This action might not be possible to undo. Are you sure you want to continue?
AERIAL ARCHAEOLOGY IN THE LANDSCAPE OF CARNUNTUM
Introduction - Potential and Need of Archaeological Prospecting in Carnuntum The archaeological landscape of Carnuntum is located 45 km E of Vienna, close to the Slovakian border (ﬁg. 1). As the Roman capital of the province Pannonia superior, Carnuntum was an important town during the ﬁrst four centuries of the ﬁrst millennium AD. Since it was abandoned in the ﬁrst half of the 5th century AD, only little building activity has taken place in this area. Hence, large parts of the buried remains are under arable land not covered by modern buildings. Only two small villages, Petronell and Bad Deutsch-Altenburg are built on top of the Roman ruins. In many areas of Carnuntum, the Roman ruins remained untouched for centuries and were only removed during the modern age, when cheap building material was needed. During the 19th century Carnuntum was called the “Pompeii at the doors of Vienna” due to the good preservation of the Roman ruins. In the meantime, the situation has changed drastically. Both aerial photography and geophysical data show that the archaeological heritage has suffered severe damage because of the local agricultural policy of the last few decades. Erosion processes accelerated by modern agricultural methods take away centimetre by centimetre of the archaeological layers. Many ﬁelds have been ploughed deeply resulting in a large-scale destruction of the antique features. A large stone quarry already destroyed the Roman sanctuary on top of the Pfaffenberg. The economic and housing development of the modern villages, which are located in the archaeological zone, pose another threat to the cultural heritage. To be able to address this destruction, the archaeologist has to know the exact location and extent of the threatened structures. During the second half of the 19th century and the ﬁrst decades of the 20th century, knowledge about the layout of the antique city was gained by large-scale excavations. Between 1877 and 1917, more ruins were uncovered than during the period since then1. Today, the archaeologists are aware of the fact, that excavation, even if it is done scientiﬁcally, is yet another form of destruction. Therefore, the process of excavation and documentation has to be carried out with a high degree of care and meticulousness. In Carnuntum, archaeological remains spread over an area of a few square kilometres. The time and expense needed to excavate areas that large, using up - to - date methods, cannot be funded. Above all, 150 years of excavation did not provide a clear picture of the layout of Carnuntum. Fortunately, the last decades saw an increased de-
* Universität Wien, Institut für Ur- und Frühgeschichte (Austria). 1 JOBST 1983, p. 23.
velopment of non-destructive prospection methods. Among these, aerial archaeology and geophysical prospection are best suited to give detailed insight into the buried archaeology of large areas within short time and few expenses. Both methods are non-destructive and therefore repeatable. Only a systematically applied prospection strategy will be able to provide comprehensive knowledge of the location, extent, and state of preservation of the archaeology in Carnuntum and its hinterland – knowledge, which is highly desirable for monument preservation. The results can consequently be used to plan further excavation strategies to investigate eg. chronological aspects of the development of the settlement. The ﬁrst aerial archaeological photographs over Carnuntum date back almost 70 years to the late 1930s. However, a systematic aerial archaeological approach, which has to incorporate image interpretation and interpretative mapping in an annual ﬂying program is still lacking. In the 1970s for the ﬁrst time attempts were made to map the archaeological evidence from aerial photographs to some extent. Due to a lack of ﬁnancial means, the plans had to be abandoned a few years later. The matter stands similar with geophysical prospection. Due to constant (re)development of methods and instruments, it is possible to investigate large areas within a comparatively short time2. Although this could already be demonstrated several times in Carnuntum, a geophysical prospection strategy to ﬁnance the desirable measurement of the entire area is still lacking, too3. In 1997 the decision was made to create a city map of ancient Carnuntum based on aerial photographs. It can be seen as a ﬁrst step for a systematic prospection of the archaeological landscape of Carnuntum. Aerial photographs of the last 50 years are currently used to create a highly detailed map of the archaeological features of Carnuntum. At the same time, large-scale geophysical prospections are intensiﬁed. Measurements are being carried out over the most important and most threatened areas and in ﬁelds, where the results from aerial archaeology are unsatisfactory. The preparation of the city map is an ongoing internal joint project between the Institute for Prehistory of the University of Vienna and the Vienna Institute for Archaeological Science (VIAS). Additionally, in November 2003, a complementary project began. Its aim is to shed light on the settlement pattern of the late Iron Age and early Roman time in the hinterland of Carnuntum. This paper will focus on the aerial archaeological aspects of creating the city map of Carnuntum and interpretation of its hinterland. The applied methods and the results obtained so far will be demonstrated.
A Brief History of Carnuntum During antiquity, Carnuntum was situated on a strategically important location. Here, the so-called “Amber-Road”, a trade route connecting the Adriatic and the Baltic Sea, crossed the river Danube. The ﬁrst date for Carnuntum mentioned by the Roman writer Velleius Paterculus is the year 6 AD4. During a military offensive, in this year, General Tiberius ordered to establish a winter camp near Carnuntum5. Around 40/50 AD a fortiﬁed camp was built by the legio XV Apollinaris between today’s villages Petronell and Bad Deutsch-Altenburg. It was the core of the large Roman city, which was divided in two major parts
2 NEUBAUER 1990; NEUBAUER, MELICHAR, EDERHINTERLEITNER 1996. 3 KANDLER 1999; NEUBAUER, EDER-HINTERLEITNER 1997. 4 FRIESINGER, KRINZINGER 1997, GENSER 1986, JOBST 1983, HUMER 2004. 5 PATERCULUS, Historia Romana, II 109, 5.
