Purpose and Scope

The purpose of this report is to assess salinity retention or mobilization by sediment-retention ponds in the GGNCA near Delta, Colorado. Conceptual models are presented of how such ponds potentially influence the retention or mobilization of salinity in small, ephemeral stream channels. Field data were collected in 2019 to assess environmental patterns and evaluate those patterns relative to the conceptual models (Richards, 2022). Field data included an inventory of conditions at 69 sediment-retention ponds, including observations of underlying geologic unit, vegetation, soil appearance, check dam condition (intact versus breached), and water holding capacity. Relative ground conductivity was assessed at 62 sediment-retention ponds via linear (one-dimensional) geophysical surveys, as ground conductivity can reveal spatial patterns in water, salt, and clay content. At a subset of 29 sediment-retention ponds, two-dimensional geophysical surveys were completed to assess patterns in how the sediment-retention ponds and channels related to the surrounding uplands. At a different subset of 17 sediment-retention ponds, soil-profiles were collected and characterized for gravimetric water and salt content. The aggregated data are compared to the conceptual models to evaluate salt retention and mobilization in the sediment-retention ponds.

Description of Study Area

The GGNCA is 62,844 acres of public land managed by the BLM, located to the east and south-east of Delta, Colorado (figs. 2, 3A) (BLM, 2021a). The Black Canyon of the Gunnison National Park and Gunnison Gorge National Conservation Area Act initially designated the GGNCA in 1999. It was expanded to its current size in 2003 with the Black Canyon of the Gunnison Boundary Revision Act (BLM, 2021b). The GGNCA encompasses diverse ecosystems, ranging from high desert to piñon juniper forest (BLM, 2021c). There are several sensitive species of plants and animals that live within the GGNCA, including Eriogonum pelinophilum (clay-loving wild buckwheat), Cirsium perplexans (Rocky Mountain thistle), Sclerocactus glaucus (Colorado hookless cactus), Centrocercus minimus (Gunnison Sage-Grouse), Haliaeetus leucocephalus (Bald Eagle), and Lontra canadensis (river otter) (BLM, 2021c). The Gunnison River is also designated as a Gold Medal Trout Fishery (BLM, 2021c).

Temperature and precipitation in the GGNCA are typical of semiarid areas of western Colorado but also reflect the large elevation range of the landscape. The weather station DELTA 3E in the nearby town of Delta, Colorado and at an elevation of 5,027 ft, represents conditions similar to those where the bulk of sediment-retention ponds in the GGNCA exist. There, mean monthly temperatures ranged from a minimum of 26.4 degrees Fahrenheit (°F), during January, to a maximum of 75.8 °F, during July from 2000 to 2020 (High Plains Regional Climate Center, 2023). In contrast, mean monthly temperatures near the rim of the Black Canyon of the Gunnison at station MONTROSE 11 ENE. with an elevation of 8,424 ft, ranged from a minimum temperature of 22.7 °F, during January, to a maximum of 65.7 °F, during July from 2000 to 2020. Annual precipitation at Delta, Colo., station DELTA 3E, Colo., averaged 6.45 inches (in.) per year from 2000 to 2020, with September as the wettest month. Precipitation at station MONTROSE 11 ENE, Colo., averaged 20.3 in. per year from 2000 to 2020, with September as the wettest month (High Plains Regional Climate Center, 2023). High temperatures and low humidity in this semiarid climate cause potential evapotranspiration to greatly exceed precipitation, resulting in minimal infiltration and deep percolation. However, rainfall that exceeds the infiltration capacity can generate runoff from hillslopes into stream channels (Branson and others, 1972). In particular, high intensity rainfall events generated by convective thunderstorms during the summer generate overland flow and runoff. The likelihood of infiltration increases where runoff becomes concentrated or held on the surface.

