The distributions and variations of Quaternary Thames River Terrace deposits of Greater London

This Technical Note documents the compilation, evaluation and analysis of a database of borehole records that intersect Quaternary sedimentary intervals, from across Greater London. This work complements the analyses and publications of many researchers and experts in the Quaternary of London and the Thames Basin.

Gibbard (1985, 1994) provided one of the first in-depth lithological and stratigraphic descriptions of the Quaternary River Terrace Deposits (RTDs) of the Thames Basin by mapping and connecting outcrops from the Upper and Middle Thames to those of the Lower Thames and Thames Estuary, and by discussing earlier courses of the rivers that now make up the Thames drainage system within London. His results are displayed in the various maps within his books (e.g. Gibbard 1994).

A great wealth of additional information is presented in other and subsequent works on the Quaternary geomorphology and geology of the Thames basin, such as those by Wiseman (1978), Devoy (1979), Gibbard (1983), Nunn (1983), Barton (1992), Whiteman and Rose (1992), Bridgland (2000), Maddy et al. (2001), Bridgland et al. (2004) and Gibbard and Clark (2011).

The British Geological Survey (BGS) holds a vast database of borehole records, dating back to the 1800s, for boreholes that have penetrated both solid and drift geology of London and surrounding areas and has underpinned the production of modern analogue and digital maps of the region.

In consideration of the above, it would be fatuous to attempt to replicate, in this document, the plentiful excellent Quaternary mapping work already undertaken by other researchers. Rather the authors considered it important and relevant to make available this large database of raw geological information from boreholes across London, in a form that has been compiled, quality assessed and analysed, and that is the purpose of this note. The borehole log data are available as raw scanned borehole logs and in electronic form from the BGS. In this Technical Note the compiled database up to August 2019 is presented as summary descriptions and borehole point distribution maps. The spreadsheets, which are the source of this information, are large and are not published here but are provided as Supplementary material.

The original purpose of the work was to assist those who wish to date the movement of known structures, but, as the work progressed, it became apparent that these descriptions and presentations would provide any interested parties, be they geologists or engineers, from industry or academia, with a representative selection of borehole data from both important Quaternary stratigraphic units (the river terrace gravels and associated regional silt deposits) and areas of important and potentially economic Quaternary deposits.

Summary details of Quaternary chronostratigraphy are illustrated in Figure 1. The uppermost 781 000 years are expanded and cover the stratigraphic interval in which the Quaternary deposits of the Thames Basin (within the M25) fall. The bulk of the RTDs fall within the Thames Valley Formation (TVF). The boreholes in the database are representative of most of the Quaternary TVF stratigraphic members, as defined by the BGS Lexicon of Named Rock Units (https://webapps.bgs.ac.uk/lexicon/). It is assumed that the information taken from that Lexicon comprises the latest stratigraphic information available. Hence the ages quoted below and the stratigraphic intervals assigned to the rock intervals within the spreadsheet relate to those of the BGS.

A generalized lithostratigraphic framework for the Quaternary and Neogene for Great Britain and the Isle of Man has been formulated by McMillan et al. (2011) and adapted by Banks et al. (2015) and is shown in Table 1 for reference. Other notable work on the stratigraphic framework and chronology includes that by Bridgland (1994, 2000), Ellison et al. (1994), Jürgen and Gibbard (2004), McMillan (2005), Bondevik et al. (2006), Bridgland and Schreve (2009), Cohen and Gibbard (2011), Lee et al. (2011), Böse et al. (2012), McMillan and Merritt (2012) and Aldiss (2014), and most recently the chronological work on the tectono-climatic evolution of the landscape of England by Lee et al. (2018).

The relationship of each Quaternary member to another, in terms of precise age, is extremely difficult to determine, owing to the absence of any reliable palaeontological information, and thus the Member's absolute stratigraphic position is not always precise. Consequently, the Quaternary stratigraphic members as encountered in the London BGS and Crossrail borehole records are shown in approximate stratigraphic sequence, with oldest at the bottom, in Table 2.

The older Quaternary sections are not well represented in the database, so they have short descriptions. Each of the Anglian BPGR, WOGR, DHGR and GCGR sections is encountered in only a few boreholes (a total of c. 59), LOFT is present in very few and the Cromerian STGR in c. 25 boreholes.

The main stratigraphically distinct RTDs are well represented and described in detail. For instance, the Devensian KPGR is present in over 2000 boreholes, the Wolstonian TPGR in over 1000 boreholes and HAGR in over 250. The LHGR deposits are recorded in over 400 boreholes and the older Hoxnian–Wolstonian BHT in over 250.

In over 25% of the boreholes (c. 1300 boreholes) it has not been possible to determine which RTD Member the intervals represent; these are termed RTDU (River Terrace Deposit, Undifferentiated). The main bulk of these lie downstream of Woolwich, along the lower Thames floodplain and into the Thames estuary; that is, within TQ47NE, TQ48SE, through TQ58SW and TQ57NE to block TQ67. These boreholes are generally below sea level and thus ground observations for mapping and continuity are absent.

The interglacial silt deposits are also represented: Flandrian ESI (49 boreholes), Devensian LASI (over 200 boreholes), Ipswichian CFSI (nine boreholes) and Wolstonian ILSI (16 boreholes). Brief descriptions are given of each of these.

In some boreholes, Quaternary RTDs are absent but it is not known whether these were ever deposited, have been deposited and subsequently eroded, or have been removed through extraction or construction. The thickness of Made Ground may help in the analysis of the reason for the absence of the RTD; also, the lack of Made Ground should indicate a genuine absence of the RTD, but it would not be known whether this was due to non-deposition or subsequent removal by erosion or faulting. Faulting should be invoked only where there is no other credible explanation or where there is evidence of fracturing and/or disturbance.

The database comprises records extracted from over 5781 borehole log records, the majority of which are from boreholes drilled since 1985. The raw information for each borehole record has been subdivided and presented according to BGS-defined lithostratigraphic intervals. All borehole records used in this note are provided in a shared folder from which users are free to download the data \Quaternary_BH_database.

Each lithostratigraphic interval and the depths at which it is observed forms a single row in the main reference spreadsheet entitled BGS-CR_RTD-BH_ALL.xlsx (in \BGS-CR-BH-data\Spreadsheets\All-BH-data); thus the borehole with ID 31922 (Fig. 2) has four recorded lithostratigraphic intervals, to depths of 22.91, 21.36, 18.86 and 15.25 m Ordnance Datum (OD), respectively. This spreadsheet contains 27 549 records from 5781 boreholes, and has two sheets (Microsoft Excel tabs), which present the entire lithostratigraphic records in ‘All data’ and a ‘Fissures’ sheet containing the extracted records from boreholes, which contains descriptions of fissures, faults or fractures; a section of this file is illustrated in Figures 2 and 3.

A further file in the same folder contains the outcrops of Quaternary lithostratigraphic intervals (BGS-CR_RTD-BH_outcrop-for-maps.xlsx) and in a sheet named ‘For-GIS’, in a ‘map-ready’ format suitable for display as points in 2D map form in ArcGIS (or other similar GIS platform). In this case, each row contains a unique record for each of the 5781 boreholes where the following information is given: DATAcat (record quality), Hole_ID (unique borehole number), Supplementary hole ID number, relevant Ordnance Survey TQ reference number, BGS number, easting coordinate (m), northing coordinate (m), borehole ground elevation (+OD(m)) and the Quaternary outcrop lithostratigraphic code (Main_Quat_Strat) for each borehole (one per row). A section of this file is illustrated in Figure 4.

