JEQ Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

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 Google Scholar
Google Scholar
Right arrow Articles by Langlois, J. L.
Right arrow Articles by Mehuys, G. R.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Langlois, J. L.
Right arrow Articles by Mehuys, G. R.
Agricola
Right arrow Articles by Langlois, J. L.
Right arrow Articles by Mehuys, G. R.
Related Collections
Right arrow Water Quality
Right arrow Watershed and Landscape Processes
Right arrow Surface Hydrology
Right arrow Nitrogen
Right arrow Phosphorus
Published in J. Environ. Qual. 32:2301-2310 (2003).
© ASA, CSSA, SSSA
677 S. Segoe Rd., Madison, WI 53711 USA

TECHNICAL REPORTS

Surface Water Quality

Intra-Storm Study of Solute Chemical Composition of Overland Flow Water in Two Agricultural Fields

Jacques L. Langlois* and Guy R. Mehuys

Department of Natural Resource Sciences, McGill University, Macdonald Campus, 21111 Lakeshore Road, Sainte-Anne-de-Bellevue, QC, Canada H9X 3V9

* Corresponding author (jlangl2{at}po-box.mcgill.ca).

Received for publication September 7, 2002.

   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Few studies have assessed the transport of dissolved nutrients at the field scale under natural rainfall conditions. Hysteresis between dissolved nutrients and discharge behavior can complicate such assessments and this effect has only been examined qualitatively. In this study, we investigated factors contributing to short-term variations of dissolved cation (Ca, Mg, Na, and K) and anion (soluble reactive phosphorus [SRP], NO3, and SO4) concentrations in runoff water and developed a quantitative method to study their hysteretic behavior. Within-storm variations of dissolved nutrient concentrations were determined in two agricultural fields during four natural rainfall events along with discharge, sediment, antecedent soil water conditions, and nutrient contents. For each event, nutrient loads were plotted against discharge during the rising and falling limb of the runoff hydrograph. The resulting hysteresis curves were characterized by an index H, which is the ratio between the integrated areas under the rising and falling curves of the hydrograph. Results showed that nutrient concentrations increased with time during each event. Counterclockwise (H < 1) hysteresis, occurring when the falling limb had larger loads, was found when soils were initially dry whereas clockwise hysteresis (H > 1) was associated with prior wet soil conditions. Two hypotheses are suggested to explain these variations. First, suspended sediments could have acted as a sink for dissolved nutrients and the sensitivity of nutrients to hydrological conditions was determined by their preferential sorption on these sediments. Second, movement of nutrients into runoff occurred more readily as soils became wetter during an event.

Abbreviations: C/SS, ratio of concentration of dissolved chemicals to concentration of suspended sediments • H, hysteresis index • Q, total volume of runoff • R, total volume of rain • SRP, soluble reactive phosphorus • SS, suspended sediments • w, gravimetric water content


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
NUTRIENT POLLUTION of surface waters can accelerate the natural eutrophication of lakes and streams by promoting algal growth, which in turn depletes dissolved oxygen content in the body of water (Gregorich et al., 2001). Among nutrients, excessive P is the most common cause of eutrophication (Correll, 1998). In the province of Quebec (Canada), 15 years of water quality monitoring in 40 watersheds has demonstrated that surface runoff from agricultural land often contains total P concentrations above Quebec's legal eutrophication threshold (0.03 mg L-1), suggesting that agricultural land is a major contributor to surface water pollution (Painchaud, 1997). Research is underway to estimate the potential risk of P movement from agricultural fields to surface waters using a P index method (Bolinder et al., 2000). In their index, Bolinder et al. (2000) give the "overland flow potential" component the greatest relative weighting because P movement from agricultural to aquatic ecosystems is thought to occur along surface transport pathways. However, as Haygarth and Jarvis (1999) pointed out, the effects of hydrology on SRP transfers are rather complex because of its temporal variability at a given scale.

Temporal variability at the field scale is due to variations in rain intervals, rainfall intensity, and rain duration because not only do these factors influence surface runoff discharge, but also physicochemical reactions between soils and runoff water (Haygarth and Jarvis, 1999). For instance, under simulated rainfall conditions on soil boxes, Sharpley (1980) found that mean concentrations of SRP in surface runoff increased as the event interval increased from 1 to 6 d, and attributed this to mineralization of organic P during longer intervals. Within a given rain event, Sharpley et al. (1981) found that SRP decreased with the logarithm of time. This trend has also been reported for ions such as Br (Ahuja et al., 1983) and NH4 (Torbert et al., 1999). The decrease in concentration with time can be explained by the initial movement of soil solution into overland flow water during the first stage of runoff generation. Thereafter, concentrations are more controlled by desorption of ions from surfaces of soil particles (Römkens and Nelson, 1974; Sharpley, 1980).

