JEQ Grow Your Career With ASA
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 9 August 2005
Published in J Environ Qual 34:1600-1609 (2005)
DOI: 10.2134/jeq2004.0306
© 2005 American Society of Agronomy, Crop Science Society of America, and Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Related articles in JEQ
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via ISI Web of Science (1)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Sánchez, M.
Right arrow Articles by Boll, J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Sánchez, M.
Right arrow Articles by Boll, J.
Agricola
Right arrow Articles by Sánchez, M.
Right arrow Articles by Boll, J.
Related Collections
Right arrow Surface Water Quality
Right arrow Biogeochemical Processes
Right arrow Nutrients
Right arrow Water Pollution
Right arrow Phosphorus

TECHNICAL REPORTS

Surface Water Quality

The Effect of Flow Path and Mixing Layer on Phosphorus Release

Physical Mechanisms and Temperature Effects

M. Sáncheza and J. Bollb,*

a BLUE: Land, Water, Infrastructure, 1271 Old Highway #1 South, Southern Pines, NC 28387
b Department of Biological and Agricultural Engineering, University of Idaho, Moscow, ID 83844-2060

* Corresponding author (jboll{at}uidaho.edu)

Received for publication August 9, 2004.

   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Soil phosphorus (P) concentrations typically are greater in surface soils compared with subsurface soils. Surface soils have a greater chance to interact with runoff leading to P transport to streams. The thin surface layer where P concentrates is referred to as the mixing layer denoting where water and chemicals mix during transport. The objective of this study was to evaluate the effect of hydrologic flow paths on soluble reactive phosphorus (SRP) loss at two temperatures. Laboratory flumes were built to simulate infiltration, return flow, saturation excess, and interflow, and subsequent interaction with the mixing layer. The sandy loam soil in the flumes was kept at saturation throughout all experiments, so that biochemical effects were normalized. Flow through the flumes was maintained at 3.6 mm/h for 24 to 99 h (at 6 and 25°C) with water entering and exiting the flumes at different ports (to simulate different flow paths) or as low intensity rainfall. Experiments were performed with and without an artificially created P-enriched surface layer (5 mm thick, total P increased from 1010 mg/kg in the original soil to 2310 mg/kg by addition of dissolved phosphate). Results indicated that (i) SRP release was greater in soil with a mixing layer than in soil without a mixing layer; (ii) SRP release was greater during experiments at 25°C than at 6°C; (iii) at 25°C, SRP release was greatest when water traversed the mixing layer in the upward direction (i.e., in return flow), and by flow parallel to the mixing layer (i.e., surface runoff); and (iv) at 6°C, SRP release in subsurface flow following rainfall was slightly greater than in return flow and infiltration. Our results confirmed the presence of a variable, temperature-dependent desorption process when runoff water interacted with the mixing layer. Our findings have important implications for how different water flow paths in and over the soil interact with P in the soil, and what the ultimate concentration will be in runoff and interflow.

Abbreviations: Ib, inflow port at the bottom of the flume • It, inflow port at the top of the flume • Ob, outflow port at the bottom of the flume • Ot, outflow port at the top of the flume • RM, water input from rainfall simulator • SRP, soluble reactive phosphorus


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
HIGH CONCENTRATIONS of P in surface waters are linked to eutrophication of fresh water bodies (Sharpley et al., 1994), which degrades water quality and is undesirable to the public (Foy and Withers, 1995; Pote et al., 1996). Understanding the transport of P from soils to surface water is critical to reduce eutrophication problems. Soils tend to sorb P, so that it generally accumulates more at the soil surface than in subsurface soil, thus increasing the risk of P in runoff. High P concentrations in streams, therefore, are often attributed to contributions from surface runoff (Gburek and Sharpley, 1998; Sharpley et al., 1994). Some field studies have shown, however, that subsurface flow should not be ignored, especially in the presence of preferential flow paths connected to artificial drains (Scott and Weiler, 2001; Gächter et al., 1998; Simard et al., 2000; Steenhuis et al., 1994).

