Geophysical Research Letters

Effects of hyporheic processes on streambed vertical hydraulic conductivity in three rivers of Nebraska



[1] Streambed hydraulic conductivity is of great importance in the analysis of stream-aquifer interactions and stream ecosystems. We investigated streambed vertical hydraulic conductivity (Kv) with two connected depths in three rivers of Nebraska. Our results demonstrated that streambed Kv in the upper sediment layer was much higher than that in the sediment of the lower layer. We speculate that hyporheic processes can result in larger streambed Kv in the upper layer. Specifically, water exchange through upwelling and downwelling zones can lead to bigger pore spaces and a more unconsolidated structure of sediments in the upper layer. The upward movement of gas produced by redox processes can loosen the sediments and further enlarge pore spaces in the upper layers. Also, permeability can increase as a result of expanded pore spaces caused by invertebrate activities in the upper part of streambed. The higher Kv will likely enhance exchange processes between stream and sediments.

1. Introduction

[2] The hyporheic zone is becoming an important consideration for researchers. It can be defined as a spatially fluctuating ecotone between the stream and the deep groundwater where important ecological processes and their requirements and products are influenced at a number of scales by water movement, permeability, substrate particle size, resident biota, and the physiochemical features of the overlying stream and adjacent aquifers [Boulton et al., 1998]. As a medium of interaction between hyporheic and surface system, hydrologic exchange can affect the physical (substrate composition, porosity), chemical (nutrient environment), and biological (population density and distribution) conditions under which stream ecosystem recovery occurs [Valett et al., 1994]. The need for additional information describing the role of river stage and riverbed permeability in controlling flow into and out of the hyporheic zone has been well recognized [Arntzen et al., 2006]. Edwardson et al. [2003] reported that the extent of exchange between channel and hyporheic water was positively correlated with apparent streambed hydraulic conductivity. Ryan and Boufadel [2006] found that high tracer concentrations were observed in the shallow depth of a streambed (7.5–10 cm below channel surface), which had a mean value of horizontal hydraulic conductivity that was statistically higher than that of the next depth of the streambed (10–12.5 cm below channel surface). Their measurement locations of the shallow depth and the next depth were, however, not co-located. Chen and Shu [2002] explained that vertical hydraulic conductivity (Kv) plays an important role in stream-aquifer interactions. Information on the variability of streambed Kv with depth is not well known.

[3] In this study, we focused on determining variations in streambed Kv within two connected sediment columns of streambed at the same measurement points. Commonly, settling of suspended sediments on the streambed may cause a substantial reduction of the conductivity in the uppermost layer of the streambed material. This process is usually referred to as colmation [Brunke and Gonser, 1997]. It has been shown that the proportion of colmation in the top 2 cm was very large but much lower below that depth and nonexistent deeper than 50 cm [Blaschke et al., 2003]. Colmation can result in a decrease of hydraulic conductivity [Blaschke et al., 2003] and can thus hinder exchange processes between surface water and groundwater [Brunke and Gonser, 1997]. Following this logic, the upper layer of streambed sediments may have a smaller mean Kv than the immediate lower layer because of the colmation phenomena. However, for porous streambed sediments, we hypothesize that water-flow into and out of the hyporheic zone forms regional and local flow paths, and these hydrological processes, combined with other processes such as aquatic invertebrate activities, may affect streambed hydraulic conductivity. Since hyporheic processes, including invertebrate activity, may vary with depth, so too may streambed Kv. Because of this, the Kv in the uppermost layer of streambed sediments may be higher than that in the deeper layer. On the basis of this hypothesis, we investigated the streambed Kv for two connected sediment columns in three rivers and then used the calculated results of the Kv to verify our hypothesis. Finally, we provide reasonable explanations of the change in Kv with depth.

2. Study Area and Methods

2.1. Study area description

[4] Tests were conducted on three rivers in Nebraska, USA. Two test sites were located along the Platte River, three test sites were located along the Big Blue River, and four test sites were located along the Little Blue River, all of which are in southeast of Nebraska (Figure 1). Channel sediments consisted largely of sand and gravel, with a very small amount of silt and clay. US Geological Survey stream gauges adjacent to our study sites provided information on real-time and historical streamflow. Table 1 summarizes the hydrologic conditions of the three rivers in the study area.

Figure 1.

(left) In-situ testing processes in river channels to estimate streambed Kv. (right) The study area in southeastern Nebraska.

