Fluvial deposits form globally important aquifers (e.g. Fitts 2013) and oil and gas reservoirs (Keogh et al. 2007), as well as hosting mineral deposits such as uranium (e.g. Turner-Peterson 1986) and exotic copper (e.g. Maiden et al. 1984). Exploitation of these resources requires understanding of regional fluid flow pathways within fluvial successions. Because of the typically limited availability of subsurface data, controls on regional fluid flow cannot necessarily be determined directly. To determine subsurface fluid flow pathways an understanding of facies distribution is crucial, as this controls sandstone connectivity, permeability and porosity (Renard & Allard 2013).

Objectives and method

We aim to document the relationship between uranium mineralization, facies distribution and fluvial architecture in the Upper Jurassic Salt Wash distributive fluvial system, SW USA. We use the distribution of mineralization as a proxy to assess controls on basin-scale porosity and permeability distribution. These observations have important implications for understanding controls on subsurface fluid flow and will affect the exploration for and exploitation of aquifers, hydrocarbon reservoirs and sandstone-hosted strata-bound mineral deposits.

Uranium mineralization in sandstone-hosted deposits is considered to have been controlled by subsurface fluid flow and is closely related to sandstone body connectivity, porosity and permeability (Sanford 1982, 1992). Uranium-enriched fluids migrate through porous and permeable sandstone strata until precipitation occurs at an interface between oxidized and reduced rocks where two chemically different fluids meet (Abzalov 2012). Massive sandstone bodies are considered to be effective flow conduits, and therefore possess good reservoir qualities, with mineralization mainly limited to areas where permeable and impermeable strata interfinger (Gabelman 1971; Abzalov 2012).

Uranium distribution in the Salt Wash distributive fluvial system provides a proxy for understanding subsurface fluid flow in an outcrop example at a basin scale. The extensive exposure (100 000 km²; Fig. 1) and trends in alluvial architecture (Owen et al. 2015c) provide a well-constrained framework in which to conduct such a study. We integrate facies distribution, alluvial architecture and uranium deposit distribution to assess controls on uranium mineralization. Uranium deposit distribution (Fischer 1968) is plotted against sandstone and channel-belt percentage (Owen et al. 2015c) and compared with variations in fluvial architecture from field observations. An amalgamation ratio (A/R) (Zhang et al. 2013) is calculated to quantify and compare the degree of connectivity present at each location.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

(a) Paleogeographical map of the study area with broad paleocurrent direction (modified from Owen et al. 2015a). (b) Schematic illustration of the lacustrine-humate (Peterson & Turner-Peterson 1980) and the brine interface (Sanford 1992) models. Modified from Sanford (1992). Fluid migration through facies belts within the Salt Wash should be noted. Line of cross-section can be seen in (a).

The Salt Wash distributive fluvial system

The Salt Wash Member of the Late Jurassic Morrison Formation was deposited in a foreland basin (DeCelles 2004) as a distributive fluvial system (for details of key distributive fluvial system trends see Weissmann et al. (2013) and Owen et al. (2015c)). The apex of the Salt Wash system is predicted to be located in present-day NW Arizona (Fig. 1a) (Owen et al. 2015a). The Salt Wash distributive fluvial system is composed lithostratigraphically of relatively proximal facies (Salt Wash Member) that prograded into the basin over the distal facies (Tidwell Member), which underlie the Brushy Basin Member, completing the Morrison Formation (e.g. Owen et al. 2015b). Overall the system shows typical characteristics of distributive fluvial system deposits such as a downstream decrease in sandstone percentage (70–8%), channel presence (67–0%) and channel thickness (15–3.8 m to the last measurable channel) with a concomitant increase in floodplain (38–94%) and lacustrine facies (0.1–7%) from proximal to distal (Owen et al. 2015a,b,c). A downstream change in deposit architecture is also evident. Proximal areas are dominated by amalgamated channel-belt complexes, which become increasingly separated by floodplain deposits downstream, and then pass into floodplain fine deposits with sparse isolated channels (Owen et al. 2015b,c).