Geology in the study area includes Cretaceous sedimentary rocks and Quaternary alluvial deposits (fig. 3B). The Lower Cretaceous Burro Canyon Formation and Dakota Sandstone are exposed at higher elevations along the east edge of the study area, but few sediment-retention ponds were constructed on these exposures. The Upper Cretaceous Mancos Shale dominates the geology where sediment-retention ponds were constructed in the study area (fig. 3B). Units contained in the Mancos Shale in the study area span the Graneros and Hartland Shale Members to the Juana Lopez and Montezuma Valley members and the Smoky Hill Shale Member is also included in some classification schemes (Noe and others, 2013). The Mancos Shale has been documented as a regional source of salinity, selenium, and uranium (Schumm and Gregory, 1986; Tuttle and Grauch, 2009). Pyrite in the Mancos Shale weathers to produce gypsum and other salts that can reside in the subsurface, until mobilized by infiltration of water (Tuttle and others, 2014a). Surface soils generally contain lower concentrations of salt, as leaching by infiltrating water redistributes them downward a few tens of centimeters (cm). In other locations in the landscape, lateral water movement transports salts to the surface where evaporation causes them to accumulate as efflorescent salts, but they are vulnerable to subsequent mobilization by runoff (Laronne and Schumm, 1982; Whittig and others, 1982). Soils in the study area that developed from the Mancos Shale generally support sparse vegetation and are prone to erosion and in some places are classified as badlands, which are defined as landscapes with little or no vegetation on unconsolidated or poorly cemented materials (U.S. Department of Agriculture, 2015; Soil Survey Staff, 2021).

The Mancos Shale predominately consists of shale and mudstone, that were deposited in marine environments during the Late Cretaceous (Weimer, 1960; Leckie and others, 1997). Four members of the Upper Cretaceous Mancos Shale, following the classifications from Morgan and others (2007) and Noe and others (2013), are recognized and mapped in the GGNCA, in stratigraphic order from oldest to youngest: (1) Graneros Shale and Hartland Shale Members, undivided (unit Kmg); (2) Bridge Creek Limestone Member and the informal Fairport member of Noe and others (2013), undivided (unit Kmfb; herein the Bridge Creek Limestone and Fairport members); (3) the informal Blue Hill member of Noe and others (2013) (unit Kmb); and (4) Juana Lopez Member and the informal Montezuma Valley member of Morgan and others (2007), undivided (unit Kmj; herein the Juana Lopez and Montezuma Valley members). In addition, the Upper Cretaceous Smoky Hill Shale Member of the Niobrara Formation (unit Kms), which intertongues with the Mancos Shale, is recognized in the GGCNA (fig. 3B); the Smoky Hill Member of the Niobrara Formation is contained within the Mancos Shale in Noe and others (2013).

The Graneros Shale and Hartland Shale Members consist of dark-gray to olive-green calcareous shale interbedded with limestone, with distinct bedding planes coated by gypsum crystals (Morgan and others, 2007; Noe and others, 2013). The Bridge Creek Limestone and Fairport members consist of light-brown, pinkish-gray to very-pale-orange calcareous shale, calcarenite, and bentonite. There is a distinct difference in fossilized oyster species between the Bridge Creek Limestone and Fairport members (Noe and others, 2013). The Blue Hill member is a medium-gray, olive-green to black noncalcareous shale. This unit is mostly nonfossiliferous, with bedding planes and fractures containing gypsum. Glauconite and sulfides within the Blue Hill member create an olive-green appearance along weathered surfaces (Morgan and others, 2007; Noe and others, 2013). The Juana Lopez and Montezuma Valley members are brown to medium-gray calcaranitic, and calcareous shale. The Juana Lopez Member weathers to a light-red to orange shale. The lower portion of this member is highly fossiliferous, containing cross- and laminar bedding. The upper part of the member is gray, fissile, and contains concretions that can reach a few feet in diameter (Morgan and others, 2007; Noe and others, 2013). The Smoky Hill Shale Member of the Niobrara Formation is a dark-gray to light-gray, slightly calcareous shale that is fossiliferous with most fossils occurring along bedding planes. The unit has a very distinctive pale-yellow or pale-brown weathering appearance, sometimes referred to as “Mancos blonde.” The member is interbedded with limestone and bentonite and numerous seams of gypsum are also present throughout the unit (Morgan and others, 2007; Noe and others, 2013).