Aerial archaeology in the landscape of Carnuntum 217 during the antique period: the military camp with the surrounding settlement (canabae legionis) and the so called “civil settlement”. the last building activities found so far date to the ﬁrst half of the 5th century AD. next to the recently discovered Forum7 the public baths and a public market hall (macellum) were excavated in 1939 and between 1957-1977 (“Palastruine”). Together with the suburban settlements south and west of the town’s wall the entire territory of the civil settlement might have been 300 hectares. http://www. Carnuntum became the capital of the province Pannonia superior. Parts of the civil settlement (2. DONEUS.26 ha) within a residential quarter were recovered and partly reconstructed (“Spaziergarten”). MELICHAR 1999.at. Septimius Severus was proclaimed Emperor on 9th of April 193 AD. the governor of Upper Pannonia. the heart of the town was fortiﬁed by a massive 2 m thick wall. capital of Pannonia.carnuntum. in the garden of Traun castle. EDER-HINTERLEITNER. In the 3rd or 4th century AD. The archaeological remains spread over today’s villages of Petronell (west) and Bad Deutsch-Altenburg (east). NEUBAUER. The civil settlement was raised to the status of a municipium – an autonomous Roman city – by Hadrian (117138 AD). a conference to solve the problem of succession was held in Carnuntum held by the former caesar Diocletian.6 ha of the town was partly excavated in 1892 and reﬁlled again. The existing map of this excavation can not yet be precisely located within the possible area of the “Tiergarten”.7 ha) have been excavated and are now open to the public as part of the Archaeological Park Carnuntum6 (see also ﬁg. 1). In the year 308 AD. The migration period brought Rome’s inﬂuence in Pannonia to an end. six insulae (about 1. In the military camp of Carnuntum. East of this area. Another 0. The top of the image points towards north. The captions refer to the text (chapter “Brief History of Carnuntum”).co. Some time between 103 and 107 AD. Thus the civil settlement received a new impetus and became Colonia Septimia Aurelia Antoniniana Karnuntum. In Carnuntum. SEREN. 1 – Relief map of the eastern part of Austria and aerial photograph covering the area of Carnuntum. 7 6 . 1 Fig. In the heart of the civilian settlement.
a sanctuary had been built on top of the Pfaffenberg. The only preserved building of the canabae legionis is the military amphitheatre northeast of the legionary camp (“Amphitheater I”). East of Carnuntum. 40. which forms the valley for a streamlet. it was reopened in autumn 20018. p. erected by Emperor Constantius II. the “Hundsheimer Berge” deﬁne the eastern border of the ancient city. It had a capacity of 8000 visitors. 172 AD against the Quads9. and working areas. After restoration. Carnuntum is always a worthwhile target. interpreted as forum. the river Danube cuts through foothills of the Carpathian Mountains (“Hundsheimer Berge”) forming the so called “Hainburger Pforte”. East of Carnuntum. On top of the terrace. FRIESINGER 1987. The Landscape of Carnuntum The landscape of Carnuntum is characterised by the river Danube (ﬁg. p. It is said that it was built in wood in the ﬁrst half of the 1st century AD and rebuilt in stone after the Marcomannic wars (167-180 AD). At a height of 480 m. 55. The river’s ﬂoodplain forest covers mainly the area north of the Danube. which lies some 180 m above sea level. The Roman remains extend over an area of 5 km2. It was a 15 m tall quadrifrons. In the village of Bad Deutsch-Altenburg a thermal spa is built around sulphurous hot springs. outside the town wall an amphitheatre was excavated in 1920s. The topography is ﬂat to slightly hilly. 10 BUNDESMINISTERIUM FÜR LANDWIRTSCHAFT 1980. The legionary camp with its surrounding canabae legionis was also subject to several excavations. 11 THENIUS 1974.218 MICHAEL DONEUS South of the municipium. the ﬂoodplain is limited by the 30-45 m high and steep slopes of the wide gravel terraces reaching some 20 kilometres to the west10. The climate is warm and dry. The area of the camp and its fortiﬁcations can still be seen from topography. Most of them are reﬁlled and not visible to the public. a monumental bath. It was probably erected in the 2nd century AD and had a capacity of 13000 spectators (“Amphitheater II”). the famous “Heidentor” has survived from the Roman times. within the municipalities of Petronell and Bad Deutsch-Altenburg. of which 4 km2 are undeveloped agricultural landscape. Today. For the aerial archaeologist. an auxiliary camp is being excavated. Large parts of the legionary camp were excavated between 1877 and 1914 and 1968-1977. Today’s village of Bad DeutschAltenburg is situated in a hollow. Between the canabae legionis and the civil settlement. the sanctuary has been completely destroyed by a stone quarry. the remains of the ancient city of Carnuntum are buried. This valley can be regarded as the potential route of the “AmberRoad”11. In the south. It is situated at the eastern border of Petronell and was used by the cavalry (around 500 horsemen) between the end of the 1st and the 3rd century AD. South of the amphitheatre. 9 8 UND FORST- . whose springs are only 5 km south in the area of the so called Brucker Pforte. The northern border of the ancient city of Carnuntum is deﬁned by the steep slope towards the Danube. 2). Various other excavations revealed a large space. p. It was devoted to Jupiter after he had miraculously helped a Roman legion on the 11th June. 177. The high JOBST 2002. They consist of chalk and were therefore used repeatedly as stone quarries. two holy districts.
2 – Visualization of the digital terrain model. © Fliegerhorst Langenlebarn. Fig. 2 3 . 3 – Enlarged clipping of a vertical photograph from June 1968. View from southwest. Parts of the canabae legionis can be seen in high detail.Aerial archaeology in the landscape of Carnuntum 219 Fig. The grey areas give a rough idea of the Roman structures.
