This is my coursework for PartIB Paper 5 Quaternary Climte and Environment, which gained a first-class with distinction (83/100).

I. Abstract

Breckland, UK, preserves extensive relict patterned ground from the last glacial cycle, yet key uncertainties remain. Existing surface morphology studies provide rough statistical descriptions, but the evaluation of pattern variability or controlling factors are lacking. Geophysical and sediment studies have examined isolated sites, yet comparative analyses between different pattern types are missing, and no micromorphological studies have provided direct, quantitative evidence of periglacial processes in this region. This study integrates Unmanned Aerial Vehicle (UAV) mapping, Ground Penetrating Radar (GPR) imaging, and sediment analysis to address these gaps. UAV surveys will produce high-resolution optical imagery and elevation models to quantify pattern variability and its controls. GPR will reveal differences in subsurface structures within and across pattern types. Sediment analysis, including micromorphology studies, will provide direct microscopic evidence of various periglacial processes. By linking surface morphology, subsurface structures, and fine-scale sediment properties, this study aims to quantify past periglacial dynamics, including their controls and impacts, in the Breckland area. It also contributes to the refinement of permafrost dynamics models, which is essential land use and infrastructure planning in contemporary periglacial regions.

II. Research rationale

Patterned ground is a surface feature characterized by regular, symmetrical morphological patterns. It is formed in frost-susceptible soils through freeze-thaw processes, including frost heave and solifluction (Ballantyne, 2025; Permafrost Subcommittee, 1998, p.61). East Anglia is one of the few regions globally where patterned ground from the last glacial cycle is well preserved (Nicholson, 1976), with Breckland offering exceptional examples due to controlled grazing and minimal human disturbance (Boreham and Rolfe, 2017; Figure 1 & 2).

Patterned ground study benefited greatly from high resolution remote sensing mapping, especially optical imagery and elevation models. However, existing optical datasets – such as aerial photographs from the Royal Air Force, the Landscape Unit at the University of Cambridge, and Google Earth satellite imagery – lack the resolution and scale needed for precise analysis of patterned ground in this area. The morphological maps based on these optical datasets (e.g. Nicholson, 1969, 1976) are qualitative and not to scale. Similarly, previous elevation models lack the accuracy to resolve key issues, such as the mismatch between stripes and slope (Watt et al., 1966; Nicholson, 1976). This study will use UAV mapping to address these limitations, providing the first detailed quantitative mapping of Breckland’s patterned ground. These datasets will improve site selection, enable statistical analysis of pattern variability, and support global comparisons with other high-resolution patterned ground studies.

While remote sensing provides valuable insights into patterned ground spatial distribution, understanding its formation processes and mechanisms requires integration with field investigations. For example, Optically Stimulated Luminescence (OSL) dating has shown that patterned ground in Breckland were formed during the last glacial cycle, with most sites experiencing permafrost-related frost action during Greenland Stadial 2 (GS2) and the Younger Dryas (GS1) (Bateman et al., 2014), when coldest monthly temperatures dropped by over 28°C (Atkinson, Briffa and Coope, 1987; Huijzer and Vandenberghe, 1998). However, the chronological sequence of pattern formation—including the initial mechanisms, subsequent modifications, and interactions between different periglacial processes—remains uncertain due to the resolution constraints of OSL dating.

Boreham and Rolfe (2017) partly addressed this uncertainty by GPR scans beneath stripes at Grimes Graves. They proposed a tripartite depositional model, composited of brecciated chalk rubble, which was resulted from initial frost-heave-induced bedrock fragmentation, and overlying chalk pellets and gravelly diamicton, which were associated with solifluction—the downslope movement of soil driven by frost heave and thaw consolidation cycles (Murton, 2021). However, studies at other sites reveal variations in subsurface structures: some stripes contain chalky rather than gravelly diamicton (Bateman et al., 2014), while others lack diamicton altogether (Linford et al., 2009). These variations will be assessed by this research. This study will also extend GPR analysis beyond stripes to include polygons and transitional forms, aiming to assess subsurface variability across pattern types and contribute to the globally limited dataset of polygon GPR images (Forte et al., 2022).

