negotovosti analiz in modeliranja kulturne krajine: (1)
vpliv spreminjanja višine morja pri interpretaciji preteklih človekovih
dejavnosti, (2) analize algoritmov za iskanje optimalnih poti in zane-
sljivosti virov podatkov za interpretacijo starodavnih poti med naselbi-
nami in (3) analize vidnosti z značilnih točk ter interpretacija vizualne
komunikacije. Izbrana so bila naslednja študijska območja: Hrvaška
(Srednja Dalmacija), Slovenija (Dolenjska) in Mehika (Jukatan,
Campeche).
Ključne besede: negotovost, kakovost podatkov, modela in inter-
pretacije, poti, vidnost, simulacije
1.01 Izvirni znanstveni članek
Podobnikar1:master test 3+1  2.12.2009  19:50  Page 23

24
Consequences of Data and Information Incertainties for Spatial Modelling of the Cultural Landscape
archaeology are exposed: (1) influence of the sea level
changes on the interpretation of various ancient human
activities, (2) analyses of the optimal paths algorithms
and reliability of the data sources for interpretation of the
ancient paths between settlements, and (3) analyses of
the visibility from the significant points and visual com-
munication interpretation. The case study areas are from
Croatia (Central Dalmatia), Slovenia (Dolenjska) and
Mexico (Yucatan, Campeche).
Problem 1: Information on sea level – Central
Dalmatia, Croatia
The estimation and simulation of the sea-level regarding
present situation for the selected historical periods:
Upper Palaeolithic, Neolithic, Bronze Age, and the
Roman period, was performed. The main objective of
the unpublished paper (Podobnikar et al. 2003) was to
discuss sea-level data for the Eastern Adriatic coast,
Central Dalmatia, including its uncertainty and devia-
tions, for the last 15,000 years. The paper focused on
modelling the average sea-level of the Adriatic basin in
order to recognize if and how its dynamics could alter
the human activity areas from the Neolithic, through the
Bronze and Iron Ages, to the Roman period. The aim
was twofold: to observe settlement patterns over their
likely past sea-level shape, and to verify regularity of
spatial analyses and re-evaluate corresponding hypothe-
ses regarding data sources that correspond to recent
environment and water source data.
Many different spatial analyses have been performed
and interesting results were obtained for the area of
Figure 1. Coastline variation of simulated models of palaeo-landscapes for the Roman period, –1m (a), Bronze Age. +8m (b), Neolithic, –6m (c)
and Palaeolithic, –70m (d). The darker area represents the area, in which the sea level was not modelled.
Slika 1. Spreminjanje obalne črte na podlagi modelov paleokrajin za (a) rimsko dobo, -1m, (b) bronasto dobo, +8m, (c) neolitik, -6m in za paleo-
litik, -70m. Višina morja (obalna črta) ni bila modelirana na potemnjenih območjih.
Podobnikar1:master test 3+1  2.12.2009  19:50  Page 24

25
Arheo 26, 2009, 23-32
Central Dalmatia (e.g. Stančič, Kvamme 1999). The
analysis and results were promising; nevertheless these
spatial analyses were all based on the present coastline.
Our further idea was that a paleo-landscape– more par-
ticularly the coastline changes–might be an important
factor of reliability of the archaeological spatial analysis
(Podobnikar et al. 2003). For modelling of the sea-level
changes, a digital terrain model (DTM) and a model of
the sea bed surface (bathymetrical DTM) were acquired
and produced from different data source. These were
then integrated into one digital surface model (see the
technique in Podobnikar 2005). The global sea-level
model indicates 14 m of difference between the
Neolithic (–6m) and Bronze Age (+8m on average).
According to the nowadays sea-level we can – according
to Rohling et al. (1998) – note variations from 1 and up
to 8m. With such amplitudes, we can expect substantial
landscape modifications, which can alter our perception
of territorial dynamics in the past. The result of the sea
level change on the coast line is presented in Figure 1. 
