Fenland Lidar
Learn more about the Fenland Lidar project and how the discoveries are being used.
Introduction to Fenland Surveys
In contrast to the Hadrian’s Wall survey, which relies on recent NLP (National LIDAR Programme) data, the Fenland surveys have a much longer history. These surveys began with data for the Witham Valley, which was initially acquired from the Environment Agency back in 2005. The survey coverage expanded gradually to cover the entirety of the Lincolnshire Fenland by 2012. Later on, Cambridgeshire and Norfolk were included, although still before the adoption of NLP data. At that time, complete coverage was not available in these counties, prompting a plan to revisit the entire area using the most up-to-date dataset. This updated data is now accessible through the map below.
Processing the Lincolnshire Fenland Lidar Plot
The processed Lincolnshire Fenland lidar plot eventually encompassed approximately 2250km², and sharing this data posed significant challenges. It is currently stored in various repositories in 5km- and 10km-square blocks. Despite advances in storage, the full-resolution 3GB TIF image remains cumbersome to transmit. However, a zoomable, pannable map is now available via Leaflet. This map can be accessed by using the map on the left, following this link to a bigger version of the map or by reading further to learn about the project, the processing involved, and the interpretation of the color scales used.
Exploring the Map and Lidar Layers
The Leaflet map initially loads with a straightforward OpenStreetMap background. Users can select lidar and other map layers from a drop-down menu. There are two lidar layers available:
- Fens2012: This layer contains data originally served from the earlier version of the site, covering Fens, Marsh, and Cambs in a non-seamless map of approximately 5250km².
- FensNLP: This is the updated and more seamless version, covering some 7800km² with a 1m resolution, totaling around 40GB.
Archaeological Studies of the Fenland
The archaeological study of the Fenland has a rich history, though too extensive to cover in detail here. Key studies include The Fenland in Roman Times (Phillips 1970) and the eleven Fenland Project volumes published between 1985 and 1996. These studies, along with significant works on sites like Flag Fen, Must Farm, and the Car Dyke, have emphasized the impact of fine topography on settlement patterns in these low-relief landscapes.
Mapping Roddons with Lidar
A defining aspect of the Fenland Project mapping was the inclusion of roddons—extinct, silt-filled creeks that provided firmer ground within the peat fen. These were originally mapped using aerial photography and field survey, but the lidar dataset allows for consistent mapping across the entire Fenland, providing a level of topographic detail unmatched by other sources.
The lidar maps reveal the extensive and detailed cropmark landscapes of the fens, illustrating how the locations of settlement sites, trackways, and watercourses are influenced by the fine-scale topography. This topography is revealed in detail for the first time through the lidar data. The images below show cropmarks in Rippingale and Dunsby Fen
The Genesis of the Fenland Lidar Project
The genesis of the Fenland lidar project began with the offer of newly captured EA lidar data to the Witham Valley Archaeological Research Committee in 2005. Recognizing its potential for mapping the Fenland landscape, the project embarked on an ambitious series of data requests. Initially unfunded,(although Heritage Lincolnshire/ Archaeological Project Services did pay for some parts) the project eventually received support from Historic England, resulting in the project reports below. The addition of the Lincolnshire Marsh and Cambridgeshire/Norfolk fens was an unfunded and rather piecemeal project undertaken once the data became more freely available. Subsequent projects, particularly through York Archaeology, have refined workflows for large mosaic production and the generation of the National Lidar Programme Fenland mosaic. Any lack of seamlessness in the data often results from varying ground conditions across survey seasons.
The original processing style was governed by the essentially topographic aims of the project (to define the landscape setting of archaeological features, rather than necessarily to prospect for the unknown). This was created using MapInfo, which had a unique approach to hill-shade production. The NLP dataset has been processed and assembled using the Relief Visualization Toolbox and QGIS, from which the XYZ tiles have been exported. It comprises a simple hillshade (A315 H35 Ve10) merged with a custom colour ramp focusing on fine-scale topography.
The original mosaic, based on a 2m-centered Digital Surface Model dataset, allowed mapping at 1:5000 across the entire area. DSM was initially preferred since tree cover is hardly an issue in this landscape and it provides better mapping of contemporary settlement (the pattern of which is equally dependent on the underlying topography). The updated mosaic uses the 1m NLP DTM dataset, which offers higher resolution and a refined focus on archaeological features.
Adjustments in Color Ramps and Interpretation
The original colour ramp was designed to make a sharp distinction at the fen edge, effectively the 3m contour, land above this shading from green (3m) to yellow (5m) to brown (10m). The white-blue-cyan scale represents everything below, with the lowest values white. The use of a continuous rather than discrete colour scale makes for a less objective plot but is much more expressive of the landscape.
One unfortunate aspect of this early approach was that the colour scales defined for Lincolnshire didn’t work as well in Cambridgeshire where elevations fall below -2m. Adjustments to the white-blue gradient meant that the map was no longer completely seamless, but now that it is superseded by the NLP mosaic it will have to stay as it is. One problem with the earlier colour ramp was that areas of the fenland settled in Roman times nevertheless fell within the blue-cyan scale, perhaps giving a misleading impression of what was (or wasn’t) ‘dry land’. The revised colour ramp therefore uses different intervals with green coming in at 2.5m and yellow at 4m OD. The different impression this gives can be seen in some of the figures below.
References
Bennett, R., Welham, K., Hill, R.A. and Ford, A. 2012 ‘A Comparison of Visualization Techniques for Models Created from Airborne Laser Scanned Data’, Archaeol. Prospect. 19, 41–48.
Doneus, M. 2013 ‘Openness as Visualization Technique for Interpretative Mapping of Airborne Lidar Derived Digital Terrain Models’, Remote Sensing 5(12), 6427-6442
Historic England 2018 Using Airborne Lidar in Archaeological Survey: The Light Fantastic, Swindon
Kokalj, Ž. and Somrak, M. 2019 ‘Why Not a Single Image? Combining Visualizations to Facilitate Fieldwork and On-Screen Mapping’, Remote Sensing 11(7), 747.