Although GPR is an effective tool for imaging high-resolution subsurface structures, linking these structures to periglacial processes remains inferential and requires further validation through sediment analysis. Studies in other regions have shown that chalk microstructure and aggregation undergo significant transformations under cryoturbation and solifluction (e.g., Kovda et al., 2021). Analyzing these changes will allow for a quantitative assessment of periglacial processes, a key objective of this research.

In short, the extent and diversity of relict patterned ground make Breckland a key site for studying past periglacial environments. This research combines UAV mapping, GPR imaging, and sediment analysis to establish direct links between surface morphology, subsurface structures, and sediment-scale features, aiming to quantify key evidence of past periglacial processes and identify the main controls on pattern evolution. Although this study does not include direct dating, previous research has shown that most sites in East Anglia were active around GS2 and GS1 (Bateman et al., 2014). Therefore, this study provides new insights into Breckland’s environmental and climatic conditions during these periods. Meanwhile, the findings also have implications for agriculture, forestry, and infrastructure planning in contemporary periglacial regions (Rowley et al., 2015).

Figure 1 Map of Breckland area. (a) Present-day geology of East Anglia with the Breckland area outlined in red (adapted from Bateman et al., 2014). (b) Distribution of patterned ground in East Anglia as mapped by Nicholson (1969) (adapted from Bateman et al., 2014). (c) Study area with 14 UAV survey sites labeled. Drone restriction zones are shown as dashed orange circles or polygons. Primary field investigation sites are marked with red stars. Green rectangles highlight areas with extensive patterned ground, which are enlarged in Figure 2. This map is based on satellite imagery, aerial photos, and previous literature and serves as an indicative, non-scaled reference.

Figure 2. Detailed distribution of patterned ground types in the areas highlighted by green rectangles in Figure 1c, around (a) Weather heath, (b) Elveden and Berner’s heath, (c) Snarehill, Barham spinney and Euston.

III. Methodology

This study will be completed in 24 months with a budget under £10,000. All fieldwork, lab work, and data analysis are conducted by two researchers.

UAV Mapping

Using aerial photos and satellite images, the patterned ground distribution in central Breckland is preliminarily mapped, identifying 14 sites covering 55 km² (Figures 1c, 2). UAV surveys will be conducted at all sites (Table 1). Four sites—Grimes Graves, Lakenheath, Eriswell, and Euston—lie on the edge of a drone restriction zone (Figure 1c, Table 1), requiring flight permissions from the UK Civil Aviation Authority.

Table 1. Site selection for UAV and field surveys. All sites will undergo UAV mapping, while only a subset will be selected for field investigations. Primary field survey sites are highlighted in brown.

A DJI Mavic 3M drone will be used for its advanced multispectral imaging capabilities, including red (~650 nm) and near-infrared (NIR, ~860 nm) spectral bands. This combination efficiently detects vegetation patterns, as chlorophyll absorbs red light, while mesophyll layers reflect NIR (Mather et al., 2019; Agapiou, 2020). This drone also employs integrated photogrammetry for ground elevation measurement. Its standard spatial resolution is ~5 cm per pixel, which can be enhanced to sub-centimeter accuracy (~1 cm) with an RTK (Real-Time Kinematic) module (DJI Specs, 2017, link: https://www.dji.com/uk/d-rtk/info). The survey will be conducted in May, when vegetation patterns are distinct and wind conditions are stable. The 2024 monthly average wind speed of 4.7 mph (Breckland Weather, 2025, link: https://www.brecklandweather.com/) falls within the safe operating range of the DJI Mavic 3 (DJI Specs, 2024, link: https://ag.dji.com/mavic-3-m/specs). The drone will fly at 100 meters above ground, optimizing image resolution and mapping efficiency.

Collected images will be stitched in Pix4D Mapper to generate orthomosaics and Digital Surface Models (DSMs). Pattern delineation and classification will be performed using Mask R-CNN, a deep-learning object segmentation method (Zhang et al., 2018), followed by manual validation. Vectorization in QGIS will extract attributes such as size, shape (roundness, regularity), orientation (aspect ratio, elongation index), and density (compactness). Elevation data will be refined from DSM to DTM using Progressive Morphological Filtering (PMF) to remove vegetation, enabling the extraction of topographic variables (slope, aspect, curvature, elevation). These elevation data, combined with external datasets (bedrock type, subsurface drainage), will be analyzed alongside extracted morphology data to determine key controls on pattern variability.

