Chloride and nitrate were coinjected into the surface waters of a third-order stream for 20 d to examine solute retention, and the fate of nitrate during subsurface transport. A series of wells (shallow pits) 0.5-10 m from the adjacent channel were sampled to estimate the lateral interflow of water. Two subsurface return flows beneath the wetted channel were also examined. The conservative tracer (chloride) was hydrologically transported to all wells. Stream water was > 88% of flow in wells < 4 m from the wetted channel.The lowest percentage of stream water was 4 7% at a well 10 m perpendicular to the stream. Retention of solutes was greater in the hyporheic zone than in the channel under summer low-flow conditions. Nominal travel time (the interval required for chloride concentration to reach 50% of the plateau concentration) was variable by well location, indicating different flow paths and presumably permeability differences in subsurface gravels. Nominal travel time was <24 h for wells <5 m from the wetted channel. Coinjected nitrate was not conservative. Two wells were significantly (P < .05) higher in nitrate-N than would be predicted from chloride, while four were significantly lower. Wells 2.0-4.0 m from the wetted channel tended to have higher nitrate concentration than predicted, whereas nitrate sink locations tended to have transport distances >4.3 m. The capacity of the hyporheic zone for transient solute storage and as potential biological habitat varies with channel morphology, bed roughness, and permeability. A conceptual model that considers the groundwater-stream water interface as the fluvial boundary is proposed. Emerging paradigms of the riverine network should consider the hyporheic zone and associated nutrient cycling as an integral component of fluvial structure and function.

Organic solute sorption by hydrous iron and aluminum oxides was studied in an acidic, metal-enriched stream (the Snake River) at its confluence with a pristine stream (Deer Creek). From 1979 to 1986, typically 40% of the dissolved organic carbon (DOC) was removed from solution by sorption onto aluminum and iron oxides, which precipitate as the two streamwaters mix. Upstream DOC concentrations, which increase during snowmelt, were identified as the most significant variables in a multiple regression for determining the DOC concentration below the confluence, and the extent of A1 and Fe precipitation was much less significant. On hourly timescales, removal of A1 and Fe varied erratically but DOC removal was steady, indicating that "sorbable" organic solutes are sorbed either by precipitating oxides or by oxides on the streambed. Characterization of two reactive DOC fractions (fulvic and hydrophilic acids) showed that sorption results in chemical fractionation. Molecules with greater contents of aromatic moieties, carboxylic acid groups, and amino acid residues were preferentially sorbed, which is consistent with the ligand exchange-surface complexation model.

We demonstrate that for losing reaches with significant diurnal variations in stream temperature, the effect of stream temperature on streambed seepage is a major factor contributing to reduced afternoon streamflows. An explanation is based on the effect of stream temperature on the hydraulic conductivity of the streambed, which can be expected to double in the 0 ø to 25øC temperature range. Results are presented for field experiments in which stream discharge and temperature were continuously measured for several days over losing reaches at St. Kevin Gulch, Colorado, and Tijeras Arroyo, New Mexico. At St. Kevin Gulch in July 1991, the diurnal stream temperature in the 160-m study reach ranged from about 4 ø to 18øC, discharges ranged from 10 to 18 L/s, and streamflow loss in the study reach ranged from 2.7 to 3.7 L/s. On the basis of measured stream temperature variations, the predicted change in conductivity was about 38%; the measured change in stream loss was about 26%, suggesting that streambed temperature varied less than the stream temperature. At Tijeras Arroyo in May 1992, diurnal stream temperature in the 655-m study reach ranged from about 10 ø to 25øC and discharge ranged from 25 to 55 L/s. Streamflow loss was converted to infiltration rates by factoring in the changing stream reach surface area and streamflow losses due to evaporation rates as measured in a hemispherical evaporation chamber. Infiltration rates ranged from about 0.7 to 2.0 m/d, depending on time and location. Based on measured stream temperature variations, the predicted change in conductivity was 29%; the measured change in infiltration was also about 27%. This suggests that high infiltration rates cause rapid convection of heat to the streambed. Evapotranspiration losses were estimated for the reach and adjacent flood plain within the arroyo. On the basis of these estimates, only about 5% of flow loss was consumed via stream evaporation and stream-side evapotranspiration, indicating that 95% of the loss within the study reach represented groundwater recharge.

