A total of 45 samples were analyzed for their bulk mineralogy. Of those, 39 were additionally analyzed for clay mineralogy. Rock colors are described according to the Munsell color chart (Supplementary Table S1). When possible, subsamples of intercalated red and green or red and bleached were collected from the same sample.

Bulk random mounts to quantify whole-rock mineralogy were prepared generally following Eberl (2003). Sample preparation involved crushing with a hammer or percussion mortar until 1–2 g passed through a 0.4 mm sieve, followed by 5 min of micronizing with Y-Zr-oxide and 5–7 mL of methanol and drying. Once dry, the sample was stirred to randomize orientation and filtered through the micronizing sieve once more. Finally, the sample was poured into a glass cavity mount flush with the rim while avoiding compaction into the cavity to minimize preferred orientation and analyzed.

Separation of the clay-size fraction followed the methods of Moore and Reynolds (1997). Rock samples were crushed, and 1–2 g were filtered through a 0.4 mm sieve (Eberl, 2003). The material was then transferred to a glass beaker, combined with 150 mL distilled water and a “pinch” of sodium (hexa)metaphosphate, sonicated for 5 min to disaggregate the rock fragments, and then centrifuged at 800 rpm for 3.5 min. While avoiding the pelleted material at the bottom, the liquid containing the clay-sized (2 µm) fraction was decanted after centrifuging to make the oriented mount. Several samples were also processed by the standard Citrate–Bicarbonate–Dithionite (CBD) treatment to dissolve the iron oxide cements (Soukup et al., 2008) for particle size analysis. Although the laser particle size results were not generally successful, this process created a suspended gel of clay in some cases that could be collected with a spatula for XRD analysis.

Random clay mounts are used primarily to elucidate the di-versus trioctahedral character of phyllosilicate minerals using (hk0) reflections that indicate b-axis dimensions or discriminate between polymorphs, neither of which can be determined in an oriented clay mount. In this case, the liquid containing the 2 µm fraction was dried in the drying oven, stirred to a powder, and mounted into a bulk random mount glass holder.

Oriented clay films were prepared on fused silica slides according to the filter peel method of Drever (1973) as described in Moore and Reynolds (1997). The clay solution was vacuum-filtered onto a 0.2 µm filter and calcium-saturated with 1M CaCl₂.

Samples were analyzed using a Rigaku Ultima IV powder X-ray diffractometer using Cu-Kα radiation with a voltage of 40 kV, current of 44 mA, and curved graphite diffracted-beam monochromator. Data were analyzed using Jade Pro with the ICDD-PDF4+ database. Whole-pattern Rietveld fitting was used for quantitative weight percent determinations. Oriented clay mounts were scanned with a 2Ɵ range from 2°–32° for a count time of 2 s and a step size of 0.02°. Samples were analyzed after air-drying (AD) “glycolation” via exposure to ethylene glycol (EG) vapor in a desiccator overnight (∼24 h but <72 h due to the possibility of total dissociation of swelling clays if present) and following heat treatment at 550°C (HT550) for 1 h.

Biscaye factors (Biscaye, 1965) were applied to all spectra of clay fraction-oriented mounts to convert peak areas to semiquantitative estimates of weight percent (Supplementary Table S4) and then assigned into categories with values of 1 (trace-minor) to 3 (abundant).

Statistical evaluation of the XRD data included: 1) comparison of population means for each mineral using GraphPad Prism to perform ANOVA with a post hoc Tukey’s test, 2) Hierarchical clustering of normalized bulk XRD weight percent data via k-means clustering in Tibco Spotfire, and 3) principal components analysis of semiquantitative mineral abundances in GraphPad Prism. Prior to cluster calculation, we removed trace minerals from the sample sets and normalized weight percent values from the whole-rock XRD to 100%.