Release of P to runoff and subsurface water involves the desorption, dissolution, and extraction of P from the soil (Sharpley et al., 1994; Steenhuis et al., 1994). This release depends on the contact of source water (e.g., rainfall, snowmelt, runoff, subsurface water) with the soil-water in the surface soil. Depending on the climate, soil type, and land use, the source water, in turn, can follow different flow paths (Fig. 1) before reaching a water body (Dingman, 2002; Gburek and Sharpley, 1998). In arid climates and/or in areas having compacted surface soils (i.e., skid trails, some bare plowed fields) or frozen soils, surface runoff occurs when the rainfall intensity exceeds the infiltration capacity of the soil (Horton, 1933, 1940), known as "Hortonian" flow. In this case, runoff water has limited interaction with subsurface soil and passes to a water body as overland flow. In humid regions on vegetated, permeable soils, runoff is limited by the storage capacity of the soil (Dunne et al., 1975), known as saturation excess flow. In this case, water reaches a surface water body as interflow (subsurface flow), return flow (subsurface flow that resurfaces in converging areas or low slope areas), and/or overland flow (i.e., direct precipitation onto saturated areas). After infiltrating the soil, interflow interacts with subsurface soil, and return flow flows upward through the surface soil and continues as overland flow, which has limited interaction with surface soil.



View larger version (26K):
[in this window]
[in a new window]
  		Fig. 1. Schematic showing the mixing layer and possible water flow paths through the soil-water system.

 
Previously, P transport in overland flow has been considered as a film of water flowing over the soil surface interacting with a thin (1–37 mm) layer of soil (Sharpley and Smith, 1989; Sharpley, 1985; Ahuja, 1982; Sharpley et al., 1981; Ahuja et al., 1981). Interaction of chemicals with the surface soil layer was considered in vertical transport of chemicals by Steenhuis et al. (1994). They referred to this layer as the mixing layer where any chemical is adsorbed to the soil before rainfall or snowmelt and desorbs once the layer becomes saturated with water (Scott and Weiler, 2001). Other flow paths, and in particular, the return flow path, where water traverses the mixing layer in the upward direction has not received any attention with respect to P release.

In this paper, we hypothesize that P release in runoff is greater when water traverses the mixing layer in the upward direction than when water flows over or below the mixing layer. Therefore, the greatest P release will occur during return flow rather than during overland flow or interflow. In addition, we hypothesize that the effect of the mixing layer on P release is temperature dependent because the desorption process is temperature dependent (Gardner and Jones, 1973; Cho and Ponnamperuma, 1971). To test these hypotheses, we determined the effect of different water flow paths on P release and on transport from the soil-water system. We compared surface soil with and without a mixing layer at two temperatures. For the duration of the experiments, biochemical effects were normalized by creating saturated conditions in the soil-water system at the onset of flow.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
We built laboratory flumes to simulate six flow paths, which represent SRP release via overland flow, return flow, and interflow (Table 1 and Fig. 1). Overland flow was simulated as water running over saturated soil, or as low intensity rainfall on saturated soil. Return flow consisted of subsurface water that resurfaced at the flume outlet. Interflow was generated as (re-)infiltrated water, as subsurface inflow from upslope areas, and as infiltration of low intensity rainfall.


View this table:
[in this window]
[in a new window]
  		Table 1. Sequence of water flow paths through Flumes 1 and 2 showing selected flume inlet and outlet ports, hydrologic processes, and the degree of soil surface interaction.

 
Flume Description
Two welded aluminum flumes (1.0 m in length, 0.30 m wide, and 0.25 m deep) (Fig. 2) were built to study SRP release. Each flume was divided into waterproof 0.10-m-wide compartments. Two compartments were used in each flume so each run was performed in duplicate. The inside of each compartment was coated with polyester resin to prevent contact of soil and water with the aluminum and subsequent release of aluminum.



View larger version (121K):
[in this window]
[in a new window]
  		Fig. 2. Top view of flume design showing feeding ports, sampling ports, and dimensions. Ib, inflow port at the bottom of the flume; It, inflow port at the top of the flume; Ob, outflow port at the bottom of the flume; Ot, outflow port at the top of the flume.

 
Each compartment was equipped with inlet (I) and outlet (O) ports, one at the top of the flume (subscript t) and one 0.2 m below the soil surface (subscript b). Port It supplied water from a reservoir to simulate surface run-on water. Port Ib was used to simulate subsurface inflow. Port Ot was a V-shaped overflow for sampling surface runoff water. Port Ob was used to sample subsurface outflow. Ports Ib and Ob were covered with a fiberglass mesh filter and landscape fabric to prevent soil from leaving the flumes.

Water Used in Experiments
Water was supplied to each flume compartment at a constant rate of 6 mL/min (or 3.6 mm/h) using Masterflex peristaltic pumps with modular drive type U-07753-80 (Cole-Parmer Instrument Company, Vernon Hills, IL). These drives were mounted on a common shaft to introduce the same flow rate. Tubing for the pumps was Masterflex Norprene, precision tubing type U6404-24. The flow rate matched approximately the flow rate during flood irrigation observed in the fields where the soils originated.