Table 1. Test Date and River Hydrologic Conditions in the Study Area
RiverNearby Stream Gauge NumberAverage Stream Stage, m, amslStream Gradient m/km,Average Stream Flow,a m3/sAverage Stream Flow During Test, m3/sAverage Channel Width, mAverage Water Depth at Test Spots, mTest Date
  • a

    For the years 1984–2004.

  • b

    For 7 braided channels.

Platte River06770500559. 14–15, 2005
Big Blue River06881000402.50.612. 2–4, 2006
Little Blue River06883000 06884000445. 28, Mar. 1, 3, and 7, 2006

2.2. Testing Methods

[5] At each test location, two layers of streambed sediments at different depth in streambed were chosen for Kv tests. A plastic tube, 147 cm long and 5.1 cm in diameter, was pressed vertically into the streambed. Sediment filled the lower part of the tube as it penetrated into the streambed materials. The tube penetrated the streambed to a depth of around 50–60 cm, and a column of unconsolidated channel sediments with a length of d1 (Figure 1) was formed inside the tube. A falling-head standpipe permeameter test, as described by Chen [2000] and Chen [2004], was conducted to estimate vertical K for this sediment column. After this test, the tube was pressed in succession into the deeper streambed and a longer column of unconsolidated channel sediments (d, see Figure 1) was formed inside the tube. This sediment column consisted of the upper layer (d1) and lower layer (d2). A second falling-head standpipe permeameter test was conducted to estimate the vertical K for d. During the test, water was poured to fill the upper part of the tube to form a hydraulic head. The head was then allowed to fall in the tube. Hydraulic head measurements were collected at regular intervals and used for calculating Kv [Chen, 2000; Chen, 2004]. The first depth for our tests ranged from 50 to 60 cm. This division of the upper and lower layers was not conformed to specific sedimentary structures. At these depths, sediment colmation is usually nonexistent [Blaschke et al., 2003] and invertebrate richness decreases largely [Storey and Williams, 2004].

[6] After the standpipe permeameter tests, sediment samples were collected as follows. The top opening of the tube was sealed using a rubber cap to disconnect it from the atmosphere, and then the test tube with sediments was pulled out. This procedure prevented sediments from exiting at the bottom end of the tube. The two layers of sediments were separated and placed into sampling bags based on the thickness of the upper layer d1 and the lower layer d2. Sediments samples were sent to a professional laboratory for grain size analyses using the sieving method. Samples were poured into a roto-tap for shaking and each sediment sample was separated into 14 grades.

[7] The upper streambed with thickness of d1 was marked as the 1st layer and the deeper streambed with thickness of d2 was marked as the 2nd layer. After the vertical K of d1 and d were calculated, the vertical K of the 2nd layer could be calculated using the following formula [Freeze and Cherry, 1979]

equation image

where Kv2 is the vertical hydraulic conductivity for sediment column d2; Kv1 is the vertical hydraulic conductivity for sediment column d1, determined during the first test; and Kv is the vertical hydraulic conductivity for the sediment column d, determined during the second test at each location (Figure 1).

3. Results

[8] For each test location from the three rivers, the Kv value for the first layer was consistently larger than the Kv value for the second layer of streambed sediments. Table 2 summarizes the number of tests, the range of the thickness of measured sediments formed in the tube, and the Kv values at each site. The average thickness from all the sites was 55.2 cm for d1 and 30.1 cm for d2. For each test location, no colmation conditions existed because the amount of silt in the sediment columns from each river was very small (Table 3). However, at locations A13, B6 and B7 in the Platte River and G2, I3, and I2 in the Little Blue River, hydraulic head losses in the tube during the second test were very small and the Kv was not calculated because sediment column d2 contained clay. Those clay lenses appeared to occur sporadically and did not form a continuous layer across the channel at the tested depths. Figure 2 shows the calculated Kv value in the 1st and 2nd layers of sediments for representative sites in the three rivers. The average calculated Kv for d1 from all sites was 43.6 m/d, compared to 24.1 m/d for d2. The ratio of the calculated Kv value in the 1st layer of sediment to that in the 2nd layer for individual test locations at each site ranged largely from 1.03 to 51.7. The ratios of the average Kv value in the 1st layer of sediment to that in the 2nd layer for each site varied from 1.44 to 4.84. Also, the streambed Kv values differed largely among the nine sites (Figure 2), indicating great spatial heterogeneity.