220 MICHAEL DONEUS Fig. Between the roads several buildings can be discerned. The main road. is leading towards the amphitheatre. On both sides smaller roads branch off in right angles. They bend away in a rounded corner. 4 – Oblique photograph from June 1998. The ﬁeld in front of the military amphitheatre I is full of crop marked archaeological features. In the background a rescue excavation is carried out by members of the ÖAI (Austrian Archaeological Institute) in-between recently built houses. . In the foreground. 4 5 Fig. 5 – Oblique photograph from June 1998. two parallel ditches of the auxiliary camp can be seen as positive crop marks. interrupted by several drains.
we have several vertical coverages of Carnuntum from various years and seasons giving us a very good overall view on topography. One of these pictures. 49-51. the aerial archaeologist is able to choose the parameters of the ﬂight and during the reconnaissance he has the possibility to decide upon which sites to photograph. 55. NEUBAUER. The major part of the aerial photographs date from the years since 1960. when the area was perennially frozen and the active layer at the surface was subject to regular seasonal thaw followed by re-freezing (cryoturbation). Today. the buried walls are covered only by a thin layer of loamy-sandy soil. NEUBAUER 2001. used his contacts in the air force to get aerial photographs from this area. shows the Roman military camp. a former member of the ﬂying corps during World War I. made on the 28th April 1939. which dries out quickly. landscape changes and archaeology in the whole area. Both vertical and oblique photographs are used15. the buried walls and roads of Carnuntum show relatively clear as moisture. In most areas of Carnuntum. Aerial Archaeology over the Roman city of Carnuntum Available Photographs and Data The ﬁrst aerial photographs made over the area of Carnuntum date back to the 1930s.UND FORSTWIRTSCHAFT 1980. Today the Aerial Archive of the Institute for Prehistory. p. even in years with high rainfall. 13. 14 DONEUS. The vertical ﬁlms are produced by the Austrian Air Force at Langenlebarn.and crop marks. Abb. Hence. The road system and ground plans of hundreds of buildings in the canabae legionis can be clearly identiﬁed. Therefore. the Institute for Prehistory has permission to use all of the photographs produced in Langenlebarn for scientiﬁc purposes. Swoboda. SCHARRER 2000. At that time. pp. when the crops reveal the archaeological structures with ﬁne detail. Another important data source besides vertical photographs is oblique photographs. Due to a contract between the Ministry of Science and the Ministry of Defence. Fig. Both the enclosure walls and the system of the inner buildings can be seen clearly13. 15 DONEUS 1996. when systematic aerial archaeological prospection in Austria started14. Calibrated medium format cameras (Hasselblad) are used with black and white and 12 BUNDESMINISTERIUM FÜR LAND. 1. 3 shows an enlarged clipping of a vertical photograph from June 1968. 10). the best time for the ﬂights can be chosen. Here. Vienna. The scales range from 1: 8000 to 1:15000. loess and loam can be recognised on air photographs as well as in magnetogrammes and make archaeological interpretation difﬁcult (see also ﬁg. reveal plenty of archaeological features. p. results frequently in long periods of drought12. During the ﬂight the crop-marked archaeological features are photographed from the most suitable position and angle of view. p. the gravel shows fossil patterns caused by geological processes that occurred during the period of glaciation. The photographs are exposed using a Zeiss RMK (“Reihenmesskammer”) both on black and white and infrared falsecolour material.Aerial archaeology in the landscape of Carnuntum 221 potential evapotranspiration which is increased by frequent winds. These are made using high-wing airplanes (Cessna 150 or 172). E. EDER-HINTERLEITNER. . 13 DONEUS. 124. The resulting patterned ground and frost cracks ﬁlled with sand. Vertical aerial photographs have a high resolution and when made at ideal times and conditions. has several hundred photographs of the area of Carnuntum. Therefore.
The main road. as well as digital SLR cameras. In the background a rescue excavation is carried out by members of the ÖAI (Austrian Archaeological Institute) in between recently built houses. a lot of additional data sources are constantly used. Fig. It is interrupted by several channels that carried the waste water towards the Danube. Within this project. They bend away in a rounded corner. This results in high quality documents which make important contributions to the interpretation. 4. The third example shows part of the burial road leading away from the legionary camp towards the village of Bruck an der Leitha. 5 from the same ﬂight depicts parts of the auxiliary camp next to the village of Petronell. The following three examples of oblique aerial photographs should give a small impression of the variety of buried structures in the area. between 1978 and 1984. Next to the road ground plans of various burial buildings are visible as negative crop marks (ﬁg. In the foreground. 6 – Oblique photograph from June 2000 showing a small section of the Roman burial road accompanied by burial buildings. coming from the legionary camp touching the southern edge of the amphitheatre leads further to today’s village Bad Deutsch-Altenburg. Two decades ago already. 6).222 MICHAEL DONEUS Fig. depicts the area west of the military amphitheatre I. a project was . two parallel ditches of the camp can be seen as positive crop marks. Fig. photographed in June 1998. 6 colour slide ﬁlms. The area in between the side roads is ﬁlled with negative crop marks that indicate ground plans of buildings of the former canabae legionis. south of Carnuntum.