Petrography is used primarily to understand mineral associations and alteration in the Cutler Formation. Splits of the same Cutler rock samples used for XRD were sent to an outside vendor for preparation of standard microprobe-polished thin sections. Thin section samples were specifically selected to capture the red/green and red/bleached redox interfaces. A Zeiss Axio Imager.Z1 stereo microscope was used to analyze all 22 thin sections via plane-polarized light and cross-polarized light to determine grain-size variations, sorting, roundness, overall mineralogy, and mineral associations. The thin sections were carbon-coated and examined via scanning electron microscopy (SEM) to identify minerals and identify mineral textures, associations, and morphology. SEM was performed on an FEI Quanta 250 with a Bruker energy-dispersive spectrometer (EDS) at the University of Oklahoma. Imaging was performed in backscatter mode for all samples. Selected carbon-coated thin sections were subsequently analyzed on the electron microprobe for a qualitative interpretation of iron oxides, micas, and chlorites. Energy-dispersive X-ray analysis (EDS) elemental mapping was performed at the University of Oklahoma on a CAMECA SX100 electron probe micro-analyzer. Analysis and mapping employed a 15 kV accelerating voltage and a 20 nA beam current (measured at the Faraday cup). Elemental mapping conditions were set to 100 frames at 200 s per frame for a total acquisition time of 20,000 s.

Selected mineral grains from rock samples and thin sections were analyzed with an InVia Mapping Raman Spectrometer integrated with a Leica DM2500M optical microscope located in the OU School of Geosciences. Spectra were collected at 532 nm and/or 785 nm and analyzed using WiRE Chemometrics software and in comparison with the RRUFF database (Lafuente et al., 2015).

Mössbauer spectra were collected at room temperature (RT), 77 and 12 K using a Web Research Company instrument that included a closed-cycle cryostat SHI-850 obtained from Janis Research Company, Inc., a Sumitomo CKW-21 He compressor unit, and a Ritverc NaI detection system. A 57Co/Rh source (75-mCi, initial strength) was used as the gamma energy source. The raw data were folded into 512 channels to provide a flat background and a zero-velocity position corresponding to the center shift (CS) of a metal Fe foil at RT. Calibration spectra were obtained with an Fe foil placed in the same position as the samples to minimize any geometry errors. Data were modeled with Recoil software using a Voigt-based structural fitting routine (Rancourt and Ping, 1991). The sample preparation and sample holder were identical to the procedures reported in Peretyazhko et al. (2012).

Whole-rock geochemistry was performed by ALS Chemex on 11 representative rock samples to determine major, trace, and rare Earth elements (REEs). Samples were pulverized, then (ALS method ME-MS81d) fused with lithium metaborate at 1,000°C, cooled, and then dissolved in 4% nitric/2% hydrochloric acid solution. Digested samples were analyzed by inductively coupled plasma mass spectrometry (ICP-MS) and ICP-atomic emission spectroscopy.

The red, green, and bleached rocks contain varying amounts of chlorite, smectite, illite + mica, plagioclase, K-feldspar, calcite, and quartz plus minor gypsum and epidote (Figure 7; Supplementary Table S2). The proximal Cutler Formation has abundant mica, dominated by biotite. Identification of the heavy minerals is primarily from thin-section (and some XRD) analysis due to their relatively low abundance (<1%) in most samples. Epidote is the most abundant heavy mineral, detected via XRD in one red sample and three green samples at 0.3–7.6 wt% (Supplementary Table S2). Additionally, the two sampled Precambrian basement rocks contain detectable epidote (1218-26D and 1218-28VMQM) with lower epidote weight percentages (0.3% and 1.3%, respectively) than most red and green Cutler samples.

Counterintuitively, both red and green samples host ferric iron oxide (hematite), verified using XRD and Raman spectroscopy (Supplementary Figure S3). Hematite is the only truly red ferric [Fe(III)] iron oxide, while many Fe(II)-bearing minerals exhibit green colors (Cornell and Schwertmann, 2003). X-ray diffraction subtraction patterns of associated red–green and red-bleached samples revealed high similarities in the red vs. green and red vs. bleached rocks, with similarity percentages ranging from 81% to 97.1%. Hematite abundance decreases as red > green >> bleached. Moreover, chlorite, which is commonly green and typically associated with Fe(II), occurs in both red and green layers.