Under these conditions, we can assume that sea-level
change has had a significant impact on settlement organ-
isation, especially in the present relatively shallow sea
bed areas. This analysis was performed within a geo-
graphical information system (GIS). The results verified
that the settlement pattern was not noticeably affected by
sea-level changes in the last 10,000 years due to the fact
that most of the islands had rather steep slopes which
resulted in a low coastline variation. However, in the
case of the Starigrad plain on the island of Hvar (Figure
2), with a large plain with a low elevation in respect to
the sea, a significant impact of sea-level change could be
observed. A different fractal dimension of the modelled
coast lines that is visually notable (Figure 2) is addition-
ally partly resulted from various qualities of the input
data sources for the DTM production that is also a poten-
tial source of the uncertainties in the modellin
However, it was discovered that the estimated sea-levels
based on the global model were not valid within the proj-
ect area, before the paper was published. This error in the
model negated the neat archaeological theory concerning
past human activities in the paleo-landscape in the
Central Adriatic Islands for the past 15,000 years, which
was constructed from the results (see Podobnikar et al.
2003). The problem was that no archaeological studies
were found during further investigations, which support-
ed a sea-level of 8m higher than the present
Mediterranean coastline during the Bronze Age. It is
confirmed by historians and archaeologists that the
recent coastline had effectively been reached by the Late
Neolithic in the east Mediterranean. Nevertheless, we
should consider minor or occasionally dramatic, local
variations (Mitrovica 2003), which are also evident from
observation of the remains of certain Roman villas. The
main problem in the presented modelling and interpreta-
tion failure of the presented study seemed to be that a
global model of the sea-level changes was applied,
which deviates significantly from the local situation
(Adriatic Sea). However, for the more detailed analysis
of the ancient coastline, the tectonic movements and
dynamics of the sediments depositions of Neretva River
should at the very least be analysed using different tech-
niques.
Problem 2: The influence of DTM quality on
path simulation – Dolenjska, Slovenia
Within this study the simulation of possible courses of
the ancient paths regarding environmental characteristics
were performed (Podobnikar et al. 2007). The simulation
enables the variation of different parameters, for calcu-
lating the path between selected points. The algorithms
were applied with spatial analysis in GIS. On the basis of
similarities of archaeological remains, we assume strong
contacts and exchanges between the settled centres. Our
assumptions of the optimal paths modelling were the fol-
Figure 2. Coast line variation of different periods in the central part of
the island of Hvar.
Slika 2. Spreminjanje obalne črte za različna obdobja, za primer osred-
njega dela otoka Hvara.
Podobnikar1:master test 3+1  2.12.2009  19:50  Page 25

26
Consequences of data and information uncertainties for spatial modelling of the cultural landscape
lowing (Podobnikar et al. 2004, 254): (1) predominant-
ly hilly terrain as the most important factor, (2) climate
regime and other environmental conditions such as
marshy soil, size of the rivers, vegetation, geology could
also be of importance; however, we suggest that these
conditions have not been changed drastically to the
present day, (3) evidence of boats for river movement
and wagons are not available, so we assume the settlers
used to walk between the chosen settlements within the
landscape, and (4) the impact of cultural and socio-eco-
nomic factors on the decision making process. 
The chosen study area was the central-southern region
of Slovenia, Dolenjska (Lower Carniola), which is lim-
ited by natural boundaries of the river Krka on the
south-west side and the river Sava on the north-east. 14
Iron Age settlements from a total of 38 documented in
this area were selected and the presences of 18 paths
that run between them were studied. The paths’ simula-
tion was performed just based on the DTM which
seemed to be the most important factor. All of the algo-
rithms base on least-accumulative-cost distance over a
cost surface. In order to evaluate the results nine sets of
cost surface parameters within the simulation were
selected (Podobnikar et al. 2004, 255). Some results of
these simulations can be inspected in Figure 3.
The analysis of the results indicated that it was possible
to recognise that simulated paths on the hilly areas are
less dispersed and denser than in less hilly areas as such
Figure 3. Simulation of ancient paths using nine different cost surface models. On the flat areas, the possible courses are more dispersed. The
bold white line illustrates paths that were determined by the subjective knowledge of archaeologists, based on standing monument survey.