Distilled water was used for all experiments because it resembled P desorption of rain water better than tap water (Fig. 3) . Distilled water also compared well to rain water based on the results from a water quality screening to determine ammonia, inorganic anions, and dissolved metals (data not shown). The distilled water supplied to the flumes was pumped from a polyethylene, 55-gallon (208-L) drum housed in the constant temperature chamber (at 6 and 25°C, respectively).



View larger version (17K):
[in this window]
[in a new window]
  		Fig. 3. Phosphorus desorption cycles for tap, distilled, and rain water following procedures described by Oloya and Logan (1980). SRP, soluble reactive phosphorus.

 
Rainfall Simulator
A rainfall simulator (RM) was mounted 0.75 m above the flumes. Since two flumes were simulated at once, a partition was mounted between the flumes to prevent cross-contamination from splashing. The rainfall simulator consisted of two sheets of 6.4-mm-thick aluminum plates (1 x 0.6 m) separated by polymer spacers (0.05 m) and bolted together creating a watertight reservoir. A fitting was mounted in the top plate for water entry. To avoid clogging a 10-µm particle filter was installed on the supply side of the water feed. In the bottom plate 6.4-mm holes were threaded and fitted with a polymer adaptor and 0.5-mm syringes equally spaced 75 mm apart. Twenty-six syringes supplied water to each flume compartment. To avoid soil disturbance due to raindrop splash, rainfall intensity was maintained at a low intensity of approximately 3 mL/min (or 1.8 mm/h), half of the flow rate of 6 mL/min used in other flow configurations. Runoff volumes were comparable with other flow configurations by adjusting the duration of rainfall simulations (Table 2). Before and after the rainfall simulations, a period of 24 h for flow stabilization was allowed. During these periods no sampling was performed. To maintain saturated conditions in the flumes at all times throughout the simulations the water level was maintained at the surface of the soil. At the inlet, this was accomplished by maintaining the level in the water reservoir at the height of the soil surface. At the outlet, when surface runoff was simulated, port Ob was closed. When subsurface flow was simulated, a silicone tube (i.d. = 0.79 cm or approximately 5/16 inches) with a polymer "T" was attached to port Ob and its end fixed at the level of the soil surface (Fig. 2).


View this table:
[in this window]
[in a new window]
  		Table 2. Duration of each water flow path configuration{dagger} and water volumes passing through the flumes at 6 mL/min.{ddagger}

 
Soil and Mixing Layer Preparation
A mixture of Roseberry loam (sandy, mixed Humic Cryaquepts) and Donnel sandy loam (coarse-loamy, mixed, superactive Humic Dystrocryepts) soils was obtained from a pasture field near Cascade, ID. The soil was taken from eight random locations, where the soil was removed from below the grass sod (approximately 0.1 m) to a depth of 0.45 m. The Roseberry loam consisted of 48% sand, 33% silt, and 19% clay, and the Donnel sandy loam consisted of 61% sand, 28% silt, and 11% clay (S.L. McGeehan, personal communication, 2004). The soil was sieved through a 2-mm mesh and packed in the flumes at an average bulk density of 1350 kg/m3 and a porosity of 48%. A control flume (Flume 1) contained the original soil as retrieved from the pasture field. The other flume (Flume 2) contained the same soil plus a 5-mm mixing layer of soil mixed with soluble P. After packing the soil in the flumes, the soil was allowed to reach full saturation by letting water flow from port Ib to Ob for approximately one week.

Total P of the original soil was increased from 1010 to 2310 mg/kg in the mixing layer used in Flume 2. Phosphorus was added as dissolved phosphate by mixing 675 g of soil with 350 mL of dissolved potassium phosphate monobasic (KH2PO4) at 2500 mg P/L. The amount of P added in the mixing layer in Flume 2 was 870 mg (equivalent to a surface application of 42 kg/ha). Once the mixing layer was prepared, it was allowed to dry for 2 d and then carefully applied on the top of the soil in Flume 2 approximately 2 d before the start of the first flow path.

Temperature Control
Experiments were conducted at 6 and 25°C by placing the flumes in a walk-in constant temperature chamber (2.5 x 2.0 x 2.2 m). Air was circulated by a heat pump specifically sized for the volume and temperature range of the chamber. The heat pump also held the humidity at a constant 25% of saturation. A digital controller monitored the temperature and humidity.