Statistical comparisons of weight percent means between populations of red (23 samples), green (16), and bleached (4) samples were performed via ANOVA (Figure 7). No statistically significant comparisons could be made with bleached samples due to the low number of samples. Mean chlorite wt% values are significantly higher in the green samples (11.1%) compared with the red samples (1.6%, p-value 0.0283). Other sample means that are not significant by ANOVA are still worth noting, having p-values near 0.2. Hematite averages 3.5% in the red samples, 2.0% in the green samples, and zero in the bleached (Figure 7). There is more illite + mica on average in the red samples (30.2%) than in the green (21.7%) or bleached samples (27.9%). Smectite exhibits the highest abundance in the green samples (7.5%), followed by the bleached (3.7%) and red samples (2%). The mean plagioclase content is highest in the bleached (16.5%) and green samples (14.7%) and lowest in the red samples (12.0%). Potassium feldspar is similar across all rock colors. The quartz content averages lowest in the green samples and highest in the red samples. Epidote occurs primarily in green samples (Figure 7). Calcite is most prevalent in the red samples collected farthest (∼8 km) from the surface expression of basement onlap but remains a minor constituent throughout the formation (present in 21 samples).

The two Precambrian rocks were plotted on a Streckeisen QAP diagram to classify each rock (Streckeisen, 1974). The QAP of samples 1218-26D and 1218-28VMQM are 37:60:3 and 34:34:32, respectively. Sample 1218-26D is identified as an alkali-feldspar granite and is one of the siliciclastic Cutler source rocks in the study area, as determined by the mineralogy and direct contact of this basement rock with the Cutler Formation. The bulk mineralogy of this sample contains 4.6% chlorite, 10.2% illite + mica, 1.3% hematite, and 1.3% epidote. Sample 1218-28VMQM is a potential Cutler Formation source rock identified as a monzo-granite that does not host chlorite or hematite but contains 5.9% illite + mica and 0.3% epidote. Full wt% values are reported in Supplementary Table S2.

Cluster analyses attempted to generate patterns connecting rock sample color, bulk mineralogical compositions, and spatial distribution. Ten clusters were generated, although three were outliers with only one member (Figure 8; Supplementary Figure S4). Perhaps the strongest link between spatial location and mineralogy is cluster 10, in which chlorite is >20% in each sample, with a corresponding decrease in illite + mica. All samples in cluster group 10 are in sections V–VII (highly proximal), have coarse grain sizes, and are described as some variation of green. The inverse relationship between abundances of chlorite and illite + mica indicates that mica was altered to chlorite. Cluster 3 includes two samples that are spatially separated but related along the strike with high smectite:illite + mica ratios, suggesting mica alteration to smectite. Calcite is present from 0% to 10% throughout the samples and only exceeds 20% in five samples, cluster group 1, which are primarily red samples farthest from the basement contact. A general trend of high illite+mica:chlorite ratios occurs in cluster groups 2, 8, and 9, with no general spatial commonalities among the three clusters. Group 2 represents red bed samples from section XII, about 8 km from the onlap contact; groups 8 and 9 are proximal-most samples that trend southeast to northwest from the mouth of Unaweep Canyon to the Colorado–Utah border.

The clay mineralogy of the Cutler Formation consists of abundant illite/smectite, illite, smectite, corrensite, and chlorite plus chlorite/smectite in one sample (53S V-1 Green) without any evidence of kaolinite (Supplementary Tables S3, S4). Examples of representative traces with mineral phase assignments are provided in Supplementary Figure S5. Previous studies of the proximal Cutler also found abundant chlorite, smectite, and illite but suggested the presence of minor kaolinite throughout the formation and lacked discussion on mixed-layer clays (Werner, 1974; Dutta and Suttner, 1986). In some cases, the dominance of smectite or chlorite was inconsistent between whole-rock and clay-size fractions. These differences can be reconciled due to the greater sensitivity of clay fraction analyses to phyllosilicates specifically, along with the likely exclusion of many clay minerals associated with larger biotite grains during clay size separation.