Slika 3. Simulacija zgodovinskih poti pri implementaciji devetih različnih modelov stroškovnih ploskev in uporabi DMR-ja. Poteki možnih poti
so bolj razpršeni na ravnih območjih. Poudarjena bela črta predstavlja poti, določene na podlagi znanja arheologov.
Podobnikar1:master test 3+1  2.12.2009  19:50  Page 26

27
Arheo 26, 2009, 23-32
areas offer fewer possibilities for economical transport.
In less hilly areas, other factors are more important than
terrain. The archaeologist can subjectively interpret
which simulated algorithm or which paths are more rel-
evant. On the base of the presented simulation some seg-
ments of ancient path in Požarnice were found by
archaeologists soon after implementation of the algo-
rithm. However, the excavated material has not yet been
completely analysed, therefore they are not yet confident
that the segments definitely represent an Early Iron Age
path (Tica 2003; Topličanec 2006; Dular, Tecco Hvala
2007, 229–230). This approach shows the importance of
parameter variation. The visualisation of all results with-
in one figure allows inspecting possible areas of uncer-
tainties in the simulation results.
However, it has to be considered that for a more reliable
study the significance of further factors influencing the
run of the paths might be necessary. The importance of
the factors fluctuates in respect to the different geomor-
phology of the area. Within this study we could show
with the empirical tests that simulations allow detecting
the paths quite reliable. Some of the significant influence
factors are: (1) stability of the starting point, depending
on the geomorphology; in some cases small changes can
lead to a different path, (2) selection of the appropriate
algorithm for path calculating, (3) selection appropriate
cost surface weights in the algorithm and other con-
strains (e.g. what kind of path is it, what kind of trans-
port was there), (4) the path from one point of the other
is newer the same to the reverse path direction
(anisotropic), (5) gross errors on the DTM can play an
important role (Podobnikar 2008, 146), (6) the various
quality of applied DTMs (considered without of gross
errors) and chosen resolution can lead to differently sim-
ulated paths (cf. Figure 4).
Figure 4. Simulated paths determined from DTMs with different quality using the same path determination algorithm and parameters.
Slika 4. Simulacija poti na podlagi DMR-jev različnih kakovosti pri uporabi istega algoritma in parametrov.
Podobnikar1:master test 3+1  2.12.2009  19:50  Page 27

28
Consequences of data and information uncertainties for spatial modelling of the cultural landscape
Problem 3: Influence of height uncertainty of
pyramids and vegetation cover on visibility
studies – Yucatan, Mexico
The visibility studies of the archaeological sites can con-
tribute to the understanding of several questions: Is visi-
bility from archaeological sites connected with the land-
scape perception? Are natural aspects of site locations
significantly different from other (average or randomly
selected) areas? How did the visibility and inter-visibili-
ty influence the location choices? Are there any signifi-
cant spatial patterns? Was it possible to communicate
between the neighbouring centres or to control others
from the tops of the sites? How much of the area was
controlled from the sites? Are the answers to those ques-
tions clear and significant?
Our primary assumption was that visibility played an
important role in settlement patterning in the Maya
Lowlands central parts of the Yucatan Peninsula in
Mexico (Podobnikar, Oštir 2008, 254). Since the area
has so far been practically unexplored, the final results
represented a significant contribution to the understand-
ing of the Maya culture in central lowlands of the
Yucatan peninsula. Visibility studies have been per-
formed and intervisibility between centres has been
determined. Viewsheds from all of the sites, with differ-
ent offsets of heights for observation and observed (tar-
get) points, helped us to understand the terrain character-
istics, possible visual communications between different
centres, reasons for the chosen centres locations, etc. We
have proven that positions of the centres chosen by the
Maya lie significantly in areas that are more visible than
the average (random) study area.
