Hinterland environments of the Late Jurassic northern Weald Basin, England

Reconstructing provenance in sandstones can be challenging, especially when the hinterland palaeogeology is unknown due to burial, diagenesis or weathering of the original outcrops. As sedimentary processes alter the distribution of minerals in the depositional environment, the use of multiple provenance methods reduces uncertainties, because, together, they can account for depositionally‐controlled textural and mineralogical re‐distribution in a basin. Here, we have applied petrography, X‐ray fluorescence geochemistry and sedimentology to understand the provenance of shallow marine Corallian sandstones with the aim of deducing the palaeogeology, palaeoenvironment and sediment distribution within the northern parts of the Upper Jurassic Weald Basin, onshore UK. The Corallian sandstones had a mixed mafic‐felsic (intermediate), metasedimentary recycled orogen source. The hinterland experienced significant physical and chemical weathering under humid conditions. Corallian sandstones were relatively more chemically mature up‐dip and more texturally mature down‐dip. Chemically unstable grains and heavy minerals were relatively concentrated down‐dip. Heterogeneous, sedimentologically‐controlled mineral distribution patterns highlight potential errors which may be made in deriving source‐area maturity. This study is significant as it illustrates the combined roles of provenance and deposition in controlling primary mineral distribution that then influenced the style of burial diagenesis. The work presented here emphasizes the importance of a multi‐proxy approach to improve provenance analysis.


