USGS Home
Contact USGS
Search USGS

U.S. GEOLOGICAL SURVEY
Scientific Investigations Report 2007–5239

Back to Table of Contents

Discussion

Distribution of Dissolved Oxygen and Nitrate in Ground Water in the Near-River Environment

Oxic ground water was uncommon in transect samples, in spite of a sampling design that emphasized identification of oxic ground water. Some well nests at transects LD16.1 and LD24.9 yielded oxic ground water in shallower wells and suboxic ground water in deeper wells. In these well nests, shallow, oxic ground water was attributed to hyporheic processes, whereas deeper, suboxic ground water represented a NO₃^-–reducing zone between the river and terrestrial recharge areas. Other occurrences of oxic ground water are discussed in the following three paragraphs.

Oxic ground water from the well at the left edge of water at D198.5.2 appears to have originated in terrestrial recharge areas. The specific conductance, 170 μS/ cm at 25°C, and Cl^- concentration, 5.30 mg/L, suggest such an origin. The concentration of Cl^- could indicate a possible septic tank influence on the ground water. However, the NO₃^- concentration, 0.131 mg N/L, does not support such a link. Furthermore, the region upgradient from this transect is one of the few sewered neighborhoods in the La Pine study area (fig. 1). Wastewater is treated nearby; wastewater lagoons are within 0.25 mi of the transect, and land-based wastewater disposal occurs adjacent to the lagoons. Such wastewater would be expected to be low in N but still retain wastewater Cl^-. This could be a source of the slightly elevated Cl^- concentration for water from the well at the left edge of the transect. Although specific conductance and Cl^- data indicate that this oxic sample has a terrestrial recharge origin, the sample possibly represents a mixture of (1) ground water that is relatively enriched in solutes but is suboxic, and (2) river water that is relatively dilute but is oxic. However, the Cl^- concentration, 5.30 mg/L, is typical of wastewater-impacted ground water in the study area (Hinkle and others, 2007), indicating minimal dilution, whereas the O₂ concentration, 5.1 mg/L, would require a considerable component of river water.

Oxic ground water in two wells near the left edge of water at LD37.8 (LD37.8.2 and LD37.8.4) could have originated as terrestrial recharge or as hyporheic water. The wells were screened at 2.0 ft, and deeper wells were not installed. The dilute nature of the water from these wells—specific conductance 46 and 42 μS/cm at 25°C and Cl^- concentrations 0.59 and 0.57 mg/L, respectively—would be consistent with water originating as hyporheic water. Alternatively, such dilute ground water could have been recharged locally, such as ground water originating from near-river flood plains that receive recharge from precipitation or dilute river water during floods. The dilute nature of this ground water indicates that distal terrestrial recharge areas are a less likely source of this water than local or hyporheic recharge areas. The presence of oxic but dilute ground water at these locations, without additional data from deeper wells, leaves delineation of the oxic/suboxic boundary for the left side of the transect at LD37.8 ill-defined because the redox state of ground water upgradient (earlier in the flow paths) from the dilute, oxic ground water is not known.

Oxic ground water in samples collected from all three wells (LD24.9.7, LD24.9.8, and LD24.9.9) in the nest at the right edge of water at LD24.9 probably originated in terrestrial recharge areas. Specific conductance values for water from these three wells, 178–184 μS/cm at 25°C, and Cl^- concentrations, 6.39–6.74 mg/L, were similar. Dissolved oxygen (and NO₃^-) concentrations were lower at the 1.0-ft zone than at the 3.0-ft zone (table 2). This redox sequence is inverted from the sequence typically observed in recharge areas, where ground-water flow has a downward component and redox conditions tend to progress from less reduced at shallow depths to more reduced at greater depths. The physically inverted redox sequence at this well nest is consistent with the typical mechanism of discharge through a riverbed, in which ground water moves through the near-river environment toward the river with an upward component of flow, and in which redox conditions evolve from less reduced toward more reduced conditions along the flow path. In this case, redox conditions in some ground water sampled from this transect remained oxic to the end of the flow path at the river, or at least to within 1 vertical foot of the river edge.

Four transect wells at LD24.9 yielded ground water with NO₃^- concentrations greater than 0.2 mg N/L. All four wells were at LD24.9, a transect near the outskirts of the city of La Pine where residential development has been in place for several decades. These four wells included the nest of three at the right edge of water (LD24.9.7, LD24.9.8, and LD24.9.9), and the well 13 ft from the right edge of water (LD24.9.3). Concentrations of NO₃^- ranged from 0.701 to 1.57 mg N/L (table 2). The three wells at the right edge of water were oxic, but well LD24.9.3 was suboxic.