Uranium distribution

Uranium in the Salt Wash distributive fluvial system is largely considered to be of the tabular type but roll type deposits are also recognized (Dahlkamp 2010). A description of the ore mineralogy has been given by Thamm et al. (1981). Two modes of ore formation are suggested (Fig. 1b): (1) the lacustrine-humate model (e.g. Peterson & Turner-Peterson 1980); (2) the brine interface model (e.g. Sanford 1982, 1992). For both models it is clear that understanding controls on subsurface groundwater movement within the Salt Wash is critical.

The relationship between known uranium deposit distribution and sandstone percentage is shown in Figure 2, with 92% (108/117) of uranium localities restricted to the 40–55% sandstone contour line and little or no uranium present below 40%. A broader relationship is present when uranium distribution is plotted onto a channel-belt percentage map, with 90% (105/117) of uranium localities falling between the 20 and 50% channel-belt percentage contour lines (Fig. 2b and d).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

(a) Uranium distribution plotted onto sandstone percentage map (modified from Owen et al. 2015c). The majority of uranium falls between 55 and 40% sandstone. (b) Uranium distribution plotted onto channel-belt percentage map (modified from Owen et al. 2015c). The majority of uranium falls between 50 and 20%. (c) Uranium distribution plotted against distance downstream and sand percentage intervals (grouped into 10% intervals). (d) Uranium distribution plotted against distance downstream and channel-belt percentage. Uranium distribution taken from Fischer (1968).

From the 40–55% sandstone percentage and 20–50% channel-belt percentage zones a change in architecture is observed (Fig. 3). The gross-scale architecture at Atkinson Creek is typical of medial distributive fluvial system facies (Fig. 3a), where channel-belt deposits are separated by laterally extensive floodplain deposits. Channel-belt deposits form 27.8% of the successions and average 4.5 m in thickness (maximum 8 m), and are up to 1.3 km in width (Owen et al. 2015c). Storey thickness within the channel-belts ranges from 0.7 to 5.3 m (Owen et al. 2015c). Using methods of Zhang et al. (2013), an A/R of 12% was calculated for Atkinson Creek, suggesting that there is limited but potentially important connectivity between channel-belt packages.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

(a) Architectural panel of Atkinson Creek. (b) Architectural panel of Little Park. The difference in architectural styles should be noted: (a) contains laterally extensive channel-belt deposits that are separated by floodplain fine deposits, which do at times amalgamate, whereas (b) is dominated by floodplain fine deposits with rare channel-belt presence and connectivity. (See Fig. 1a for location of panels.)

Further down-system, a distinctive change in architecture associated with increased floodplain fine material is observed at Little Park (Figs 1a and 3b). Amalgamated channel-belt deposits form 16.3% of the succession, and are on average 3.8 m thick and 800 m wide. An A/R ratio of 0% was calculated, indicating that effective connectivity has been lost at this point in the system.

Discussion

A clear relationship is present between uranium distribution, sandstone percentage, channel-belt percentage (Fig. 2) and fluvial architecture in the Salt Wash system (Fig. 3), indicating a sedimentological (i.e. facies) control on the distribution of uranium. We postulate that uranium distribution is related to down (depositional) dip variations in porosity and permeability, controlled by facies distribution.