| INTRODUCTION
In ancient sandstones, direct evidence for palaeoenvironmental conditions in the sediment source area is typically scarce and hence is commonly inferred, for example, from the characteristics of sediment from surrounding basins (Sladen & Batten, 1984). Sandstone mineralogy and textural maturity are functions of their hinterland mineralogy, the dispersal paths that link the hinterland to the depositional basin, climate and the sedimentary processes within the depositional basin (Dickinson & Suczek, 1979;Griffiths et al., 2019;Pettijohn et al., 1973;Potter, 1978;Tobin & Schwarzer, 2013). It, therefore, follows that understanding sandstone provenance will potentially unlock information about conditions in the source-area, especially when the palaeogeology cannot be directly studied due to continued hinterland outcrop weathering or burial diagenetic obliteration of primary sedimentary character (Dickinson, 1988;Odom et al., 1976). Understanding sediment source-areas is important as it enhances knowledge of the palaeogeography of sediment's hinterland, as well as palaeocurrent and palaeoslope directions which helps prediction of sediment movement (Pettijohn et al., 1973). Provenance studies can be practically applied to problems including basin evolution and tectonics (Dickinson, 1985), climate and relief (Folk, 1980;Ruffell & Rawson, 1994), discrimination of sandstone petrofacies and source-areas (Ingersoll, 1990), as well as palaeo-drainage and sediment dispersal patterns (Lowe et al., 2011).
Many provenance analysis studies have been undertaken using element geochemistry (Middelburg et al., 1988;Middleton, 1960;Nesbitt et al., 1996;Potter, 1978) and detrital mineralogy (Basu et al., 1975;Blatt, 1967;Dapples et al., 1953;Dickinson, 1985;Lowe et al., 2011;Nesbitt et al., 1996;Reinson, Bureau of Mineral Resources, & Geophysics, 2015) complemented by other techniques such as facies analysis, wireline log analysis and palynology. Most studies have focused heavily on mineralogical and bulk geochemical analyses, for example, Basu et al. (1975), Dickinson and Suczek (1979), Nesbitt et al. (1996). Sedimentary processes, such as hydrologic sorting which can cause mineralogical as well as textural redistribution of sediment (Dokuz & Tanyolu, 2006;Kroonenberg, 1992), are typically overlooked in analysing provenance and source area evolution; it is significant that there have been few studies incorporating sedimentological processes within the depositional basin in F I G U R E 1 Geological map of the Weald Basin showing surface geology major structural trends as well as the London-Brabant Massif. Figure  1a Shows the location of the Palmers Wood and Bletchingley fields which are separated by a fault to the north of the Weald Basin. Figure 1b Schematic cross-section of (a) showing faults and subsurface geologic units. Figure 1c Palaeogeographical map of Upper Oxfordian areas of southern England after Brookfield (1973) showing some facies distribution in areas including the Weald Basin and London-Brabant Massif to the south up to the Pennine Massif in the north.
provenance studies (Odom et al., 1976). This oversight may potentially lead to errors in provenance analysis because sedimentological processes continue to modify the sediment mineralogy derived from a source area (Odom et al., 1976). In this study, we have applied mineralogical, geochemical, wireline and sedimentological analyses to investigate the provenance of the Weald Basin's Upper Jurassic Corallian sandstones. These techniques have been chosen to help determine the mineralogy, degree of weathering and climate in the sediment source-area (Nesbitt et al., 1996;Pettijohn et al., 1973;Potter, 1978;Ruffell & Rawson, 1994). Sedimentological analysis has been used to deduce potential controls of sediment redistribution and mineralogical enrichment in the basin. Wireline logs have allowed an unbiased separation of sand-rich from clay-rich stratigraphic zones.
This study focused on the sand-rich portions of the Corallian sandstones, as defined by neutron-density cross-over plots because precise identification of detrital components by point counting is best achieved in sand-grade sediments compared to mud-grade sediments (Dickinson, 1985). Data  The Corallian contains producible oil in these two accumulations and several wells have been cored; most wells have good quality downhole wireline log data. The LBM has been identified by several authors as the likely source-area for the Weald Basin (Butler & Pullan, 1990;Hawkes et al., 1998;Sladen & Batten, 1984), but the Oxfordian section and sediment supply to the Weald Basin have not been extensively studied. The Palmers Wood and Bletchingley fields are separated by a synsedimentary fault and seem to have subtly contrasting high and lower energy hydrodynamic regimes, sedimentology and sediment dispersal patterns updip and downdip of the fault (Barshep & Worden, 2022). The two fields are relatively close to the London-Brabant Massif (15 km) ( Figure 1). The Corallian (Oxfordian) sandstones are at shallow presentday depths (< 1000 m true vertical depth) and have not been buried deeper than about 1500 m, hence they are not significantly altered by mesodiagenesis (Barshep & Worden, 2021. The lack of burial diagenesis and relative proximity to the sediment source terrain make the F I G U R E 1 (Continued) case study suitable for the reconstruction of the palaeogeology, palaeoclimate and palaeotectonic setting of the Jurassic LBM area, as well as for the investigation of the effects of sedimentological controls on provenance analysis. Our objective is therefore to understand the palaeogeology, palaeoclimate and palaeotectonic setting of the Jurassic LBM area, as well as to investigate the effects of sedimentological controls on provenance analysis. To achieve this, the study will answer the following to the west and the Portsdown-Middleton trend and to the south (Butler & Pullan, 1990;Hansen et al., 2002;Lake & Karner, 1987) ( Figure 1). The geological history of the Weald Basin can be summarized into three phases (i) Palaeozoic deformation of basement rocks (ii) Early Jurassic to Lower Cretaceous basin subsidence and (iii) Cenozoic uplift.
The Palaeozoic deformation of basement rocks occurred during the Variscan (also known as the Hercynian) Orogeny (Hansen et al., 2002;Trueman, 2003) with resulting east-west trending compressional fault development (Figure 1). These faults evolved to become extensional faults resulting in the Mesozoic extensional Weald Basin due to thermal subsidence accompanied by block faulting (Butler & Pullan, 1990). Basin formation was accompanied by the Rhaetic transgression and the deposition of the White Lias, Lower Lias, Middle Lias and Upper Lias units in the Lower Jurassic ( Figure 2) (Andrews, 2014;McLimans & Videtich, 1989;Sellwood et al., 1986).
The Inferior Oolite and Great Oolite Formations were deposited in the Bajocian and Bathonian, respectively, on a carbonate platform formed due to uplift (Lake & Karner, 1987;Talbot, 1973). The Callovian to Lower Oxfordian saw continued rifting and deposition of the transgressive Oxford Clay Formation, Corallian Group, Kimmeridge Clay Formation and the Tithonian Portland sandstones (Figure 2). The sea level then fell causing the sabkha-type Purbeck Group sediments to be deposited at the end of the Jurassic through to the Lower Cretaceous (Butler & Pullan, 1990;Hansen et al., 2002;Radley, 2006).
Deposition of the Purbeck Group was followed by Valanginian Wealden Group clastic sediments (Radley, 2006). It is likely that there was Chalk, Upper and Lower Greensand and Palaeogene clastic sediment deposition, subsequently removed due to uplift and inversion of the E-W inherited Palaeozoic previously extensional faults.
Cenozoic uplift was caused by compressive forces from the south of the basin during the opening of the North Atlantic (Jones, 1999). This episode has been attributed to the Pyrenean Orogeny (Parrish et al., 2018), based on the dating of calcite veins. Cenozoic uplift caused erosion and removal of sediments younger than the Wealden Group (Sellwood et al., 1985), including Chalk, Upper and Lower Greensand and Palaeogene clastic sediments.
The main source-area for sediment in the Weald Basin throughout the Mesozoic was reported to be the London-Brabant Massif (LBM), also known as the London Platform (Figure 1a, c) (Allen, 1975;Butler & Pullan, 1990;Hawkes et al., 1998;Sladen & Batten, 1984).
Rocks from the LBM were weathered and sediments transported southward into the Weald Basin (Allen, 1975;Butler & Pullan, 1990;Hawkes et al., 1998;Sladen & Batten, 1984). The LBM consists of Precambrian crystalline basement and metasedimentary rocks, as well as Ordovician and Silurian volcanic and clastic sedimentary rocks (Rijkers et al., 1993). The LBM is an east-west trending stable structure located beneath southeast England, extending across the southern North Sea and into Belgium (Pharaoh, 2018;Rijkers et al., 1993).
In contrast to the Weald Basin to the south, the LBM did not subside in the Mesozoic, as it had minimal Variscan-induced internal structures, in contrast to the Weald Basin (Chadwick, 1985). The LBM thus remained as a structural high to the north of the Weald Basin and has been reported to be the primary sediment source for the Upper Jurassic Corallian sandstones (Hawkes et al., 1998).