Figure 5. Viewsheds from Yaxnohcah Structure B-3 with different offsets, showing the visibility potential of the site (presented as a rectangle with
dimensions of approx. 60 by 40km). Offsets (from/to) are (a) 2/2m, (b) 10/2m, (c) 30/30m and (d) real heights of all selected pyramids (20)m for
Yaxnohcah Structure B-3) with offset 2/2m.
Slika 5. Izračuni vidnosti iz Strukture B-3 Yaxnohcaha ob upoštevanju različnih nadvišanj nad terenom, kar kaže na potencial vidnosti z določene
točke na območju, predstavljenem kot pravokotnik z merami pribl. 60 krat 40 km. Nadvišanja (od/do), so (a) 2/2 m, (b) 10/2 m, (c) 30/30m in (d)
dejanske višine izbranih piramid (20 m za Yaxnohcah Struktura B-3) z dodatnim nadvišanjem 2/2m.
Podobnikar1:master test 3+1  2.12.2009  19:50  Page 28

29
Arheo 26, 2009, 23-32
The visibility analysis requires a high quality DTM,
information about heights of the buildings – mostly pyr-
amids (53 from 40 major, medium and minor centres)
that are used for observations – artificial outlooks and
information about vegetation cover. Visibility maps from
all 53 structures were calculated. Figure 5 presents the
results of these simulations using different parameters
for Structure B-3 of Yaxnohcah. Figure 5a shows visibil-
ity from the ground with offsets of 2m (2/2m – for obser-
vation/target or from/to points) applied to the observa-
tion and target point; it reveals that the site’s visibility
potential was rather poor before the pyramid was built.
Figure 5b with offsets 10/2m presents visibility from
small pyramids or trees about 10m high; we can see that
large areas to the NW have become visible. Picture c
shows an extreme case with offsets of 30/30m, when the
observer is located in a tree top and observes the tops of
other trees. The examples illustrated in Figure b and c
are useful for a better understanding of the natural terrain
characteristics regarding chosen observation points and
data quality. They can also help us to understand the why
the founders of the particular sites have chosen their
positions and what they might have possibly seen from
the trees before they built the settlements. Figure 5d,
shows a situation after the pyramid with a height of 20m
was built and furthermore assumes that the vegetation
was cut down in order to produce viewsheds that are as
realistic as possible.
Visibility analysis helps to understand the importance of
visual communication. This seemed to be very important
for the Maya settlements on Yucatan peninsula. The
basic data were the DTM and heights of the most domi-
nant pyramids. The most uncertain variable was vegeta-
tion. The visibility was simulated with or without vege-
tation consideration from the areas of pyramids or their
tops. The study shows on one hand a quite high sensitiv-
ity to possible errors on DTM (see also Podobnikar
2008) or heights of pyramids for some sites and on the
other hand it describes the intervisibility between some
pyramids that can not be tested in reality, due to the
dense modern vegetation that cover the selected area.
Namely it seems that some pyramids are positioned on
special positions that allowed visual communication, as
it was shown by different analyses. However, it is impos-
sible to validate these results in field. The main problems
that increase uncertainties of the model interpretation are
therefore potential lack of quality of the DTM, inade-
quate classification of the pyramids according to their
functionality, and uncertain or non-stop changed vegeta-
tion coverage.
Conclusion
In the study, three different examples were explained and
possible problems of the results were analysed. (1) The
analysis of the Bronze Age settlements and their cultural
landscape show that just slightly wrong information on a
sea level in the past, can produce unpredictable results
on spatial analysis and at the same time deny all the reli-
able looking hypotheses behind. The main problem was
therefore the sources of the information and additional
potential the unpredictable influences of the DTM qual-
ity to the coastline modelling that should be both careful-
ly evaluated for the reliable results of such studies. (2, 3)
Analysis of the ancient paths or visual communications
shows potentially unexpected results, therefore they
need additional verification. In our case the simulation of
different possible parameters was performed. The results
show that in some cases just slightly different coordi-
nates from the present locations of starting and target
points of paths or points of observation can introduce
completely different solutions. Additional problem can
be uncertain quality of the used DTMs. One of the reli-
able solutions of such problem might be in (Monte
Carlo) simulations of different parameters and finding
the most feasible result(s).