Chloride was injected as a conservative tracer with nitrate to examine nitrate retention (storage plus biotic uptake) and transport in a 327-m reach of a third-order stream draining a forested basin in northwestern California. Prior to injections, die! patterns of nutrient concentrations were measured under background conditions. Nitrate concentration of stream water increased downstream, indicating that the reach was a source of dissolved inorganic nitrogen to downstream communities under background, low-flow conditions, despite uptake by photoautotrophs.At the onset of continuous solute injection over a I 0-d period, timing the passage of the solute front indicated that storage dominated nitrate retention. Instantaneous concentration differences at the base of the reach at hour 24 indicated that biotic uptake accounted for 13% of the nitrate amendment while hydrologic storage constituted 29%. Corrected for groundwater dilution (11.7%), saturation of the stream's channel and hyporheic zones was not complete until 6.8 d of continuous injection. By day 3 nitrate retention was dominated by biotic processes. Biotic uptake was greatest during daylight hours indicating retention by photoautotrophs, but also occurred during darkness. After 10 d of continuous injection, mass balance calculations indicated that 29% of N (339 g) was retained from the total injected (1155 g), while the balance of injected nitrate was transported downstream. Storage of NOr N was II 7 g or 10% while biotic uptake was 222 g or 19%.Periphyton biomass on slides, chlorophyll a both on slides and on natural cobbles, and net community primary production all indicated a lag in periphyton response to nitrate amendment. Earliest indicators of a biotic response to nutrient amendment were decreases in both tissue C/N and epilithic respiration.

No abstract

Physical parameters characterizing solute transport in the Snake River (an acidic and metal-rich mountain stream near Montezuma, Colorado) were variable along a 5.2-kin study reach. Stream cross-sectional area and volumetric inflow each varied by a factor of 3. Because of transient storage, the residence time of injected tracers in the Snake River was longer than would be calculated by consideration of convective travel time alone. Distributed inflows along the stream were a significant source of in-stream chemical variations. These transport characteristics of the Snake River were established on the basis of the assumption of lithium as an ideally conservative tracer and use of simulations of advection, dispersion, and transient storage. Evaluations of the validity of this combined tracer and simulation approach lend confidence to the estimation of the physical transport parameters, but further development is warranted for methods of onsite transport experimentation in hydrologically complex, chemically reactive envixonments. INTRODUCTIONWith the current interest in determining the environmental impacts of acid mine drainage and acid rain, basic hydrology is a necessary component in the study of geochemistry in acidic, upland surface waters. Physical transport of solutes along streams can influence the potential for geochemical reactions to occur. The task of establishing "physical transport characteristics" seems initially to be entirely straightforward. It would appear that physical determinations could be readily and routinely obtained. However, recent studies by Thorne and Zevenbergen [1985] show that reliable, process-based equations for flow resistance in mountain rivers require additional development. Furthermore, Marchand et al. [1984] indicate that conventional current meter measurements of discharge do not properly account for hydraulic conditions present in high-gradient, shallow depth streams common to mountainous regions. This is because stream flow is over and around cobbles, resulting in rapid changes in direction and unsteady velocity distributions, which make the assumptions of conventional measurements inappropriate. Also, in gravel bed and cobble bed mountain streams there may be significant flow of "streamwater' through the streambed [Bencala, 1984]. Such flow is contributing to the transport of solutes; however, conventional current meter measurements would not register this flow component [Zellweger et al., 1989]. Finally, physical prope•ies of mountain streams (e.g., cross-sectional area) may vary on distance scales as short as meters [e.g., Kennedy et This paper is not subject to U.S. copyfight. Published in 1990 by the American Geophysical Union. ?•per number 89WR03235. al., 1984]. As a practical matter, physical properties must be characterized on a scale of hundreds of meters to be applicable to analysis of chemical reactions occurring during trans-port. The concern with conflicting distance scales is also an issue when attempting to quantify the multiple inflows of solutes in upland...