Soil Redox Potential and pH
Soil redox potential (Eh) and soil pH were measured using a portable redox–pH meter (Model 250A; Orion, Beverly, MA) with an ATC probe. Measurements of Eh were made using an Ag/AgCl single junction reference electrode (Orion 900100) corrected to the standard hydrogen electrode (+199 mV) (Patrick et al., 1996). Platinum probes were placed at depths of 50 and 150 mm below the soil surface, spaced equally at 0.20 m along the flume length (Fig. 4) . Both Eh and pH probes were calibrated regularly as suggested by the manufacturer.



View larger version (28K):
[in this window]
[in a new window]
  		Fig. 4. Flume showing the inlet and outlet ports to manipulate the different flow paths that occur in the soil, rainfall simulator, and soil redox potential (Eh) and pH electrodes. Ib, inflow port at the bottom of the flume; It, inflow port at the top of the flume; Ob, outflow port at the bottom of the flume; Ot, outflow port at the top of the flume.

 
We assumed that redox potential was a measure of reducing conditions. Since aerobic and anaerobic cycling is responsible for SRP release (Ponnamperuma, 1972) and the purpose of our experiments was to evaluate the influence of water flow paths through the soil on SRP release, constant saturated conditions were established from the onset of the experiments and fluctuations in Eh were measured. If fluctuations in Eh were responsible for SRP release from the mixing layer (Flume 2), we expected to see similar SRP release in the control flume (Flume 1).

Flow Path Sequence and Duration
Six flow paths were established using the terminology, sequence, and duration shown in Tables 1 and 2: (1) Ib -> Ob, (2), It -> Ot, (3) Ib -> Ot, (4) It -> Ob, (5) RM -> Ot, and (6) RM -> Ob. Flow Configurations 1 to 4 simulated upslope surface and subsurface flows while 5 and 6 simulated flow following low intensity rainfall. These flow paths were run in Flumes 1 and 2 concurrently. Each flow path was run in duplicate. Flow Paths 1 and 2 tested flow parallel to the soil surface, simulating interflow and overland flow, respectively. Flow Paths 3 and 4 traversed the soil surface representing return flow and infiltration followed by interflow, respectively. Flow Paths 5 and 6 tested flow parallel to the surface following rainfall. Between Flow Paths 4 and 5, flow was temporarily set to go from It to Ot while the rainfall simulator was installed and prepared.

The flow path sequence was the same at both temperatures, but the flow durations were slightly different. Table 2 summarizes the flow durations for experiments at 6 and 25°C along with approximate water depths (mm) and pore volumes. Return flow (Ib -> Ot) was repeated at the end of the sequence at 6°C to assure repeatability of this flow path.

Sampling and Laboratory Analyses
The duration of a sampling period was approximately 14 h each day. A water sample from the respective sampling port and records of Eh and pH were taken approximately every 2 h for the first 8 h of each flow path configuration. After the first 8 h, a sampling interval of 4 to 6 h was chosen so that approximately 8 to 10 samples were obtained for each flow path configuration. All water samples were filtered immediately after collection using a 0.45-µm pore size filter, and transferred to 100-mL plastic bottles, which were acid washed in a 12 M HCl solution and rinsed thoroughly with distilled water before use. Before filling a bottle with sample water, the bottle was rinsed three times with the water to be collected. These water samples were stored at 4°C (APHA 1060-C; American Public Health Association, 1995) to be analyzed the next day for SRP using the molybdate-blue method (APHA 4500P. E; American Public Health Association, 1995). Total P in soil was determined using the total colorimetric, automated method (EPA 3050 365.4) at the Analytical Sciences Laboratory at the University of Idaho.

Data Analysis
The time-weighted SRP concentration per configuration was calculated as follows:

where SRPweighted,j is the average SRP concentration in configuration j, SRPi is the SRP concentration in sample i, ti is the time interval between sample i and i + 1, and Tj is the duration of the flow path configuration j. Statistical differences in SRPweighted,j (n = 2) were determined on log-transformed data. A two sample t test with equal variances (using the two-tailed hypothesis µ1 = µ2) was applied to Flume 1 (control) and Flume 2 (mixing layer) at 6 and 25°C, respectively, and between Flume 1 at 6°C and Flume 1 at 25°C, and Flume 2 at 6°C and Flume 2 at 25°C. A paired two sample t test also was applied to SRPweighted,j (n = 2) of different flow path configurations in Flume 2 at 25°C (Snedecor and Cochran, 1976). We considered sample means statistically significant if p < 0.05, and marginally statistically significant if p < 0.1.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Saturated Conditions
All experiments were conducted while the flumes were saturated causing the soil environment to reach reducing conditions from the onset of the flow experiments. Changes in Eh and pH were monitored and did not appear to fluctuate widely. On average, the Eh was between –11 and –41 mV for all experiments, while pH was between 5.26 and 5.44 (Table 3). The maximum Eh value was 57 mV and the minimum Eh was –139 mV. These Eh values fall in the upper range of stable potentials reached after several weeks of submergence (+200 mV to –300 mV) (Ponnamperuma, 1972; Sah and Mikkelsen, 1989) and with Fe reduction (+50 mV to –200 mV) (Atlas and Bartha, 1993).


View this table:
[in this window]
[in a new window]
  		Table 3. Soil redox potential (Eh) and pH measurements during the flow path experiments.{dagger}

 
During the flow configurations Ib -> Ob, It -> Ot, Ib -> Ot, and It -> Ob, reducing conditions were most consistent. When using the rainfall simulator in configuration RM -> Ot, Eh increased (except in Flume 1 at 6°C) and pH increased. Apparently, rain splash across the entire soil surface aerated the soil more than water entering the flume from the inlet reservoir. During the flow configuration RM -> Ob, the change in Eh and pH reversed. Eh controls Fe reduction, whereas pH governs dissolution and precipitation of Fe compounds and, consequently, P sorption–desorption in waterlogged soils (De Mello et al., 1998). In this study, while Eh certainly affected SRP release, the relative effect was not drastic as seen in results from Flumes 1 and 2.

Soluble Reactive Phosphorus Release
Soluble reactive P concentrations measured at the outlet of Flumes 1 and 2 as a function of time since onset of the flow experiments are shown in Fig. 5 for both 6 and 25°C. Figure 5 shows data for two replicates and the average concentrations. Flow path configurations are shown at the top of Fig. 5. Cumulative amounts of SRP removed (mg) with the standard error as error bars for each flow path configuration are shown in Fig. 6 .



View larger version (29K):
[in this window]
[in a new window]
  		Fig. 5. Effluent concentrations of soluble reactive phosphorus (SRP) versus time showing the sequence of water flow paths in Flumes 1 and 2. Experiments at (a) 6°C and (b) 25°C. During flow stabilization periods, flow was set to inflow port at the top of the flume to outflow port at the top of the flume (It -> Ot). Single dashed lines indicate beginning of new flow path configuration. Hashed lines indicate small flow stabilization period during installation or removal of rainfall simulator. Open symbols are averages of two observations, indicated by small vertical or horizontal lines.

 


View larger version (32K):
[in this window]
[in a new window]
  		Fig. 6. Weighted soluble reactive phosphorus (SRP) concentrations per flow path configuration in Flumes 1 and 2. Experiments at (a) 6°C and (b) 25°C. Ib, inflow port at the bottom of the flume; It, inflow port at the top of the flume; Ob, outflow port at the bottom of the flume; Ot, outflow port at the top of the flume; RM, water input from rainfall simulator. Diagonal lines for Flume 2; vertical for Flume 1.

 
The total amount of SRP removed during the experiments at 6°C was 2.5 and 3.6 mg for each replicate in Flume 1 and 8.6 and 11.1 mg for each replicate in Flume 2. At 25°C, the total amounts removed were 7.1 and 6.5 mg for each replicate in Flume 1 and 55.4 and 29.8 mg in Flume 2. In the latter, 49.5 mg (89.4%) and 25.4 mg (85.1%), respectively, for each replicate, was removed while water exited at port Ot. Both the Roseberry loam and Donnel sandy loam can desorb P in excess of SRP removal found in our experiments. McGeehan (1996) performed cumulative P desorption experiments on these soils after adding 1000 mg P/kg and subjecting them to flooding periods of 0, 10, 20, and 30 d. For samples flooded 0 d, he reported cumulative P desorption of 153.1 and 120.7 mg P/kg soil for Roseberry loam and Donnel sandy loam, respectively. For samples flooded 10 and 20 d, cumulative desorption was 110.6 and 99.3 mg P/kg for the Roseberry soil, and 66.8 and 72.1 mg P/kg for the Donnel soil. In percent, the cumulative P desorption was between 7 and 15% of the initial P added. In our experiments, we added 1289 mg P/kg to 0.675 kg of soil in preparing the mixing layer. Cumulative SRP desorption for the mixing layer in our experiments could have amounted to 61 to 130 mg.

