GIS-based Analysis of Large-scale Historical Maps and Archival Sources to Reconstruct Former Features of the Cultural Landscape Near Early Celtic "Princely Seats" in Southern Germany

Christof Schuppert

Andreas Dix

Abstract:

English

The article points out how an archaeological question can be approached by the use of historical-geographical methods with the help of a geographic information system. First results are presented on the basis of an example from the Heuneburg at the upper Danube, an early Celtic "Princely Seat" that represents one of the best-researched archaeological sites of this period (650 to 475 B.C.).

German

Der Artikel zeigt auf, wie eine archäologische Fragestellung durch den Einsatz der historisch-geographischen Methode der Archivrecherche unter Zuhilfenahme eines Geoinformationssystems bearbeitet werden kann. Erste Ergebnisse werden am Beispiel der Heuneburg an der oberen Donau vorgestellt, die eine der am besten untersuchten archäologischen Stätten des Zeitraums 650-475 v. Chr. darstellt.

Keywords

Historical geography, historical maps, archival research, cultural landscape, landscape history, GIS, georeferencing, princely seats

.01. Introduction

This research project is designed for interdisciplinary purposes. The methods applied originate from historical research. The actual question comes from archaeology. Moreover the topic represents a geographical area of research to which traditionally historical geography dedicates itself. The spatial aspect of the investigation becomes apparent in historical sources (maps, descriptive records) and is the reason for using a geographic information system.

.02. Historical Background

The central zone of the Celtic world originally was within the area of today's southwest Germany, the east of France and the northwest of Switzerland. During the late Hallstatt period (650 - 475 B.C.) the so-called "Fûrstensitze" (princely seats) emerged as a result of centralisation processes, which led to the development of proto-urban central places north of the Alps. [1] This happened approximately 500 years before the late Celtic Oppida-Culture, which created the first centers similar to cities today.

For a long time little was known about settlement structures in the late Hallstatt period, but in the last 30 years an extensive discussion started, caused by spectacular finds (e.g., the Heuneburg on the upper Danube, the Glauberg in the northeast part of Frankfurt/Main, and the Mont Lassois on the upper Seine). Latest finds from the Heuneburg (first stone gate on the northern side of the Alps) in Southern Germany, the Mont Lassois (first Mediterrean influenced, palace-like building in Central Europe) in Burgundy, and the realization that most of these sites were much bigger than expected confirm the enormous meaning of the early Celtic "Princely Seats" for Southern Central Europe during the early Iron Age. [2]

However, some questions are still unsettled: They concern the still unknown reasons for the centralisation processes. Political, economic or security factors are conceivable.

.03. The Research Problems

First, due to the evolving cultural landscape, which has changed the face of the land substantially since the beginning of industrialization, there are some historic facts that cannot be understood with today's archaeological methods of excavation or geophysical inspection. For this research cartographic sources can help reconstruct the landscape's history. And thus, through the historical reconstruction of the landscape, it is possible to recognize the effect to the Princely Seats created by the intensification of agriculture.

Second, in the past there has been a lack of appropriate tools for collecting and evaluating together both textual and spatial artefacts for historical arguments. Little evaluation has been done in this area of historic research.  [3] Using historic maps in conjunction with geographic information system is today possible, and considered to be the right tool for such investigations.

.04. Research Goals

The methodological goal is to apply the classic approach of historical geography (i.e., archival research) in connection with modern GIS technology regarding archaeological questions. By systematically developing and evaluating a new use of GIS technology, a contribution is to be made for research about the early Celtic "Princely Seats" and their surroundings.

In the context of the investigations two levels are to be considered. On the first investigation level, which covers the site and its surrounding field to approximately two kilometres, stands the identification of the remainder of presumably prehistoric structures (e.g., ramparts, ditches, burial mounds). On a second level, which is based on a regional scale, there is the analysis of traffic routes and occurrences of natural resources in a periphery of approximately 20 kilometres. Thus, in this two-step process, shifts in the cultural landscape are just as important as the identification of prehistoric structures.

.05. Methods

Archival Research

In order to discover suitable historical sources for the research project, extended archival research is necessary, since the critical texts are in many cases distributed over several archives due to former territorial splintering of southern Germany. Furthermore, written records from the 17th to 19th century are often difficult to read and need to be transliterated. [4]

Pre-Processing Historical Maps

First of all, the historical maps which should be used in a GIS must be scanned with high resolution. This task was carried out by the responsible institutions with the presence of appropriate technical equipment. The coloured scans were done with a resolution of 600 dpi. A higher resolution would have produced unnecessarily large image files (.TIF), which would have made a further partitioning of the individual sheets necessary, since the processing of large image files in a desktop GIS can cause several problems.

Georeferencing and Processing Historical Maps in GIS

Subsequently, the historical maps are loaded as raster files into a GIS. There they can be georeferenced by the use of reference points. [5] The presence of current geodata as a reference dataset (e.g., vector data like digital cadastral maps, raster data like digital topographic maps, or digital aerial views) is a condition for historic georeferencing. Finding suitable reference points can be difficult, especially if the respective historical map contains only few suitable points, which can be securely assumed as stationary in a spatiotemporal manner. These could be significant buildings or way crossings. The reference points should be evenly distributed over the mapped area. [6]

If enough reference points are selected, the transformation of the historical map can be calculated. The more reference points selected the smaller the risk of error becomes. By means of a transformation equation a re-organization of the data of the input picture can be accomplished into the matrix of the output screen. The assignment of the grey scale value can take place according to different rules. The applied "nearest neighbour" method stresses the smallest cost of computation. The grey scale value of the pixel is assigned, which is next to the computed coordinates. [7]

Georeferencing makes an overlay of the historical map with current geodata (raster or vector data) possible. Thus, changes of the land use can be mapped and visualized in the GIS. Currently missing landscape features such as ramparts, ditches, burial mounds, roads or bridges are mapped on the digital historical map as points, lines or areas, and put down in a GIS database (Geodatabase).

