LiDAR: the archaeologist’s tool for woodland prospecting
LiDAR has the potential to show many archaeological features previously hidden from aerial reconnaissance by woodland cover (Figure 1).
Hillshaded model of ground surface
Hillshaded LiDAR image of woodland
The technique uses an ‘eye-safe’ laser in an aircraft which flies over a site and rapidly scans the terrain below (at a rate anywhere between 20,000 and 300,000 pulses a second). As the aircraft travels forward, the laser scans an oscillating sideways pattern to scan a wide path of landscape beneath the aircraft. For every individual laser pulse, some of the energy is reflected back to the aircraft; the time taken for the reflection to be detected is recorded.
The reflection times are combined with very sophisticated GPS and navigational equipment and computers on the aircraft to provide 3D spatial data for the surface struck by each laser pulse. By overlapping parallel flight paths, it is possible to survey large areas, producing millions of 3D data points known as ‘point clouds’. The density of the point clouds equates to the number of laser pulses per square metre of landscape surface and the resolution of the survey.
These point clouds are used to generate detailed 3D models of the landscape surveyed. The model is used to create a ‘hillshaded’ image – a flat, black and white picture like a photograph which uses computer-generated highlights and shadows to emphasise changes in relief.
3D model of woodland
3D model of ground surface
Example of raw data point cloud
Unlike other aerial surveys such as photography, lidar has the potential to ‘see through’ woodland and produce 3D models of the forest: the canopy, undergrowth and with optimum conditions, the floor itself.
The woodland canopy is effectively porous. Some of the energy from each laser pulse will be reflected back from the upper canopy, lower down the crown or the understory vegetation. Some however, may reach the true ground surface.
Lidar detectors record the times for the first and last reflections for each single laser pulse, corresponding to the top of the canopy and potentially the woodland floor respectively. However, with modern detectors, it is now possible to record a whole series of intermediate reflections (ranging from an additional two to hundreds) for every single laser pulse, helping to build up a substantial and complex 3D model of the forest structure.
Sophisticated computer processing filters the data to remove all the reflections from the canopy and undergrowth, but it requires specialist expertise to generate the best results. Different species of tree and vegetation absorb or reflect laser pulses to varying degrees. Surveys in summer produce different results to those in winter; especially for broadleaf woodland where the vast majority of the laser energy can pass through the canopy to the floor when the trees are without leaf cover.
A good terrain (or ‘bare earth’) model can reveal archaeological earthworks and historical features that have never previously been recorded in aerial photographs as they were hidden beneath the canopy.
Lidar is excellent for detecting linear features, revealing subtle earthworks that are often difficult to see on the ground. Under optimum conditions (well-spaced broadleaf woodland with minimal understorey vegetation), the method can show smaller features such as charcoal platforms and saw pits. Nonetheless, even in stands of conifer where laser penetration is restricted, some larger earthworks may still be evident in the final data.
This technique is not necessarily the answer to every archaeologist’s dreams, as it does have limitations:
- Many laser configurations and flight variables must be optimised to obtain the best results for this type of survey
- Where little light naturally penetrates to the forest floor, lidar may be of little use as the laser will also be blocked by the dense vegetation.
- Smaller features may not be discernible, especially under sub-optimal conditions
- Data filtering or visualisation may remove or hide features of interest, so specialist interpretation and experience is required
- Image features of potential archaeological interest may actually be the result of modern drain clearance, woodland thinning along a roadside or dead bracken and occasionally an artefact of the data processing – some ground-based verification is usually required.