On average (Fig. 5), SRP concentrations during all configurations at 6°C were less than at 25°C with the exception of subsurface runoff following rain (RM -> Ob). In flumes with a mixing layer, for It -> Ot, and Ib -> Ot, the time-weighted SRP concentrations were statistically significantly higher at 25°C (Table 4) than at 6°C. For RM -> Ot they were marginally statistically significantly higher. Further results for release at 6 and 25°C are provided below.


View this table:
[in this window]
[in a new window]
  		Table 4. Statistical significance of time-weighted log soluble reactive phosphorus (SRP) (t statistic and p value, n = 2) based on a two-sample t test with equal variances (two-tailed hypothesis) for Flume 1 (control) versus Flume 2 (with mixing layer) at 6 and 25°C, for Flume 1 at 6°C versus Flume 1 at 25°C, and Flume 2 at 6°C versus Flume 2 at 25°C.

 
Experiments at 6°C
Soluble reactive P concentrations in Flume 2 (with mixing layer) were always greater than in Flume 1. In Flume 1, SRP concentrations ranged from 0.01 to 0.045 mg/L, while in Flume 2 SRP concentrations ranged from 0.03 to 0.20 mg/L. In the first flow path configuration (subsurface flow, Ib -> Ob), SRP concentrations were nearly the same (0.02 mg/L) in both flumes. These concentrations are similar to those observed in shallow ground water in the field where the soil was obtained (Sánchez, unpublished data, 2002). No clear desorption trends were observed in any flow configuration at 6°C when the y axis was scaled up (not shown). Time-weighted log SRP concentrations were (marginally) statistically significantly (p < 0.1) greater in Flume 2 than in Flume 1 for six flow configurations (Table 4).

In Flume 2, SRP release was greater when flow traversed the mixing layer than when flow was parallel to the mixing layer, but the differences were small. Among the former flow configurations in Flume 2, SRP release was greatest in subsurface flow following rain (RM -> Ob) compared with return flow (Ib -> Ot) and infiltration (It -> Ob). Release of SRP in RM -> Ob coincided with the strongest reducing conditions during the experiment when Eh reached its smallest values during that time (Table 3). Among the parallel flow configurations in Flume 2, the surface runoff path (It -> Ot) had the greatest SRP release followed by surface runoff following rain (RM -> Ot) and interflow (Ib -> Ob). The rerun of return flow path Ib -> Ot was not statistically different from the earlier one based on a paired two sample t test.

Experiments at 25°C
Soluble reactive P concentrations in Flume 2 were greater than in Flume 1 except on a few occasions in configuration Ib -> Ob and RM -> Ob. In Flume 1, SRP concentrations ranged from 0.03 to 0.12 mg/L with the smallest concentrations occurring when flow exited at surface port Ot and the greatest concentrations when flow exited at subsurface port Ob. In Flume 2, SRP concentrations ranged from 0.06 to 1.9 mg/L with the greatest concentrations occurring when flow exited at surface port Ot and the smallest concentrations when flow exited at subsurface port Ob. Curvilinear desorption trends were observed during surface runoff (It -> Ot) and return flow (Ib -> Ot) in Flume 2. Time-weighted log SRP concentrations were statistically significantly greater (p < 0.05) in Flume 2 than in Flume 1 when water exited at port Ot (Table 4).

In Flume 2, the greatest SRP concentrations were observed when water traversed the mixing layer in the return flow configuration (Ib -> Ot). The second greatest SRP concentrations were observed in the overland flow configurations (It -> Ot and RM -> Ot). Among the flow configurations crossing the mixing layer in Flume 2, the return flow (Ib -> Ot) clearly generated the greatest SRP release compared with infiltration (It -> Ob) and interflow following rain (RM -> Ob). Among the flow configurations parallel to the mixing layer in Flume 2, SRP release during surface runoff (It -> Ot) and surface runoff following rain (RM -> Ot) was greater than during the interflow (Ib -> Ob) configuration. It is notable that the spread in the data was greatest during the return flow (Ib -> Ot) and surface runoff following rain (RM -> Ot). Statistical intercomparison of flow configurations in Flume 2 showed that configurations where water exited at Ot were marginally (p < 0.1) or significantly (p < 0.05) different from configurations exiting at Ob (Table 5).


View this table:
[in this window]
[in a new window]
  		Table 5. Statistical significance of time-weighted log soluble reactive phosphorus (SRP) (p value) based on a paired two-sample t test (two-tailed hypothesis) for flow configurations{dagger} in Flume 2 at 25°C.