Figure 2.

Paired Kv values of the streambed in the 1st layer and the 2nd layer of individual tests from (a) site A in the Platte River, (b) three sites (C, D, and E) in the Big Blue River, and (c) two sites (F and G) in the Little Blue River. Also, at the other sites (B, H, and I, results not shown in Figure 2), the calculated Kv value is consistently larger in the first layer for each individual test.

Table 2. Thickness of Measured Sediments and Respective Kv Values at Test Sites
Test SitesThickness of Measured Sediments, cmKv, m/dNumber of Tests>
1st Layer d12nd Layer d21st Layer2nd Layer 
Platte RiverA40.6––48.742.36.3–59.634.40.4–41.623.915
Big Blue RiverC61.0–64.762.917.2–27.922.13.2–32.315.01.3–7.93.14
Little Blue RiverF46.5–63.753.317.1–31.423.79.7–22.517.91.4–20.98.78
Table 3. Average Weight Percentages of Streambed Sediments From the Three Rivers
SedimentsPlatte RiverBig Blue RiverLittle Blue River
Site ASite BSite CSite DSite ESite FSite GSite HSite I

4. Discussion

[9] Although the average streamflow differed largely among the three streams, our tests showed the same results in each stream; the value of streambed Kv in the 1st layer of sediment was consistently larger than that in the 2nd layer of sediment at each test location. This supports our hypothesis that the Kv in the upper part of the streambed is larger than that in the deeper streambed at our study sites and possibly in similar river types in other regions. Permeability is considered to be controlled mainly by grain size. Low Kv values are thought to be due to a large amount of small-size particles (fine sand, silt, and clay) that fill the pore space between coarser sediments [Ryan and Boufadel, 2006]. However, in our study, we found at some test locations that the weight percentages of gravel and coarse sand relative to total sample weight in the 1st layer of the streambed was significantly lower than that in the 2nd layer of the sediment. For example, at test location E5, the weight percentage of sediments with particle diameter >0.50 mm (coarse sand and gravel) in the 1st layer of the streambed was 50.0% while that in the 2nd layer of the streambed was 70.1%. Yet, the streambed Kv value in the 1st layer of sediment was much higher than that in the 2nd layer at this test location (see Figure 2b). The averaged weight percentage of coarse sand and gravel for five samples of the 1st layer from site E accounted for 54.1% of the total sample weight, compared to 57.6% of the total sample weight for the five samples of the 2nd layer from the site E. Thus, the decrease in streambed Kv with increasing depth did not fully result from the size of sedimentary particles.

[10] A proportionality expression of hydraulic conductivity K to the diameter of pore space in porous media can be written as [Freeze and Cherry, 1979]

equation image

with k = Cd2. Here, k is intrinsic permeability, C is shape factor, d is the mean pore diameter, ρ is the density of the fluid, g is the acceleration of gravity, and μ is properties of the fluid. We speculate that the hyporheic processes had a greater impact on the vertical variation in permeability, resulting in variable Kv values with depth.

[11] In the hyporheic zone, groundwater and surface water exchange. River water can locally recharge the subsurface and be routed via flow paths that return to the river a short distance downstream. Therefore, the hyporheic zone can be regarded as “a subsurface flowpath along which water ‘recently’ from the stream will mix with subsurface water to ‘soon’ return to the stream” [Bencala, 2000]. Upwelling and downwelling zones represent windows of reciprocal influence through which interstitial and surface subsystems interact [Valett et al., 1994]. Generally, river water enters the subsurface where the stream gradient increases and re-enters the river where the stream gradient decreases [Hurlbert, 1984]. Groundwater discharge to a river can further enhance upwelling processes [Chen, 2007]. Either downwelling zone or upwelling zone fluctuates with stream gradient and also over time, resulting in the extent of water exchange varying more widely in the upper layer of sediment than that in the lower layer of sediment [Ryan and Boufadel, 2006]. Inflow and outflow appeared to occur more frequently in the upper part of sediment than in the lower part. A high frequency of hyporheic water exchange can possibly expand pore space within streambed sediments. Thus, the sediments in the upper layer become more unconsolidated and permeable than those in the lower layer.

