Gravhøje set fra luften

Burial Mounds seen from the airA source critical investigationDenmark was systematically surveyed in secret by means of aerial photography in May 1954. The operation, codename Basic Cover, was executed by the US Air Force and resulted in approximately 42.700 orto-photographs taken from a fixed height of 10.000 feet and rendering a national coverage of 99.6%. The series is now declassified and, as numerous types of archaeological features are apparent on the individual pictures, it is drawn upon with increasing frequency as part of the daily routine of ­archaeological institutions around the country. As a consequence of the aerial survey having taken place relatively early in the year, archaeological monuments such as burial mounds stand out against the newly ploughed fields primarily as differences in the colour of the soil (fig 1). In 2005 the increasing general interest in the use of aerial photography prompted a group of colleagues to create the archaeological aerial photography network (LAND). COWI A/S are presently in the process of making the entire series available digitally which will doubtless lead to a further increase in its usage. Despite a general awareness of the numerous burial mounds visible on Basic Cover, little information is available concerning its properties as a systematic source. The purpose of this article, based on the results of an empirical study, is to provide tangible data on Basic Cover as a general source for use in burial mound listing and, more specifically, as a supplement to the mounds listed in the national database. The starting point was a systematic recording of burial mounds apparent on Basic Cover images from West Jutland. This took place within the framework of a project with the aim of studying the barrow line phenomenon and its links with a number of archaeologically known fords and bridges dating to the Iron Age and medieval times.The recording of burial mounds from Basic Cover was carried out for an area of roughly 1270 km2 around Skern Å (River Skjern) (figs. 2-3). The area is divided in two by the river valley; the landscape in the northern part is relatively undulating with a maximum height of 86 m above sea level, whereas the southern area is predominately flat with long smooth ridges running north/south. The river, Skern Å, which has one of Denmark’s largest river catchments, constitutes a formidable ­obstacle in the landscape. Due to contrasting soil conditions caused by variations in local glacial deposits, any comparison with Basic Cover as a source for burial mound recording in the eastern parts of Denmark should be approached with caution.An unconventional approach was employed in recording to facilitate a comparative study of different sources for burial mound recording (fig. 4). Initially all visible burial mounds were recorded from Basic Cover regardless of previous listing in the national database. The criteria for a positive recording comprised the presence of either a circular white spot or a distinct shadow relief. As a supplement to Basic Cover, burial mounds were recorded from the highly detailed historic map Generalstabens Høje Målebordsblade from 1871, which subsequently became standard issue for many of the archaeological surveyors from the National Museum. All recorded burial mounds were finally correlated with the mounds listed in the national database. The data were then transformed into a ­single set of digital points where presence or absence on Basic cover and the 1871 map were indicated along with – if any – listing in the existing national database.The compiled results of the study are presented in figure 5a. At the start of the project, the national database contained records for 2872 burial mounds from the area. Identification on the 1871 map and Basic Cover resulted in the recording of 2186 and 2209 burial mounds, respectively. The total number of positive recordings was therefore 7267, whereas the number of unique burial mounds was 3983. This adds a total of 1111 to the number of listed mounds, equivalent to an increase of 39% or about 1.1 burial mound per km2 of dry land. The mutual correspondence in percentage coverage between mounds recorded from Basic Cover, the 1871 map and the mounds listed in the national database is shown as a graph in figure 5b. It can be seen that 69% of the listed mounds already appeared on the 1871 map prior to the ­national archaeological survey. The effects of various biasing factors, for example scheduled (and thus well preserved) mounds versus ploughed-over examples, are discussed in an assessment of the strengths and weaknesses of Basic Cover as a source. Less than 50% of the mounds listed in the database and appearing on the 1871 map are evident on the aerial photos. However, the photos still make a notable contribution to the record because they “capture” the very faint traces of the almost completely destroyed mounds that were not detected by the surveyors of the other sources. The newly recorded burial mounds have a significantly positive effect on the clarity of the linear structures in the distribution of burial mounds in the area (fig. 6). Based on an hypothetical, but not unsupported, statistic calculation it is ­argued that as much as 80% of the original population of larger burial mounds has been recorded following the present study.The distribution of mounds was ­explored on the basis of agrarian land use categories shown on the Vidensskabernes Selskabs kort (map) from 1800, in order to evaluate the situation prior to the introduction of major agricultural reforms, (fig. 7) .The distribution of the mounds recorded from the various source categories on the different area types is presented in fig. 8a. When only area types heath and open land are taken into consideration it is clear that mounds recorded from Basic Cover are under-represented on heathland, whereas listed mounds and those on the1871 map are under-represented on open land (fig. 8b). The patterns are to be seen in conjunction with the long-term destructive effect of agrarian land use on open land throughout historical times. In order to investigate the effect of vegetation cover on the visibility of mounds on Basic Cover, the areas covered by forest and heath around 1950 have been added to the map. To understand the effect of agrarian land use on the records from the different sources, the open land category has been divided into two further categories, old open land and new open land, respectively. The latter represents the parts of the heath that were reclaimed for cultivation between about 1800 and 1950 (fig. 9). In order to understand the factors influencing visibility on Basic Cover the distribution of mounds has been studied with respect to heath/forest and old open land/new open land (fig. 10a). When attention is turned to the representation of mounds in old open land and new open land it can be seen that there is no significant difference in the distribution of mounds recorded on Basic Cover (fig 10b). This suggests that there was only a small difference in the density of the prehistoric settlement ­between these two areas. Furthermore, it indicates that the under-representation of mounds recorded in the open land on the 1871 map and in the listed mounds in fact mirrors a bias resulting from historic land use rather than an actual prehistoric pattern. The fact that Basic Cover makes its most significant contribution concerning new mounds in the old open land is thought to be a product of the time when photos were taken. When surveyors from the National Museum visited these areas of long-term ploughing in the late 1900s many of the mounds were already too ploughed-out to be recognisable as such. However, due to the late introduction of mechanical cultivation they were still visible as white spots when Basic Cover was executed in 1954. The aerial photos thus constitute a excellent source for supplementing existing records, but due to a weakness in identifying mounds in areas of dense vegetation cover, compensation is naturally only partial (fig. 10c).The project has generated some factual information on Basic Cover as a source for the recording of burial mounds:1. Basic Cover provided an increase of 32% in the number of recorded mounds compared with existing records.2. 45% of the listed mounds could be identified. This illustrates Basic Cover’s weakness in areas with dense vegetation cover. However, mounds have fre­quently already been listed from these areas3. There is a systematic negative bias in the national database concerning listed mounds in the cultivated areas of histor­ically open land.4. Basic Cover provides a good coverage in historically cultivated areas and it can thus productively be used to comp­ensate for this bias in existing records.5. The soils and topography in the study area are ideal for aerial photography and the results can, therefore, not be transferred directly to the rest of the country where conditions may be less ideal. Basic Cover was carried out at a fortuitous time; after the heath had been reclaimed but before mechanical cultivation had fully run its destructive course. The regrettable destruction of monuments had accelerated but was by no means complete. Fortunately, Basic Cover took place at a time when ploughing had exposed the features without destroying then altogether. Systematic recording of mounds from the Original-1 maps is also recommended ­because data from this source can compensate for the biases seen in the listed mounds from the open land. A study of place names connected to mounds has indicated that many more mounds have disappeared from the open land through historical times.Kasper Lambert JohansenDanmarks MiljøundersøgelserAarhus UniversitetSteffen Terp LaursenMoesgård Museum