Professor J. Crow, Dr. S. Turner, Dr. Athanasios Vionis
Characterizing the historic landscapes of Naxos

HLC is a generalising technique that seeks to present a broad-brush characterisation of an area’s historic landscape. As such, it does not normally provide a detailed description, though such detail can be added to the GIS from other sources. HLC maps differ from traditional methods of storing and presenting records about historic landscapes in several important ways, though like them HLCs are used for both landscape management and research (Turner 2006a). Such ‘traditional methods’ include archaeological databases or inventories of sites and monuments [e.g. the UK’s county-based Historic Environment Records (HERs)]. In Turkey the current TAY project provides an overview of regional and period inventories available on the web. Generally speaking, archaeological databases provide lists of archaeological sites together with relevant information – location, period, extent, date identified and so on. Some of these databases are now very sophisticated: for example, many UK HERs are web-based and available in whole or in part to interested members of the public. As inventories of sites they are crucial tools for research, landscape management and planning, particularly where the preservation and enhancement of individual sites and monuments is concerned. However there are limitations to such datasets and Historic Landscape Characterisation provides one way to help deal with these problems. Unlike an archaeological inventory, HLC does not map individual archaeological features. Instead, it groups together features like field boundaries, lanes and farms that are linked by their historical development and then maps them as areas. To do this, the HLC researcher needs to understand how patterns in the landscape reflect its historical development, and how the physical features that make up the landscape relate to one another. So like all landscape archaeology, HLC mapping involves a partly subjective process of interpretation that is informed by the physical landscape.

Two principal sources were used to inform our characterisation:
1. IKONOS 1m black and white and 4m multispectral satellite data supplied by European Space Imaging (LLC), Munich.
2. Historic air photography. This comprised RAF air photos taken during sorties in 1943. Most of the study area was covered. The photographs were scanned and supplied by the Keele University Air Photo Unit (this collection has since moved into the care of the Royal Commission on the Ancient and Historic Monuments of Scotland (RCAHMS) in Edinburgh).

In our research on Naxos, we are particularly interested in the research applications of HLC, and how it can help us understand past landscapes and societies. One particular research question has been the continuing debate concerning the origin of terraced field so characteristic of the eastern Mediterranean rural landscape. In places, the location of dated Byzantine monuments hints at the antiquity of Naxiot terrace systems. Although the relationship cannot be proved absolutely without fieldwork on the ground, many Byzantine churches appear to stand on terraces within braided terrace systems. Examples include the early Byzantine churches of the Taxiarchis Rachis and Ag. Isidoros in Rachi, where both monuments perch on long terraces constructed along the hillside. On the opposite side of the valley below the middle Byzantine church of the Panagia Rachioditissa great oaks that must be several hundred years old stand on similarly massive terraces that can run for at least 800m. If earlier than the churches, the Rachi terraces must be late Antique or classical. Similar long, slightly sinuous terraces run along the hillsides below the classical temple of Demeter near Ano-Sangri.

Archaeological field survey also hints at the antiquity of these terrace systems. Around the church of Ag. Kyriaki, an early Byzantine monument north-east of Apeiranthos, analysis of fieldwalking data by Vionis et al. (Forthcoming)suggests that up to 70% of the ancient finds collected belong to the 7th–9th centuries AD. Curving drystone walls enclose small fields here that only partially and untidily enclose the terraces; the latter are probably related to the early Byzantine settlement. It seems likely that whatever the original date of Naxos’ braided terrace systems, the vast majority would have existed in or before the 17th century. Further details of this research will be available shortly on a dedicated website (see http://www.shc.ed.ac.uk/archaeology/ ) and in a number of articles.

We are grateful for the interest shown in our project by Dr Charalambos Pennas of the Second Ephorate of Byzantine Antiquities, and to the Arts and Humanities Research Council for their financial support.

