Mapping missing roads

As in any city or built up area, it can be difficult to determine exactly where features such as Roman roads once stood. So while Stewart Ainsworth was running around with his maps, landscape surveyor Henry Chapman was trying out a new computer analysis system to see if his surveys can help with the investigation.

'I've been doing "cost surface analysis",' says Henry. 'Sometimes it's also called "cost path analysis". It's basically a way of finding likely routes across the landscape.'

'My digital model turns the landscape into a virtual wire-frame,' Henry continues. 'Each square on the frame is equal to 10 metres square in real life. The software gives each square a value (or price) depending on where it is and the amount of slope on the ground. For example, a steep bit of ground will have a high value and a flat piece of ground will have a low value.

'What we then do is plot the known Roman roads onto the wire-frame map. There are plenty of gaps where we're not sure which roads continue or connect to others. The computer analyses the value (or cost) of all the land between the end of one road and the beginning of another, and then offers us a proposed route and the cheapest (or easiest) way to connect the two. Hopefully the archaeology in the trenches then matches up with the prediction on the computer.'

Other formulae can also be added to refine the results. But how accurate can this system be? Wouldn't it be easier simply to follow the contours on a map?

'Of course something like this can't tell you everything,' says Henry. 'But it does give us another clue to what could have been here in the past. These are all components that help put the story together. And it looks like Stewart's conclusions from his paper maps are matching up nicely with my computer results.'

Back to Bath

Text only