In addition to exposures of Upper Cretaceous Mancos Shale, sediment-retention ponds evaluated in this study are located on two Quaternary deposits: (1) middle to late Pleistocene gravel deposits, undifferentiated (unit Qg) that are a mix of gravel, clay, silt, and sand with small boulders in some locations, all emplaced as debris flows (Noe and others, 2013); and (2) Quaternary alluvial, mud flow, and mud fan deposits (late Pleistocene to Holocene) (unit Qamf) derived from Mancos Shale representing multiple events, the earliest of which may have occurred 10,000 years ago (Morgan and others, 2007; Noe and others, 2013).

Pond Site Selection and Inventory

Sites were selected using the BLM check dam inventory of the GGNCA. The check dam inventory consists of latitude and longitude locations of sediment retention ponds that are unpublished data compiled by David Sinton and Hilary Kruger, held by the Uncompahgre Field Office of the BLM, and communicated to the USGS via Jedd Sondergard on June 25, 2018. Please contact the BLM Uncompahgre Field Office for more information. The BLM check dam inventory was conducted in 2008 and 2009 and provided information on 1,217 check dams in the GGNCA. From the 1,217 check dams inventoried by the BLM, a subset of associated sediment-retention ponds was selected for this study based on input from USGS and BLM personnel. Broadly, sediment-retention ponds were selected to represent various members of the underlying Mancos Shale and pond condition. A total of 69 sediment-retention ponds were selected for an inventory assessment that included visual observations of geology, vegetative cover, soil condition, pond condition, check dam condition, water holding capacity and incoming and outgoing channel condition.

Geophysical Data Collection

Geophysical data were collected to understand relative pond-to-pond variability in ground conductance, as this assists in understanding the range of pond conditions regarding subsurface water and salt content. One-dimensional surveys were conducted along channels spanning multiple ponds and covered 62 ponds. Two-dimensional surveys were conducted across ponds and into surrounding channel and upland areas at 29 ponds. Geophysical surveys were conducted using the GEM-2, a hand-held, multi-frequency broadband electromagnetic induction sensor, manufactured by Geophex, Ltd. (2021). The portable and lightweight nature of the GEM-2 electromagnetic instrument allows for a single person to operate the instrument (fig. 1) (Won and others, 1996; Geophex, Ltd., 2021). The GEM-2 is capable of recording measurements at multiple frequencies (30 Hz to 93kHz), with lower frequencies penetrating deeper into the ground. The goal of geophysical data collection was to assess variations in relative ground conductivity, which is derived from the in-phase and quadrature components of the electromagnetic response (Huang and Won, 2000). Conductivity in soils and sediments are primarily influenced by water, salt, and clay content of the material, with conductivity positively influenced by greater contents of each.

Because only patterns of relative differences were sought, qualitative geophysical data were sufficient for study needs, and data were not extensively processed. Instrument drift within surveys was assessed by collecting 3 minutes of data at the same designated location at the start and end of each survey. All surveys were conducted in less than 50 minutes and drift ranged between 0 to 3 percent. No within-survey drift corrections were applied.

In-phase and quadrature components of GEM-2 data can be inverted to yield apparent ground conductivity values, although these values may be inaccurate and require confirmation and calibration by other methods like direct measurement (Abraham and others, 2006; Minsley and others, 2012). Ground conductivities in this study, as estimated by inversion, range from about 12 to 220 milliSiemens per meter. Ground conductivity affects the depth of investigation in a manner that is frequency dependent. The depth of investigation is the depth at which the instrument responds most to ground conductivity. Maximum values of response from the 48 kilohertz quadrature data were extracted for each pond and the downstream area immediately below its check dam. These response values were used to evaluate the range in ground conductivities among sediment-retention ponds so that sediment-retention ponds selected for soil sampling spanned that range.

Soil-Profile Sampling and Analysis

A subset of 17 sediment-retention ponds were selected for soil-profile sampling. The geophysical data assisted in selecting sediment-retention ponds that represented qualitative variability in ground conductivity, and sediment-retention pond selection also considered geologic variability. Sampling consisted of vertical soil profiles within a selected sediment-retention pond and at a location downstream from the pond’s check dam to understand how sediment-retention ponds have influenced the movement of salts. Soil samples were analyzed for water and salt content to determine the distribution of those properties.