For map plotting, ESRI point shapefiles are also provided (in \BGS-CR-BH-data\Shapefiles and in \CR-BH-data\CR-clast-shapefiles) for the borehole locations and separated clast types and other features.

Notes on the name and significance of each spreadsheet (and other files) can be found in the Supplementary material: \BGS-CR-BH-data\README-BGS-spreadsheets.docx and \CR-BH-data\README_CR_spreadsheets.docx

The cores and samples recovered from boreholes are most commonly obtained in southern England from either cable percussion or rotary drilling. The former need not use a drilling fluid whereas the latter does, and the circulation of this fluid through the drilling bit and against the sample within the core barrel can wash away fine material. A borehole log will normally record the method of drilling used. Descriptions and their implication, particularly of mixtures (some, occasional, with, etc.) and material strength (strong, weak, etc.) have evolved over the past 25 years and subtle differences now exist between old and more recent descriptions of the same material. Users of this database are advised to check such details when comparing logs.

The database described in this note has been compiled from two distinct sources:

1. borehole logs held in the BGS digital database available online via the BGS GeoIndex web portal (https://mapapps2.bgs.ac.uk/geoindex/home.html?layer=BGSBoreholes), which intersect and are representative of the main Quaternary stratigraphic intervals;

2. trial-pit and borehole logs captured during the Crossrail project (CR), which also included some analytical results (grain-size and engineering in situ tests).

The most obvious anomalies and inconsistencies have been corrected for this note. Typically, these have included incorrect ground level entries, which, after reference to the Ordnance Survey (OS) Digital Terrain Model (DTM, Terrain 5 product), have been corrected to reliable values above OD. Similarly, ground surface heights, which are missing entirely from the original borehole data, have been estimated from the Terrain 5 DTM using the borehole coordinate location.

The CR borehole positions were recorded using a unique ‘local’ CR georeference system, which is related to the UK National Grid (OSGB30) by a transformation in x, y and z; these CR positions have been transformed to six-digit British National Grid (BNG) coordinates and are included in this database The six-digit BNG coordinates (e.g. [559490/193760]) can be used as a search key on GeoIndex. Online records include scanned copies of the original borehole log (pdf or multi-layer tiff files), which contain the raw, often handwritten records of the driller's or geologist's observations. This online database has been used to quality assess and confirm the stratigraphy of many (but not all) boreholes.

Where there are multiple (different) borehole records with identical grid references, the borehole with the best or most complete information and fullest log descriptions has been used in any further analysis.

The quality of the borehole records can thus vary greatly, so each record has been quality assessed using the simple schema described here. Where anomalies, discrepancies or ambiguities were identified, the data were downgraded; for example, from best to good, or from good to useful.

1. Best quality data: these are the most recent boreholes and they consequently benefit from modern or improved drilling techniques, more accurate sample collection, better or more detailed descriptions and representation of the lithologies encountered, and better retention of ‘fines’ material. These tend to include logs recorded since 2000, which benefit from borehole positioning recorded using Global Positioning Systems (GPS) and are believed to be accurate to within 10 m. They are shown as: B-1, best quality data from BGS website source; C-1, best quality data from the CR source. The latter include some analytical results (i.e. grain-size distributions, water contents and mechanical properties), and these have largely been collected from records ascribed to the Lynch Hill Gravel Member (LHGR). These analytical data are included in CR-LHGR_grainsize-analysis.xlsx (in directory \Quaternary_BH_database\CR-BH-data\CR-LHGR-grainsize-data).

2. Good quality data: these boreholes are older, but have been drilled within the last 25 years; the sample collection techniques and/or lithological descriptions may be of reduced quality; the grid references may be less accurate but are still believed to be accurate within 50 m (acquired using early GPS in the era of US selective availability) but at worst within 200 m. It might be that the data are as good as those in B-1 category, but where the ground level or grid reference has had to be modified. These are shown as B-2 and C-2.

3. Useful data: these boreholes have information that is generally good, but the quality for mapping is poorer because: the boreholes are a lot older and grid references are less accurate or absent (pre-GPS); these are generally within 100 m, but at worst within 500 m; ground level may be approximate or uncertain; generally, the lithological descriptions are less detailed. Nonetheless, they would be helpful in any mapping or analytical exercises. These are shown as B-3 and C-3.

The dataset described in this note does not include any records from boreholes marked on BGS GeoIndex as ‘confidential’, as far as is known.

The lithostratigraphic codes used in the spreadsheets are those defined and provided by the BGS Lexicon (https://www.bgs.ac.uk/technologies/the-bgs-lexicon-of-named-rock-units). For the majority of boreholes, either the authors or MSc students at Imperial College London (during project work) have correlated and confirmed the lithostratigraphic interpretation in the log with BGS mapped outcrops, or have identified the stratigraphy of bodies logged as ‘uncertain’ or ‘undifferentiated’ or ‘unidentified’ gravel bodies. No subcrop information was provided with the Crossrail data; many of the ‘RTDU’ gravels interpreted in BGS logs indicate that the borehole did not penetrate the subcrop. In some cases, the authors have been able to identify a subcrop from the logged descriptions and notes.

The majority of borehole records are organized and described with the following information: location (TQ square number, easting and northing), ground level (in metres above OD), lithostratigraphic interval thickness and interpreted lithostratigraphic code of the main Quaternary interval, plus a description where present. Where two Quaternary lithostratigraphic units have been recorded as present, these are entered separately (e.g. LOFT with overlying BPGR or KPGR with overlying LASI).

The varied lithostratigraphic units and detailed descriptions have been supplemented, where possible and in separate additional columns, by codes representing the presence of peat (P), cobbles (L), shell clasts (SH), chalk clasts (CH) and fissures (F)/fractures (Fr) and faults (Fa). These again have been derived from the descriptions provided by the drilling contractor or geologist and have been incorporated into BGS-CR_RTD-BH_ALL.xlsx by the authors.

A series of five separate spreadsheets (prefixed Clast-Map-) provides separated records of boreholes where cobbles, shell and chalk clasts are identified, and which are also provided in map- or GIS-ready format and used to produce maps of the spatial distributions of the various clasts (in directory \Quaternary_BH_database\BGS-CR-BH-data\Spreadhseets\Clast-separation). A sixth sheet is provided for borehole records where fissures and fractures have been observed and described in the log descriptions; this also includes the locations where faults are suspected to be present, or where faults are identified by the authors.

The borehole ID numbers in this Technical Note and database all refer to areas by reference to the BNG ‘TQ’ square number, which relates to the Ordnance Survey (OS) classification, using metric six-figure BNG grid references (e.g. [559490, 193760]). The OS division of the UK into 100 km × 100 km squares, where London lies in square TQ, and further subdivision into 10 km × 10 km squares (e.g. TQ 06), is illustrated in Figure 5. Each of the squares, such as TQ 06, is finally subdivided into four geographical regions: NW, NE, SW and SE.

The London Basin Forum (LBF) study has concentrated on the area within the M25 (within OS square TQ) but this study includes several data points located just north of the M25, in OS square ‘TL’, which lies to the north of OS square TQ (these are not shown in Fig. 5).