In the redox boundary approach developed to represent the occurrence of denitrification in the La Pine study area, Hinkle and others (2007) observed that the oxic/suboxic boundary closely matched the denitrification boundary, and as such, was a useful representation of denitrification at the aquifer scale. However, a redox boundary that appears relatively sharp at the aquifer scale may appear more gradational at smaller spatial scales. Some NO₃^- may be denitrified on the oxic side of the boundary in response to denitrification in reduced microenvironments, and some NO₃^- may be advected beyond the oxic/suboxic boundary as microbial systems adjust from O₂-reducing to NO₃^--reducing conditions. Thus, the occurrence of NO₃^- in suboxic ground water in one well is reasonable, especially in light of the geochemical structure at this site, where oxic, NO₃^--bearing ground water apparently migrated to the right edge of the river without evolving to suboxic conditions. The redox sequence at the right edge of water, in which both O₂ and NO₃^- concentrations decreased with increasing proximity to the riverbed, also is consistent with redox boundary concepts in that some NO₃^- may undergo denitrification at microsites of anoxia within the oxic zone prior to arrival at the redox boundary.

Samples collected from the four wells at and near the right edge of water at LD24.9 were analyzed for stable isotopes of NO₃^- (fig. 5). The mean and standard deviation of a set of 27 samples of septic tank effluent collected after percolation through unsaturated volcanic sand in the study area, approximating the isotopic character of septic tank NO₃^- loaded to the aquifer, also are shown in figure 5. The enrichment in both δ¹⁵N–NO₃^- and δ¹⁸O–NO₃^- in the transect samples is indicative of fractionation during denitrification (Groffman and others, 2006), and indicates that NO₃^- in these sediments had been partially denitrified. Thus, the oxic/ suboxic boundary can be seen as an imperfect representation of the denitrification boundary, insofar as denitrification actually can begin before NO₃^- reaches the boundary (as evidenced by the isotopes of NO₃^-) and can continue to occur downgradient from the boundary (as evidenced by the presence of NO₃^- in ground water beyond the oxic/suboxic boundary). In locations where the oxic/suboxic boundary is near the river, some NO₃^- can migrate beyond the boundary and potentially discharge to the river. On the other hand, as NO₃^- approaches the oxic/suboxic boundary, some denitrification can occur prior to arriving at the oxic/suboxic boundary, lessening the mass of NO₃^- available for discharge to the river.

Integration of Ground-Water Flow Paths at a Cut-Off Slough

The slough that enters the Deschutes River at river mile 194.8 (fig. 1) provided an opportunity to characterize surface water that likely received sizable contributions from multiple ground-water flow paths. Slough water may have contained a component of river water that entered the slough during a period of high river stage and (or) a component of surface runoff from snowmelt. However, slough water contained 1,650 μg/L of Fe, indicating a component of suboxic ground water. (Oxic ground water may also discharge to the slough.) The balance between the kinetically controlled nature of Fe oxidation and precipitation, and the relatively short residence times of most surface water relative to most ground water, accounts for the presence of both O₂ and Fe in the slough water. Slough water contained 0.278 mg N/L of NO₃^- and 0.63 mg N/L of NH₄⁺, also indicating non-equilibrium redox conditions (table 2).

Nitrate in slough water could have been present in ground water that discharged to the slough (the area near the slough was developed earlier and more densely than most other locations in the study area). Alternatively, NO₃^- in the slough could have been produced by nitrification of NH₄⁺ in the presence of O₂ in the slough. The δ¹⁵N–NO₃^- value, 11.9 ‰ (table 2), suggests that NO₃^- in the slough water primarily was from ground water (with possible denitrification of some NO₃^-). The Cl^- concentration (13.2 mg/L) also is consistent with a ground-water source for the NO₃^-. Although slough water was not analyzed for δ¹⁵N–NH₄⁺, other ground-water δ¹⁵N–NH₄⁺ values measured in the La Pine area ranged from 2.5 to 3.9 ‰ (Hinkle and others, 2007), and NO₃^- derived from nitrification of NH₄⁺ would be isotopically depleted (lighter) than source NH₄⁺ (Hübner, 1986). The δ¹⁵N–NO₃^- value for the slough water indicates that nitrification of NH₄⁺ was not a dominant source of NO₃^- in the slough at the time of sampling. The occurrence of both NO₃^- and NH₄⁺ in the slough water reflects distinct contributions of N from various sources: NO₃^- apparently from septic tank effluent (based on Cl^- and stable isotopes of NO₃ ^‑) and NH₄⁺ apparently from natural sources (based on the presence of Fe).