| Core log analysis
High-resolution sedimentary core logging, recording cm-and mmscale features, was undertaken on core from three wells, Bletchingley

| Wireline log and core analysis
A suite of modern conventional wireline logs was available for each well including calliper, gamma, compressional sonic, bulk density, neutron and deep and shallow resistivity logs. Wireline data were reported every 15 cm, as this reflects the minimum stratigraphic (vertical) resolution of the suite of logging tools. Core analysis data were also available for each well, including porosity, permeability and grain density data. Neutron-density cross-over diagrams were plotted using Rstudio to define sand-rich areas, that is, sand packages in the logs similar to pay zones in conventional petrophysical analysis (Tiab & Donaldson, 2012). Core-to-log depth shifts were corrected by optimizing the match of both core porosity to density logs and the sedimentary log-determined presence of mudstones to gamma ray logs ( Figure 3).

| Handheld XRF analysis
One hundred and eighty-nine sample points were analysed from the sand packages in BL5, PW3 and PW7 using a Thermofisher Niton XL3T GOLDD (Geometrically Optimized Large Area Drift Detector) handheld X-ray fluorescence (XRF) device. The XRF sample points include 37 from PW7, 60 from BL5 and 27 from PW3. XRF analysis points were directly matched to core analysis plug points to allow XRF data to be related to core analysis data.
Before measurements, the handheld XRF was tested for precision by taking ten readings on the same sample point for a duration of 50 s, 100 s, 150 s and 200 s and the data were compared for repeatability. Repeatability was calculated by comparing the standard deviations for all ten readings for silicon, iron, aluminium, calcium, zircon and titanium. The data showed the least deviations at ≥150 s of analysis duration; consequently, a sampling time of 150 s was chosen for the whole core measurements to optimize precision and measurement duration. Forty-three elements were analysed although some readings, especially trace elements, were below detection.
Data from XRF analyses were assessed using geochemical proxies for lithology, provenance and weathering for clastic sediments.

| Optical petrography
Optical microscopy and point counting were carried out on 51 polished sections from PW7, PW3 and BL5 using an Olympus BX51 transmitted light microscope. Analyses were undertaken to determine mineralogy, textural relationships and cement types. Thirty-three samples from the F I G U R E 3 Sedimentary and wilogs from PW7 (left) PW3 (centre) and BL5 (right). The sedimentary logs show the lithology and sedimentary structures in the Corallian sandstones. The NPHI_RHOB (neutron porosity and bulk density) cross-over plots separate the sand-rich from the clay-rich sections of the sedimentary logs. The sand-rich sections are labelled as PW7-1, PW7-2, PW3-1, PW3-2, BL5-1, BL5-2 and BL5-3. Lithostratigraphic correlation shows that none of the sand packages correlates with sand package BL5-3 which suggests that BL5-3 has a different genetic relationship to the other sand packages.
sand-rich zones were selected for point counting using an Olympus BX53M microscope aided by Conwy Valley Systems Petrog software with samples mounted on a Petrog stage. The samples analysed include 18 polished sections from BL5, five from PW3 and ten from PW7. Three hundred points were analysed for each polished section using a x10 objective, for a statistically representative analysis of each polished section. Point counting focused on differentiating diagenetic minerals from detrital minerals with emphasis on polycrystalline and monocrystalline quartz, feldspars, lithic grains and cement types.