Field Sampling Techniques

Soil samples from vertical profiles were collected using a bucket auger. Profiles were collected at the lowest point in the sediment-retention pond and at channel-center locations immediately downstream from the check dam. The goal was to collect vertical profiles to at least 150 cm depth. If the depth of auger refusal was shallower owing to rock or hard soil, a second profile was attempted. Vertical profiles range between 50 and 270 cm. Soil cores were extruded onto a plastic sheet next to the auger hole so that visual changes by depth could be examined and documented (fig. 5). Soil sample depth increments were divided based on visible changes in texture, color, apparent moisture content, and the presence of fine roots or salt crystals. Two splits were collected for each depth increment. The first split, for measurement of GWC, was immediately placed in a steel can and sealed with electrical tape. The second split, for measurement of salt content and properties, was placed in a plastic bag.

Pond Inventory

Of the 69 sediment-retention ponds inventoried, check dams averaged 2.5 ft high with a range of 0 to 10.5 ft on their upstream sides. Low heights likely reflected infill by retained sediments or general deterioration of the check dam. Check dam heights on downstream sides averaged 3.5 ft and ranged from 0.5 to 13 ft. Owing to incision of the check dams or adjacent soil by overtopping flows, water holding capacities were generally less than upstream dam heights. At the time of measurement, the maximum water holding capacity was zero for 26 sediment-retention ponds (38 percent) and at least 17 of these showed some incision on the downstream side of the channel or dam. Another 14 sediment-retention ponds (20 percent) had maximum estimated water holding capacities of less than 1 ft. The remaining 29 dams (42 percent) had maximum estimated water holding capacities that averaged 2.5 ft and ranged from 1 to 6 ft (Richards, 2022). The projected life expectancy of sediment-retention ponds like those studied here was 14 to 33 years based on sediment infill rates, but erosion of earthen structures commonly ends their functional life before complete infill (Gifford and others, 1977; Gellis and others, 1995).

No substantial salt accumulations were noted in or on the surface soil of the inventoried sediment-retention ponds. No salt crusts, efflorescent salts, or substantial salt along edges of mud cracks were observed. In some cases, pond boundaries were apparent based on color or surface cracking differences relative to surrounding soils. Such differences might indicate import of sediments from up channel, or simply repeated ponding and drying of water. In other cases, vegetation within pond boundaries was thicker or composed of different species compared to upstream in the channel or in surrounding uplands, indicating an effect of greater water availability even if holding capacity was minimal. All visual observations of sediment-retention ponds are available in Richards (2022).

The inventory data indicate sediment-retention ponds most likely fulfilled their intended purpose of sediment retention during their early existence. Decades later, and with no intervening maintenance, these earthen structures have deteriorated. Over time, it can be expected that segments of ephemeral stream channels, separated by sediment-retention ponds, will reconnect and a natural stream gradient will re-establish itself by incision through dams and the accumulated sediments. Once the dam structures are incised and no longer function to retain water, more natural rates of incision will resume, uninterrupted by the retention structures in these ephemeral streams. Such natural incision will potentially mobilize sediments to downstream water bodies, along with any salts associated with those sediments.