GPR Imaging

A GSSI UtilityScan DF GPR system will be used, equipped with both 300 MHz and 800 MHz antennas to optimize depth penetration and near-surface resolution. The system offers a maximum penetration depth of ~7 m (GSSI, https://www.geophysical.com/products/utilityscan-df), which is sufficient for imaging patterned ground structures (chalk rubble reaches the rockhead at ~3.5–4.5 m; Boreham and Rolfe, 2017). The 300 MHz antenna will capture deeper subsurface features, while the 800 MHz antenna will improve near-surface resolution, addressing previous limitations in distinguishing coversand from chalk pellets (Boreham & Rolfe, 2017). Data from both frequencies will be integrated later.

Three sites are selected as the primary choices for field investigations, namely Weeting, Weather Heath, and Berner’s Heath (marked with red stars in Figure 1c). These sites see most extensive distributions of patterned ground, encompassing all four types of patterns, covering large areas (>6 km²), and are free from anthropogenic disturbances. If access to these sites is restricted, secondary options with more diverse patterns but smaller coverage areas will be considered (Table 1). The expected number and orientation of transects for each pattern type are summarized in Table 2, with specific transect locations determined wih UAV mapping results. The location and elevation of all transects will be recorded using a Leica GNSS Smart system during field work. Hand-augered boreholes will validate GPR results and provide soil samples for laboratory analysis (see below), and will be refilled after sampling to minimize site disturbance.

Table 2. Transect allocation by pattern type, specifying the number per site and orientation criteria.

GPR data will be processed using RADAN 7, applying automatic gain control to recover far-offset amplitudes, signal-to-noise ratio filtering. Radar surfaces (bounding surfaces) and radar facies (bed assemblages) will then be classified using metrics such as amplitude, depth, reflector continuity, and geometry, with reference to literature (e.g. Boreham and Rolfe, 2017; Keskinen et al., 2017; Jeffery et al., 2020). Interpretation will compare subsurface structures within and across pattern types. First, stripe transects will be compared to assess questions such as whether diamicton existence correlates with upslope bedrock lithology. Second, transects of different pattern types will be compared to test if the tripartite deposit model of stripe troughs and ridges (Boreham and Rolfe, 2017) also applies to polygon troughs and centers, and how they manifest in intermediate zones. Special focus will be given to chalk pellets and diamicton. These are proposed to be solifluction products (Boreham & Rolfe, 2017), which in theory are prevalent on slopes (stripes) than flat terrains (polygons) (Matsuoka, 2001). Finally, GPR data will be integrated with UAV measurements to assess macro-scale relationships, such as whether overlying coversand depth influences patterned ground size (Watt, Perrin and West, 1966).

Sediment Study

Field sampling will integrate excavation pits and auger cores, strategically aligned with GPR transects to directly correlate surface morphology with subsurface structures. Excavation pits (~1m depth) will allow for undisturbed sampling of coversand, chalk pellets, diamicton, and rubble, as augers may compress fragile sediments (especially chalk pellets). Hand augers (to 2m or more) will supplement pits, capturing deeper stratigraphy, particularly chalk rubbles. A total of eight pits will be excavated: one stripe and one polygon per site, plus one vermicular and one elongated pattern across all sites. Corresponding auger cores will be placed adjacent to each pit.

After excavation, the pit section will be imaged and features such as layering, bedding, sediment contact conditions will be described. Each pit will provide six samples each of coversand, chalk pellets, and diamicton, along with three samples of chalk rubble. Undisturbed block sampling (Kubiena tins) will be used to ensure minimal disturbance during transport, with the orientation marked on each sample. Auger cores will be extracted using liners to prevent collapse, with top-bottom orientation recorded for later analysis. Macroscopic analysis of these samples will be conducted on-site to document clast-matrix relationships, angularity, and orientation.