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
CONCLUSIONS
 REFERENCES
 
Influence of Flow Path Configuration on Soluble Reactive Phosphorus Release
Release of SRP was greater in Flume 2 (with a mixing layer) than in Flume 1 (without a mixing layer) when water exited at Ot (Fig. 6), with the exception of RM -> Ob at 6°C. Statistically, the effect of the mixing layer was most significant at 25°C. Our findings support our hypothesis that when a mixing layer was present, SRP release was greater if water traversed the mixing layer in the upward direction than if water flowed over or below the mixing layer. This can be explained by more contact of water with the P enriched surface layer in the cross flow paths compared with the parallel flow paths. When water flowed across the mixing layer and exited at port Ob, SRP released from the mixing layer most likely was adsorbed in the subsurface soil. The large spread in replicates during Ib -> Ot and RM -> Ot (at 25°C) may have been caused by a small rill that developed causing reduced contact of water with the mixing layer. At 6°C, while differences between flow paths were smaller overall, SRP release was greatest when water traversed the mixing layer (Fig. 6). So, while earlier studies (e.g., Sharpley and Smith, 1989; Ahuja, 1982) focused on the interaction of overland flow with the mixing layer to estimate P release during rainfall, this study suggests that during rainfall and after rainfall has ceased, P release in other flow paths also must be considered. We note that our findings should be interpreted qualitatively, because a different flow path sequence may have resulted in different SRP concentrations and SRP release.

Two-Stage Decay Process
Desorption and release of SRP in configurations It -> Ot and Ib -> Ot in Flume 2 at 25°C was similar to results of Ahuja et al. (1981) as discussed in Wallach et al. (1988). A plot of log(SRP) versus time for these configurations clearly shows two characteristic decay rates (Fig. 7) . A rapid decay rate shortly after changing the flow configuration is followed by a slower decay rate. Both rates are linear on the semilogarithmic plot. It is not obvious why the rapid decay was not observed in the RM -> Ot configuration. Soluble reactive P concentrations during the slow decay rate in It -> Ot and RM -> Ot were similar in magnitude.



View larger version (22K):
[in this window]
[in a new window]
  		Fig. 7. Two-stage decay process of the logarithm of soluble reactive phosphorus (SRP) at 25°C during (a) It -> Ot flow path configuration and (b) Ib -> Ot flow path configuration. Ib, inflow port at the bottom of the flume; It, inflow port at the top of the flume; Ot, outflow port at the top of the flume.

 
Observing these different decay rates, Ahuja and Lehman (1983) concluded that the concept of an effective complete mixing depth of rainwater and soil solution was not strictly valid. They tested a first-order approximation by incorporating an exponential decrease in the degree of mixing with depth, and a piston displacement of soil solution for infiltration. Wallach et al. (1988) were able to simulate the time dependent chemical release observed by Ahuja and Lehman (1983) using a diffusion and transfer model in the absence of infiltration. Wallach and van Genuchten (1990), considering rainfall-induced runoff with infiltration, explained the chemical transfer at the early stage of the runoff process as direct mixing between the runoff water and the soil solution. The later stage was considered convective mass transfer by overland flow water. Their theory explained results of Ahuja and Lehman (1983) well, except for the higher rainfall intensities. The two-stage release of a chemical from the mixing zone also was observed in subsurface flow by Steenhuis et al. (1994) where the early stage was shown in macropore flow and the later stage as a mix between macropore flow and matrix flow. Normalized runoff losses of single super phosphate in consecutive irrigations also followed an exponential decay (Austin et al., 1996). Our results (Fig. 7) indicate that SRP release during overland flow on a saturated area (It -> Ot) and during return flow (Ib -> Ot) underwent the same two-stage decay process as observed during rainfall-induced runoff. More replications of these processes are warranted to test which theoretical explanation best fits the SRP release trends seen in Fig. 7.

