Kortlægning af marksystemer fra jernalderen – En kildekritisk vurdering af luftfotografiers anvendelighed

Resumé

Mapping Iron Age field systems
An assessment of the applicability of aerial photography

There is little doubt that agriculture constituted the fundamental activity in prehistoric Denmark following its introduction 6000 years ago. Traces of cultivation are, however, almost solely preserved in the form of ard marks on surfaces sealed beneath barrows or layers of aeolian sand. Only one period in prehistory shows coherent traces revealing how field systems were formed and how they fitted into the landscape. During the course of the Late Bronze Age (1000-500 BC), a system of cultivation was introduced over large parts of NW Europe in which the individual fields or plots were separated from one another by low earthen banks and terrace edges or lynchets. These field systems could extend over several hundred hectares.

These cultivation systems appear primarily to have been in use between 500 BC and AD 200. Research into prehistoric field systems has a long tradition extending all the way back to the 1920s in England, The Netherlands and Denmark, whereas in NW Germany and on Gotland work took place during the 1970s, with the Baltic Countries being involved in the 1990s. Early research was directed in particular towards mapping the field systems which, at that time, lay untouched in agriculturally marginal areas such as heath and woodland.

In Denmark, Gudmund Hatt was a pioneer in this field. During the course of several campaigns, especially during the 1930s, he recorded 120 occurrences of field systems, primarily on the heaths of Northern and Western Jutland. These were published in 1949 in his major work Oldtidsagre (i.e. Prehistoric Fields). His work was continued by Viggo Nielsen who recorded 200 field systems in the forests of Zealand and Bornholm, largely between 1953 and 1963. In the former Aarhus county, the record has subsequently been augmented by a systematic reconnaissance of the forests which took place between 1988 and 1992. Subsequently, this led to the extensive investigations of field systems at Alstrup Krat near Mariager. As early as the 1920s, English researchers were aware of the fact that both ploughed-down and preserved field systems were visible on aerial photographs. However, the method was first applied in Denmark, The Netherlands and NW Germany in the 1970s, leading to a several-fold increase in the number known localities. In Denmark, P.H. Sørensen recorded 447 field systems in the former Viborg and North Jutland counties alone. P. H. Sørensen has published a series of articles dealing with various aspects of aerial photography in relation to ancient field systems. For example, the colour and origin of the various soil marks, the shape and size of the plots, different types of field systems and the relationship with soil type. He has also published several surveys of individual field systems. A significant problem with P.H. Sørensen’s work relates to the very few published plans showing the field systems and to the fact that these are based exclusively on a single series of aerial photographs.

The main aim of this article is to demonstrate the potential for mapping field systems on the basis of not one but several series of aerial photographs. This is done through the detailed survey and mapping of three individual field systems and access to a series of data sources with respect to the interpretation of information contained in the aerial photographs. These comprise an interpretation of the origin of soil marks of banks and lynchets and an evaluation of the degree to which this interpretation is influenced by subjectivity. It is beyond the scope of this investigation to locate the field systems within a settlement and landscape context.

Sources and study area
In order to explore the problems and questions outlined above, three field systems were chosen in the central part of Himmerland: Skørbæk Hede, Gundersted and Store Binderup (fig. 1). This selection took place on the basis of an examination and assessment of almost all recorded field systems in Himmerland evident on several series of aerial photographs. These three field systems chosen are among those best preserved and also the most cohesive. Furthermore, all three have been mapped previously: Skørbæk Hede by Hatt on the basis of field survey, and the two others by P.H. Sørensen on the basis of aerial photographs. This provides the opportunity to evaluate any possible subjectivity in the procedure employed. Hatt makes a distinction between field boundary banks and lynchets. This opens up the possibility of evaluating how the two forms of boundary appear on aerial photographs. At Gundersted Hatt cut two sections through boundary banks. These, together with sections from other of Hatt’s excavations and more recent examples from the investigations at Alsing Krat, form the basis for an investigation of how soil marks arise and develop over time. In this investigation, use has also been made of historical maps in order to reveal the influence of historical cultivation on the presence/absence of soil marks. The earliest maps are from c. 1780. The primary source remains, however, series of vertical aerial photographs. Access to the latter has become considerably easier in recent years. A large proportion is now accessible via various web portals, and recently an overview became available of the contents of private and public archives. For the purposes of this investigation, use has been made of scanned contact copies of aerial photograph series from 1954, 1961 and 1967. From digital archives, use has been made of aerial photographs from 1979 and 1981 and the orthophoto maps from 2007 and 2008, respectively.

Digitalisation and rectification of aerial photographs
Previously, mapping on the basis of aerial photographs was a laborious process involving tracing paper and the transfer of features to topographic maps. The introduction of GIS has, however, eased the process considerably and has also made it easy to compare various map themes such as soil-type, land-use, and digital finds databases. Before mapping can commence, the aerial photograph must be scanned, rectified and geo-referenced. rectification was carried out using the programme Airphoto, while geo-referencing and drawing in of the features were done in MapInfo. An example is shown in figure 2.