Professor J. Crow, University of Edinburgh
Dr S Turner, University of Newcastle upon Tyne
Dr Athanasios Vionis, University of Leuven

This year’s field survey in Thrace combined two distinct projects, both developing from the established programme of work on the late Roman and Byzantine monuments commenced in 1994 (Crow 2007; Crow et al 2008). The first was a new programme, supported by the AHRC (Crow and Turner), concerned to evaluate the potential of remote sensing satellite images and their application for the study of ancient landscapes applying the technique of Historic Landscape Characterisation (HLC) in the region around Silivri (for a discussion of the technique see the outline in Crow, Turner and Vionis on Naxos). The second project, in collaboration with Professor Derya Maktav of ITU and supported by TUBITAK, will apply satellite data and other digital map data to develop further research into the Byzantine water supply system (Çeçen 1996) and to provide an extensive digital terrain model and GIS to document the complex system of Byzantine hydraulic monuments in the region west of Istanbul.

Survey in the Silivri District was directed by Prof. Crow and Dr. Turner and in Catalca by Professors Crow and Mektav.

Silivri
The aim of the Silivri fieldwork was to test the application of Historic Landscape Characterisation for Mediterranean Landscapes, it forms part of a parallel prject on the Greek island of Naxos funded by the Arts and Humanities Research Council. The study is coincident with the area of research concerned with the southern part of the Anastasian Wall and the territory of the medieval town of Silivri. The main focus of our study was the use of IKONOS Satellite images to recognise significant patterns of past agricultural field present within the modern agricultural landscape extending from the sea shore as far as the edge of the forested hills defined to the north by the main railway line to Istanbul.

In the survey are we were able to discern several distinct patterns of fields surviving before the creation a large agro-fields which are a feature of part of the current landscape. Further research is needed before we are able to establish a clear chronology for these field systems, but in places we can suggest they date back to the later medieval period. In places we were able to identify a farm or villa of Roman date located east of Fener village and a number of mounds or tumuli. We noted two areas of significant archaeological damage along the line of the Anastasian Wall. First at south of Kurfalı, at Çilingir Tepe, a prominent mound had been badly damaged by treasure hunters revealing parts of a massive tower (burc) at a turn in the Anastasian Wall. Second at the small medieval settlement located due north of Pinarca we encountered treasure hunters at work. The Jandarma at Catalca were informed and seven men were arrested. There was limited damage to some burials of Byzantine date.

Kurşunlugerme K 20
We visited the bridge to familiarise the ITU members of the project the structure and retake a number of GPS points using the new ITU equipment. We were able to follow the line of the lower, broad channel to the N, this is marked as a prominent terrace contouring around the hill slopes. We observed a number of channel openings and took new GPS points. On the return to the bridge it was possible to follow the higher, narrow channel which is clearly defined as a ledge, further points were recorded. At the bridge we were disappointed to observe that further disturbance had occurred since Crow’s last visit in July 2005. On the east face it was noted that the inscription to the Augusti recorded by Oreshkov and published in Crow et al. 2008 had been destroyed by treasure hunters. This is the location of a previous attempt to damage the bridge in 2004 when the Jandarma at Binkılıc were called and visited the site. The cross marked stone beside the inscription remained undamaged, but the inscription was destroyed.

Karamandere
South-east and East of Karamandere Village

Çüçürüdere Germesi (K20.12) first bridge after Kurşunlugerme (700m due north), the line of the channel can be followed as a clear path, in places on top of the channel, elsewhere as a ditch where the channel has collapsed. Traces of the lower bridge survive in the stream bed 4.70 m wide with a core of long irregular blocks. (ITU 2). Remains of a terrace wall support the channel as it leads back towards Kurşunlugerme.

Kayınlık Tarla Germesi (K20.11) First main bridge after Kurşunlugerme
Single arch surviving to an estimated height of 9.20 m, the bridge is 4.70 m wide and the archway is 2.30 m across. It is constructed of regular blockwork 0.40 x 0.70 with an even face without any bosaj, except in few paces, (JC 129). No clear evidence for spring courses or springing. Masonry is similar to Elkafdere 2, but without the chamfered springing. Probably phase 3 (6th cent.) rebuild of phase 2. [Visited 11/07/07]

Another valley with a possible surviving bridge lies to the south-east of Kayınlık Tarla Germesi, the stream leads into the Ceviz Dere, but this site was not visited. The area of forest was however exceptionally thick in this area and it would be difficult identify small structures in the dense woodland.