Rectiﬁcation of vertical and oblique photographs. KAGER. 18 See also JOBST 1983. there are plans to combine all available excavation maps with our interpretation results.Interpretation of orthophotographs. rectiﬁed. NEUBAUER.Measuring a digital terrain model and digitizing the cadastral map. EDER-HINTERLEITNER 1997. . where aerial photographs of smaller areas within Carnuntum are oriented. As a standard procedure. Mapping can. WALDHÄUSL 1983. In a further step. mapping is especially desirable in countries like Austria. EDER-HINTERLEITNER. VOZIKIS. Unfortunately. SEREN. JOBST. The single steps cannot be seen as a rigid working process. . Among these. DONEUS. In Carnuntum. 17 NEUBAUER. dates and scales can be compiled as maps. Consequently.Orientation of vertical photographs used for the digital terrain model. mapping all of the photographs is like bringing pieces of a puzzle together and the only way. this project was not continued. to realise distribution patterns of the features. The procedure of mapping Roman Carnuntum can be divided into several steps: . KANDLER. and pedological maps. the resulting orthophotographs are still available. both amphitheatres. the thermal bath. the aerial photographs have to be mapped. which differ from year to year. which otherwise could not be seen from a single oblique or vertical photograph. geological. It was a joint venture between the Austrian Archaeological Institute and the Institute for Photogrammetry and Remote Sensing of the Technical University of Vienna and can be seen as a predecessor of our efforts. Besides the aerial photographs. The idea was to create a city map at the scale of 1:2000. HAITZMANN. Mapping strategy To understand landscapes and to combine the evidence that may be recorded in hundreds of archived photographs. and parts of the civil city are most useful18.e. MELICHAR 1999. resistivity and ground penetrating radar leads to highly detailed 2-dimensional and 3-dimensional images of the buried structures17. Evidence from photographs taken at different times. the results are combined with the rectiﬁed aerial photographs in GIS. However. which usually have different crops. The detailed geophysical images give additional information which help get the maximum of information from the aerial photograph. 16 SCHEDIVY 1986. Most important are the cadastral maps of the modern villages Petronell and Bad Deutsch-Altenburg with scales of 1:2000 and 1:1000. . geophysical prospection methods are applied in the area of Carnuntum by ArcheoProspections® in close co-operation with the Aerial Archive. and depiction of interpreted detail) as are appropriate to study different questions. different kinds of maps are available.Aerial archaeology in the landscape of Carnuntum 223 set up to rectify aerial photographs from the area of Carnuntum16. where small and narrow ﬁelds. the forum. The application of magnetics. high-precision mapping was desirable to be able to link aerial archaeological evidence with results from excavation and geophysical measurements. ranges of precision. Mapping of aerial photographs rather has to be seen as an iterative process. More than a hundred years of excavation have resulted in numerous maps of excavated features. . be undertaken at different levels (i. at scales. . Many sites are visible only in small parts.Combination of the results with other data. however. They show parts of the Canabae and the military camp and are also incorporated into the ongoing project. are very common. the Roman military camp. Since 1990. Therefore.
For the measurements a total of 45 vertical aerial stereo 19 DONEUS 1994. DONEUS 1996. positions and real world co-ordinates can be calculated from oriented images. Right from the beginning. a block of 46 vertical photographs covering the whole area with an overlap of 60% was available. In that way. So far. which are photographed afterwards can be included rapidly and often give new clues and details to the already interpreted area. each of the 5 steps has been realised to a different degree. excavation maps. we decided to measure them ourselves. or older orthophotographs is consequently realized. topography of the terrain. where also a combination with all other data sources. Sometimes. Depending on the photographic material (vertical stereo pairs. it is possible to mathematically simulate the reverse light ray projection from the image through the lens. through a lens. different methods can be applied19. The ﬁrst step of the project was to obtain a digital terrain model (DTM) to be used later for the rectiﬁcation of the aerial photographs. For each of these parts. Orientation of photographs During the process of orientation. The other steps have been fulﬁlled to a degree of about 10 percent. the exact spatial position and viewing angles of a photograph at the time of exposure is mathematically reconstructed. To get good results within a short time. . The block was manufactured in 1976 at a scale of approximately 1:5000 by the Bundesamt für Eich. the area was divided into smaller parts. the DTM would have to be a fairly accurate representation of Carnuntum’s topography. Therefore. and desired quality of the results. they even make a re-interpretation of the whole area desirable. To be able to also rectify oblique photographs within a sufﬁcient error tolerance. Interpretation is done using the orthophotographs in a geographical information system (GIS). onto the image plane in the camera. single oblique photographs). available geophysical data or excavation results are used to verify and reﬁne the interpretation results. Aerial photographs. both analytical and digital photogrammetry is used for obtaining both a digital terrain model and the rectiﬁcation of the aerial photographs. Each photograph is the result of light rays reﬂected from the terrain. the work can be done only sporadically. Methods Used The project started in autumn 1997. Since there is no additional funding. the whole block was already oriented via aerotriangulation. done by the Institute for Photogrammetry and Remote Sensing of the Technical University of Vienna.und Vermessungswesen on behalf of the Institute for Photogrammetry and Remote Sensing. The ﬁrst two points are almost ﬁnished. Knowing the orientation of a photograph.224 MICHAEL DONEUS and interpreted. Additionally. To meet the broad spectrum of the different data sources. the best aerial photographs were chosen for mapping. such as geophysical prospection results. From the former mapping project from the 1970s mentioned above. instead of purchasing the terrain data.