Integrating Information from written records into GIS

To these data extracted from large-scale historical maps, further information comes from written records. This information is mapped on the basis of current geodata. Therefore the spatial information contained in the records is used. In the act of georeferencing this written spatial information, several levels of accuracy are distinguished: land parcel accuracy, farm accuracy or rough accuracy (e.g., between two villages). Thus every cultural landscape element, which is mapped on the basis of written information, receives information about the spatial accuracy of the data. The database also contains a table classifying all used sources. This information about dating, author, point of origin as well as exact specification and reliability of a source can be linked with the particular culture landscape element as metadata.

Mapping of the Results

On the basis of these data land use maps of different temporal levels of the respective investigation area can be created. By historical blending (meaning a retrogressive investigation in order to describe the history of the cultural landscape) those maps make the mapping of changes in the culture landscape possible. [8] Tools in the GIS are used for visualization and cartographic representation of the results.

.06. First Results

The study starts with the Heuneburg at the upper Danube. This site takes a special position among the early Celtic "Princely Seats," since it represents one of the best-researched archaeological sites of this period. Today it still has an impressive appearance.

After intensive preparations, which covered the research history of the site and territorial development of the area, extensive archival research was tackled, which has not yet been finished. Within this research both cartographic and written records were discovered, which contain interesting information about the landscape history. The focus was first put on the direct environment of the Heuneburg. Figure 1 shows a cut-out from a geo-referenced cadastral map (1:2500) of 1824, on which the prehistoric fortification of the Heuneburg with gate openings still visible. These gate openings were leveled in the middle of the 19th century for gaining access to agricultural crop land and have to a large extent disappeared today.

On the basis of a late medieval text from 1390 the location of a former bridge in direct proximity to the Heuneburg was reconstructed. A geo-referenced cadastral map (1:2500) from 1822 and recent vector-based topographic data identify this bridge crossing the Danube which was situated in a different location than today. This bridge's ramps with a gap in between can still be seen today (Figure 2). This assumption is supported by the digital elevation model from airborne laserscanning, which was likewise included into the GIS analysis. These data clearly show an old Danube creek in the appropriate place.

.07. Conclusion & Perspective

The summarised representation of the preliminary results of archival research and their integration into a GIS are shown in the map draft in Figure 3, which represents missing cultural landscape elements within today's topography. The detailed interpretation is still pending, but the discussion of the past results with archaeologists showed that new explanatory models are necessary concerning the temporal classification of the levelling of the fortifications, route relocations, and relocations of old Danube creeks. Some results are to be verified by field survey.

The next step, besides the continuous build-up of the database (including still unconsidered field forms), is going to be an investigation on a regional level. In the case of the Heuneburg the research will be expanded to neighbouring hill settlements, which originate from the same period. In addition to this the investigation of further "Princely Seats" (Hohenasperg, Glauberg and Ipf) is scheduled.

.08. Notes

1. Kimmig, Wolfgang. Zum Problem späthallstättischer Adelssitze. Siedlung, Burg und Stadt. Studien zu ihren Anfängen. In: Festschrift Paul Grimm. Deutsche Akademie der Wissenschaften Berlin Schr. Sekt. Vor- u. Frûhgeschichte 25 (1969), 95-113.

2. Kurz, Siegfried. Zentralort und Umland: Untersuchungen zur Struktur der Heuneburg-Außensiedlung und zum Verhältnis der Heuneburg zu den um gebenden Höhensiedlungen. http://w210.ub.uni-tuebingen.de/volltexte/2005/2076/ [accessed 10/07/07].

3. Madry, Scott. The Integration of Historical Cartographic Data within the GIS Environment. In Steven N. and Kevin M. Bartoy (editors): 'Between Dirt and Discussion: Methods, Materials, and Interpretation in Historical Archaeology'. (New York 2006), 33-60.

4. Schenk, Winfried. Historische Geographie. In: Schenk, Winfried / Schliephake, Konrad (editors): Allgemeine Anthropogeographie. (Gotha, Stuttgart 2005), 220-230.

5. Gregory, Ian N. A Place in History. A Guide to using GIS in Historical Research (Oxford 2003).

6. Domaas, Stein Tage. Structural analyses of features in cultural landscapes based on historical cadastral maps and GIS. Acta Universitatis agriculturae Sueciae vol. 2005:100 (Uppsala 2005). http://diss-epsilon.slu.se/archive/00000900/ [accessed 10/07/07].

7. Plöger, Rolf. Inventarisation der Kulturlandschaft mit Hilfe von Geographischen Informationssystemen (GIS). Methodische Untersuchungen fûr historisch-geographische Forschungsaufgaben und fûr ein Kulturlandschaftskataster (Bonn 2003). http://hss.ulb.uni-bonn.de/diss_online/phil_fak/2003/ploeger_rolf/ploeger-abstract-dt.htm [accessed 10/07/07].

8. Jäger, Helmut. Historische Geographie (Braunschweig 1973).