All geological map information presented in this Technical Note has been extracted from BGS digital 1:500 000 scale published digital solid and drift geological sheets, available on the BGS website at https://www.bgs.ac.uk/information-hub/bgs-maps-portal, and from Digimap (via subscription, https://digimap.edina.ac.uk), as ESRI shape files. The BGS map sheets of outcrops of Mesozoic and Quaternary sediments and subcrop strata are used in the discussion below on the various Quaternary sedimentary packages. The Quaternary lithostratigraphic units that crop out in the Greater London area are illustrated in the simplified map in Figure 6.

The borehole record database ‘cleaned’ as described above and in ‘map-ready’ form was imported into ESRI ArcGIS. The data were then split according to lithostratigraphic code to identify the boreholes in which specific mapped lithostratigraphic units had been interpreted from the logs. These were then plotted as point entities in 2D map form for the 2D visualization of the spatial distribution of each unit. The point features were overlain on the outcrop of the relevant lithostratigraphic unit to which they were likely to be related or sourced from, or to which they had been classified. For example, all boreholes in which KPGR terrace deposits had been interpreted at some depth were plotted and overlain on the surface outcrop of the mapped KPGR, so that any likely mis-identifications could be spotted, and so that likely cases of subcrop KPGR could be discriminated from the outcrop of KPGR.

All the plotted lithogratigraphic units are shown overlain on a basemap consisting of linear features representing the M25 motorway and the tributary rivers of the Thames plus the Thames itself, along with the BGS mapped lithostratigraphic unit(s) that crop out shown as filled (grey) polygons.

As the database comprises exclusively borehole data, it is considered worth providing the reader who is new to Quaternary deposits with descriptions of typical examples of the sedimentary facies, as well as the top and bottom surface features and any internal disturbances, as these features are not always determinable from any particular borehole description. All these features have been well described by Gibbard in the treatise of 1985 from his field observations (Gibbard 1985).

The surface on which the deposits now rest (the bench), the deposits themselves and the top terrace surfaces may or may not be discernible. Apart from the obvious pre-depositional subaerial erosion of the underlying strata and post-depositional erosion of the RTDs themselves, there are many other reasons for local variations in thickness. The RTDs have frequently been disturbed by post-depositional processes, particularly cryoturbation and solifluction. The user therefore needs to be aware of the possibility of such occurrences.

The facies associated with Quaternary deposition, which would probably be present within the RTDs of the River Thames and its tributaries, include the following types: massive or crudely bedded gravel (most common facies); interbedded, tabular, cross-bedded gravel and sand; trough cross-bedded pebbly sand alternating with gravel; scour fill sand (scours are <50 cm deep and <1.5 m wide); and horizontally bedded, often massive, sand (very local extent and <20 cm thick).

Also recorded is the relative morphology, on a local scale, of the basal surface of the gravel or sand deposit (the bench), and the degree of unevenness of that surface. Similarly, the upper surface of a unit can be very variable. The relevant descriptions of these phenomena, summarized below, are from Gibbard (1985):

1. fluvial scouring and channelling:

   •    • ‘irregular channelled surface of the Reading Beds’;

   •    • ‘over 2 m of poorly stratified gravel filling channels scoured into the Reading Beds’;

2. solution of the bedrock and subsequent collapse of the overlying gravels, particularly dolines or funnel-shaped hollows:

   •    • ‘large infilled funnel-shaped hollow contained pebbly clays as well as disturbed gravel and sand. The hollows … extend through … and into the bedrock beneath … [suggesting] that they originated by bedrock solution collapse’;

   •    • ‘the gravel bedding was much disturbed by collapse into a solution hollow in the underlying Chalk’;

3. doming, diapirs and pillars of bedrock protruding up into the overlying gravels:

   •    • ‘the gravel being penetrated’ ‘seemingly from below by irregular pinnacles (up to 2 m high) composed of disintegrated Chalk … diapiric-like structures’;

4. erosion along fault or fracture systems:

   •    • see descriptions below in the section ‘Disturbances to the internal structure of the RTDs’, solution collapse.

Gibbard (1985) has provided useful descriptions of the following as evidence of degradation and dissection of Quaternary deposits: ‘the surface is dissected by dry valleys and degraded by solution collapse’, ‘degraded remnants of a previously thicker aggradation’, or ‘been dissected by later stream erosion’. These descriptions have been used to identify borehole records and lithostratigraphic intervals that show signs of such activities.

The following descriptions are recorded by Gibbard (1985); again, these have been used to identify borehole records and lithostratigraphic intervals that show signs of such activities:

1. cryoturbation (disturbance owing to frost action) causes disturbance within the RTDs:

   •    • ‘The … contact between the deposits being heavily disturbed by cryoturbation’;

   •    • ‘the upper 60 cm of the gravels … was disturbed by cryoturbation’;

   •    • ‘the uppermost 1.8 m was much disturbed’;

2. solifluction (slow downhill flow of soil, in periglacial regions) causes disturbance within the RTDs:

   •    • ‘[the gravels] included large rafts of Reading Beds’;

   •    • ‘the brown silt … appears to represent a slope wash deposit, … from its indistinct stratification and variable thickness’;

3. olution collapse of the underlying bedrock and disturbance of overlying sediments is also common, particularly above Chalk bedrock:

   •    • ‘Observations of steeply dipping strata … showed evidence for large-scale collapse of the deposits into bedrock solution cavities … some of the collapse was post-depositional … shown by a large thickness of the silt … filling a funnel-shaped hollow in the gravels at least 3 m deep and 6 m wide in one place’;

   •    • ‘this suggests that solution occurred during or even prior to gravel and sand deposition;

4. solution collapse associated with faulting and dolines (sinkholes):

   •    • ‘Throughout the pit, the sediments were disturbed by much small-scale normal, step faulting. The faulting increased in intensity downwards and indicated collapse by release from beneath. … It appears … that this and probably the other collapse features result from syn-depositional bedrock collapse’;

   •    • ‘the upper part of the sequence is highly fissured … from its form, position and size, … the basin [probably] originated by bedrock solution. Its funnel-like shape and proportions closely resemble those of dolines … [which are] thought to form by intense solution activity along joint planes or fissures in the rock’;

   •    • ‘The faulting … suggests that the basin was actively subsiding during this time’;

5. cryoturbation and solifluction may be described together in the same deposit:

   •    • ‘The upper 1 m of the gravels was cryoturbated and overlain by 20 cm of very clayey, probably soliflucted gravel’;

6. cryoturbation, solifluction and solution hollows may be described together in the same deposit:

   •    • ‘the upper 1 m of the gravel was cryoturbated, beneath it was undisturbed except by two large hollows filled with brown pebbly clay. These features, up to 2 m wide and 3 m deep, penetrated the entire thickness of the bedded deposits and extended into the Chalk beneath. The form of these hollows suggests that they are solution features.’

Small disconnected areas of RTDs have been identified and described by Gibbard (1985) as ‘remnants on either side of the … valley’, ‘extensively worked outliers’, ‘the dissected surface of the outlier’ and ‘terrace remnants’. The original extents of such areas cannot be determined but the spatial distribution of their occurrence may hint at the course of the palaeo-fluvial channel(s) that once connected them.