In most cases secondary ground control. DONEUS. A total of 60 land registery maps at a scale of 1:2000 covering an area of 4250 hectares of today’s villages Petronell and Bad Deutsch-Altenburg were digitized and geo-referenced using a CAD program. 6. the accuracy lies between 0. scale. Additionally. To allow a flexible presentation. which were used for the calculation of the orientation values. vegetation. SCHARRER 2000. cadastral borders.5 m in plan and height. The stereo models were set up on our analytical plotter DSR 14 using the initial values of the aerotriangulation. an accurate digital terrain model is needed. The block was produced by the “Kommando Luftaufklärung des Fliegerhorstes in Langenlebarn” in June 1998 at a scale of 1:12000. Although a digital photogrammetrical station (SoftplotterTM on a Silicon Graphics O2 workstation) was available. The orientation of oblique aerial photographs is usually calculated using space resection (for single images) or bundle adjustment 21.25 m in height. whereas the automated extraction using digital correlation techniques would measure the DTM on top of crops. 21 DONEUS 1996. we decided to measure the digital terrain model “manually” using the analytical plotter DSR14. the procedure of aerotriangulation was performed in a digital environment (SoftplotterTM 3. This procedure is done digitally using the software ERDAS OrthobaseTM 8. 20 See also. several other vertical stereo pairs were oriented using up to 10 control points. buildings. p. Additionally. 63. .6. and measurement of the ground control. 3D-breaklines were extracted. approximately 127000 points were measured.1)20. In this case. which is derived from the previously oriented verticals. and annotations were placed on different layers. NEUBAUER. To verify the result and to identify landscape changes since the 1970s.Aerial archaeology in the landscape of Carnuntum 225 pairs were analysed. the vectorized cadastre is put on top of the rectified images to immediately estimate any geometric errors in the orthophotos. The ground control points. In some areas. trees and houses.30 m in plan and 0.25 and 1.5 m. The reason was that we wanted to obtain the DTM at ground level. quality of the distribution. To be able to visualize these still “prominent“ features. roads. ancient features such as Roman roads or the legionary camp and its fortifications have not been completely flattened by the plough and therefore can be seen in the topography. Abb. lots. Due to the smaller scale. Digital cadastre and digital terrain model The digital cadastre is used to identify single fields from the interpretation map. Additionally. the accuracy was approximately 0. Depending on the used camera. Altogether. That way 3D-points of a surface covering approximately 2434 hectares (also the surrounding topography of Carnuntum was considered) were recorded from the vertical photographs using a 30 m grid. coordinates. ancient ruins. The average model accuracy is approximately 0. is used. were measured in the field using a total station. The model is also a prerequisite during the rectification of the aerial photographs. another block of vertical photographs was oriented.
The DTM was measured at ground level. During the interpretation process. single features can be automatically colour-coded according to the entries in the database. which can be individually deﬁned. In that way. visible marks. or orientation can be derived automatically. and number of the interpreted photograph will be the initial input during an interpretation. Using a vertical or horizontal swipe line. each feature from the drawing is given a meaning. 22 SCOLLAR. However. using the outer orientation values and the digital terrain model. Oblique slides with a format of 6 by 6 cm are scanned using our Polaroid SprintScan 45 with a resolution of 2000 DPI or 12 µm. To exclude this error. WEIDNER. The interpretation is done image by image on screen in separate layers using different colours and attributes for different features. . which is a geo-referenced photomap. and legends can be automatically added. HUANG 1984. The resulting orthophotographs usually have a pixel size of 0. the archaeological features are usually visible in the crop which is somewhat higher than the ground. The photographs are scanned with a pixel size of 15 µm (black and white negatives) or 30 µm (infrared falsecolour). On the resulting interpretation map. This is very useful if one wants to look “behind“ the interpretation drawing to see whether or not the features drawn from one orthophoto are also visible in other images. the resulting orthophotograph would have an error in the viewing direction. Wallis ﬁlter. At the beginning.2 meter. The ﬁle format of the orthophotographs can be directly read by the GIS. The result is an orthophoto. 7). Usually what kind of features. or crispening22. the orthophotographs are treated with contrast stretch. It is also a very good way to quickly estimate the geometrical accuracy of both images. each image is rectiﬁed.226 MICHAEL DONEUS Rectiﬁcation of vertical and oblique photographs Vertical and oblique photographs are digitally rectiﬁed using ERDAS OrthobaseTM 8. each drawing layer has an attached attribute table. as size and area of the object. In GIS. soil type. an adequate value is added to the terrain data depending on the height of the photographed crops. Additionally. During the process of drawing the interpretation. All of the georeferenced orthophotographs and their ﬁltered variations are compiled in a layer stack of the viewer. Using image enhancement techniques. Today. landscape changes are identiﬁed easily using this tool. only portions of the top and bottom image can be viewed. Other attributes. the scanning was done on a PS1 scanner provided by the Bundesamt für Eichund Vermessungswesen and the Technical University of Vienna. Therefore. depending on the obliqueness of the photograph’s viewing direction. the orthophotos are enhanced to make the archaeological features better visible. They are stored with a corresponding “world ﬁle” containing the georeferenced information for further use in geographic information systems. SCOLLAR 1990.6. descriptive values are added in the corresponding database(s) (ﬁg. Interpretative mapping of orthophotographs The interpretation is done using GIS. Finally. a Vexcel Ultra Scan 5000 is used.
. 24 DONEUS. Every added orthophoto may show new details and therefore has to be checked with the temporary interpretation result. The resulting magnetogram is a geo-referenced digital image. On the other hand. a prerequisite that is already fulﬁlled with our data. If necessary. They are ﬁxed in a Plexiglas tube for gradiometer array measuring the difference between the total and the local magnetic ﬁeld.Aerial archaeology in the landscape of Carnuntum 227 Fig. Combination of the results with geophysical prospection and excavation maps The magnetic scanning system developed and used in Austria by ArcheoProspections® is mounted on two completely non-magnetic wooden handcarts. the forum. MELICHAR. The aerial overview facilitates the estimation of the geophysical details (ﬁg. EDER-HINTERLEITNER 1996. The main cart carries the two caesium sensors. geomagnetic images and vectors. The excavation results – mainly the military camp. the detailed geophysical images are a very important source during the aerial archaeological interpretation. Lower Austria. The combination of the orthophoto and vectorized data as well as the combination of the orthophoto and geomagnetic results is again done using GIS. the interpretation map will reveal a more and more detailed image of Roman Carnuntum. The achievable accuracy lies at about 10 cm as could be veriﬁed by excavations23. Each feature is automatically colour-coded according to the entries in the database. they have to be set up in a uniform system of coordinates. Results are always provisional. 8). NEUBAUER 1998. 7 Interpretative mapping is the most time-consuming process. already used orthophotos have to be re-interpreted. To be able to overlay orthophotos. 7 – Detailed interpretation results of orthophotographs from a false-colour vertical and an oblique aerial photograph in the area of the canabae legionis of Carnuntum. They give additional information which helps retrieve the maximum of information for a photograph24. the two amphitheatres. Step by step. and parts of the civil city – were scanned from 23 NEUBAUER.