Rainfall intensity influences the transfer of ions into runoff water because the upward movement of ions from the soil results from turbulent mixing and dispersion caused by raindrop impact (Ahuja, 1986). At high rainfall intensity, these effects occur to a greater depth and thus more ions are contributed to runoff water (Sharpley, 1985). The depth of this mixing zone is also influenced by soil aggregation, which mediates the soil's exposure to raindrops. For instance, Ahuja et al. (1983) found that cloddy surfaces released more Br in surface runoff due to higher turbulent release of Br as a result of greater raindrop impact on the soil surface. Rainfall intensity also influences dissolved nutrient losses because of its impact on soil erosion and suspended sediments (SS) in runoff water. Sharpley et al. (1981) showed that SRP might be adsorbed onto sediments during overland flow transport. He suggested that SRP released from the soil surface was then adsorbed to particles in runoff that were finer than those dominating at the soil surface. In contrast, Yli-Halla et al. (1995) showed that suspended sediments from agricultural plots could be a source contributing up to 38% of SRP in runoff water.

Because most field-scale studies have focused on SRP, Durand et al. (1999) believed that monitoring other dissolved nutrients could clarify the effects of suspended sediments on nutrients in runoff. They examined the use of ratios between dissolved nutrients and suspended sediment concentrations (C/SS ratio). During two rain events, they observed the highest ratio for K, meaning high nutrient concentration with low suspended sediment concentrations, followed by Ca, Na, and Mg for cations, and Cl, SO4, and NO3 for anions. Furthermore, these ratios were not constant within a runoff event and their variations were attributed to runoff dilution and depletion of available solutes. However, neither discharge nor other measurements to separate these effects were recorded. Pote et al. (1999) showed that for a given level of water-extracted P in Ultisols, plots producing more runoff had higher SRP concentrations in runoff. Working on the same fields as Yli-Halla et al. (1995), Ekholm et al. (1999) were not able to correlate SRP concentrations with runoff volumes.

These varying relationships between solute concentrations and discharge could be attributed to hysteresis. This occurs when nutrient concentrations or loads at a given discharge vary between the rising and falling limbs of the runoff hydrograph. Williams (1989) proposed to classify the hysteretic relationship using the direction of the loops. For a given nutrient, if the concentrations are higher on the rising limb of the hydrograph, the loop is classified as clockwise whereas if the concentrations are higher on the falling limb, the loop is classified as counterclockwise. A third type of hysteretic relationship classified by Williams (1989) is a figure-eight loop in which rising and falling limbs cross one another. Most studies looking at hysteretic relationships of dissolved nutrient losses from agricultural lands have been performed in streams (House and Warwick, 1998). In such catchments, hysteresis may result from near-stream sources being chemically different from distant sources, and to flow partitioning (Webb and Walling, 1992). Ulén and Persson (1999) showed that, even at the field scale, both clockwise and counterclockwise loops can be observed for SRP losses via subsurface tile drains. However, to our knowledge, no work has assessed nutrient-discharge hysteretic relationships in overland flow at the field scale. Here, we report short-term variations of dissolved nutrients in runoff water at the field scale resulting from natural rainfall. We also present an exploratory technique to quantify hysteretic relationships of nutrient and sediment losses to compare results and help evaluate impacts of hydrology on nutrient transfers from fields to surface waters.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Study Area
Surface runoff samples were collected from an agricultural watershed in St-Bruno-de-Montarville (near Montreal, QC, Canada). The study area covers 1.9 ha and consists of two undrained, raised beds (280 x 35 m) separated along their length by a ditch (Fig. 1) . These fields have the same general shape, but Side 2 has a steeper slope than Side 1. Along the ditch, the plots can be divided into upper (terrace with 1.6 and 2.4% slopes), middle (steep 9.3 and 12% slopes), and lower (2 and 5.7% slopes) sections. All soils at the site are Typic Humaquepts associated with the Providence series, though Aston and Laplaine series are also present (Martin and Nolin, 1991). In 2001, corn (Zea mays L.) was planted with 65 kg P ha-1 applied as triple superphosphate, 35 kg N ha-1 applied as urea, and 20 kg K ha-1 applied as potash in early May. Another 115 kg N ha-1 as urea was applied at the beginning of June.