Gabelman (1971) noted that areas of high permeability are not the most effective sites for uranium precipitation, as internal porosity and permeability barriers are required for concentration of uranium-enriched fluids. Fluid barriers also need to occur in conjunction with the reducing conditions necessary for uranium mineralization. The lack of uranium in the proximal part of the Salt Wash distributive fluvial system (Fig. 2) is in part considered to be related to the high connectivity of channel-belts, owing to repeated avulsions, channel occupation and reworking (Weissmann et al. 2013; Owen et al. 2015b). An exception to this occurs in the Henry Mountains district (Fig. 2), where <6% of uranium sites occur owing to local variations in subsidence that deflected regional flow (Sanford 1992). Downstream, avulsions occur over a larger area and together with reduced sedimentation rates and channel bifurcation result in separation of the channel-belt sandstones by floodplain deposits (baffles), reducing vertical and lateral channel-belt connectivity (Fig. 3). A lack of uranium NE of Atkinson Creek suggests that channel-belt connectivity, and therefore large-scale (system-scale) fluid flow connectivity, dissipates close to the 40–45% sandstone contour (Fig. 2a). This coincides with a change in regional-scale architecture and a facies transition from medial to distal distributive fluvial system deposits, resulting in compartmentalization of fluid flow in sandstone bodies and precipitation from uranium-rich fluids (Fig. 1b).

Once fluid flow is compartmentalized into discrete channel-belts, internal heterogeneities will play a key role in baffling fluid flow. Meander-belt deposits within channel-belt complexes are reported to be key sites for mineralization in the Salt Wash (Stokes 1954; Ethridge et al. 1980). Sanford (1992) related uranium distribution to a combination of a regional change in sandstone:mudstone ratio, a change from low- to high-sinuosity channels and a change in total thickness. We concur that a large-scale change in sandstone percentage plays a crucial control (Fig. 2a), and here provide quantification of the precise location. However, we relate this to system-scale changes in fluid flow, owing to channel-belt connectivity and architectural changes across a distributive fluvial system rather than changes in sinuosity. Hartley et al. (2015) showed the preservation of an amalgamated meander belt, up-dip of the uranium belt, suggesting that sinuous features are ubiquitous across the system. Trends observed in the Salt Wash are also apparent in the Westwater Canyon Member of the Morrison Formation, which is also interpreted to be a distributive fluvial system (Turner-Peterson 1986), where all the major uranium occurrences are located in mid-fan facies (Kyser & Cuney 2009).

Larue & Hovadik (2006) provided a theoretical model in which reservoir sandstone body connectivity is considered to be good (>90%) when the sandstone percentage is >30%. It is important that the geometry and form of the deposits is also considered, which the amalgamation ratio helps us achieve. We therefore suggest that a higher cut-off of 40% should be used as our data from a rock record example show that effective connectivity between channel-belt deposits starts to diminish at 55% and that by 40% an A/R of 12% is present. However, internal permeability within the channel-belt must be considered and further statistical analysis is needed to test this robustly.

Understanding system-scale porosity and permeability variations is crucial when exploring and understanding migration pathways of key resources. Although other post-depositional factors such as cementation or compaction (Haszeldine et al. 2000) need to be considered, we provide an understanding of primary basin-scale trends and controls. Our unique dataset relating uranium distribution to sandstone percentage allows context to be given to the uranium deposits, improving understanding of fluid flow in distributive fluvial system deposits. Because of their quantified nature, results from this study can be related directly to subsurface datasets, aiding exploration and recovery of key resources.

Conclusions

We suggest that uranium distribution within the Salt Wash distributive fluvial system can be used as a proxy for understanding basin-scale porosity and permeability variations. Clear relationships are present between uranium mineralization and sandstone and channel-belt percentage maps, with 92% of mineralization concentrated at the 40–55% sandstone and 90% in the 20–50% channel-belt percentage contours respectively. The amalgamation ratio and field evidence indicate that this is a critical point at which effective connectivity is lost, as a drop from 38% in the proximal region to 12% at Atkinson Creek to 0% at Little Park is observed, allowing internal porosity and permeability variations to concentrate uranium-bearing fluids, with precipitation occurring when reducing conditions are met. We relate changes in channel-belt connectivity to predictable downstream facies variations in the distributive fluvial system model, providing a system-scale model of subsurface fluid flow in a distributive fluvial system. Our results will aid prediction of uranium occurrence in similar settings, and the addition of statistics such as sandstone and channel-belt percentage makes this study directly applicable to subsurface successions.