If the maps had no coordinates.228 MICHAEL DONEUS Fig. Fig. 8 9 publications. 8 – View of the civil settlement based on an orthophotograph with excavated insulae and the monumental bath. control information was taken from the digital cadastre or measured from the aerial photographs. 9 – a) Aerial archaeological interpretation of the area around the legionary camp. magnetic survey. b) Aerial archaeological interpretation of the area of the civil settlement of Carnuntum. Geophysical data: ArcheoProspections®. imported into the GIS. resistivity mapping and GPR images and aerial archaeological interpretation. . and geo-referenced.
E. They could belong to the large building. because both had been already excavated. Northwest of the amphitheatre. At least one of the roads crosses the town wall leading to the large graveyard south of the civil settlement. because the area is used as a paddock. could be mapped. map 3. it ﬁts the system proposed by E. The area within 150 m of the legionary camp is almost free of archaeological features. SCHEDIVY 1986. the composite map already shows a considerable degree of detail (ﬁg. This very rough descrip- 25 26 SCHEDIVY 1986. archaeological structures belonging to the auxiliary camp. a large graveyard with hundreds of burial monuments is spreading over an area of at least 10 hectares. Civil Settlement In the area of the civil settlement 40 hectares could be mapped so far (ﬁg. p. lining up with the Via Decumana. Somewhere close to the west entrance of the camp. however. Although only about 10 percent of the available photographs have been rectiﬁed and interpreted. 9). 9a). The town was enclosed by a large wall with a ditch in front. BECKEL 1973. as to be expected27. where one of the gates can be assumed. With a few exceptions. forum. In the area in front of the Porta Decumana the layout of the road network of the canabae legionis is almost rectangular in shape. 111. Canabae Legionis Only a few photographs covering the area between Petronell and Bad Deutsch-Altenburg could be rectiﬁed so far (ﬁg. aerial photographs showing archaeological features in an area of 200 hectares have been mapped. where the cavalry was situated. 98.Aerial archaeology in the landscape of Carnuntum 229 Preliminary Results So far. The layout of a few hundred houses with inner structures and sunken ﬂoors can be identiﬁed in detail. Swoboda found in 194126. the burial road diverts at an angle of 15 degrees from the Via Principalis and leads straight outside towards southwest. Both can be seen clearly. The crossing of this road with the town wall. walls of a large building complex are visible. 27 VORBECK. The layout of the area southeast of the camp seems to be deﬁned by topography. Some of the buildings intersect with burial monuments and are therefore not contemporary. The visibility of the building structures within the road network is far from perfect. South of the town. Northeast of the amphitheatre. . The areas of the legionary camp and the military amphitheatre were not mapped. It can be followed 13 kilometres until it reaches today’s village Bruck an der Leitha. The camp is already partly destroyed by the expansion of the village Petronell. Schedivy25. p. that the archaeological structures spread over a much larger area. Nevertheless. It is not intended to give a full description of the mapped features which would be beyond the scope of this article. is covered by a modern road. Within the civil settlement the road network is very well identiﬁable. 9b). burial road. and large parts of the canabae legionis spreading over 160 hectares could be mapped. a large building complex is situated. Further west several ditches belonging to different phases of the auxiliary camp. Aerial photographs from the surrounding area indicate.
Orientation. the layout of the buildings could be seen like never before. but the knowledge of its exact location was lost afterwards. The project is still at the stage of data collection. The resulting orthophotos function as a ground control source for the orientation and rectiﬁcation of the oblique and single vertical photographs from the other ﬂights. which mi- 28 29 NEUBAUER 1997. The rectiﬁcation is done using the specialized Software Airphoto29. The following two examples of oblique aerial photographs should give an impression of what can be expected from this area. One of the vertical aerial blocks of photographs covering the total project area is currently being oriented and rectiﬁed using aerotriangulation. It was rediscovered by aerial archaeology in 1993. the data will be the basis for the assessment of settlement patterns and social. 11 shows the Roman villa rustica of Arbesthal. which was programmed to aid aerial archaeologists with the rectiﬁcation of oblique photographs. size. 12 shows an example of the already interpreted archaeological landscape around the well known Roman Villa of Bruckneudorf30. The team consisting of 2 full-time and 3 part-time researchers tries to meet two project aims. systematic aerial archaeology and ﬁeld surveys. In the course of this project. it would be the ﬁrst known archaeological site from this time in this area. Fig. Neustadt and Bruck an der Leitha. Even at this early stage of the project we have already found more than 200 archaeological sites. SCOLLAR 1998. ZABEHLICKY 1996. In November 2003.230 MICHAEL DONEUS tion gives an impression of the great variety of structures and the high degree of detail possible. a research project ﬁnanced by the Austrian Science Fund was launched by the author which deals with the settlement pattern of the late Iron Age and Early Roman times in the hinterland south of Carnuntum (“Celts in the Hinterland of Carnuntum”). If this hypothesis proves true. systematic data collection of all sites will be carried out in a 10 km wide corridor along the river Leitha between Wr. we want to produce a detailed archaeological map of this 600 km2 large region derived from literature. and layout of the graves suggest that the burials could be Avar. 10 shows one of two newly discovered graveyards in the vicinity of Roman Carnuntum. Comparatively little is known about the hinterland. 30 LANGMANN 1979. Aerial Archaeology in the Hinterland of Carnuntum 150 years of archaeological research have been uncovering several hectares of archaeological structures within the civil settlement and the canabae legionis of Carnuntum. Fig. Secondly. at least one block of vertical infrared photographs is produced and several reconnaissance ﬂights with documentation on oblique photographs are carried out. and geographic conditions which were substantial for the Romanization of the Carnuntum area. Due to the extremely dry weather conditions in 2003. Because of the large area and the limited time. . A small part of this villa was already excavated in 1902 by M. economic. Fig. which are already partially published28. Each year. All of the relevant aerial photographs will be rectiﬁed. Even the aerial archaeologist is attracted by the detailed and widespread crop marks indicating the buried archaeology of Carnuntum and tends to forget to also ﬂy beyond the densely settled area. Firstly. Even more details can be seen in the results of the geophysical prospection. a slightly different mapping approach is used than that described above. von Groller.