Factors that may have affected the original depositional thickness of any RTD unit include, as a minimum:

1. scouring during deposition and post-depositional effects of erosion; many pre- and post-depositional solution and collapse features noted in (4) above;

2. pre-, syn- and post-depositional effects of faulting;

3. effects of extraction, locally for construction projects and regionally by extraction companies.

The lithostratigraphic distribution maps are composed from BGS 1:50 000 solid and drift geological digital map sheets, over which any boreholes with records including the relevant lithostratigraphic unit are overlain. The locations of the boreholes that include specific stratigraphic or lithological features are overlain on relevant outcrop lithostratigraphy in a series of maps that are referred to in each subsection below. All mapping is necessarily limited by the database available:

1. base Quaternary outcrop map;

2. distribution of borehole data against specific lithostratigraphic units (e.g. boreholes encountering KPGR over the KPGR outcrops);

3. distribution of specific lithological features: cobbles, chalk clasts, shell clasts against relevant outcrops;

4. distribution of Buried (Drift-Filled) Hollows (hereafter DFHs) and thick sand and gravel sequences;

5. distribution of tectonic features: fissures, fractures, faults.

The following sections describe each of the RTD stratigraphic units in turn, using first the description from the BGS Lexicon of Named Rock Units, then pertinent descriptive facts from analysis of the database. These facts are based on the raw information provided by the operations team who drilled the borehole and information and/or analysis from the Crossrail contractors. Their descriptions are summarized, from oldest to youngest, in the following sections.

The STGR is classified and characterized by the following summary descriptions of age, lithostratigraphy, thickness distribution, parent unit, previous names and type area and/or section.

Other database information considered relevant to this unit includes the number and location of boreholes that encounter it, data quality, encountered unit thicknesses and the driller's or geologist's log descriptions and comments, where these contain other useful information. Such information is summarized below.

The LOFT is classified and characterized by the following summary descriptions of age, lithostratigraphy, thickness, distribution, parent unit, previous names and type area and/or section.

Other database information considered relevant to this unit includes the number and location of boreholes that encounter it, data quality, encountered unit thicknesses and the driller's or geologist's log descriptions and comments, where these contain other useful information. LOFT is present in about five boreholes in TQ58 NW and 59SE, where it is up to 3.4 m thick. It is described as boulder clay: firm, brown and grey, sandy clay with gravel; no cobbles or shell clasts are recorded but chalk clasts are present in TQ58NW. It overlies London Clay and is overlain by BPGR in TQ58NW.

The BPGR is classified and characterized by the following summary descriptions of age, lithostratigraphy, thickness, distribution, parent unit, previous names and type area and/or section.

Other database information considered relevant to this unit includes the number and location of boreholes that encounter it, data quality, encountered unit thicknesses and the driller's or geologist's log descriptions and comments, where these contain other useful information. Such information is summarized below.

The WOGR is classified and characterized by the following summary descriptions of age, lithostratigraphy, thickness, distribution, previous names and type area and/or section.

Other database information considered relevant to this unit includes the number and location of boreholes that encounter it, data quality, encountered unit thicknesses and the driller's or geologist's log descriptions and comments, where these contain other useful information. Such information is summarized below.

The DHGR is classified and characterized by the following summary descriptions of age, lithostratigraphy, thickness, distribution, parent unit, previous names and type area and/or section.

Other database information considered relevant to this unit includes the number and location of boreholes that encounter it, data quality, encountered unit thicknesses and the driller's or geologist's log descriptions and comments, where these contain other useful information. Such information is summarized below.

The GCGR is classified and characterized by the following summary descriptions of age, lithostratigraphy, thickness, distribution, parent unit, previous names and type area and/or section.

Other database information considered relevant to this unit includes the number and location of boreholes that encounter it, data quality, encountered unit thicknesses and the driller's or geologist's log descriptions and comments, where these contain other useful information. Such information is summarized below.

The BHT is classified and characterized by the following summary descriptions of age, lithostratigraphy, thickness, distribution, parent unit, previous names and type area and/or section.

Other database information considered relevant to this unit includes the number and location of boreholes that encounter it, data quality, encountered unit thicknesses and the driller's or geologist's log descriptions and comments, where these are particularly detailed. Such information is summarized below.

The LHGR is classified and characterized by the following summary descriptions of age, lithostratigraphy, thickness, distribution, parent unit, previous names and type area and/or section.

Other database information considered relevant to this unit includes the number and location of boreholes that encounter it, data quality, encountered unit thicknesses and the driller's or geologist's log descriptions and comments, where these are particularly detailed. Such information is summarized below.

The HAGR is classified and characterized by the following summary descriptions of age, lithostratigraphy, distribution, previous names and type area and/or section.

Other database information considered relevant to this unit includes the number and location of boreholes that encounter it, data quality, encountered unit thicknesses and the driller's or geologist's log descriptions and comments, where these are particularly detailed. Such information is summarized below.

The TPGR is classified and characterized by the following summary descriptions of age, lithostratigraphy, thickness, distribution, parent unit, previous names and type area and/or section.

Other database information considered relevant to this unit includes the number and location of boreholes that encounter it, data quality, encountered unit thicknesses and the driller's or geologist's log descriptions and comments, where these are particularly detailed. Such information is summarized below.

The KPGR is classified and characterized by the following summary descriptions of age, lithostratigraphy, thickness, distribution, parent unit, previous names and type area and/or section.

Other database information considered relevant to this unit includes the number and location of boreholes that encounter it, data quality, encountered unit thicknesses and the driller's or geologist's log descriptions and comments, where these are particularly detailed. Such information is summarized below.

The SHGR is classified and characterized by the following summary descriptions of age, lithostratigraphy, thickness, distribution, parent unit, previous names and type area and/or section.

Other database information considered relevant to this unit includes the number and location of boreholes that encounter it, data quality, encountered unit thicknesses and the driller's or geologist's log descriptions and comments, where these contain other useful information. Such information is summarized below.

Some 1461 borehole logs record the presence of undifferentiated river terrace deposits, dominantly along the lower Thames floodplain but also in central and north London. By comparing their distribution with the mapped superficial outcrop geology, it was hoped to reveal the likely parents of these undifferentiated RTDs and there appears to be a close association with the TPGR (see Fig. 14).

Database information considered relevant to this unit includes the number and location of boreholes that encounter it, data quality, encountered unit thicknesses and the driller's or geologist's log descriptions and comments, where these contain other useful information. Such information is summarized below.

It should be noted that many of the intervals described as ‘brickearth’ are probably associated with one of the following silt members.

The ILSI is classified and characterized by the following summary descriptions of age, lithostratigraphy, thickness, distribution, parent unit, previous names and type area and/or section.

Other database information considered relevant to this unit includes the number and location of boreholes that encounter it, data quality, encountered unit thicknesses and the driller's or geologist's log descriptions and comments, where these contain other useful information. Such information is summarized below.

The CFSI is classified and characterized by the following summary descriptions of age, lithostratigraphy, thickness, distribution, parent unit, previous names and type area and/or section.

Other database information considered relevant to this unit includes the number and location of boreholes that encounter it, data quality, encountered unit thicknesses and the driller's or geologist's log descriptions and comments, where these contain other useful information. Such information is summarized below.

The LASI is classified and characterized by the following summary descriptions of age, lithostratigraphy, thickness, distribution, parent unit, previous names and type area and/or section.

