Mapping Roman Roads in Southwest England

BAS Lecture on 18th January 2025 – Beyond Isca: the Roman Road Network in Devon and Cornwall in context

by Dr Chris Smart – University of Exeter 

This lecture described the results of a project that used LiDAR data to identify sections of a likely Roman road network that extended across southwest England, particularly Devon and Cornwall. Previous research has suggested Exeter in Devon was the most westerly extent of the Roman road network, but this study revealed roads much further west, particularly connecting recently identified Roman military settlements.

The ‘Understanding Landscapes’ project ran from 2017 to 2024 and was funded by the Heritage Fund (previously the National Lottery Heritage Fund). It was resourced by students from the University of Exeter and community groups from Newton Abbott in Devon and Calstock in Cornwall. The work involved desk-based studies during the covid pandemic to identify characteristic features of Roman roads from geophysical surveys and LiDAR data.

LiDAR (LIght Detection And Ranging) is a remote sensing technology that emits infra-red pulses from an aircraft that bounce back, from which topographical maps are produced. It reveals the surface of the ground that would not otherwise be seen from normal aerial photographs where they are beneath habitats such as woodland. LiDAR data belongs to the Environment Agency and until 2018/19 mainly covered river valleys and floodplains where there is a flood risk, but since then the National LiDAR Programme has seamless coverage across England. There is now open access to this data which is free to download.

Between 2020 and 2021, volunteers analysed the LiDAR data for Devon and Cornwall using Google Earth Pro, which is also free to use. They were able to identify archaeological features such as earthworks and field boundaries, as well as more subtle features that would not be observable on the ground. The newly discovered Roman roads featured in the local and national press, providing a much-needed good news story to boost morale during the pandemic. 

LiDAR data was also used for the ’Unlocking Landscapes’ Heritage Fund project that ran from January 2022 to February 2023, the aim of which was to undertake a systematic search for new archaeological sites of 10,270 sq. km. across 638 parishes in Devon and Cornwall. A total of 68 volunteers with a core group of 15 regular volunteers carried out this work, with regular training and support. Their results were cross-checked against the local Historic Environment Records. They identified 278 new archaeological sites in Cornwall and 312 new sites in Devon, resulting in the production of the Unlocking Landscapes Story Map, that includes interactive maps. The two Heritage Fund projects and other research have mapped an additional 100km of probable Roman roads in southwest England.

Linear features were particularly clear and after eliminating modern features such as pipelines, the most likely explanation was that most were lost routes. Previous research, including ‘Roman Roads in Britain (Ivan Margary, 1967), and ‘Roman and Medieval Exeter and their Hinterlands (Stephen Rippon and Neil Holbrook, 2021), showed no evidence of Roman roads much further west than Exeter. ‘The Small Towns of Roman Britain’ (Burnham and Wacher, 1990) also showed no Roman roads or towns west of Exeter, and whist excavations have found roads associated with Roman military and civilian sites in SW England, no long-distance routes had been identified.

However, using the LiDAR data, Chris has identified a network of  Roman roads across Devon and Cornwall as well as parts of Somerset, many of which are much further west than Exeter, and others not connected to Exeter at all. Linear features have also been identified west of Nanstallon, Cornwall, but further work is needed to determine whether or not these were Roman roads.  

Many of the newly discovered Roman roads link known military sites, and the talk went on to describe several of these sites, some of which were known sites and others newly identified from the LiDAR data analysis: 

  • Restormel (Cornwall), situated at the navigable limit of the River Fowey, was a Roman fort from 55-80 AD accommodating approximately 500 soldiers. A previously unknown road system has been identified from the periodic quarry pits (where material for the roads was won) along its length, which is a diagnostic characteristic of early Roman road systems. 
  • A Roman road was discovered at Davidstow (Cornwall), north of Bodmin Moor. Numerous animal tracks criss-crossing the area made interpretation tricky, but sections of the road can be clearly seen following the contours,  disappearing beneath the village and then continuing to Camelford. 
  • At Dawe’s Wood, Thrushelton (Devon), a site west of Exeter, there is a known Roman road, but this project has identified flanking boundary ditches, suggesting possible Roman buildings alongside the road.
  • There is relative dating evidence that supports the likelihood of these features being Roman roads. For instance, in the Tamar Valley southwest of Dartmoor (Devon), which is also west of Exeter, there is a linear earthwork with associated quarry pits that were later transected by leats that fed water to the town of Plymouth (Devon). One is known as Drake’s Leat and dates from 1585, so we know that the quarry pits must pre-date this. 
  • At another potential new Roman road site south of Dartmoor a metal detectorist found Roman coins.
  • On the development site of a new town being built east of Plymouth, archaeologists found a 5-7m wide agger and nearby quarry pits, dated to the early Roman period. 