Redox Controls at the Scale of River Reaches

The widespread occurrence of suboxic ground water beneath the Deschutes and Little Deschutes Rivers represents an extensive denitrifying zone between rivers and upgradient recharge areas that appears to remove much of the NO₃^- being transported from upgradient recharge areas. However, oxic ground water does occur in sediments directly beneath these rivers in places, and these oxic zones are potential pathways for NO₃^- transport from terrestrial recharge areas to rivers.

Oxic ground water, where detected, tended to be present near the outside edge (edge closest to the margin of the valley floor) of a river meander bend (pl. 1). The three transects where oxic ground water was detected (excluding sites where oxic hyporheic water was underlain by suboxic water) were: (1) D198.5 (left or western edge), (2) LD24.9 (right or eastern edge), and (3) LD37.8 (left or western edge; oxic, but possibly hyporheic or otherwise of local origin). The pattern of oxic ground water occurrence relative to river meander bends is apparent on plate 1: oxic ground water was detected at the western edge of D198.5, where the river bends toward the west, in the eastern part of LD24.9, where the river bends towards the east, and in the western part of LD37.8, where the river hugs the western edge of the flood plain. Oxic water, when detected, was near the outside edge of meander bends, but not all ground water near the outside edge of meander bends was oxic.

Occurrence of oxic ground water also was related to riparian zone extent. Oxic ground water was detected in zones where the width of the riparian environment adjacent to a site on the outside edge was narrow relative to the overall width of the riparian zone in the valley at the site (pl. 1). Plate 1 shows the riparian zone extent using a geographical information system (GIS) representation the cryaquoll soils (Natural Resources Conservation Service, 2006). GIS coverage is available for 9 of the 10 transect sites. Relations between ground water redox state and riparian zone size are hampered because transect D198.5 is not represented by GIS coverage. However, field observations indicated that the west bank (but not the east bank) at this site had essentially no riparian zone; the west bank consisted of a beach of eroded aquifer gravel and a near absence of vegetation. Some transects with a relatively narrow riparian zone extent between the river and the outside edge (for example, LD2.4), or even, in the case of D194.1, essentially lacking a riparian zone, were largely free of oxic ground water (no oxic ground water was detected). As a consequence, establishing quantitative estimates of riparian zone width associated with suboxic (or, alternatively, oxic) ground water in the near-river environment is difficult. The inability to draw quantitative conclusions reflects the limited number of transects available for making such an analysis. However, it also reflects the facts that the near-river geochemical environment is highly heterogeneous, and that ground water redox conditions in the near-river environment are controlled not only by near-river redox processes, but also by redox processes occurring in parts of the aquifer upgradient from the near-river environment.

Although oxic ground water appears to be relatively uncommon in sub-river sediments, the redox state of ground water discharging to rivers through springs and seeps in the study area has received little attention. Seeps usually were located near the outside bends of river meanders (near the edge of the flood plain, which closely corresponds to the edge of the riparian zone). Although seeps were not represented in this analysis, they tended to be located in the same types of sites (outside bends of river meanders) as sites where oxic sub-river ground water was detected.

Particle tracking analysis can be used in combination with an understanding of occurrence patterns of near-river ground water redox conditions to infer relative degrees of river reach vulnerability to ground-water NO₃^-. Particle tracking results delineate areas where NO₃^- reaching the water table from septic tanks could be transported to the near-river environment, subject to the aquifer-scale redox controls discussed earlier. In this analysis, pathlines for particles transported into the suboxic part of the system (most pathlines) were removed. Only pathlines for particles that remain in the oxic part of the ground-water system were retained, providing a delineation of oxic ground-water contributing zones to rivers as represented by the aquifer-scale model (pl. 1). (Pathlines for particles that travel through suboxic zones, although not shown on plate 1, generally lie beneath pathlines for particles in oxic zones.) Because the thickness of the oxic zone in the near-river environment was represented in the aquifer-scale model by a 10-ft minimum thickness, the suite of particle pathlines shown on plate 1 should be considered to represent the maximum number of pathways for NO₃^- transport into rivers.

Particle tracking results provide insights at scales ranging from the scale of transects to the scale of the aquifer. At the aquifer scale, oxic pathlines primarily occur near the rivers (pl. 1). Oxic ground-water pathlines extend farther from the river in the northern part of the study area, where (primarily forested) volcanic uplands contain thick oxic zones. These results indicate that rivers generally are less vulnerable to residential development farther from rivers than development closer to rivers.