[12] We did not examine the biological activities in the upper part of streambed, but numerous studies have demonstrated that the hyporheic zone is habitat for many organisms [Stanford and Ward, 1988; Valett et al., 1994; Boulton et al., 1998; Varricchione et al., 2005]. The interstitial spaces among sedimentary particles in the hyporheic zone are occupied by a diverse array of aquatic invertebrates, which includes many types of crustaceans (such as cyclopoida, amphipoda and isopoda), segmented worms, flatworms, nematodes, rotifers, water mites, and juvenile stages of aquatic insects [Stanford and Ward, 1988; Boulton et al., 1998; Hancock et al., 2005]. These invertebrates might influence Kv. Benthic invertebrate activity (e.g. burrowing), along with the movements of small hyporheic invertebrates and feeding activities of benthic invertebrates and fish, may create new pore spaces, thus increasing sediment permeability [Danielopol, 1989] and Kv. Hyporheic invertebrate density and richness commonly decreases with depth [Pugsley and Hynes, 1983; Marchant, 1995; Varricchione et al., 2005], which likely results in the effect invertebrates on Kv decreasing with depth.

[13] Furthermore, redox processes (such as respiration and denitrification, which come mainly from bacterial and invertebrate activities) cause gas production and diffusion in the sediments. Upwelling subsurface water provides stream organisms with nutrients while downwelling stream water provides dissolved oxygen and organic matter to microbes and invertebrates in the hyporheic zone [Boulton et al., 1998]. The organic matter (carbon and nutrients in the form of carbohydrates, proteins, fats and nucleic acids) present in the sediment can be converted into carbon dioxide and methane under anaerobic conditions, and nitrate can be reduced to nitrogen gas under low dissolved oxygen conditions [Durako et al., 1982; Marshall and Hall, 2004]. The gas produced can dissolve in the pore water where it accumulates. With continued gas production, pore water becomes saturated and bubble nucleation might occur, and then oversaturation occurs. The transition of gas from the dissolved phase into the vapor phase is usually accompanied by a large increase in volume. This may result in a swelling of sediments, as found by the Dutch Ministry of Public Works (Gas production in sediment layers,, 1998). The sediments become loose and the pore space can expand, resulting in increased permeability. It is well known that the gas produced in sediments can migrate upward to form the ebullition zone (the vertical range of sediment from where gas bubbles are released) beneath the water-sediment interface [Brennwald et al., 2005]. Also, organic matter concentrations [Hancock et al., 2005] and redox potential have been found to decline with depth [Franken et al., 2001]. The extent of this redox reaction was likely higher in the upper layer of sediment than that in the lower layer of sediment.

[14] It is, therefore, all the above-mentioned hyporheic processes together that might result in the higher Kv values in the upper layer of sediments than in the lower layer of sediment. While other research commonly shows that the shallow streambed region has low hydraulic conductivities, which may have resulted from the surface colmation by fines [Packman and Mackay, 2003], our study did not observe the effect of colmation on Kv. It is known that high streamflow can scour fine sediments from streambeds. This process can increase hydraulic conductivity in the upper layer of sediment. However, our tests were conducted in low-flow periods and the effect of scouring was probably small.

5. Conclusions

[15] Although the average streamflow differed largely among the three rivers, the same pattern of Kv variation with depth was found in our study; the Kv values were consistently higher in shallow streambed sediments than in deep streambed sediments at test locations in each river. It can be concluded from our grain size analyses that the higher Kv in the shallow streambed sediments did not fully result from the larger particle size. The colmation process, which is considered ubiquitously in streambeds, does not necessarily lead to smaller Kv values in the upper part of the streambed, compared to the Kv values in the deeper part of the streambed. We found support for our hypothesis that the larger Kv value in the upper streambed was mainly due to hyporheic processes as follows. First, the sediment in the upper layer was more unconsolidated and permeable, which probably resulted from hyporheic water exchange through upwelling and downwelling processes. Second, invertebrate activities are known to expand pore space and increase Kv, while invertebrate density and richness decrease with increasing depth and their effect on Kv decreases with increasing depth as well. Third, gas production and diffusion can cause swelling and higher permeability of streambed sediments. The extent of this reaction likely increases in the upper part of the streambed because organic matter concentrations and redox potential are highest near the surface.


[16] Funding was provided by a USGS 104(b) grant, a USEPA grant, and the Central Platte, the Upper Big Blue, the Lower Big Blue, and the Little Blue Natural Resources districts of Nebraska. We thank property owners who allowed us to perform our tests on their properties. Mark Burbach reviewed the manuscript.