Influence of Soil Temperature on Soluble Reactive Phosphorus Release
Comparison of flow configurations at 6 and 25°C confirms our second hypothesis that average SRP release is temperature dependent (Fig. 6). Total amounts of SRP removed over the entire experiment at 6°C were smaller (3.0 mg on average for Flume 1 and 9.8 mg on average for Flume 2) than at 25°C (6.8 mg/L for Flume 1 and 42.6 for Flume 2). However, the comparison of time-weighted SRP release was only statistically significantly different in configurations at 25°C where water exited at the top (Ot) and for Ib -> Ob at 6°C. Temperature effects on SRP release can be attributed to differences in P sorbed, P desorbed, and the rates of sorption and desorption (Sallade and Sims, 1997). Gardner and Jones (1973) showed that temperature had an immediate effect on the rate of P sorption. On a moderately acid soil, the amounts of P sorbed after 11 d at 5°C and 3 d at 20°C were not different. At low temperatures, they also found consistent extraction of small amounts of P, and lower equilibrium solution concentrations when P was desorbed. Similar effects of temperature on P sorption and desorption were observed by Sharpley and Ahuja (1982) and Barrow and Shaw (1975). Mixed results in this study suggest that the effect of temperature on P sorption–desorption processes may need to be accounted for when modeling P release and transport in runoff and subsurface flow, but further research is warranted.

Applicability of Results to Field Conditions
Results of SRP release in this laboratory study apply to field conditions where the soil profile has experienced saturation on the order of days to weeks. Saturated conditions occur in soils in low-lying areas near the stream (Frankenberger et al., 1999), soils used to grow rice (Chung et al., 2003), or during flood irrigation (Sánchez, unpublished data, 2002). In areas where surface runoff is limited by soil water storage capacity, soils also can become saturated for longer periods of time (Gburek and Sharpley, 1998). In these soils, although it is most common to find return flow on toe slope positions of a hillslope (Frankenberger et al., 1999; Brooks and Boll, 2002; Johnson et al., 2003; Mehta et al., 2004), return flow and overland flow can also occur on relatively flat terrain and is not confined to soils near the channel. Sánchez (unpublished data, 2002) observed return flow and overland flow in flood-irrigated fields in central Idaho where subsurface flow resurfaces before entering drainage ditches. Once the surface soils become saturated then any additional precipitation falling on these areas will become overland flow. We hope that results from this study will lead to additional studies that will determine P release for other soil conditions, such as shorter saturation periods before return flow and overland flow, and different mixing layer depths.


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
REFERENCES
 
This study shows the importance of different flow paths and temperature on P release from soils to runoff and interflow. In flumes with a P-enriched mixing layer, SRP release was greater than in flumes without a P-enriched mixing layer, if water exited as surface flow. In our laboratory study, the mixing layer was only 5 mm thick, but in field soils this layer may be thicker. Release of SRP in such soils may be different, especially when water traverses this layer, because the mixing length will increase. At 25°C, SRP release generally was greater than during experiments at 6°C, showing the potential importance of accounting for temperature effects on P sorption and desorption. At 6°C, differences in SRP release between flow paths were smaller than at 25°C. In our experiments at 25°C, SRP release was greatest when water traversed the mixing layer in the upward direction (i.e., in return flow), followed by flow parallel to the mixing layer (i.e., surface runoff). This is an important finding with respect to soil settings where runoff is limited by storage capacity and/or in hilly terrain where a substantial lateral flow component is present. Our findings have important implications for how different water flow paths in and over the soil interact with P in the soil, and what the ultimate concentration will be in runoff and interflow. Further research is needed to determine if the influence of the mixing layer on P release during return flow and runoff also holds for soils that experience shorter periods of saturation, such as found in variably saturated areas where soil water storage capacity limits surface runoff, and for other soil types. Studies on the kinetics of P release similar to those of Sharpley et al. (1981), Ahuja and Lehman (1983), and others for overland flow should be applied to flow path configurations where water traverses the surface soil (e.g., return flow).


   ACKNOWLEDGMENTS
 
This work was funded by USGS Grant ID#01HQGR0147. We thank Dr. Chris Scott (International Water Management Institute, India) and four anonymous reviewers for constructive comments on an earlier version of this manuscript.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 


Related articles in JEQ:

This Issue in Journal of Environmental Quality

JEQ 2005 34: v. [Full Text]  




This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Related articles in JEQ
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via ISI Web of Science (1)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Sánchez, M.
Right arrow Articles by Boll, J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Sánchez, M.
Right arrow Articles by Boll, J.
Agricola
Right arrow Articles by Sánchez, M.
Right arrow Articles by Boll, J.
Related Collections
Right arrow Surface Water Quality
Right arrow Biogeochemical Processes
Right arrow Nutrients
Right arrow Water Pollution
Right arrow Phosphorus

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Crop Science
Vadose Zone Journal Journal of Plant Registrations
Journal of Natural Resources
and Life Sciences Education		 Soil Science Society of America Journal