| SEM analysis
A bench-top Hitachi TM3000 scanning electron microscope was used to undertake Back-Scattered Scanning Electron Microscopy (BSEM) at the University of Liverpool's Central Teaching Labs (CTL). The BSEM analysis was carried out to determine mineralogy, mineral/cement types and textural relationships.

| SEM-EDS
Scanning electron microscope-energy dispersive spectroscopy (SEM-EDS) imaging and analysis was carried out for automated quantitative mineral analysis . The SEM-EDS instrument used in this study was an FEI WellSite QEMSCAN, at the University of Liverpool, using a tungsten-filament, operating at 15 kV and equipped with two Bruker EDS detectors .
The technique makes use of a Scanning Electron Microscope (SEM) equipped with two or more high-speed Energy Dispersive X-Ray Spectroscopy (EDS) detectors (Armitage et al., 2011;Barshep & Worden, 2021;Worden & Utley, 2022;Worden, Utley et al., 2018) to give a quantitative proportion of mineral assemblages. In this study, the ideal minimum practical spacing resolution of 2 μm was used for higher resolution analyses and 20 μm spacing was used where average mineralogy across a whole polished section was required. The identification and quantification of minerals were undertaken using a predefined mineral database, known as Species Identification Protocol T A B L E 1 Summary of selected geochemical proxies for lithology, provenance and weathering.

Indicator
Elements The Si/(Si + Al) index is a useful indicator of clastic or shale/ clay content. A high index is indicative of elevated sandgrade sediment content and a low index is indicative of elevated clay mineral content.
Herron, (1986), (1988), Herron & Herron, (1990) Fe, K A high Fe/(Fe + K) index is indicative of elevated Fe-rich clay relative to K-rich clays. In sand-rich sediment, Fe and K oxide ratios have also been applied to differentiate arenaceous and lithic sandstones from arkosic sandstones.
2 Provenance Mafic indicators: Ti, Fe, Va, Cr, Ni, Nb, Co, Sc Clastic sediments with high mafic concentrations are indicative of mafic-dominated provenance while elevated felsic concentrations suggest felsic provenance.