Chlorite/smectite with regular interstratification occurs in a range of samples from across the entire study area (53 III-14C green, 53 IV-244.5 red, 1218-15 red, 1218-10 red and bleached, 63 XII-16.5 red, and 63 XII-37.5 red), only entirely absent from section VI–VIII (Supplementary Table S3; Supplementary Figure S5). A green clay sample, MC-01-1078, collected from a fracture at 328.6 m (1,078 ft) subsurface in the gneiss basement of Massey Core #1 is regularly interstratified chlorite/smectite (Supplementary Figures S6, S7). All samples in sections VII–X plus the western proximal samples contain some amount of smectite. The distal-most section XII samples instead host corrensite and illite/smectite. Smectite distribution in the proximal-most samples at the mouth of Unaweep Canyon is not consistent throughout the samples. Smectite in the air-dried Cutler samples is primarily present in a minimally hydrated state, giving AD peaks around 12.5Å that swell to 16.9Å upon exposure to ethylene glycol. The clay fractions of green and red sections VI–X all have chlorite, smectite, and illite with no mixed-layer clays. Illite/smectite occurs only in proximal sections III and IV and in the distal-most section XII but is non-unique to the red or the green rocks. No bleached layers have illite/smectite, but all have abundant smectite.

Random clay fraction mounts were used to discriminate dioctahedral versus trioctahedral sheets based on the (060) d-spacings. In many cases, the results suggest mixtures of trioctahedral and dioctahedral phyllosilicates and/or are ambiguous. Peaks associated with trioctahedral phyllosilicates are expected, given that biotite is a dominant trioctahedral primary mineral. In two cases, single-phase clay fractions allowed unambiguous assignment of (060) peaks to a single phase. Dioctahedral smectite clay was extracted from sample 1218-10 Green, with a d(060) of 1.50 Å (Supplementary Figures S8, S9). The sample of mint-green ‘corrensite’ basement fracture fill (MC-01-1078) contains two peaks in the d(060) region; one at 1.538Å, which indicates a trioctahedral clay, and one at 1.507Å, which indicates a dioctahedral clay; thus, this green regularly interstratified chlorite/smectite is interpreted to be tri-di-octahedral (Supplementary Figure S10).

Principal component analysis (PCA) was applied to the semiquantitative dataset to assess potential spatial or stratigraphic patterns of clay distribution across the study area (Figure 9). The first two principal components account for 53% of the variance. Generally, samples from within ∼0.5 km of the basement onlap contact (samples with Arabic numeral labels and from sections III and IV) show a strong influence of smectite (lower left quadrant with yellow ellipse, Figure 9) or a mixture of chlorite and illite/smectite and chlorite (right-hand side and upper-right quadrant, Figure 9). Defying simple generalization, two samples from section XII, 8 km from the onlap, also plot along the I/S vector (Figure 9, right-hand side, gray ellipse). Most samples collected farthest distally show the strongest influence of illite+mica and chlorite, with a smaller smectite contribution (Figure 9, upper-left quadrant, gray ellipse). Finally, two red samples 1218-10 (“10”) and 1218-15 (“15”) that were collected more distally/up-section plot strongly toward the corrensite and C/S vectors.

Grain-size variations between the red, green, and bleached sediments were analyzed via optical petrography and SEM. Sections that include red–green and red-bleached redox interfaces consistently demonstrate that the green and bleached areas exhibit a coarser-grained texture relative to finer-grained red regions (Figure 10). For example, sample 53 III-14A has similar volume percentages of biotite in red and green regions, but the biotite grains are ∼1,000 µm in the green and 100–200 µm in the red (Supplementary Figure S11).

Minerals observed in the Cutler samples via petrography are consistent with Werner (1974) and primarily include quartz, micas, rock fragments, feldspars, sparse laumontite, iron oxide grains, various clays, gypsum, and authigenic calcite, as well as heavy-minerals: epidote (var. allanite in some samples), chlorite, tourmaline, rutile, titanite, ilmenite, zircon, magnetite, manganese oxide, and garnet. Micas are primarily biotite or a biotite alteration to clay; however, minor muscovite is present in some samples. Fe oxide grains are pervasive throughout all samples, including hematite ranging up to 500 µm (Supplementary Figure S3). Gypsum cement is present in one sample.