Fine-grained particulate material can pass through a 0.45-#m membrane filter and cause major errors up to an order of magnitude or more in the determination of AI, Fe, Mn, a•d Ti dissolved in natural waters. Other elements enriched in clay-size sediments but normally very low in solution concentration may also be affected. Membrane filters have come into widespread use in recentyears for the filtration of water samples prior to analysis.Although filters of various pore sizes are available, membranes with 0.45-urn nominal pore size have been most commonly used. However, clay minerals of the type found in stream sediments can be much smaller than 0.45 #m [Kennedy, 1965;Grim, 1968], and such materials may pass the filter in sufficient quantity to influence significantly the reported concentration of some elements in solution. The constituents most affected are AI, Fe, Ti, Mn, and those asmciated trace elements that are concentrated in sparingly soluble oxyhydroxides and clay minerals. Erroneous values for concentrations of these elements in true solution are of special concern when mineralogic controls on water chemistry are being evaluated.When filtered water samples are analyzed by spectrographic or neutron activation methods, dissolved constituents and suspended particulates are determined simultaneously in a composite, and all elements in suspended material will be reported as in solution. Many wet chemical techniques determine only dissolved constituents, but it is common practice to acidify filtered samples to prevent precipitation or adsorptive losses, and this acid could be expected to attack any dispersed fine particulates in suspension, thereby yielding added dissolved constituents.The purpose of this study was to obtain more information on the extent to which material passing through filters might be affecting the reported concentrations of dissolved constituents in stream water. This was done by passing separate aliquots of various water samples through filters of different pore sizes and analyzing the resultant tiltrates chemically or, in some instances, spectrographically. Detailed evaluation of data obtained by Durum and Haffty [1961] using 2-•tm filters was included with the information gained in the present study. PREVIOUS WORKThere has been appreciation of the effect of filters on the results of analyses of trace constituents for more than 25 years, but much of the concern has been with chemical interaction between the filter and solutions being filtered rather than with the possible effects of any solids passing the filter, as long as the tiltrate appeared clear. Sandell [1944] called atten-Copyright (D !974 by the American Geophysical Union. tion to the fact that filter paper may sorb such metals as Pb and Cu from solution and recommended that inorganic filter media be used. Marvin et al. [ ! 970] found that several types of filters are contaminated with Cu, which can be partly removed or added to depending upon the ch.aracter of the sample being filtered. Others [Robertson, 1968; Spencer and Ma...

An experimental injection was performed to study the transport of stream water solutes under conditions of significant interaction with streambed sediments in a mountain pool‐and‐riffle stream. Experiments were conducted in Little Lost Man Creek, Humboldt County, California, in a period of low flow duringwhich only a part of the bank‐full channel held active surface flow. The injection of chloride and several trace cations lasted 20 days. In this report we discuss the results of the first 24 hours of the injection and survey the results of the first 10 days. Solute‐streambed interactions of two types were observed. First, the physical transport of the conservative tracer, chloride, was affected by intergravel flow and stagnant watt, zones created by the bed relief. Second, the transport of the cations (strontium, potassium, and lithium) was appreciably modified by sorption onto streambed sediment. In the stream the readily observable consequence of the solute‐streambed interactions was an attenuation of the dissolved concentration of each of the tracers. The attenuation in the stream channel occurred concurrently with the storage of tracers in the streambed via both physical and chemical processes. All tracers were subsequently present in shallow wells dug several meters from the wetted part of the channel. Sediment samples collected approximately 3 weeks after the start of the injection contained increased concentrations of the injected cations.