Quality of the applied data, either spatial datasets or
archaeological evidences is a crucial factor for the reli-
able results in modelling. The chosen modelling tech-
niques are therefore less important than the quality of the
data. Significant and not enough respected value in spa-
tial modelling are availability of the data/information
sources and especially their quality. Cost of the data is
nowadays much higher than of the hardware or software
and even of the management of the whole system.
What can be the solutions of the presented problems?
The described examples show the importance of under-
standing the data that are used for the certain application
and algorithms behind the modelling. To enhance the
understandability, the data/information and their imple-
mentation (procedures) to the models should be careful-
ly studied. Not perceived gross or systematic error or not
assessed random error may cause unexpected and even
wrong results of the particular study. Decision maker
should be then able to do a relevant interpretation of his
work. An important problem presented in this study is
Podobnikar1:master test 3+1  2.12.2009  19:50  Page 29

30
Consequences of data and information uncertainties for spatial modelling of the cultural landscape
availability of appropriate quality data/information. It is
much bigger than the quality of the numerical algo-
rithms, models or simulations. There are many empiri-
cal and numerical tools and techniques for quality
assessment, which are presented in the introduction, but
one of the first techniques that might be applied is “rule
of thumb” that or the best implementation needs a lot of
experiences. Further complex techniques includes study
of the metadata with the data profiling, including sam-
pling, applying the ETL (extract, transform, and load)
functionalities with quality assessment/control tech-
niques or even total quality management approaches.
Podobnikar1:master test 3+1  2.12.2009  19:50  Page 30

31
Arheo 26, 2009, 23-32
Posledice negotovih podatkov in informacij
pri prostorskem modeliranju kulturne krajine
(Povzetek)
Kakovost podatkov in informacij je odvisna od velikega
števila elementov, ki lahko pomembno vplivajo na
rezultate prostorskega modeliranja in interpretacije v
arheologiji. Vsi elementi potrebujejo skrbno obravnavo,
ki nam omogoča celovito spoznavanje narave obravna-
vanega problema. V študiji so bila analizirana izbrana
območja kulturnih krajin Hrvaške (Srednja Dalmacija),
Slovenije (Dolenjska) in Mehike (Jukatan, Campeche).
V okviru arheologije krajine smo razčlenili tri probleme
negotovosti rezultatov analiz in modeliranja kulturne
krajine. (1) Analizirali smo vpliv spreminjanja višine
morja pri interpretaciji preteklih človekovih dejavnosti,
kot so bronastodobne naselbine. Prikazali smo, da je
rezultat na podlagi analiz prostorskih podatkov lahko
nepredvidljiv in popolnoma napačen, četudi je informa-
cija višine morja za obdobje bronaste dobe le malo
različna od pravilne. Dodatno vpliva tudi kakovost digi-
talnega modela reliefa (DMR) na podlagi katerega
lahko določamo obalno črto v povezavi z višino morja.
Pod današnjo morsko gladino je DMR praviloma nena-
tančno izdelan, poleg tega je bilo paleo površje druga-
čno od današnjega. Ob vsem tem se na podlagi napačnih
rezultatov lahko še tako verodostojna hipoteza izkaže za
neprimerno.
Nadalje smo analizirali (2) algoritme za iskanje opti-
malnih poti in zanesljivosti virov podatkov za interpre-
tacijo starodavnih poti med naselbinami in (3) vidnost z
značilnih točk ter interpretirali vizualne komunikacije.
Pri omenjenih nalogah so se izkazali za pomembne
izbor parametrov algoritmov, natančnost določitve
koordinat začetka in konca poti oz. stojišč (še posebej
na razgibanem reliefu) ter kakovost DMR-ja. 
Opisani primeri kažejo na pomembnost razumevanja
podatkov, ki jih uporabljamo za izbrane aplikacije in
algoritme v sklopu modeliranj. Kakovost uporabljenih
prostorskih podatkov ali arheoloških dokazov so porok
za zanesljivost rezultatov prostorskih analiz.