Texturally, samples are immature, with angular-to-sub-angular grains (Figure 12; Supplementary Figure S11, 12). In addition to their generally finer grain size, the red samples are more compacted than the green samples, with oriented biotites, for example (Supplementary Figure S12). In general, the most predominant mineral alteration is the transformation of biotite to smectite or chlorite. We found this relationship holds for all samples investigated in this study: those with lower-than-average illite + mica contents had either elevated chlorite or elevated smectite via bulk XRD (Supplementary Figure S13). Therefore, additional details regarding biotite alteration are warranted.

Biotite ranges in chemical and textural alteration from relatively pristine to completely replaced by clays and Fe-Ti oxides and includes additional mineral intergrowths. Oxide replacement of biotite ranges from a “dusting” of nanoscale islands to micron-scale growths in interlayers to complete replacement by iron oxide (Supplementary Figure S14). Other biotite intergrowths include quartz, apatite, titanite, rutile, fluorite, fluorapatite, barite, and calcite (Supplementary Figure S15). The iron–titanium oxide, titanite, apatite, and quartz replacements occur in the red, green, and bleached samples. Calcite intergrowths occur between biotite interlayers in proximal sample 53 III-14A in both the red and green samples with and without iron oxide replacement. On the other hand, rutile, barite, fluorite, and fluorapatite intergrowths occur only in the green/bleached samples.

Many clays are associated with biotite alteration that is evident via iron and magnesium content plus the loss of K+ ions and addition of Al. As an example, altered biotite from sample 53-14C Green has composition of K_0.8(Mg_1.2, Fe_1.2, Al_0.6)(Si_3.2, Al_0.8)O₁₀(OH)₂ as determined by EPMA (Supplementary Table S5). We observed authigenic smectite and potential epidote formation via SEM in sample 1218-03 Green. The Al-rich smectite is typical of smectites found in many other altered samples (Figure 11). Supplementary Figure S16A illustrates the only example we found to date demonstrating potential epidote authigenesis. The epidote hosts no signs of weathering and occurs within the Na-Mg clay matrix of the sample in tandem with altered biotite (Supplementary Figure S16A). Typically, epidote occurrences we observed in Cutler Formation samples exhibit fractures from mechanical weathering.

Samples in contact with or near the basement at the western mouth of Unaweep Canyon have the highest chlorite content. These samples also exhibit evidence of biotite chloritization, but other samples noted to contain chlorite via XRD simply host spherulitic chlorite throughout the sample, typically in association with biotite (Supplementary Figure S16B). This does not discount detrital chlorite, since chlorite also occurs in the potential source rock, 1218-26D.

Titanium is pervasive in different forms throughout the samples, including evidence for Ti-bearing clays and altered biotites in the green and bleached samples. Ti-oxides occur only as biotite intergrowths in the green and bleached samples. Sample 1218-15 Bleach hosts Fe-Ti-oxide rims around quartz and biotite in the presence of Fe-rich clay that follows an apparent preferential flow pathway for alteration fluids (Supplementary Figure S16C).

Some samples exhibit evidence of a past sulfate-bearing fluid alteration. Sulfur-rich sample 1218-03 Green contains gypsum cement, sulfur-bearing clays, and thenardite, an anhydrous sodium sulfate (Supplementary Figure S17). Sample 53S V-1 has barite crystals plus barite intergrowths in biotite (Supplementary Figure S18). Sample 63-27 includes regions of apparent hematite replacements of magnetite (martite).

Three proximal samples (1218-03 Green, 1218-15 Bleached, and 1218-15 Red) were selected for Mössbauer spectroscopy, providing additional information about the mineralogy and oxidation state of iron in these samples. Therefore, we also include whole-rock geochemistry results for these samples and separately present the XRD results of these three samples. According to XRD bulk mineralogy, sample 1218-03 green contains almost 10 times more smectite but less illite + mica than the bleached and red samples (Figure 13; Supplementary Table S5). The XRD patterns from Sample 1218-15 red and bleached were 96.5% similar, differing mostly in the presence of hematite in the red sample, which is absent in the bleached sample (Supplementary Figure S19, left). The green and bleached XRD patterns were 93.6% similar, with differences clearly visible in the low-angle phyllosilicate region, where the green sample exhibits smectitic >10Å d-spacings compared to the bleached sample dominated by 10Å phases (illite + mica, Figure 13; Supplementary Figure S19, right). The red sample overall has more illite + mica and hematite than the other two samples. The elevated smectite in the green and bleached samples corresponds to the lower amount of 10Å phases (illite + mica).