Soil marks – how do they originate?
In order to understand how the boundary banks and lynchets between plots appear as soil marks on the aerial photographs, it is necessary to examine how these boundaries were built up and also the influences to which they have been exposed from their creation and up until the time when they are visible on aerial photographs. Figures 3 and 4 show sections through two boundary banks at Gundersted These were carried out by Hatt at the beginning of the 1930s, just prior to the area coming under cultivation again and 20-25 years before the first aerial photographs revealed pale traces of boundary banks. As the area had not been cultivated since the Iron Age, the stratigraphy is the result of natural soil-formation processes: a podsol has been formed, comprising a heath mor layer uppermost, beneath this a bleached sand layer and an iron pan, and at the base the old cultivation layer and the topsoil core of the boundary bank, consisting of brown and grey sand. Ploughing of the boundary banks will, initially, not result in significant soil marks as the three uppermost layers are of equal thickness along the whole length of the section. A pale soil mark will, however, appear when the boundary bank has been levelled out and the plough begins to turn up material from the light topsoil core. This soil transport can in some instances continue for more than 70 years, but the soil marks will as a consequence also become wide and fragmented. This account of the processes leading to the appearance of the pale soil marks is completely different from the only other theory proposed in this respect, i.e. that of P.H. Sørensen. He describes a development involving three phases, beginning with the ploughing up of the bleached sand horizon which generates a pale soil-colour trace. Later in the development there is a shift to a dark trace, when the material in the topsoil core becomes ploughed up. In the final phase, the trace shifts again to a pale colour, when the plough begins to bring up the subsoil. However, these two sections show neither a bleached sand horizon nor a darker topsoil core. Furthermore, no colour changes have been observed at any of the localities. The fact that the boundary banks are apparent as pale soil marks is not due to ploughing up of the bleached sand layer but of the topsoil bank core. Ploughing down of the other boundary form, the terrace edge or lynchet, as shown in figure 5, will similarly result in the formation of a pale soil-colour trace through material being brought up from the pale topsoil core. P.H. Sørensen was also fully aware of this situation, and it can be confirmed by comparing Hatt’s map of the Skørbæk Hede site, where a distinction is made between boundary banks and lynchets, with the soil marks apparent on the aerial photograph series Basic Cover 1954 (fig. 6).

Dark vegetation marks and pale erosion marks
Almost all the soil marks that form a basis for the mapping of the three field systems appear pale in relation to the surroundings. There are, however, occasional exceptions to this rule in the form of dark marks in areas of heather heathland and newly-ploughed heath. On the aerial photograph of Skørbæk Hede from 1954, a few dark marks can be seen directly south of Trenddalen (fig. 6) which correspond with the results of Hatt’s survey. These lie in an area which was cultivated between 1937 and 1954. In 1961, the area was taken out of cultivation and became covered with small trees. A corresponding phenomenon can be observed to the west of the settlement where the heather heathland was cultivated between 1954 and 1961 (fig. 7). These marks probably arise from the vegetation as a consequence of better growing conditions over the topsoil cores of the boundary banks. The fact that lynchets and boundary banks offer different growing conditions has been documented at Alstrup Krat where it could be seen that in several places anemones grew on the lynchets. Differences in the vegetation on the field surfaces and the boundary banks have also been observed on aerial photographs showing the scheduled examples of field systems at Lundby Hede and Øster Lem Hede.

The final type of soil-colour trace to be dealt with here comprises the very pale patches that occur on both sides of Trenddalen at Skørbæk Hede and on the western margins of the field system at Gundersted. These could possibly be interpreted as ploughed-up deposits of aeolian sand, but this is not the case. By comparison with the topography and through stereoscopic viewing of the aerial photographs it becomes clear that these features are located on steeply sloping terrain and that they are due to ploughing up of the sandy subsoil. They become both larger and more pronounced with time as more and more subsoil sand is progressively eroded out due to ploughing (figs. 6, 7, 8 and 9).