Sand packages
Average SEM-EDS fraction Average XRF proxies  F I G U R E 4 shows proxies for provenance, bulk lithology, mineralogy and weathering in the Corallian sandstones. The elements are expressed in atomic fractions to discount for quartz dilution and vectors have been included for granite and basalt using molar proportions calculated by Krauskopf (1979). Figure 4a Compares Ti and Al content as proxies for mafic versus felsic provenance. Sample points plot between the basalt and granite lines to indicate a source-area lithology of mixed mafic-felsic (intermediate) composition. Sample points cluster close to the origin to suggest hydrologic enrichment of quartz. Figure 4b Sample points on Al vs. Si plot predominantly away from the basalt and granite line and cluster closer to the quartz line to underscore vast quartz enrichment by hydrologic processes. Figure 4c On a plot of K vs. Al, most sample points cluster around the graph's origin to indicate hydrologic enrichment of quartz. The elevated Al content is indicative of intense weathering. Figure  4d Comparison of Al and Fe, reflects local Fe enrichment in the sediments. The many samples that plot away from the granite and basalt lines suggest Fe-enrichment by secondary processes such as weathering and diagenesis. Figure 4e The Ti-Fe plot shows many sample points clustered along the basalt and granite lines to indicate significant mafic input to the sediments regardless of the protolith. The plot indicates elevated Ferich rather than Ti-rich mafic influence on the sediments. Figure  To objectively delineate sand-rich, from clay-rich, intervals, neutron-density cross-over plots (Rider, 1986)  F I G U R E 5 Box plots show selected element indices by sand packages calculated using their ppm concentration. The grey lines in the boxes are the mean values for each well. Figure 5a shows a high silica index for all sand packages indicative of vast quartz enrichment due to hydrological sorting. Figure 5b indicates elevated Al content due to intense weathering in the sand packages similar to the chemical index of alteration of Nesbitt and Young (1982). BL5-1, 2 and 3 have elevated K values which suggests less intense weathering in the Bletchingley area. Figure 5c also emphasizes intense weathering by the enrichment of the more resistant Zr relative to the more mobile Rb. Figure 5d shows differential sedimentological Fe-enrichment in the sand packages. Figure 5e shows the Ti-Al index dominant intermediate mean values between the mafic and felsic lines. The mafic and felsic lines are drawn based on the ppm concentration of elements in basalt and granite respectively as reported in Krauskopf (1979).
F I G U R E 6 Petrographic images which highlight the main mineralogy in the Corallian sandstones. Figure 6a is a cross-polarized image from PW3 (1083.1 m) highlighting monocrystalline and polycrystalline quartz, calcite cement, pyrite cement and a bivalve bioclast. The monocrystalline quartz at the top centre-left shows undulose extinction typical of metamorphic quartz. Figure 6b is from PW7 (1131.9 m) showing pervasive calcite cement and several bivalve bioclasts under different stages of neomorphism and/or dissolution to form calcite cement. Figure 6c) is from BL5_ (2199.92 m), and it shows a pervasive siderite matrix in a fine-to medium-grained sandstone. Figure 6d is from PW7 (1138.04 m), and it shows a detrital matrix draped over a fine-grained sandstone. Abbreviations: Qtz: quartz, mqtz: monocrystalline quartz, pqtz: polycrystalline quartz, Eqtz: embayed quartz; cal: calcite, py: pyrite, ksp: K-feldspar, dol: dolomite, brt: berthierine, bi: bivalve. Optical petrography suggests a quartz-dominated felsic origin with potential significant sediment transport or reworking before deposition. The occurrence of iron-rich minerals also suggests significant input of iron from the source-area or sedimentological Fe-enrichment.

| Geochemical analysis
To compare the composition of sediments in the various sand packages, we have used numerous element cross-plots as well as geochemical indices (Table 2). Indices and ratios tend to be better than element concentrations for studying clastic sediments as they remove the effects of variable dilution by dominant minerals such as calcite and quartz. Indices are preferred to ratios as they vary from 0 to 1 rather than from infinitely small to infinitely large values. Indices provide proxies for bulk lithology, mineralogy and weathering characteristics and can complement other mineralogical analysis techniques (  (Figure 4a-e) thus revealing mafic and felsic provenance, using end-member values reported in Krauskopf (1979).
Ti and Al were compared as they are proxies for mafic or felsic provenance respectively (Andersson & Worden, 2004).
Monocrystalline quartz is present as rounded to sub-rounded grains (Figure 6a), which locally show undulose extinction under cross-polarized light (Figure 6a). Polycrystalline quartz is present as angular to sub-rounded grains ( Figure 6a) and is recognized under cross-polarized light as a single grain composed of several interlocking crystals each of which has distinct undulose extinction under crosspolarized light (Figure 6a). Quartz grains have embayments filled by early calcite (Figure 6a) which suggests pre-diagenetic or early diagenetic corrosion and embayment.
K-feldspar and plagioclase feldspars are present in small quantities as angular grains in both the coarser-grained and F I G U R E 9 (a) QtFL plot for PW7, PW3 and BL5 showing the relative proportion in per cent of total detrital quartz (Qt), feldspars (F), and lithic grains (L). The plot shows a dominant quartz-rich lithology for the sandstones with minor feldspars and lithic grain fragments. This is indicative of a mature sediment source typical of long sediment transport and/or reworked sediments, including metasediments, that have undergone intense hydrologic sorting. (b) QmFL compositional plot of the Corallian sandstones (after Dickinson, 1985) showing source-area tectonic setting. Monocrystalline quartz shows a dominant proportion above 60% relative to feldspars and lithic grain components (Igneous + sedimentary + metamorphic rock fragments). The sample points plot dominantly in the quartzose recycled orogen region with BL5 having more monocrystalline quartz than PW3 and PW7. Sandstone package BL5-3 has more feldspars and clusters close to the craton interior orogen. (c) The lithic fragments plot dominantly in the metamorphic zone (>90%) with minor sedimentary clasts (<5%) and no igneous grains. The dominance of quartzose sediments and the low feldspar content suggests an intensely weathered hinterland.
Grain size analysis indicates that the Palmer's Wood sand packages have coarser grain sizes than the Bletchingley sand packages ( Figure 10a). This systematic difference in grain size suggests different processes were responsible for the deposition of sandstones in the two fields.