In the laboratory, samples will be analyzed using macroscopic analysis, size analysis, micromorphology, and X-ray diffraction. The specific sediment characteristics examined in each analysis are outlined in Table 3. These analyses aim to identify evidence of initial frost heave in chalk rubble, and evidence of solifluction in chalk pellets and diamicton. Additionally, the chemical composition of chalk rubble and coversand will be compared with that of pellets and diamicton to determine whether they originate from the same source. Finally, samples from different pattern types will be compared to quantify variations in periglacial processes across patterned ground morphologies.

Table 3. Laboratory analysis plan and sample distribution for different sediment types. Relevant references for analytical techniques are included.

IV. Outcomes and implications

This study will produce the first high-resolution UAV mapping of Breckland’s patterned ground and the first micromorphological analysis of its sediments. By linking macro-scale morphology, subsurface structures, and microscale fabrics, it will establish a multi-scale framework for understanding periglacial processes. It will also provide insights into the controls on pattern variability and refine contemporary permafrost dynamics models. The results will not only enhance reconstructions of past environmental changes in Breckland but also serve as a reference for land use and infrastructure planning in contemporary periglacial regions.

V. References

Agapiou, A. (2020) ‘Estimating Proportion of Vegetation Cover at the Vicinity of Archaeological Sites Using Sentinel-1 and -2 Data, Supplemented by Crowdsourced OpenStreetMap Geodata’, Applied Sciences, 10(14), p. 4764. Available at: https://doi.org/10.3390/app10144764.

Atkinson, T.C., Briffa, K.R. and Coope, G.R. (1987) ‘Seasonal temperatures in Britain during the past 22,000 years, reconstructed using beetle remains’, Nature, 325(6105), pp. 587–592. Available at: https://doi.org/10.1038/325587a0.

Ballantyne, C.K. (2025) ‘Patterned ground’, in S. Elias (ed.) Encyclopedia of Quaternary Science (Third edition). Oxford: Elsevier, pp. 108–121. Available at: https://doi.org/10.1016/B978-0-323-99931-1.00002-7.

Bateman, M.D. et al. (2014) ‘The evolution of periglacial patterned ground in East Anglia, UK’, Journal of Quaternary Science, 29(4), pp. 301–317. Available at: https://doi.org/10.1002/jqs.2704.

Boreham, S. and Rolfe, C.J., 2016-2017. Imaging periglacial stripes using ground penetrating radar at the ‘GRIM’ training site, Grimes Graves, Breckland, Norfolk. Bulletin of the Geological Society of Norfolk, 66, pp.31–43. Available at: https://doi.org/10.17863/CAM.7774.

Breckland Weather (2025) Breckland Weather Data. Available at: https://www.brecklandweather.com/ (Accessed: 30 January 2025).

Cambridge Air Photos (n.d.) Cambridge Air Photos – Aerial Photography and History. Available at: https://www.cambridgeairphotos.com/ (Accessed: 30 January 2025).

DJI (2017) DJI RTK Information. Available at: https://www.dji.com/uk/d-rtk/info (Accessed: 30 January 2025).

DJI (2024) Mavic 3M Specifications. Available at: https://ag.dji.com/mavic-3-m/specs (Accessed: 30 January 2025).

Dzieduszyńska, D.A. et al. (2025) ‘Dry valleys and dells’, in S. Elias (ed.) Encyclopedia of Quaternary Science (Third edition). Oxford: Elsevier, pp. 325–355. Available at: https://doi.org/10.1016/B978-0-323-99931-1.00069-6.

Forte, E. et al. (2022) ‘Investigations of polygonal patterned ground in continuous Antarctic permafrost by means of ground penetrating radar and electrical resistivity tomography: Some unexpected correlations’, Permafrost and Periglacial Processes, 33(3), pp. 226–240. Available at: https://doi.org/10.1002/ppp.2156.

GSSI (n.d.) UtilityScan DF Specifications. Available at: https://www.geophysical.com/products/utilityscan-df (Accessed: 30 January 2025)

Huijzer, A.S. (1993) Cryogenic microfabrics and macrostructures: interrelations, processes, and paleoenvironmental significance. PhD thesis. Vrije Universiteit Amsterdam.