Other database information considered relevant to this unit includes the number and location of boreholes that encounter it, data quality, encountered unit thicknesses and the driller's or geologist's log descriptions and comments, where these contain other useful information. Such information is summarized below.

The ESI is classified and characterized by the following summary descriptions of age, lithostratigraphy, thickness, distribution, parent unit, previous names and type area and/or section.

Other database information considered relevant to this unit includes the number and location of boreholes that encounter it, data quality, encountered unit thicknesses and the driller's or geologist's log descriptions and comments, where these contain other useful information. Such information is summarized below.

Alluvium and peat is present all the way along the flood plain, but downstream of Woolwich particularly thick sequences of alluvium and peat occur overlying KPGR and RTDU sequences, as in TQ47NE and as at TQ [547286/179025] and TQ [547096/179037]; in TQ47NW, as at TQ [542925/179973]; in TQ48SW, as at TQ [542671/180031]; TQ [541423/180882] and at TQ [540306/180868]; also in TQ57NW at TQ [553550/177110] and TQ [552930/179790]; in TQ57NE at TQ [556809/176009]; also in TQ67NW, as at TQ [561709/177148] and TQ [562640/177580]. This area of thick alluvium and peat extends downstream to at least Rainham. Individual peat beds can probably be correlated between boreholes across various parts of this area, although this has not been attempted.

Shell clasts are recorded in c. <5% of the RTDs. They are found in all the RTD members, one in BHT, and several in boreholes from HAGR, LHGR and TPGR, but most are found in the KPGR and RTDU boreholes. They are absent from the earlier Pleistocene to Anglian gravels (STGR, BPGR, GCGR, DHGR, WOGR) and from boreholes with LOFT deposits.

The shell clasts normally occur at, or near, the base of the gravel, which suggests that they have been eroded from underlying shell-bearing strata and included into the gravel deposit from erosion of the subcrop. These underlying strata include London Clay Formation, as at TQ58 [551282/181697], Lambeth Group, Lower Shelly Clay Formation, as at TQ38 [537672/180131], and Thanet Sand Formation, as in TQ48 and in TQ47 as at [545931/179440].

Figure 16 shows the distribution of shell clasts against the outcrop of the London Clay Formation, Harwich Formation and Lambeth Group rocks. As can be seen, there is a very close relationship between shell clasts and these underlying fossiliferous strata, particularly over the Shelly Clays of the Woolwich Formation, Lambeth Group.

The clasts are also recorded at the top of the RTD sequences, as at TQ47 [541460/179730], where they underlie shell-bearing alluvium and may well represent ‘contamination’ during cable percussion drilling, from the overlying alluvium.

Elsewhere shell-rich gravels are present; here shell clasts are recorded throughout thick Quaternary KPGR and RTDU gravel sequences. These thick and shell-rich gravels extend from TQ48SW and TQ47NE, as at [545653/178923], eastwards through TQ58SE, TQ57NW and NE, to TQ67. The presence of shell clasts throughout the RTDU sequences would suggest that rather than being clasts of underlying strata, the shells accumulated during deposition and thus, that these gravels were laid down in estuarine, marine or possibly intertidal conditions. These latter boreholes are often underlain by Thanet Sand Formation or London Clay.

Chalk clasts are present within <5% of the RTD sequences. They are found in the glacial tills of the Lowestoft Formation and all the main RTD Member sediments younger than the LOFT. Like the shell clasts, they are absent from the earlier Pleistocene to Anglian gravels (STGR, BPGR, GCGR, DHGR, WOGR).

Chalk clasts normally occur in the basal gravel beds, and usually are present in those boreholes that overlie, or are close to, a subcrop of Chalk strata; their distribution is shown in Figure 17. Again, the presence of the clasts at the base of the RTD would suggest that it has been eroded from underlying Chalk during erosion and deposition of the gravels. Nowhere are they found at different levels through the sequence of gravels.

Cobbles are recorded in boreholes within all the main sand and gravel RTD Members and their distribution as recorded in borehole log descriptions is shown in Figure 18. They are normally present in the basal beds as would be expected in fluvial deposits. Cobbles are absent from the LOFT tills and the earliest gravels of Pleistocene to Anglian age (STGR, BPGR, GCGR, DHGR and WOGR), except for one borehole within the GCGR deposits, in TQ09NE at [507570/199200], where cobbles occur at the base of the sequence.

Cobbles are present in all the later RTD Members. They typically occur in c. 10–20% of the boreholes and are assumed to represent proximal areas of fluvial deposition, close to glacial fronts or highland areas. Many clusters of boreholes recording cobbles occur in the terrace deposits to the south of the present Thames and these suggest that the Weald–North Downs area was an important uplifted source area relative to the present River Thames area, which was depressed (probably owing to isostatic movements caused by the thick ice sheet to the north).

BHT Member. Approximately 20% of the BHT boreholes record cobbles, these are mostly (c. 80%) in TQ57SE and TQ58), along the ancestral Darent River, flowing from its Wealden source in the Westerham area through Kent to Essex. The frequency of cobbles in TQ57NE boreholes suggests that they were deposited following uplift and erosion of the Wealden area. Cobbles are also present in the BHT deposits in TQ38NE, in the Leytonstone area, between the Lea and Roding Valleys. Clearly, this is a totally separate group of sands and gravels from those of TQ57, but the source area for these deposits may well be from the north.

LHGR Member. Approximately 10% of the LHGR boreholes record cobbles, and these are located mainly (c. 80%) within the LHGR outcrops in TQ28SE, TQ38SW (Paddington to Holborn) and TQ27SE (Streatham), which may represent a lobe of an ice sheet to the north of this area. The remainder of cobbles (c. 20%) are recorded in a few boreholes within TQ57NE and TQ68SE in the Purfleet and Mar Dyke area, further supporting the notion that this area is probably part of the ancestral Darent drainage system.

HAGR Member. Also c. 10% of the boreholes with HAGR deposits contain cobbles. These boreholes are split approximately evenly (c. 50%) between two separate outcrops: that in TQ26NE (Beddington, near Carshalton, Wallington) to the east of the River Wandle and another in TQ38SW (Farringdon) to the west of the lower Lea Valley. This would suggest different source areas; the Beddington area deposits probably had a source from an uplifted southern, Wealden, area whereas the Farringdon area probably had a northerly glacial front source area. Surprisingly, none are recorded from boreholes within the outcrop area in TQ58NW.

TPGR. Approximately 10% of boreholes record cobbles, most of these (c. 70%) in the TPGR outcrop extending from west to east from TQ38SW through TQ38SE, TQ48NW and TQ48SW to TQ58SW and TQ57NE, suggesting a northerly source along a long glacial front to the north. A second cluster occurs from TQ56NE to TQ57SE, part of the Darent drainage system.

KPGR. Approximately 20% of the boreholes with KPGR deposits record cobbles. The majority of those (c. 80%) lie in the outcrop area between TQ37NW and TQ38SW (Victoria) through TQ38SE to TQ48SW (Poplar, Greenwich), to TQ47NE (Abbey Wood, Woolwich) and also in TQ37NE (Lewisham). These extend from west to east and probably reflect proximal fluvial deposition from a glacial front to the north. The remainder form a cluster in TQ06SE, in the Byfleet area on the River Wey, and probably reflect deposition from an uplifted Wealden source area, south of the North Downs.