SEM-EDS boxplot analysis confirmed the high quartz content of
≥90% of the detrital grains in all the sand packages (Figure 11a).

| Authigenic mineralogy
Authigenic mineral volumes were determined by optical point counting (Table 3). The dominant authigenic minerals in the Corallian sandstones include calcite, siderite, Fe-clay, pyrite, dolomite, authigenic illite and kaolinite. Other authigenic minerals include traces of quartz cement and authigenic apatite (Figures 6, 8 and Table 4). Calcite is the main authigenic mineral in these sandstones, comprising up to 62%, with an average of 28.7%, of total rock volume (Table 4). Calcite fills intergranular and moldic pore spaces and separates detrital grains from each other; this suggests an early diagenetic origin (Figures 6a, b, 7a) (Barshep & Worden, 2021. Calcite cement is present at different stages of neomorphism of primary bivalve shells ( Figure 6b); the outline of the bivalve shells is locally preserved by micrite envelopes.
T A B L E 4 Summary of authigenic mineralogy for the Corallian sandstones from point counting. Some close spatial relationships can be inferred between detrital and authigenic minerals; for example, authigenic berthierine occurs in close association with detrital biotite (Figure 7b). Also observed, is the presence of elevated quantities of calcite cement in bioclast-rich sections (Figures 6a, b, 7a) but much less calcite cement in less bioclastrich sections (Figure 6c, d).
The QmFL composition of the Corallian sandstones within the quartzose recycled orogen (Figure 9b) suggests a provenance of stratified rocks that were deformed, uplifted and eroded (Dickinson, 1985;Dickinson & Suczek, 1979). Sediments from a recycled orogen are susceptible to significant reworking and can be from different rock types (Dickinson, 1985(Dickinson, , 1988. The medium-grained, monocrystalline quartz grains found in these sediments (Table 3, Figures 6a, 9b) indicate a primary metamorphic or plutonic igneous source-area (Pettijohn et al., 1973). The sediment maturity observed in the Corallian sandstones (Figure 9a, b), however, is not necessarily typical of the relatively short transport distance (15 km from the LBM to the study area) for plutonic rocks as such high maturity typically results from either intense weathering, significant (100-1000 s of km) transport (Damuth & Fairbridge, 1970), or sediment recycling (Kroonenberg, 1992;Milliman et al., 1975;Pettijohn et al., 1973;Potter, 1978;Tucker, 1981). Primary plutonic rocks would require significant alteration and hydrologic sorting to reach the level of maturity observed in the Corallian sandstones.
Hydrologic sorting typically causes the separation of grain size classes of specific minerals, sorting of minerals by density and enrichment in quartz content (Kroonenberg, 1992). Evidence for hydrologic sorting in the Corallian sandstones include significant quartz enrichment Lithic grains, such as polycrystalline quartz and minor quartzite rock fragments, present in the Corallian sandstones (Figures 6a, 9c,   10c and Table 3) are indicative of metamorphic rocks in the sourcearea (Basu et al., 1975;Blatt, 1967;Pettijohn et al., 1973). Polycrystalline quartz is susceptible to mechanical weathering and typically disintegrates into monocrystalline quartz in mature sediments (Basu et al., 1975;Dickinson, 1985), hence polycrystalline quartz content is expected to be relatively low in mature sediments (Pettijohn et al., 1973). The high polycrystalline quartz content (up to 16%) in the mature Corallian sandstones suggests a much higher original polycrystalline quartz fraction, hence a significant metamorphic influence on provenance. The presence of lithic grains of sedimentary and metamorphic origin in the Corallian sandstones (Table 3) implies the reworking of sedimentary or metamorphic rocks (Dickinson & Suczek, 1979). Given the evidence for sedimentary and metamorphic input, as well as sediment recycling and the absence of igneous grains (Figure 9c), the Corallian sediments were thus most likely sourced from a metasedimentary source-area.
The Ti-Al cross-plot has most sample points between the granite and basalt trajectories indicating a predominant mixed mafic-felsic provenance (Figure 4a), which includes metamorphosed equivalents of primary mafic and felsic rocks. Ti and Al are useful indicators of mafic and felsic provenance (Andersson & Worden, 2004) as Ti is enriched in basaltic rocks and Al is enriched in felsic rocks (Bhattacharjee & Mondal, 2021;Krauskopf, 1979). Ti and Al are both relatively stable and immobile elements under conditions of weathering and transport, hence are useful indicators of provenance in sediments (Hallberg, 1984).