A small part of this villa was already excavated in 1902 by M. It was rediscovered by aerial archaeology in 1993. The background pattern is caused by geological crop marks (frost cracks). Orientation. 10 11 . 11 – Roman villa rustica of Arbesthal. and layout of the graves suggest that the burials could be Avar. Groller. but the knowledge of its exact location was lost afterwards. 10 – One of two newly discovered graveyards in the vicinity of Roman Carnuntum. size. Fig.Aerial archaeology in the landscape of Carnuntum 231 Fig.
The map shows the large Roman villa. In the hinterland. Such a prospection seems to be urgently necessary from the scientiﬁc as well as from the development-planning point of view and could give enormous input to economic strategies by bundled and focused actions for the largest archaeological zone in Austria. various graveyards and a Hallstatt Period settlement. Interpretative mapping of comparatively few photographs resulted in a detailed city map of both the civil settlement and canabae legionis. 12 – Interpretation of the archaeological landscape around the Roman Villa of Bruckneudorf. In the near future we will concentrate our work on the integration of the remaining aerial photographs and the combination of the results with geophysical prospection. Conclusion The results stated in this paper show the potential of a systematic aerial archaeological approach in the area of Carnuntum. . And we will continue ﬂying in the area of Carnuntum. various graveyards and a Hallstatt Period settlement.232 MICHAEL DONEUS ght have been the residence of the Imperial family in the autumn of 375 AD. Experience resulting from case studies and the developed evaluation and interpretation methods give the opportunity to plan a speciﬁc strategy for an overall prospection of Carnuntum. new features are still to be found in the centre as well 12 Fig. more than 200 archaeological sites were found so far on vertical aerial photographs and during reconnaissance ﬂights. Even after 70 years of aerial photography. The map shows the 7 hectare large Roman villa. We hope that the project will establish a new basis for a better comprehension of the historical processes in the Danube-Leitha area. which might have been the residence of the Imperial family in the autumn of 375 AD.
: P16449-GO2. 1986. KRINZINGER. Wien 2004. The author wants to thank him and Peter Waldhäusl (Technical University of Vienna) for making the materials (vertical ﬁlm and aerotriangulation data) and the rectiﬁed photographs of the former mapping project available.000. G. HUMER. The author is also very grateful to Ralph Totschnig (Institute of Pre. 11-34. Erläuterungen zur Bodenkarte 1:25. 8. DONEUS.UND FORSTWIRTSCHAFT 1980: BUNDESMINISTERIUM FÜR LAND. 1. Archäologische Prospektion in Österreich. NEUBAUER.Aerial archaeology in the landscape of Carnuntum 233 as in the hinterland which help to complete the puzzle. REFERENCES BUNDESMINISTERIUM FÜR LAND. 5. W. Archäologische Prospektion der Landschaft Carnuntum. 1998. NEUBAUER 1997: M. The Roman City Quarter in the Open Air Museum Petronell. KRINZINGER 1997: H. FRIESINGER. in Archäologie Österreichs. A.). 1994. EDER-HINTERLEITNER. DONEUS. DONEUS. 5. EDER-HINTERLEITNER. Heft 33. Photogrammetrical Applications to Aerial Archaeology at the Institute for Prehistory of the University of Vienna. Der Römische Limes in Österreich.) Archaeological Prospection 2001. Austrian Academy of Sciences Press 2001. NEUBAUER. Wien 1997. 29-56.. Acknowledgements The project was initiated by Manfred Kandler (Austrian Archaeological Institute). 124-129. Der Römische Limes in Österreich. DONEUS 1994: M. FRIESINGER 1987: H. A. SCHARRER 2000: M. W.4/01). in Carnuntum Jb. 2D Combination of Prospection Data. NEUBAUER.086/5-1. 76-87.and Early History) and Christian Gugl (Austrian Academy of Sciences) for fruitful discussion and review of the manuscript. NEUBAUER (eds. . Wien. NEUBAUER. 19-33. Archaeological Prospection in Austria. Vol. in M. Niederösterreich. Kulturabteilung des Landes Niederösterreich and Archäologischer Park Carnuntum. SCHARRER. DONEUS. Part B5. pp. GENSER 1986: K. 1997. which resulted in a patchwork of excavations. We hope that after more than 150 years of archaeological investigations in Carnuntum. Österreichischen Akademie der Wissenschaften. Austria. DONEUS. Möglichkeiten der Luftbildarchäologie. NEUBAUER 2001: M. DONEUS. 13.und Frühgeschichte in Wien. pp. we can summarize all the available information and produce the ﬁrst comprehensive map of the ancient city. GENSER. Wien 1980. DONEUS 1996: M. 2000. in Archaeological Prospection. DONEUS. pp. Wien 1987. 53-72. W. FRIESINGER. pp. NEUBAUER. F. Commission V. pp. Kartierungsbereich Hainburg an der Donau. in Archäologie Österreichs. HUMER 2004: F. FRIESINGER. XXXI. NEUBAUER 1997: M. Der Österreichische Donaulimes in der Römerzeit – ein Forschungsbericht. DONEUS. Luftbildarchäologie am Institut für Ur. W. grant no. DONEUS. All aerial photographs are released by the Ministry of Defence (No. The DTM was measured by Gabriele Scharrer (VIAS). in International Archives of Photogrammetry and Remote Sensing. 4th International Conference on Archaeological Prospection. DONEUS. EDERHINTERLEITNER. DONEUS. pp.UND FORSTWIRTSCHAFT (Hrsg. The project “Celts in the Hinterland of Carnuntum” is supported by the Austrian Science Fund. Die vielen Väter Österreichs. Vienna 1996. W.