SHGR. The only two boreholes with cobbles in SHGR deposits lie in TQ06NW (near the M4–M25 junction, near Harmondsworth), which is on the present River Colne.

RTDU. There are many boreholes where the RTD Member is unidentified and, of these, c. 15% are boreholes in which cobbles are recorded. Almost all of these (>95%) lie in the east, within an area defined approximately by TQ48NW (Ilford near the River Roding), TQ47NE (Woolwich), TQ58SW (Hornchurch, Rainham), TQ57NE (Dartford Tunnel area), TQ67 (Gravesend) and TQ68SE (Mucking, Stanford-le-Hope) in the Thames Estuary. Although their age is not known, the fact that they are distal to the Thames source area and are dominantly on the northern edge of the Thames suggests that they are probably related to a northerly glacial front source area.

Buried Hollow or Drift-Filled Hollow are terms given to an increasingly widely recognized geohazard in the UK and particularly in the London Basin. The ‘hollows’ are closed depressions in rockhead (often London Clay and sometimes Chalk), of varying size and shape, which are infilled by a range of Quaternary deposits, usually a non-laminated, non-structured mixture of clays, silts, sands and gravels (Hutchinson 1980; Bricker et al. 2013; Collins et al. 2015; Flynn et al. 2020). Buried Hollow is the term currently preferred by the BGS but DFH is still very widely used. In the spreadsheet BGS-CR_RTD-BH_All-master.xlsx, these features are also commonly referred to by the drilling contractor team as Scour Hollow infill (SCHO). The genesis of Buried (Drift-Filled) Hollows is still under some debate and a number of recent works have revealed structural control, palaeogeographical correlation, relation to disturbance or folding of underlying strata, to movement of water, and that they may be connected with glacial processes.

DFHs can embody a variety of different genetic processes: scour features (deep erosional bases, probably along fracture zones), pingos (a mound of earth-covered ice in periglacial regions), dissolution hollows (sinkholes), diapiric structures (from frost heave and ice wedging) and from liquefaction of sediments owing to fault activity.

Berry (1979) worked on DFHs in central London (South Lambeth in TQ37NW, Battersea, Westminster in TQ27NE, Southwark in TQ38SW), which he concluded to be ‘shallow buried “channels”, now forming elongate closed hollows’; with most of the DFHs forming in the London Clay surface. ‘They often appear to coincide with stream junctions in the Recent drainage pattern. Under-drainage may occur in some depressions through contact with underlying granular Lower Tertiary sediments. In some cases, these deposits appear to have risen above the adjacent levels as diapiric features, possibly at the time when deepening of the hollows occurred.’ Berry's work, on a DFH at Gray's Inn Road, described fine-grained alluvial sequences containing fossils, with silts and clays reworked from London Clay, along with densely packed gravels and over-consolidated reworked London Clay, some of which could represent over-bank channel deposits.

Three particular DFHs, with slopes of c. 20–25°, along the Crossrail route were mapped by Flynn et al. (2020) and this summary is shown in Table 3 (see Figs 19 and 20). Banks et al. (2015) developed an updated geohazard susceptibility map for the London region, which compiles previously mapped DFHs in addition to several newly identified features. These have been included in the map shown in Figure 21. A further DFH is proposed in the Watling Street Area, in TQ57SE, and is compared with Davis’ DFHs in Table 3.

The section shown in Figure 20 shows a very thick section of sands and gravels recorded in the borehole at TQ38SE [538642/180353], with the entire sequence forming one DFH. The log shows that this thick section of DFH sediments is described as ‘Medium dense grey occasionally brown angular to sub-angular fine to coarse flint gravel with much very soft fine to coarse sandy clay’. This sequence is interrupted at −25 m to −27.5 m and at −34.97 m (−26.87 m drilled depth) to −39.90 m (−31.80 m drilled depth) by Chalk, which appears to be in the form of large blocks ‘floating’ within the thick sand–gravel sequence.

The upper section of Chalk (c. 2 m thick) is described as ‘(recovered as) structureless Chalk, composed of fine to coarse gravel sized (weak to moderately weak) chalk and flint fragments with some soft white comminuted chalk matrix and occasional cobbles (possibly SCHO)’. Below that a further c. 5 m section of Chalk is described as ‘Off-white moderately weathered CHALK, moderately weak’. Below c. 7 m, the rock changes character even further, and some detail is provided by the presence of the following features:

• fractures, very closely to medium spaced, infilled up to 40 mm with brown clayey, silty fine sand and rare rounded fine to medium flint gravel;

• fissures at 29.85–29.98 m, 45°, rough, planar, infilled up to 10 mm with sand;

• fissures at 30.00–30.08 m, 45°, rough, planar, infilled up to 30 mm with sand;

• fissures at 30.46–30.50 m, subhorizontal, rough, planar, infilled up to 40 mm with sand;

• fissures at 30.85–31.80 m, very closely spaced, rough, irregular, locally infilled up to 20 mm with brown silty fine sand.

Faulted intervals, as recorded in logs at other locations, have descriptions that show similar patterns and they are always associated with subvertical and subhorizontal fissures and fractures; for example:

• ‘at 27.84–28.08 m, subvertical rough irregular fracture infilled with soft grey clay and at 28.80 m, subhorizontal rough irregular HV fracture’;

• ‘at 31.27 m in WRB, subhorizontal rough SL planar fracture’;

• ‘in TAB at 35.60–35.75 m, subvertical, rough, irregular, fracture; and at 39.00–39.20 m, and at 39.90–40.10 m, subvertical, rough, planar fractures’.

In this database, drift-filled hollows are described in 15–20 boreholes, within the KPGR, and all near the Blackwall Tunnel. It is noticeable that they are all closely associated with fissures and fractures described as very closely spaced, smooth, planar and often cross-cutting and dipping typically from subhorizontal at 10°, 20° or 30°, but occasionally at 45° and 80°. A further 171 borehole records include descriptions of materials that resemble the DFH descriptions and/or that directly refer to the materials as ‘Scour Hollow infill (SCHO)’; these are predominantly in central and east London. The distribution of the DFHs and SCHOs, identified in this project, in addition to a number of Buried Hollows described by Banks et al. (2015) (mainly from descriptions by Berry 1979), are shown in relation to the distribution of all RTDs in Figure 21.

Other typical descriptions of sands and gravels, described on the logs as being likely DFHs, are shown below.

• At TQ38SE [538583/180377], 16.71 m thick: loose, brown, sub-rounded to rounded, generally fine to medium, flint gravel with much medium to coarse sand. Below about 4.0 m, becoming medium dense gravel with rare cobbles. Medium dense, brown medium to coarse sand with much sub-rounded to rounded, fine to medium flint gravel. Below about 7.0 m, becoming slightly orange brown sand and gravel. Medium dense, becoming dense, below about 14.50 m, grey brown generally fine to medium silty sand, with occasional sub-rounded to rounded, fine to medium, flint gravel. Brown grey, fine to medium sand. Grey, thinly laminated, slightly sandy, very clayey, thinly laminated silt. At 16.71–16.72 m and 16.90–16.91 m, very stiff, grey, fissured clay: rounded, smooth, planar, fine. Brown grey, silty sand, firm, brown grey, fine, slightly sandy clay, locally fine to medium, very sandy clay and grey very thinly to thinly bedded, fine sand, becoming locally slightly silty, clayey, sand below 26.7 m and slightly silty with abundant pockets of light brown silty, fine sand below 29.25 m, sub-angular to angular, fine to coarse, flint gravel with cobbles with some slightly fine, sandy clay overlies Chalk.