Whole-rock geochemical analyses of splits from the same three RGB samples demonstrate that most major and trace elements occur in similar concentrations. However, there are notable differences (Table 1): 1) the green and bleached rocks have approximately half as much total iron as the red at 2.98 wt%, 3.19 wt%, and 6.49 wt%, respectively, 2) the green sample has much higher Fe(II)/Fe(III) (green 51.8% > bleached 36.3% > red 29.5%), and 3) the green sample has considerably higher uranium (green 14.35 > red 2.26 > bleached 1.82 ppm U) and vanadium (green 100 > red 44 > bleached 42 ppm V).

Splits of the same samples were analyzed by Mössbauer spectroscopy at room temperature (RT), liquid nitrogen (77 K), and 12 K temperatures to resolve the contributions of crystalline Fe oxides, such as hematite and goethite, from their nanocrystalline counterparts and Fe sequestered in various forms of aluminosilicate minerals, such as phyllosilicates and feldspars (Noel et al., 2019; Zhao et al., 2020), (Supplementary Figure S20). RT spectral modeling of the red sample unambiguously indicates that the bulk of Fe in this sample is hematite (∼60% of total Fe), the only Fe oxide evident in powder XRD (Supplementary Figure S10); additionally, ∼10% of Fe is in epidote (Grodzicki et al., 2001), and the remainder occurs in silicate minerals—most likely phyllosilicates and feldspars (Boiteau et al., 2020), given their detection in XRD. More importantly, the lack of significant difference in spectral features among RT, 77 K, and 12 spectra suggests the sample is free of goethite and easily reducible Fe oxides, such as ferrihydrite. Crystalline hematite, on the other hand, is not evident in the RT Mössbauer spectra of the green and bleached samples. The green sample also contains epidote and Fe in clay/feldspar, with a minor amount of small-particle Fe oxide (* in 77 and 12 K spectra, Supplementary Figure S20). Unambiguous attribution of the small-particle Fe oxide to a particular phase is not possible with the present data; the broad sextet in 77 K spectrum is likely a mixture of small-particle/Al-substituted hematite and maghemite with oxyanion (e.g., silicate) coatings. The proportion of Fe(II)/Fe(III) in the clay/feldspar fraction is significantly higher, approximately 50%. Such a high Fe(II)/Fe(III) ratio compared to the red sample is expected given the influence of reducing conditions. The bleached spectra are generally similar to the green but with clay/feldspar signals dominated by Fe(III).

In general, the k-means clustering analysis of the bulk XRD (Figure 8; Supplementary Figure S4) and the PCA analysis of the clay fraction XRD (Figure 9) showed that mineralogical trends varied spatially across the study area. Pairs of red/green samples were often in the same cluster/group depending on their map locality. For example, bulk XRD Clusters 9 and 10 together overlap with the “Chlorite and illite/smectite dominant” proximal-most zone in the clay fraction samples; these samples show biotite alteration to chlorite. Many of the other clusters do not show a clear correspondence to the clay fraction PCA, although many of the samples that were neither proximal-most nor distal-most occur in Cluster 8. We next discuss the observed mineralogical and textural features related to specific mineral types.

Based on our field and petrographic observations, many proximal Cutler Formation rocks in our study area experienced at least three distinct phases of mineral alteration: 1) oxidative reddening, 2) reductive fluid alteration, forming bleached and green colorations, and 3) sulfate-rich fluid alteration (Figures 13, 14). Alteration due to meteoric infiltration and flow of groundwater from the up-dip Unaweep Canyon region continues. This groundwater contains uranium from contact with the Precambrian crystalline rocks (Faluso, 2021), which can be reduced and accumulated in the Fe(II)-rich green layers.