Koserelik Germesi (Cevizlik Dere main)
Large, single arched bridge, 8.90 m wide with the arch spanning 6.15, aligned south-west to north-east, with an estimated height of over 12 m. On the downstream face on the north side of the stream there is a clear stepped offset, projecting 3 m and rising in steps to 3.10 m. Traces of secondary arch perhaps only 2.5 m wide survive on both faces of the inside of the arch. South-west of the bridge there are the remains of the broad channel turning on to the bridge with a width of 1.9m and surviving depth of 1.65 m. The masonry of the bridge is of long quarry-faced blocks with drafted margins similar to the large bridges such as Kumarlıdere. There is a large pool on the down- stream side of the bridge.

This bridge was first visited by Crow and his team in 1997 but we were able to make more detailed measurements and GPS readings in 2007. The bridge and its environs have become very overgrown in the last ten years, but it is a good example of a large single arched bridge of the second phase, similar to the first phase of Talas K22. Evidence for the secondary arch indicated a narrow span across the archway, although it is not clear how this was supported (see Crow et al 2008, K20.1)

Tatlıdere Germesi (Güngormez Dere)
From the configuration of the map it is likely that another aqueduct bridge was located in this valley it was not visited due to the thickness of the forest cover.

Turçine Germe (also Turçineçatak Germesi)
A long monumental bridge with a high surviving single main arch survives in a deep thickly wooded valley of the Turçine Dere. The remains of the narrow and broad are especially clear in the approach to the bridge from the SW and it is possible to walk across the bridge. The tunnel opening was seen in a number of places, but these details were not checked on the NE side. The bridge piers are 7.50 m wide and there is a single arch across the stream 6.40 m wide. The total height of the bridge was estimated by GPS at 20 m and there are square springers for the main arch. Flanking the main arch are two secondary archs, 3.60 m wide and c. 6 m in total height.with chamfered blocks at the springing. Within these arches the masonry is well preserved and shows similar reticulated work (bosaj) as seen at the other major bridges. This can also be seen in the upper face above the main arch. The west face (down steam) is poorly preserved, but there is clear evidence for the cladding of the east face with blockwork similar to Kayınlık Tarla Germesi and as seen in the Elkafdere in the valley to the east. There is evidence for the abutments of an earlier bridge 100 m upstream seen in the S face of the Vadi, but not the N. In length and height this bridge is comparable to Talas K22.

West of Karamandere Village
Taking a forest road west of Karamandere we followed the slopes overlooking the valley of the Stranja ( Binkılıc )Dere, we observed evidence for the broad channel as leads on from Karamanoğlu Tepe in the territory of Aydinlar village. The channel was 1.40 m wide and survived to a height of 1.45 m, there was a well plastered face on the N side. West of the Karaman Dere there is only a single broad channel known to the west as far as Binkılıc. Further west as far as the major springs near Vize only a narrow width of channel is known.

Testiler Germesi (K13.1)
The bridge is aligned N-S across the deep valley of the Testiler Dere. The bridge is 8 m wide and ruined. On the south side the construction can be seen to constructed of large well cut limestone blocks. The course of the towards the SE is clearly defined where the woodland has been cleared. On the N side of the valley there are traces of blockwork indicating reconstruction and on the west face the remains of a buttress, indicating later repairs.

References:
Crow, J., ‘The Anastasian Wall and the Danube frontier in the sixth century’, in L. Vagalinski, ed., The Lower Danube in Antiquity (the fifth century BC – the beginning of the seventh century AD) (Tutrakan, Bulgaria) 397-410.

Crow, J., Bardill, J. and Bayliss, R. The Water Supply of Byzantine Constantinople (JRS Monograph 2008).

Acknowledgements:

We are very grateful to the General Directorate for permission to carry out our survey. At Karamandere Köy we would like to thank Nuri Engin and Ömer Köbek who guided us through difficult forests and drove us in a trailer along forest roads. They showed an interest and enthusiasm for the monuments in their village territory. In addition we would wish to thank Celal Kolay of ITU and the students from ITU who accompanied us, Irfan Akara, Keıem Esenem, Başak Geze and Cihan Uysıl.

Professor J. Crow, University of Edinburgh
Professor D. Mektav, Istanbul Technical University
Dr S Turner, University of Newcastle upon Tyne