Properties of Soil Profiles

Patterns of GWC in soil may reflect water increase from infiltration, loss from evapotranspiration, and movement in the subsurface. Across the different geologic units and sensed ground conductivities, GWC for sediment-retention pond soils at and close to the surface (less than [<]30 cm depth) is lower than those at greater depths (fig. 6A). This pattern suggests sediment-retention ponds had not contained water for some time prior to sampling and evaporation from the soil surface and root uptake had dried the near-surface soil. GWC generally increased below the surface and two patterns were observed. First, soils with GWC values <0.15 grams/gram(g/g) had relatively unchanging GWC with greater depth (fig. 6A), indicating that relatively little water had infiltrated into the sampled depths of the profiles. Second, soils with GWC >0.15 g/g showed maximum values at intermediate depth, values then decreased with increasing depth, and in some cases to <0.15 g/g (fig. 6A). Soils with the second pattern indicate that water infiltrated, but drier soil deeper in the profile indicates that little to none of the water had penetrated to deeper portions of the profile. Similarly low GWCs have been measured in a study of upland soils under semiarid climate in Wyoming which found soil water contents of <0.2 g/g, and more commonly <0.15 g/g, along with substantial accumulations of gypsum that indicated little leaching of salts under the prevailing climate (Bern and others, 2013). Low water content beneath the GGNCA sediment-retention ponds appears to indicate insufficient infiltration for substantial leaching of salts. The relatively clay-rich nature of the Mancos Shale may be limiting the infiltration of water at shallow or more intermediate depths beneath the sediment-retention ponds. Soils that are more clay-rich have lower permeability under both saturated and unsaturated conditions, and when the water content in clay-rich soil is low, soil permeability is particularly low (Schaap and Leij, 2000).

Soil profiles downstream from check dams are also generally driest near the surface and had few samples with GWC >0.15 g/g (fig. 6B). Where values >0.15 g/g were present, they are highest at depths greater than the depths for maximum GWC values in soils directly under sediment-retention ponds. For the three downstream locations with this pattern, it is indicative of limited, but detectable, movement of water downgradient from sediment-retention ponds to downstream locations. Even for those three downstream locations, GWC is low enough (most samples <0.20 g/g) to indicate that movement of water is insufficient to generate substantial leaching of salts (Bern and others, 2013)

The general patterns described above are like those observed from two studies of similar sediment-retention ponds in Utah in the first several years following their construction. In those studies, depth of moisture penetration was commonly less than 60 cm and lateral movement of moisture was less than 120 cm (Wein and West, 1973; Gifford and others, 1978). High soil bulk density (1.5 to 1.7 grams per cubic centimeter) in undisturbed, fine textured material, located 30 cm and deeper below pond surfaces was identified as contributing to drainage restriction and evaporation rather than infiltration of water retained by sediment-retention ponds (Wein and West, 1973).

SC for the 25:1 water-soil extracts strongly correlate (R²=0.95) with TDS, an additional measure of salt content (fig. 7). Some variation in the relation between SC and TDS can be expected depending upon the predominance of different ions in solution. That variation, relative to geology, might be apparent in Figure 7. Based on the relation between SC and TDS, SC values of the 25:1 water-soil extracts were good measures of salt content.

The majority of the soil profiles from the sediment-retention ponds yield the lowest salt content at or near surface (fig. 8A), indicating no accumulation of salts owing to evaporation of waters retained in the sediment-retention ponds. The low salt content at the surface could indicate removal of salts by water entering sediment-retention ponds and lost owing to a lack of water-holding capacity. However, a similar pattern is observed in soils derived from Mancos Shale in upland locations nearby in the GGNCA, where low salt contents at the surface are attributed to leaching of salts out of the top tens of centimeters but little to no net removal from the soil profile (Tuttle and others, 2014a). Such upland locations experience little infiltration because of runoff. Observation of a similar pattern beneath the sediment-retention ponds (fig. 8A) indicates that most of them also experience little deep infiltration despite occasional inundation. The relatively unchanging salt content with greater depth below most of the sediment-retention ponds sampled do not indicate substantial redistribution of salts into or out of the soils below the sediment-retention ponds. If redistribution occurred, an increase in salt content would likely be present at intermediate depths. If leaching occurred, patterns of salt depletion would be apparent at depths other than just the surface and would likely vary among sediment-retention ponds. Neither such pattern is strongly apparent. A study of sediment-retention ponds in the western United States earlier in their lifecycle found similar steep increases in salt content from the surface to 60 cm depth; and concluded that little to no vertical leaching had been induced by the creation of the sediment-retention ponds (Wein and West, 1973). Observing the same pattern of steep increases in salt content from the surface to 60 cm depth, decades later, indicates little redistribution of salt despite longer opportunities for leaching.