2. 120. JOBST. 4. JOBST. H. THENIUS 1974: E. SCOLLAR 1990: I. NEUBAUER. pp. 66 ff. Image Enhancement Using the Median and the Interquartile Distance. pp. 121-129. VOZIKIS.234 MICHAEL DONEUS JOBST 1983: W. NEUBAUER 1997: W. LANGMANN. H. Resistivity and magnetics of the Roman town Carnuntum. P. 37-38. pp. erweiterte Auﬂage. Graphics and Image Processing. Collection. Österreichs größte archäologische Landschaft. JOBST. MELICHAR. JOBST 2002: W. in Carnuntum Jb. 179-189. Photogrammetrie in der Architektur und Denkmalpﬂege.Rom an der Donau. 28. Plan der Zivilstadt von Carnuntum. Bericht über die Grabungskampagnen 1975-1979 in Bruckneudorf. 3. TAEUBER. Bez. und 100 ff. Provinzhauptstadt Carnuntum. Bundesdenkmalamt Wien 1983. Numerical treatment of archaeological air photos: a non-mathematical introduction. S. Computer Applications and Quantitative Methods in Archaeology. AirPhoto – a WinNT/Win95 program for geometric processing of archaeological air photos. Geophysikalische Prospektion in der Archäologie. EDER-HINTERLEITNER 1997: W. 1997. WEIDNER. . EDER-HINTERLEITNER. Ballonaufnahmen. visualization and simulation of magnetic prospection data. 41. DONEUS. Himberg 1983. in P. 179-189. Archäologie und Photogrammetrie. 1990. L. in Computer Vision. pp. WEIDNER. p. Kombination archäologischgeophysikalischer Prospektionsmethoden am Beispiel der römischen Zivilstadt Carnuntum. HAITZMANN. P. pp. Festschrift für Dieter Knibbe zum 65. 65. Pläne und Orthophotos von Carnuntum und des Kultbezirkes auf dem Pfaffenberg. NEUBAUER. SCHEDIVY. pp. H. in Archaeological Prospection. in Analecta Praehistorica Leidensia. MELICHAR 1999: W. in Öjh. SCOLLAR. KAGER. WALDHÄUSL 1983: W. 16. 1999.). ZAHBELICKY 1996: H. M. in MAnthrWien. 111-118. Geburtstag. A. Neusiedl am See. in AARGnews. KANDLER. Grabungen 1995. Resistivity and magnetics of the Roman town Carnuntum. 4. HUANG. NEUBAUER. MELICHAR. Ein Führer. KANDLER. Das Heidentor von Carnuntum. LANGMANN 1979: G. EDER-HINTERLEITNER. in GOTTFRIED BOEHM (ed). MELICHAR. Das Forum der Colonia Carnuntum. in Burgenländische Heimatblätter. A. NEUBAUER. BECKEL. E. 1999. Wien 2002. Bruckneudorf. 270-271. VORBECK. THENIUS. THÜR (Hrsg. JOBST. NEUBAUER. DONEUS. Salzburg 1973. T. Erste Ergebnisse von geophysikalischen Bodenprospektionen im Tiergarten des Schlosses Petronell. Carnuntum . pp. Austria: an example of combined interpretation of prospection data. 1997. SCOLLAR. HAITZMANN. in ÖAI Sonderschriften. NEUBAUER 1990: W. Geologie der Österreichischen Bundesländer in kurzgefassten Einzeldarstellungen: Niederösterreich. BECKEL 1973: E. 1979. H. pp. NEUBAUER. 1996. HUANG 1984: I. in AAustr. pp. KANDLER 1999: M. 359-368. EDERHINTERLEITNER. EDER-HINTERLEITNER. 201-256. SCOLLAR 1998: I. 1984. 1985. SCHEDIVY 1985: E. 24f. CAA95. NEUBAUER. Wien – geologische Bundesanstalt 1974. WALDHÄUSL. SEREN. Austria: an example of combined interpretation of prospection data. KAGER. 1-60. VORBECK. B. M. EDER-HINTERLEITNER 1996: W. in Archaeological Prospection. ZAHBELICKY. KANDLER. SCOLLAR. SCHERRER. 236-251. 25. NEUBAUER. Interfacing the Past. A. SCOLLAR. VOZIKIS. pp. 1996. 32. SEREN. in Studies in modern archaeology. Burgenland. P. Steine und Wege.
This action might not be possible to undo. Are you sure you want to continue?
We've moved you to where you read on your other device.
Get the full title to continue listening from where you left off, or restart the preview.