Particle tracking results at the scale of transects are consistent with study results. For example, simulated oxic pathlines on the west at D198.5, east at LD24.9, and west at LD37.8 (pl. 1) are consistent with the measurement of oxic ground water at the same sites, as well as the detection of NO₃^- at wells near the east edge of LD24.9. Simulated oxic pathlines near the slough that enters the Deschutes River at river mile 194.8 (pl. 1), together with the observation that the area near the slough contains residential development, are consistent with the detection of NO₃^- in the slough (table 2). At other transect sites, ground water simulated to remain oxic in the aquifer scale model appears to be suboxic based on field measurements. For example, oxic pathlines were simulated for D194.1 and D206.4, but field measurements identified only suboxic ground water. These results also are consistent with study results, insofar as near-river redox processes not represented in the aquifer-scale model still are expected to occur in some areas.

The combination of particle tracking analysis with near-river redox characterization has potential application in a variety of hydrologic or environmental assessments. For example, resource managers attempting to reduce potential N loads to rivers could prioritize N-reduction efforts in recharge areas or could prioritize locations for riparian zone protection or establishment in discharge areas by focusing on oxic pathlines that link unsewered neighborhoods with river reaches that occur near the edges of riparian zones. Residential neighborhoods are approximately represented on plate 1 by the networks of roads. Alternatively, strategies for establishing river Total Maximum Daily Load (TMDL) guidelines could use an analysis in which a range of scenarios for pathline redox state, based on a range of riparian zone widths, is used to bracket likely NO₃^- loads to river reaches. Actual identification of pathlines that would likely remain oxic along their entire lengths (including during transport through near-river sediments) was not depicted on plate 1 because (1) uncertainty remains regarding the potential for ground water to remain oxic along the near-river parts of the pathlines, and (2) different potential applications of this analysis preclude the presentation of a single subset of pathlines. However, the absence of a more quantitative assessment does not prevent useful applications. For example, at LD5.7, oxic pathlines essentially do not extend beyond the riparian zone extent, suggesting a relatively small degree of vulnerability to NO₃^-. Near LD13.4, oxic pathlines do extend beyond the riparian zone, but do not appear to extend to residential areas, again suggesting a relatively small degree of vulnerability. However, large areas of the Deschutes River between river miles 190 and 200 contain residential development and large numbers of oxic pathlines, indicating a potentially large degree of vulnerability.

Effects of Ground-Water Discharge to Rivers

Ground water is a potential source of both NO₃^- and NH₄⁺ to rivers in the study area. The cut-off slough adjacent to the Deschutes River at river mile 194.8 demonstrates this potential, with 0.278 mg N/L of NO₃^- and 0.63 mg N/L of NH₄⁺ at the time of sampling. The magnitude of the NO₃^- concentration and the ratio of NO₃^- to NH₄⁺ in the slough sample indicate the existence of a considerable contribution of oxic ground water to the slough, consistent with the extension of the slough 0.6 mi away from the riparian zone (pl. 1). Most river reaches, however, likely would be less vulnerable to NO₃^- from sub-river ground water sources because of the widespread occurrence of suboxic ground water in sub-river sediments and large amounts of water for dilution.

Contributions of oxic ground water, and therefore, potential contributions of NO₃^- to rivers in the study area probably are limited primarily to areas where rivers bend toward the outside of the riparian environment. Thus, contributions of NO₃^- to rivers likely are restricted to relatively small areas of river reaches. However, the potential effects of such N loads to rivers might not be negligible because rivers can be highly sensitive to changes in nutrient fluxes (nutrient over-enrichment represents the single greatest source of impairment to rivers in the United States; U.S. Environmental Protection Agency, 1998). A seepage run (series of river discharge measurements along a section of river) with water sampling was done on the Little Deschutes River during a period of stable low flow in October 2006 (table A3, appendix A). Total N concentrations were less than 0.10 mg N/L in all samples, whereas dissolved orthophosphate concentrations were about 0.04 mg/L, indicating N-limited conditions. Generally increasing discharge, specific conductance, and Cl^- downstream (figure A1, appendix A) indicate a river receiving ground-water discharge. Concentrations of N were variable and may have been depressed by plant uptake; rivers in the study area contain extensive algae and macrophytes. Concentrations of NO₃^- in study area ground water likely will increase over time as NO₃^- currently in the aquifer is transported farther along ground-water flow paths and as continued residential development leads to increasing NO₃^- loads to the water table (Morgan and others, 2007), but potential effects of future increased ground-water NO₃^- contributions to rivers in the study area have not been investigated.

Back to Table of Contents