Podobnikar1:master test 3+1  2.12.2009  19:50  Page 31

32
Consequences of data and information uncertainties for spatial modelling of the cultural landscape
References
BURROUGH, P. A., R. A. MCDONNELL 1998,
Principles of Geographical Systems Information
Systems: Spatial Information Systems and Geostatistics.
Oxford University Press.
DULAR, J., S. TECCO HV ALA 2007, South-eastern
Slovenia in the early Iron Age: settlement, economy,
society = Jugovzhodna Slovenija v starejši železni dobi:
poselitev, gospodarstvo, družba. Opera Instituti archae-
ologici Sloveniae 12, Ljubljana.
MITROVICA, J. X. 2003, Recent controversies in pre-
dicting post-glacial sea-level change. – Quaternary
Science Reviews 22, 127-133.
PODOBNIKAR, T. 2005, Production of integrated dig-
ital terrain model from multiple datasets of different
quality. – International journal of geographical infor-
mation science 19/1, 69–89.
PODOBNIKAR, T. 2008, Simulation and representa-
tion of the positional errors of boundary and interior
regions in maps. In: A. Moore, I. Drecki (eds.),
Geospatial vision: new dimensions in cartography:
selected papers from the 4th National Cartographic
Conference Geocart'2008, New Zealand. – Lecture
notes in geoinformation and cartography, Berlin,
Heidelberg, 141–169.
PODOBNIKAR, T. 2009, Methods for visual quality
assessment of a digital terrain model. – S.A.P .I.EN.S,
Special Issue, 2/20, http://sapiens.revues.org/index738.html.
PODOBNIKAR, T, C. BRIESE, M. DONEUS, N.
PFEIFER 2007, Uncertain observations of the Past with
a modern outcome. In: W. Börner (ed.). Archäologie
und Computer, 5.-7. November 2007: Workshop 12 :
Kulturelles Erbe und neue Technologien , Wien, 1–9 .
PODOBNIKAR, T., L. NUNINGER, T. VELJANOV-
SKI, Z. STANČIČ, V . GAFFNEY 2003, Past human
activities concerning palaeo-environment in the Central
Adriatic Islands, 23 pp. (unpublished paper).
PODOBNIKAR, T., K. OŠTIR 2008, Geographical
information systems and remote sensing analysis. In: I.
Šprajc (ed.). Reconocimiento arqueologico en el sureste
del estado de Campeche, Mexico : 1996-2005, BAR
international series, Paris monographs in American
archaeology, 1742, 19. – Archaeopress 243–261.
PODOBNIKAR, T, S. TECCO HV ALA, J. DULAR
2004, Iterative approach to ancient paths modelling in
the Iron Age study of the Dolenjska region (Slovenia).
In: K. Fischer Ausserer (ed.), Enter the past: the e-way
into the four dimensions of cultural heritage: proceed-
ings of the 30th conference, BAR international series,
1227. – Archaeopress 254–257.
ROHLING, E. J., M. FENTON, F. J. JORISSEN, P.
BERTRAND, G. GANSSEN, J. P. CAULET 1998,
Magnitudes of sea-level lowstands of the past 500,000
years. – Nature 394, 162–165.
STANČIČ, Z., K. KV AMME 1999, Settlement patterns
modelling through Boolean overlays of social and envi-
ronmental variables. In J. Barcelo, I. Briz, A. Vila (eds.),
New technique for old times. – Archaeopress 16 pp.
TICA, G. 2003, Požarnice pri Družinski vasi. In: B.
Djurić (ed.), Zemlja pod vašimi nogami. Arheologija na
avtocestah Slovenije. – Vodnik po najdiščih, Ljubljana,
229-231.
TOPLIČANEC, M. 2006, Rimska in prazgodovinska
cesta pri Požarnicah: rimska cesta Emona-Neviodunum-
Siscia. – Rast (104) 17/2, 189–190.
Podobnikar1:master test 3+1  2.12.2009  19:50  Page 32