| Tectonic settings
The average Qm 76.5 F 2.3 L 21.2 composition of the Corallian sandstones is indicative of a recycled orogen, for example where a fold and thrust belt has supplied the sediment. This is consistent with the history of the London Brabant Massif (LBM). The LBM is composed of Proterozoic to the early Palaeozoic crystalline basement (Rijkers et al., 1993).
These rocks are broadly similar to outcropping folded and faulted metasedimentary rocks in north-western England, including the Silurian Windermere Super Group (WSG). The WSG is dominated by lowgrade metamorphosed clastic sediments (Kneller et al., 1994;Moseley, 1978). Sediments from BL5 show more monocrystalline quartz than PW3 and PW7 and cluster closer to the craton interior orogen boundary ( Figure 9b); this may indicate subtly different sources of sediment.
The stratigraphically youngest sand package in BL5, BL5-3, has more feldspars ( Table 2) than all the other sand packages and clusters closest to the craton interior orogeny on the provenance diagram ( Figure 9b). BL5-3 was deposited during a period of accelerated fault movement (Barshep & Worden, 2022). The higher feldspar content in BL5-3 is therefore probably the result of increased subsidence, which resulted in a steeper gradient thereby increasing erosion rates and sediment supply, leaving less time for feldspar alteration (Leeder et al., 1998;Mack, 2003).

| Climate and weathering
Mineralogical proxies in the Corallian sandstones at BL5, PW3 and Kaolinite is indicative of warm tropical to sub-tropical climates with high rainfall rates and vegetation (Burley & Worden, 2003). This is because, the decay of organic matter under humid conditions creates acidity that enhances the chemical weathering of alumino-silicate minerals to produce kaolinite (Burley & Worden, 2003;Hallam, 1975).
The presence of kaolinite in detrital clay, the near absence of smectite and the lack of evaporite minerals, such as gypsum/anhydrite (Figures 6,7,8), indicates humid conditions in the Upper Jurassic hinterland (Barshep & Worden, 2022).
Fe-dolomite, siderite, pyrite and berthierine in marine sediment typically indicate warm, humid, continental weathering as the development of these iron-rich minerals in shallow marine environments involves terrigenous weathering under humid conditions to form lateritic soils, which, when eroded, transports iron into the marine environment (Hallam, 1984;Odin, 1988;Tucker, 1981;Worden et al., 2020).
The intense chemical weathering and iron enrichment in the source areas are emphasized by the low value of the K/Al index (Figures 4c, 5b), as enrichment of aluminium is typical of advanced chemical alteration (Krissek & Kyle, 2000;Middleton, 1960). Other indicators of intense weathering include the dominance of quartz over feldspars (Figure 9a), depletion of Rb relative to Zr (Figure 4f) and the presence of altered K-feldspar ( Figure 8a) (Damuth & Fairbridge, 1970;Dickinson, 1985;Milliman et al., 1975). Rb and Zr have different mobility during weathering with Zr (in zircon) relatively more stable than Rb (which resides as a trace element in potassiumbearing minerals such as K-feldspar and mica) (Everett et al., 2019).
The elevated zircon and feldspar concentrations in BL5 compared to PW3 and PW7 suggests intense mechanical weathering which typically does not completely remove feldspars from sediments but causes a reduction in feldspar grain-size (Odom et al., 1976). The mechanical stability of zircon also causes zircon enrichment during physical weathering (Hubert, 1962) (Cleary & Conolly, 1971;Potter, 1978) as well as volcanic provenance (Folk, 1968). The sediments here do not have volcanic lithic grains (Figure 9c) and locally occur with rounded edges (Figures 6a, b, 7a, b, d) typical of sedimentary transport (Potter, 1978) and precluding volcanic origin to embayments and corrosion. Embayments are therefore interpreted to be formed in sediments weathered in tropical to subtropical conditions.