• At TQ38SE [538675/180282], 19.20 m thick: loose, becoming medium dense, below 2.50 m, brown and grey sub-angular to rounded, flint gravel with some to much medium to coarse sand. Medium dense, brown sand with some to much sub-angular to rounded flint gravel. Brown, slightly clayey, becoming clayey below 13.20 m, sub-angular to rounded flint gravel with much sand. Here the underlying beds of TAB, between 25.05 and 31.46 m, have frequent fissures described as: fissures, rounded, planar, dipping from extremely closely spaced, subhorizontal to 10°, 30° and subvertical, and fractures dipping at 45°, 80° at 31.46 m. Similarly, in the underlying Chalk, fractures are common and described as dipping at 80°, 45°, 80°, 80°, 45° or 70°; these are commonly large, open fractures, smooth, planar, very closely spaced, with infillings up to 40 mm width of comminuted Chalk.

• At TQ38SE [538840/180174], 7.4 m thick: grey clay, locally sandy, with occasional to much angular to rounded, fine to coarse flint gravel. Slight tar odour. Immediately below the gravels, at 27.84 m–28.08 m, the sequence is described as a subvertical rough irregular fracture, infilled with soft grey clay and, at 28.80 m, subhorizontal, rough, irregular fractures. At 31.27 m in Woolwich, Reading Beds, subhorizontal, rough, slickensided, planar fracture. In TAB at 35.60–35.75 m, subvertical, rough, irregular, fracture and, at 39.00–39.20 m and at 39.90–40.10 m, subvertical, rough, planar fractures. Frequent intervals of fracturing are recorded from the TAB at 40.01–42.2 m, 45.43 m and 48.5–48.65 m; these are variously recorded as subvertical or as subhorizontal, rough planar fractures, and dips of 10° and 20° are noted.

DFHs have also been recorded in the BHT. For example, within an area of c. 120 m × 350 m in TQ57SE, there are >30 boreholes that have very thick sequences of DFH material (9.7–27.95 m, average 13.2 m) and are all underlain by Chalk. Many of these boreholes have intervals described as ‘Infill Solution Hollows’; a sample of these are described below.

• At TQ57SE [555668/173665], 27.95 m thick: medium dense to dense, light brown, fine, sandy, silt with a trace of clay. Medium dense, slightly dark brown, mottled light brown, fine sandy clayey silt below 2.2 m. Medium dense, yellow brown, fine, very sandy, silt with some sub-angular to rounded flint gravel. An increasing amount of black mottling especially below 7.50 m. Occasional traces of clay. Dense, angular to rounded, flint gravel with some brown silty fine-grained sand. Medium dense, orange brown, silty, fine- to medium-grained sand with some angular to rounded flint gravel. Traces of clay especially towards top. Becoming more gravelly. Dense, sub-angular to rounded flint gravel with some orange, silty, fine- to medium-grained sand. Occasional flint cobbles.

• At TQ57SE [555675/173828], 23.8 m thick: brown, slightly clayey, silty, fine sand with angular to rounded, fine and medium gravel. Some roots. Very loose to medium dense, brown, clayey silt with some angular to rounded, fine and medium gravel. Sometimes with some fine sand. Some black staining throughout. Medium dense, brown and grey mottled clayey silt. Some black staining. Dense brown clayey silt with angular to rounded fine and medium gravel. Dense to very dense, orange brown, clayey, silty sand with angular to rounded, fine and medium gravel. At 20.50 m sand, generally medium and coarse. Overlying Chalk.

This area is clearly extensively fissured in weak, highly porous Chalk, the combination of fissuring and porosity giving rise to dissolution of the chalk and thus solution collapse structures (or sinkholes in modern parlance). Here the fissuring is important in generating a solution collapse feature, and the presence of extremely closely spaced fissuring in multiple adjacent boreholes would support a fault-related origin (see discussion below under fissures).

A recent study by Toms et al. (2016) on DFHs in the Nine Elms area finds that there are some 26 DFHs identified in the London area, which arrange in depth from a few metres to nearly 100 m, and whose origins are far from certain. This work found that the two DFHs in the Battersea area formed before the Last Glacial Maximum, contrary to published literature, and suggested that one is related to pingos in the London Clay, and the other caused by erosion of a palaeo-channel, and that both may be related to the existence of faults.

Banks et al. (2015) found that there are two groups of buried hollows: scours and ‘rooted’ hollows, and that there is rarely sufficient evidence to discriminate between them. They concluded that a nationwide study is needed to identify and characterize these and other related features.

The distribution of the boreholes with thick sand and gravel sections (>10 m thick), which may be of economic value, is shown in Figure 22. The majority of these boreholes are close to the drift-filled hollows noted above in the Canary Wharf, River Lea Mouth (Limmo) and Blackwall Tunnel areas, near Greenwich and on the southern edge of the Isle of Dogs, the last of which coincides with the folding structure or fault revealed by Crossrail. Another group of boreholes with thick RTD deposits occurs in TQ47NW in the Woolwich Reach area, Dagenham Reach area, both the north and south banks of Barking Reach and further east in the Plumstead, Abbey Wood area.

All these thick sequences are associated with the younger RTDs and lie within or along the banks of the present River Thames. Many of them include descriptions of heterogenerous deposits (i.e. sand, gravel, silt, clay pockets, etc.) and many of them overlie a subcrop (London Clay, Woolwich and Reading Beds, Harwich Formation and Chalk) that is described as having extremely closely spaced fissures and/or subhorizontal and subvertical fractures. The Chalk subcrop frequently shows expansion of the fissures to >2 cm, presumably through dissolution of the Chalk, and often these contain comminuted Chalk or other fine sediment. Thus some of these may also be drift-filled hollows, but the lack of seismic or 3D mapping prevents a certain answer.

Some of the best quality borehole log records collected during the Crossrail project contain the sorted quantitative grain-size results from particle size dimension analysis (sieving), which was largely performed on LHGR deposits. Data collected include grain-size distributions, water contents and mechanical properties. The grain-size fractions are included in the file CR-LHGR_grainsize-analysis.xlsx. This file includes two separate sheets, as follows.

• CR-LHGR_ALL_grainsize-analyses, which includes all the analyses for each sampled interval in each borehole. Some boreholes have only one sampled interval. Some have multiple sampled intervals, and in these cases, each interval is presented vertically (i.e. with the same columns) with the analyses listed below one another according to depth. The sample analyses are described and arranged by the following information: borehole ID, sample top (metres below ground level (m bgl)), sample reference (ref), and sample type (B, U, LB, D); this information appears in Row4 of this sheet.

• Top-sample – geog-sort(WtoE): here the uppermost interval sampled from each borehole has been sorted by location and arranged in a broad sequence from west to east.

Plotting reveals some complex patterns that have no particular west-to-east trend but some of the samples clearly show removal of the finer fractions of material, suggesting reworking and winnowing, whereas the majority show a much more gradual fining; some show fining upwards and other coarsening upwards.