After analysis of the LiDAR data for the entire study area, there were still large gaps with no evidence of Roman roads. Therefore, the LiDAR data results and information on known Roman sites, including military sites, were sent to a Spanish researcher, Joao Fonte. He carried out predictive modelling using GIS (Geographic Information System) tools to identify likely sites within these gaps. The model used factors such as topography and the location of wetlands to predict where major roads may have been constructed. This coincided with the Environment Agency releasing the last of the LiDAR coverage data  and resulted in three additional main roads being predicted.

Previously, a route to the west of Exeter and north of Dartmoor, as far as North Tawton in mid-Devon had been proposed. Exeter and North Tawton were the two main Roman settlements known in southwest England. Being more central than Exeter, it’s possible that North Tawton was the main road network centre, supporting strategically vital connections with tidal estuaries north and south of Bodmin and Dartmoor. North Tawton is known to have extensive Roman activity, including in the military period c. AD 50-85. This was believed to include a Roman military camp including an overwintering camp and a permanent auxiliary fort, as well as a possible fortress alongside the road to Exeter and north of this a possible amphitheatre. Early LiDAR analysis disproves the auxiliary fort interpretation as no double ditches are evident. A more likely explanation is a post-Roman attempt to enclose the land, possibly during the early Medieval period. As well as the military presence at North Tawton, many regimented buildings, crop marks and roads running N-S and W-E suggest an urban centre, i.e. a Roman town. 

If North Tawton was a Roman centre, why were there not more elsewhere in the west country? There is a Second Augustan Legion’s marching camp at North Tawton (Devon), 12.6ha in size, and campaigning appeared to have ended here. At Trerank, Roche, near St Austell (Cornwall), as part of a large development site, archaeologists identified an auxiliary camp, although it is more likely to be a marching camp. At Respryn, St Winnow (on the Nation Trust Lanhydrock estate, Cornwall) , the LiDAR analysis identified an 11ha Roman legionary camp next to the River Fowey. At Rashleigh, Eggesford (Devon), there is a temporary Roman marching camp and in 2021 Roman military coins were discovered. These marching camps appeared to be following the coast and were in pairs, which is unusual.

At Calstock (Cornwall), a Roman auxiliary site has been excavated revealing evidence suggesting that it was constructed between 55AD and the early 80sAD to house 500 soldiers. It is enclosed by a huge ditch that contained many layers of rubbish characteristic of a military profile, including a soldier’s shoe, tweezers, pottery, a gaming token with a Roman building etched onto it and a piece of concreted iron, which an X-ray showed to be a dagger scabbard. 

At Bishop’s Tawton (Devon), a military complex has been excavated on the Tay estuary. It was a Roman auxiliary site enclosed by a ditch. At Higher Woodhall, Exbourne (Devon), there is a 2.8ha site that hasn’t been excavated but there is an obvious earthwork, likely to be a Roman military site. The LiDAR data shows a second line of defence, i.e. a refortification. One possible explanation could be an uprising in the southwest of Britain orchestrated as part of the Boudican revolt of 60/61AD requiring additional marching camps to pacify this uprising. In addition, approximately fifty previously unknown square enclosures with rounded corners have been identified which may be Roman fortlets. 

In conclusion, it is likely that the proposed road network taken from the LiDAR evidence and GIS predictive modelling includes prehistoric routes and Roman military and civilian roads. North Tawton is now believed to be the likely nerve centre of this network rather than Exeter, and connections with tidal estuaries north and south of Bodmin and Dartmoor have also been identified. 

You can read more on the subject in Remote Sensing and GIS Modelling of Roman Roads in South West Britain, which has been published in the Journal of Computer Applications in Archaeology.  

As a parting gift, if BAS are interested in using LiDAR data, Chris can supply data and instruction guides.

Figure 1: Study area, showing the location of the Roman roads known or hypothesised before the beginning of this study, alongside Roman towns, forts and other potential military sites. Also new sections of Roman roads in southwest Britain identified through the 2022 National LiDAR Programme data, shown in red.
Figure 2: Primary and secondary nodes (points to be connected) in the study area.
Figure 3: Study area showing LiDAR roads and Low Cost Paths (LCPs), i.e. the optimal connections between two or more points.

report by Debbie Cousins