| Sediment reworking and redistribution
The Corallian sandstones reveal mineralogical and textural differences between the Bletchingley and Palmers Wood areas which suggest either different tectonic influences, changes in source-area weathering characteristics (Dickinson, 1985) or changes in depositional processes (Odom et al., 1976). Hydrologic sorting during transportation and deposition of sediments may have significantly affected the distribution and texture of Corallian sandstone sediments in the study area.
Compared to the Palmers Wood wells, BL5 has elevated K-feldspar  (Figure 3c) (Barshep & Worden, 2022). The textural and mineralogical differences across the fault are most likely due to the hydrodynamic differences that are capable of causing differences in sediment characteristics and distribution (Folk, 1956;Kroonenberg, 1992;Odom et al., 1976).
It is not uncommon for mature sandstones to have elevated feldspar content in their finer-grained fractions (Odom et al., 1976), as oberved in BL5, because feldspars are not destroyed when abraded during mechanical weathering but become finer-grained, hence mineralogical maturity differences might be due to depositional grain size and sorting effects rather than mineralogy or hinterland climates (Odom et al., 1976). In addition, fine-grained feldspars can be carried in suspension for long distances in large water bodies (Milliman et al., 1975). The high energy, nearshore setting of the Palmers Wood area was susceptible to intense mechanical sediment reworking, hence hydrological sorting would cause (i) deposition of coarsergrained fractions in the higher energy Palmers Wood area (Figure 10a) (ii) potential mechanical breakdown of feldspars to finer-grains which remained in suspension in the higher energy Palmers Wood area, hence lower feldspar content (Figure 11b, c) and (iii) less mica, detrital illite and kaolinite in the Palmers Wood area as they would remain in suspension under high energy conditions (Figure 11f-i). The larger grain sizes in the Palmers Wood area and elevated polycrystalline quartz content indicate intense chemical weathering, capable of retaining coarser-grained sizes, diminished feldspar content and retention of polycrystalline grains that did not mechanically disintegrate during transport (Potter, 1978). Conversely, finer-grained mechanically unstable grains such as feldspars could be transported in suspension and deposited under lower energy conditions in the Bletchingley area; this would increase the feldspar content (Figures 5b,c,11b,c) and the fine-grained mineral content present, for example, in detrital clay and including illite, muscovite, biotite and kaolinite (Figure 11f, g, h, j). Evidence of intense physical weathering in the Bletchingley sediments includes elevated monocrystalline quartz (Figure 10b), zircon ( Figure 11d) and rutile (Figure 11e). Zircon and rutile have a high resistance to weathering and, unlike unstable minerals like feldspars, can be relatively enriched, especially in the fine-grained fractions of weathered, transported and recycled sediments (Carroll, 1953;Dokuz & Tanyolu, 2006;Hallberg, 1984;Hubert, 1962;Kroonenberg, 1992;Pettijohn et al., 1973). Their abundance in the finer-grained sediments in BL5 is thus attributed to intense recycling and sorting, for example, Dokuz and Tanyolu (2006). The elevated zircon, rutile, K-feldspar, plagioclase feldspar, biotite, muscovite, monocrystalline quartz and finer-grained sizes in the Bletchingley area were therefore not caused by source-area evolution but by hydrological sorting.
In summary, the comparison of geochemical mafic-felsic endmembers to investigate provenance and hydrological sorting is novel and improves on results derived from petrographic analysis as well as highlighting hydrologic sorting. The incorporation of sedimentological analysis in this study highlighted the importance of sedimentary processes which caused redistribution of sediments in the basin, which would otherwise have been attributed to source area migration, and hence a misinterpretation of the effects of provenance on sediment supply. This work shows that it is important that sedimentological processes are included in provenance analyses.
The study is necessarily limited by its geographic extent as it is  probably enriched through transport viaFe-rich pathways and concentrated in the basin by the localised early diagenetic development of Fe-rich minerals.
5. Spatial, compositional and textural differences in sediment distribution were caused by hydrodynamic sorting, not source-area change. Tectonic evolution caused changes in depositional energy and enhanced the effects of hydrodynamic sorting.
6. Integrating depositional controls into provenance analysis has here enhanced understanding of sediment source-area geology, sediment migration, estimation of hinterland sediment maturity, and generally enhanced interpretation of results from bulk mineralogy and geochemical analysis.

DATA AVAILABILITY STATEMENT
Research data are not shared.