The influence of historic cultivation on soil marks
The fact that Hatt could still see boundary banks and lynchets in the landscape during his investigations in the 1930s was of course due to these areas not having been ploughed since they were abandoned at some time during the Iron Age. The Royal Danish Academy of Sciences and Letters’ conceptual map from the end of the 18th century shows that 30% of Himmerland was covered by heath, 42% was cultivated, 21% lay as meadow and bog and only 4% was covered by woodland (fig. 1). By comparing the identified field systems with the heath areas on the maps, an idea can be gained of the duration of cultivation and how it has influenced the soil marks. Correspondingly, by comparing plans showing soil marks with the cultivated area shown on the conceptual map, it is possible to investigate whether cultivation, presumably continuous here since the 12^th century, has erased traces of field systems dating from the Early Iron Age. Plates I-III show combined plans of soil marks from boundary banks, lynchets and recorded barrows at the three localities. The ordnance maps from the 1880s have been chosen as a background, showing contour lines, land use and wetland areas, and the cultivated areas have been added from the conceptual map. At both Gundersted and Skørbæk Hede, there are clearly no soil marks in the areas marked as cultivated on the conceptual map. Conversely, the immediately adjacent heath areas show many coherent traces. On this basis, it must be assumed that the field systems from the Early Iron Age also once extended into areas shown as cultivated on the conceptual map but that the long-term cultivation has apparently erased any trace of them. It should, however, be mentioned that Lis Helles Olesen’s investigations in NW Jutland only reveal a slight preponderance of field systems located on the old heath areas, so there may well be regional differences.

The original total extent of the field systems is of course difficult to assess, but the field system at Store Binderup provides an idea of the order of magnitude. This field system is apparent as a well-defined topographic unit surrounded by wetland areas; the latter are shown on the conceptual map to be completely covered by heath. The field system extends over c. 75% of the cultivable area. In order to examine the influence of modern cultivation on the clarity of the soil marks, plans showing traces of the boundary banks have been compared with a series of historical maps. In general, the soil marks at all three localities appear most clear in areas which were cultivated latest. Former heath areas completely lacking in soil marks have probably never been cultivated. The last 50 years of cultivation with large agricultural machinery has had a dramatic effect on the soil marks. On figures 7, 8 and 9, clear evidence of ploughing out can be seen, whereby the soil marks in several places increase from 5 to 9 m in width. The negative effect of long-term cultivation on soil marks documented here only applies to pale soil marks on sandy soils. A number of field systems are apparent as dark soil marks, the visibility of which does not appear to be affected to the same extent by long-term cultivation. These make up only 3% of those recorded by P.H. Sørensen, and no sections through boundary banks are available from any of these field systems.

Comparison of maps produced by field survey and from aerial photographs
Every map expresses an interpretation of what has been observed. This also applies of course to both Hatt’s mapping of the field systems on the ground in the 1930s and the subsequent mapping conducted on the basis of aerial photographs. Quality and credibility are, however, increased considerably, if the features observed can be confirmed by several sources or several researchers, reducing the subjective aspect to a minimum.

On figures 10 and 11, the author’s plan of Skørbæk Hede based on aerial photographs is compared with the results of Hatt’s field survey. There is no doubt whatsoever that the aerial photographs are better able to show the overall extent of the field system. Conversely, the resulting plan is less detailed than Hatt’s map. In a few cases, however, sub-divisions of the fields are seen on the aerial photographs which Hatt did not record in his survey (figs. 8-9). In order to investigate subjectivity in the interpretation of the aerial photographs, a comparison has been made between the author’s and P.H. Sørensen’s plans of the field systems at Gundersted and Store Binderup (figs. 12, 13 and 14). Good agreement can be seen in the interpretation of the soil marks apparent on the aerial photographs of both localities. This suggests that the subjective aspect of the interpretational process is not a major problem.

Evaluation of the method’s range with respect to studies of the agrarian landscape
Aerial photographs encompass a great research potential relative to studies of the arable landscape during the Late Bronze Age and Early Iron Age. They are the only source available with respect to mapping the morphology and extent of the field systems, with the exception of the few remains tangible which still exist in woodland and on heaths. Field systems are particularly important in a cultural-historical context because they constitute the sole example from prehistory of the appearance of a total integrated cultivation system and how it was adapted to the landscape.

The information contained on the aerial photographs, particularly in the form of pale soil marks resulting from the exposure or ploughing-up of the topsoil core of the boundary banks and lynchets, is a credible source relative to the mapping of the morphology and extent of field systems. Comparison between the maps and plans produced by several researchers mapping does not give cause to perceive the interpretation of the information as the aerial photographs as being particularly subjective. On the contrary, very good agreement can be seen between these interpretations.

In a mapping exercise, use should be made of a number of different series of vertical aerial photographs as this provides the most detailed picture of the morphology of the field systems.

A very significant source of error has been identified which must be taken into account when mapping the extent of the field systems, i.e. cultivation during historical times. In areas that were cultivated prior to the enclosure movement, i.e. in the very great majority of cases presumably since the 12^th century, it cannot be expected to find pale soil marks. Long-term cultivation and the consequent mixing of the upper soil layers have erased most traces of boundary banks and lynchets. Renewed cultivation within the last 100-150 years appears, conversely, only seems to have had a marginal effect on the occurrence of soil marks. As mentioned above there can, however, be marked regional differences on the influence of historical cultivation on the clarity and degree of preservation of the soil marks. This is an aspect it will be interesting to study in more detail in the future.

Michael Vinter
Moesgård Museum