Huijzer, B. and Vandenberghe, J. (1998) ‘Climatic reconstruction of the Weichselian Pleniglacial in northwestern and Central Europe’, Journal of Quaternary Science, 13(5), pp. 391–417. Available at: https://doi.org/10.1002/(SICI)1099-1417(1998090)13:5<391::AID-JQS397>3.0.CO;2-6.

Jeffery, Z.E. et al. (2020) ‘Identification, investigation and classification of surface depressions and chalk dissolution features using integrated LiDAR and geophysical methods’, Quarterly Journal of Engineering Geology and Hydrogeology, 53(4), pp. 620–644. Available at: https://doi.org/10.1144/qjegh2019-098.

Keskinen, J. et al. (2017) ‘Full-waveform inversion of Crosshole GPR data: Implications for porosity estimation in chalk’, Journal of Applied Geophysics, 140, pp. 102–116. Available at: https://doi.org/10.1016/j.jappgeo.2017.01.001.

Kovda, I. et al. (2021) ‘Microrelief and spatial heterogeneity of soils on limestone, Sub Ural plateau, Russia: Attributes and mechanism of formation’, Soil and Tillage Research, 209, p. 104931. Available at: https://doi.org/10.1016/j.still.2021.104931.

Linford, N., Martin, L. and Holmes, J. (2009) Grimes Graves, Norfolk: Report on Geophysical Survey October 2007. English Heritage Research Department Report Series No. 64-2009.

Mather, A. et al. (2019) ‘Automated mapping of relict patterned ground: An approach to evaluate morphologically subdued landforms using unmanned-aerial-vehicle and structure-from-motion technologies’, Progress in Physical Geography: Earth and Environment, 43(2), pp. 174–192. Available at: https://doi.org/10.1177/0309133318788966.

Matsuoka, N. (2001) ‘Solifluction rates, processes and landforms: a global review’, Earth-Science Reviews, 55(1), pp. 107–134. Available at: https://doi.org/10.1016/S0012-8252(01)00057-5.

Murton, J.B. (2021) ‘What and where are periglacial landscapes?’, Permafrost and Periglacial Processes, 32(2), pp. 186–212. Available at: https://doi.org/10.1002/ppp.2102.

Nicholson, F.H. (1969) An investigation of patterned ground. PhD thesis. University of Bristol.

Nicholson, F.H. (1976) ‘Patterned ground formation and description as suggested by low arctic and subarctic examples’, Arctic and Alpine Research, 8(4), pp. 329–342.

Rowley, T. et al. (2015) ‘Chapter 13 - Periglacial Processes and Landforms in the Critical Zone’, in J.R. Giardino and C. Houser (eds) Developments in Earth Surface Processes. Elsevier, pp. 397–447. Available at: https://doi.org/10.1016/B978-0-444-63369-9.00013-6.

Subcommittee, P. et al. (no date) ‘Glossary of Permafrost and Related Ground-Ice Terms’. Available at: https://edunorth.wordpress.com/wp-content/uploads/2020/12/glossary-of-permafrost-and-related-ground-ice-terms-nrc-canada-1988.pdf (Accessed: 30 January 2025).

Van Vliet-Lanoë, B. and Fox, C.A. (2018) ‘Chapter 20 - Frost Action’, in G. Stoops, V. Marcelino, and F. Mees (eds) Interpretation of Micromorphological Features of Soils and Regoliths (Second Edition). Elsevier, pp. 575–603. Available at: https://doi.org/10.1016/B978-0-444-63522-8.00020-6.

Watt, A.S., Perrin, R.M.S. and West, R.G. (1966) ‘Patterned Ground in Breckland: Structure and Composition’, Journal of Ecology, 54(1), pp. 239–258. Available at: https://doi.org/10.2307/2257670.

Wentworth, C.K. (1922) ‘A scale of grade and class terms for clastic sediments’, The Journal of Geology, 30(5), pp. 377–392.

Zhang, W. et al. (2018) ‘Deep Convolutional Neural Networks for Automated Characterization of Arctic Ice-Wedge Polygons in Very High Spatial Resolution Aerial Imagery’, Remote Sensing, 10(9), p. 1487. Available at: https://doi.org/10.3390/rs10091487.

Breckland's Periglacial Patterned Ground