Patterns of salt content in soil profiles from locations downstream from check dams (fig. 8B) are generally like those in the sediment-retention ponds (fig. 8A). The high salt contents at depth for the profile on the Graneros Shale and Hartland Shale Members of the Mancos Shale might indicate some redistribution of salts downstream from the check dam. However, the check dam for the associated sediment-retention pond (pond number 25; fig. 3) was about 1 ft tall, flow could easily bypass it in the years before this study, and no substantial leaching pattern of salts is apparent for the associated sediment-retention pond. Therefore, the vertical pattern of salt distribution in soil profiles might simply be associated with a bedding plane of gypsum, which is very common in the Graneros Shale and Hartland Shale Members (Morgan and others, 2007; Noe and others, 2013). Two downstream profiles in the Juana Lopez and Montezuma Valley members (Pond #429), and the Graneros Shale and Hartland Shale Members (Pond #170) show low salt concentrations at intermediate depths that could reflect salt removal (fig. 8B). The dam for sediment-retention pond number 170 was substantially breached and not capable of holding water, and the check dam for sediment-retention pond number 429 was mostly intact and capable of holding about 2 ft of water (Richards, 2022). Therefore, no clear association between water retention and possible salt depletions was apparent. The deduction is that there is little evidence in salt distribution patterns for sediment-retention ponds retaining or mobilizing salts, and that is consistent with results from studies of similar sediment-retention ponds on soils derived from Mancos Shale in Utah earlier in their lifecycles (Wein and West, 1973; Hessary and Gifford, 1979).

The pH values of the 25:1 water-soil extracts show some distinct differences that appear related to geology, particularly for profiles collected in sediment-retention ponds (fig. 8C). Solutions in equilibrium with carbonates and gypsum might be expected to have pH values of about 7.5 to 8.5, and values of 7.5 to 8.3 have been measured in and around other sediment-retention ponds on Mancos Shale in Utah (Wein and West, 1973). Many soil profiles collected in this study fit those ranges (figs. 8C, D). Of those that do not, such as soil profiles from the Blue Hill member of the Mancos Shale, lower pH values might be expected because of a lack of carbonates in this noncalcareous member (Morgan and others, 2007; Noe and others, 2013). Minor amount of sodium carbonate salts could cause pH values >8.5, and these values seem to be associated mostly with the Bridge Creek Limestone and Fairport members, as well as the Graneros Shale and Hartland Shale Members of the Mancos Shale (fig. 8). The influence of geology on salt chemistry is an indicator that little transport of salts into or out of these profiles has occurred, as would be associated with subsurface movement of water.

Chloride concentrations were measured for 25:1 water-soil extracts from 10 soil profiles (fig. 9). Chloride is highly soluble and exists at low concentrations in most minerals, which makes it a useful and generally conservative tracer for water movement (Hendry and others, 2000). Chloride in soils is generally derived from atmospheric deposition, and distribution patterns with depth are commonly interpreted to reflect how infiltrating water has affected its transport (Allison and others, 1994). Two patterns of chloride distribution are observed in the soil profiles assessed. In the first pattern, soils contain relatively little chloride, but concentrations are slightly greater closer to the surface (fig. 9). In the second pattern, chloride content is lower close to the surface and increases substantially between 30 and 100 cm depth, and in some cases shows relative constancy below the increase (fig. 9). The soil profiles with the first pattern of low chloride content were drier than those of the second pattern, with all values of GWC ≤0.15 g/g, and many <0.12 g/g. The soil profiles with the second pattern of greater chloride content were wetter, with some samples in each profile containing GWC >0.15 g/g, and all but one of the samples from the Juana Lopez and Montezuma Valley members profile ranges between 0.18 to 0.24 g/g (figs. 6, 9).

The drier, lower chloride soil profiles are consistent with the idea that insufficient water infiltration has occurred to transport atmospheric chloride to depths where it could be retained. Presumably, finer texture, greater density, and lower porosity may have strongly limited water infiltration in such soil profiles. The wetter, higher chloride soil profiles are consistent with the idea that sufficient water infiltration has occurred to transport chloride to intermediate depths, but not enough to flush it out of the sampled profile. Overall, the combined patterns of GWC and chloride depth distribution in the soil profiles are not consistent with infiltration of water that would have mobilized substantial quantities of salts.