We suggest that the green-bleached intercalations result from one or more episodes of reducing basinal hydrothermal fluids that migrated upward, preferentially infiltrating more permeable and coarse-grained layers near the (buried) Uncompahgre uplift bounding fault system. Bleached and green layers are inferred to be more permeable than the red layers based on grain size and field observations and to have therefore served as preferential fluid pathways. The proximal green rocks contain a greater proportion of Fe(II), whereas Fe(III) (e.g., Table 1; Supplementary Figure S20) reduced in bleached layers was likely solubilized and transported into adjacent red areas or flushed away.

An alternate model for the reddening of Cutler sediments could be 1) pigmentary hematite formed during penecontemporaneous weathering, 2) further hematization occurred during diagenesis, followed by 3) reduction via hydrothermal/reducing/basinal fluids to form green and bleached layers, and finally, 4) sulfate minerals in pore spaces in the green altered rock suggest the influence of sulfate-bearing fluids that post-date the original greening reduction event.

Conduits for reducing fluids include faults, fractures, and the basement onlap contact with overlying Cutler strata. Fault-associated fluid alteration in the Paradox basin is well established (e.g., Jacobs and Kerr, 1965; Morrison and Parry, 1986; Briet et al., 1990; Bailey et al., 2021; MacIntyre et al., 2023). However, alteration in the proximal-most Cutler is unique owing to 1) the spatial association down-dip from the Unaweep Canyon paleovalley (Soreghan et al., 2005b) and 2) the onlap contact of the Cutler Formation with the Precambrian basement here. We hypothesize that the Precambrian basement/Cutler contact serves as an additional preferential fluid pathway.

Discriminating the timing and extent of regional fluid alteration events remains an active area of research, with historical ideas summarized by Dahlkamp (2010) and Shawe (2011). Potential sources of fluid alteration that might have influenced the proximal Cutler Formation include meteoric syn-depositional and diagenetic groundwaters, Paradox basin fluids expelled during the Permo-Triassic salt tectonic valley collapse (Thompson et al., 2018), large-scale regional groundwater flow (Sanford, 1982), hydrocarbon evolution, dewatering in overlying Mesozoic mudstone during burial and compaction, fluids mobilized during the Laramide orogeny, fluids associated with multiple regional Tertiary volcanic episodes to the east, southeast, south, and southwest, basinal groundwaters brought to the surface by erosional reconfiguring of regional hydrologic gradients, and modern meteoric groundwater (e.g., Barton et al., 2018; Bailey et al., 2021; Bailey et al., 2022; Kim et al., 2022).

In particular, the La Sal Mountain laccolith, located 30–40 km southwest of the study area, caused regional hydrothermal activity during the early Tertiary (Shawe, 2011; Thomson et al., 2015). Chan, Parry, and Bowman (2000) determined the La Sal Mountain activity to have occurred from 28 to 25 Ma; thus, over a span of 4 Ma. This igneous activity heated formation waters in the Paradox Basin that subsequently followed permeable flow paths along fault zones and within the Uravan mineral belt (Shawe, 2011). The reduced (often sulfide and/or hydrocarbon-bearing) formation fluids, in tandem with the water expelled from magmas, interacted with subsurface sediments during flow and led to deposition upon mixing with other more oxidizing fluids at the chemical/redox interfaces (Shawe, 2011). More recent studies (e.g., Bailey et al., 2021; Bailey et al., 2022; Kim et al., 2022; MacIntrye et al., 2023) also highlight the importance of sulfide-bearing fluids from Paradox Basin hydrocarbon maturation as sources of sulfur-rich fluids that migrated upwards extensively along faults in the Cretaceous to Paleogene. Occurrences of sulfate minerals in the green samples from our study support this model.