There is a growing body of work in support of the notion that faults are significantly under-represented on published BGS geological maps of the London area (De Freitas 2009; Aldiss 2013), and that the geology of London is altogether more structurally complex than was previously thought, and more than was previously shown on maps and cross-sections across London (Sumbler 1996; Royse 2010; Royse et al. 2012). Consideration of the palaeo- and current stress regimes of the London area would suggest that structures of Caledonian, Variscan and Alpine origin should be present at depth and that some of these structures are likely to have propagated upwards into the younger stratigraphy (Vandycke and Bergerat 2001). This topic is the subject of other, much more detailed pieces of work (e.g. Morgan et al. 2020).

There is plentiful and widespread evidence of brittle deformation, which is highly likely to have been caused by fault movement, throughout this borehole database of Quaternary lithogratigraphic descriptions from the BGS and Crossrail borehole archive. From each borehole log examined by the authors, any descriptions considered to suggest fault deformation, whether specifically described by driller or geologist, or indirectly inferred from the detailed descriptions of the sediments affected, have been coded to allow the production of a separate spreadsheet of borehole locations where faults and/or fissures are observed; stored in BGS-CR_RTD-BH_All-master.xlsx in the ‘Fissured’ sheet. Here we provide only a selection of the relevant descriptions.

The map in Figure 23 shows the distribution of boreholes with records of fissures or fractures and the location of boreholes where faults are thought to be present, either where the strata are repeated, or where the boreholes are very close and yet the subcrop pattern is completely different.

Descriptions of open fractures within the Chalk subcrop are also common, and differences in adjacent boreholes may be due to faulting or to subcrop activity such as ‘sink-holes’.

Boreholes at TQ38SW [532420/184780] show repeated sequences of LASI over KPGR. These appear to be genuine repeated sequences and are interpreted as reverse faults. They lie on a NNE–SSW trend, which continues SSW with consistent descriptions of highly fissured clay into the Westminster Reach of the River Thames and continues NNE to South Tottenham at the sharp corner of the Lea Valley, where the London Clay is described as extremely closely fissured and where a fault has been identified in two adjacent boreholes. This trend continues NNE along the Lea Valley.

Descriptions of the London Clay Formation in the borehole logs from around the mouth of the River Lea provide clear evidence of fissuring and of mechanical deformation. In this area faults have been proven and have been observed in situ by contractors working on the Crossrail tunnel construction and in the wall of the Canary Wharf station box at c. 30 m below current ground level (R. Ghail, pers. comm. 2017). Some relevant descriptions from logs in this area include the following.

• TQ38SE [539440/180724], −6.66 m to −10.56 m: firm to stiff, grey brown, extremely closely fissured clay. Fissures are subvertical to subhorizontal, slightly undulose.

• TQ38SE [539407/180660], −7.35 m to −9.95 m: stiff, grey brown, fissured clay with occasional sub-angular medium gravel sized nodules of pyrite. Fissures are extremely closely spaced, vertical to subhorizontal, undulose.

• TQ38SE [539497/180982], −8.76 m to −14.06 m: stiff, grey fissured clay with the occasional fissure containing grey fine sand.

• TQ38SE [539497/180633], −7.69 m to −14.99 m: stiff brown fissured clay, with rare sub-rounded medium gravel sized pyrite nodules. Fissures are extremely closely spaced, vertical, subvertical and subhorizontal. Undulose. … Very stiff, grey brown fissured clay with some partings of dark grey silt. Fissures are extremely closely spaced, subvertical and subhorizontal, slightly undulose.

• TQ38SE [539213/180907], −4.04 m to −25.25 m: stiff, very stiff, grey brown closely fissured silty clay with occasional fine sandy partings.

The Chalk in this and other cases is often heavily fissured, fractured and faulted. The descriptions from the in situ Chalk include ‘fissures sub-horizontal, sub-vertical’; ‘fractures extremely to very closely spaced’; ‘fissures, rough, planar, dipping from (extremely closely spaced) sub-horizontal to 10°, 30° and sub-vertical and fractures dipping at 45°, 80°’; ‘fractures are common, dipping at 80°, 45°, 80°, 80°, 45°, 70°, these are commonly large, open fractures, smooth, planar, very closely spaced, with infillings up to 40 mm width comminuted chalk’. Similar descriptions occur in the deeper rafted Chalk block within borehole TQ38SE [538642,180353]. Similar descriptions come from the subcrops of London Clay and Woolwich Bed.

From examination of this database, the occurrence of extremely closely spaced fissures as well as associated pyrite and fissure infill in the subcrop of other closely spaced boreholes is very common. These structures appear to be tectonic in origin and thus proximity to a fault would be a reasonable interpretation in each case; from the distribution of fissuring in the boreholes of this database, faults appear to be present across much of central London.

The constraint or provision of evidence for the timing of faulting is considered of great importance for understanding of the evolution of London and southern England; there is plenty of evidence for recent tectonic activity in this area (Ghail et al. 2015). Work by Morgan et al. (2020) suggested that the London Basin is an uplifted platform, rather than a basin as such, and that its basement geology is just as fractured as elsewhere in southern Britain. The geological evidence provided in the Quaternary lithostratigraphic descriptions in this database strongly supports fault activity during the Quaternary, and this supports continuing work, which also suggests that tectonic uplift and inversion has been occurring in SE England in the last 400 kyr (Blundell 2002; Aldiss et al. 2014; Ghail et al. 2015; Morgan et al. 2020).

This Technical Note provides a database of c. 27 000 borehole records of Quaternary lithostratigraphy, as recorded from drilling through the River Terrace Deposits of Greater London, for assessing the distribution of facies and eroded and deposited materials, and their implications for the reconstruction of Quaternary depositional environments and interpretation of neotectonics, and to aid understanding of geotechnics in this region. This database has been compiled from boreholes drilled by British Geological Survey and for the Crossrail project (in preparation for the construction of the Elizabeth Line). The data presented have been quality assessed, corrected and duplicates removed (where necessary), and have been selected and graded for quality as ‘Best’, ‘Good’ or ‘Useful’ from c. 58 000 such records publicly available in the BGS archive or acquired by the Crossrail project and made available to the authors.

Detailed analysis of these data has revealed how some drainage systems evolved during the Quaternary, the likely sources of their detritus and the nature of their fluvial environments. These analyses have revealed many patterns in the data that might not have otherwise been recognized, such as particularly thick sequences of sand and gravel, or the distributions of shell and chalk clasts that point toward the source material of the deposits. Our analyses also revealed that descriptive evidence of small-scale fracturing in the bored materials is widespread, and the authors suggest that this provides compelling evidence of faulting in the subsurface of London, which supports a growing body of evidence in other recently published literature. These data therefore provide a basis against which evidence from other sources (e.g. the location and nature of Drift-Filled Hollows and presence and character of faulting) can be compared.

There are many people to thank for their help, support and unstinting efforts in gathering together the database and in the production of this technical note; too many to mention here but of particular note are David Entwistle, Ursula Lawrence, John Davis and Michael de Freitas.

EA: conceptualization (equal), data curation (lead), investigation (equal), writing – original draft (lead); PJM: conceptualization (equal), formal analysis (equal), visualization (lead), writing – review & editing (lead)

This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

All data generated or analysed during this study are included with this published article and its supplementary information files.