Clay mineralogy and chemistry in the study area of the proximal Cutler region provided both expected and surprising results. Acidic alteration typically leads to kaolinite formation, as observed across many Paradox Basin deposits influenced by fluid flow (e.g., Barton et al., 2018; Bailey et al., 2021; MacIntyre et al., 2023). In contrast, smectite, tosudite/corrensite, and chlorite are the clay products observed in the proximal Cutler Formation, indicating Mg/Ca-bearing intermediate or basic pH fluid alteration. The tri/di tosudite observed in this study also differs from the di/di corrensite reported along the Lisbon fault (Morrison and Parry, 1986), although dioctahedral smectites occur in both. Likely, the abundance of primary trioctahedral biotite in the proximal Cutler explains the difference. However, the abundance of octahedral Al in proximal Cutler/basement fracture-fill smectite and corrensite implies Al mobility and acidic conditions or perhaps breakdown of kaolinite (Morrison and Parry, 1986). We have no current evidence of kaolinite occurrence in our samples, but the presence of kaolinite during deposition would have large paleoclimatic implications.

The structural and depositional settings of the proximal-most Cutler strata contrast greatly with that of the strata within the Paradox Basin, in that proximal strata onlap Precambrian basement, and—within it—the significant paleorelief and bury the frontal fault zone along which the Uncompahgre highland rose. This setting thus conferred several additional ways in which fluids could have affected the Cutler strata. The proximal-most Cutler strata have been interpreted variously as arid or humid alluvial fan (e.g., Campbell, 1979; Dutta and Suttner, 1986) and as proglacial lacustrine and fluvial (Soreghan et a., 2009), and Dutta and Suttner (1986) interpreted the abundance of chlorite and smectite clays in the Cutler strata to bolster their interpretation of an arid setting. However, this was based on the assumption of Mg-dominated clay precipitation due to pore water concentration of solutes rather than Al-dominated, as observed in our study (e.g., Figure 11; Supplementary Figure S9). In contrast, Soreghan et al.’s (2009) inference of a humid, proglacial setting implies (by modern analog) elevated porewater solutes (e.g., Deuerling et al., 2018) and highly accelerated rates of biotite weathering (Föllmi et al., 2009). This significant contrast in depositional and paleoclimatic interpretation thus carries implications for the evolution of mineral alteration and color development. Studies of weathering associated with modern valley glaciation suggest the unique signatures of proglacial valley groundwaters are oxidative solutions with elevated sulfate and Ca (e.g., Salerno et al., 2016; Torres et al., 2017). Generally, proglacial groundwaters contain dissolved oxygen but are commonly somewhat undersaturated with respect to atmospheric O₂ (Brown et al., 1994; Tranter and Wadhawan, 2014). In contrast, weathering under larger ice sheets tends to be associated with the development of anoxia and reducing groundwaters (Wadham et al., 2010; Graly et al., 2014; Andrews and Jacobson, 2018). Although biotite alteration was observed to be significant both in this study and in modern proglacial settings (e.g., Anderson et al., 2000), the observed authigenic mineralogical signatures in the proximal Cutler Formation are not consistent with reduction in low-temperature meteoric solutions. However, the presence of anoxic groundwater flowing down-gradient along the basement–sediment contact may have uniquely preserved the reducing character of the proximal Cutler sediments. Future investigation of Cutler Formation clay chemistry up- and down-section and proximally/distally would help clarify and test the contrasting climate-related hypotheses of Dutta and Suttner (1986) and Soreghan et al. (2009).

Furthermore, the paleorelief visible in the onlap map pattern documents an embayment or small paleovalley carved into basement. Soreghan et al. (2007), Soreghan et al. (2015a) inferred a much larger (partially buried) paleovalley, occupying modern Unaweep Canyon and potentially over-deepened by Permian glacial excavation a few kilometers northeast from the outcrops of the proximal Cutler Formation. Either way, the paleorelief on the basement implies the potential for hydrologic flow along the basement–Cutler interface, and the generally coarse-grained proximal strata with muddy interbeds—whether alluvial fan or proglacial fan—would have enabled hydrothermal fluids from local faults and fractures to flow through the more permeable green-bleached sediments and strip them of some of their hematite/Fe III) oxides. The interplay of location along a tectonic boundary, proglacial(?) paleovalley sedimentary and hydrologic environment, and basinal fluids makes green/bleaching alteration in the proximal Cutler Formation a unique piece to the puzzle of ancient fluid flow, mineralization, and climate at the edge of the Paradox Basin.