View larger version (33K):
[in this window]
[in a new window]
  		Fig. 1. Topographic map of both fields and location of instruments.

 
Field Instrumentation
A trench was dug on each side of the ditch and a waterproof polyethylene membrane was placed in it. The edge of the membrane was buried in such a way that water caught by the membrane was surface runoff (Fig. 2) . At the lower end of each trench, runoff discharge was monitored continuously using a 20-L-capacity tipping bucket combined with a tray that had a flow spreader and a discharge slot whose purpose was to ensure uniform flow into the bucket (Khan and Ong, 1997). Rainfall was measured using a rain gage situated between both fields approximately 50 m uphill from the tipping buckets. Each tip of this rain gage corresponded to 0.1 mm of rainfall. Both tipping buckets and the rain gage were connected to a datalogger (Polycorder 516-C; Omnidata International, Logan, UT) to record the time of each tip.



View larger version (12K):
[in this window]
[in a new window]
  		Fig. 2. Design of the installation of the waterproof membrane.

 
Soil Analyses
To collect soil samples from all areas of each field (terrace, mid-slope, and bottom of slope), 24 soil samples were collected from each field, consisting of eight samples from each of three transects at 1, 15, and 30 m from the ditch. For each transect, soil samples were collected every 30 m starting at 15 m from the tipping buckets. All samples were collected from the top soil layer (0- to 2-cm depth), air-dried, passed through a 2-mm sieve, and stored at room temperature. Samples were analyzed for pH (1:2 ratios of water and 0.01 M CaCl2), organic carbon (wet oxidation method; Sheldrick, 1984), oxalate-extractable Al and Fe (Sheldrick, 1984), exchangeable cations (using BaCl2; Hendershot and Duquette, 1986), and particle size distribution by hydrometer (Gee and Bauder, 1986).

In addition, before each rainfall event, 10 random soil samples were collected from the soil surface, starting near the tipping bucket, in a zigzag pattern up the terrace. On each of these samples, water content was determined gravimetrically (w) and water-soluble nutrients were extracted from moist subsamples containing equal amounts of dry soil in a 1:27 soil to water ratio using Sissingh's (1971) method as modified by Beauchemin and Simard (2000).

Water Sampling
Besides the snowmelt period, only four rain events in 2001 produced surface runoff for which rainfall and surface runoff were sampled. During each rainfall event, rainwater was collected in a bottle located near the rain gage. Information on the intensity and duration of rainfall and ion concentrations in rainwater are given in Table 1. For each event, the ratio of total volume of runoff to total volume of rain (Q/R) was computed for each field. Runoff water was sampled every 3 min for the first hour when water started to flow in the membranes and at 10-min intervals thereafter. Runoff water was collected manually a few centimeters above the membranes about 2 m upstream of the tipping bucket. Because it was impossible to filter samples in the field immediately after collection, they were brought to the laboratory where aliquots of runoff samples were filtered at 0.45 µm and kept at 4°C until analysis.


View this table:
[in this window]
[in a new window]
  		Table 1. Date, amount, duration, maximum intensity for 5 min, mean rainfall intensity (MRI), and rain water chemistry for the four rain events.

 
Analyses
For filtered solutions, SRP (Quickchem Method 10-107-06-2-c) and NO3–N (Quickchem Method 10-107-04-1-c) concentrations were determined colorimetrically with a Quickchem automatic flow injection ion analyzer (Lachat Instruments, Milwaukee, WI) and SO4 concentration was measured by ion chromatography (Waters, Milford, MA). Base cation concentrations (Ca, Mg, K, and Na) were measured by flame atomic absorption spectroscopy (PerkinElmer, Wellesley, MA). Finally, pH was measured using a pH meter (Accumet Research AR10; Fisher Scientific, Pittsburgh, PA) equipped with a combination glass electrode. The dissolved nutrient load was calculated as the product of dissolved nutrient concentration with discharge at the time of sampling. Sediment concentration in runoff samples was determined by evaporating (at 105°C) a 250-mL aliquot of unfiltered runoff samples. The suspended sediment load is the product of sediment concentration and runoff discharge at the time of sampling. The C/SS ratios were calculated on a mg L-1 basis as proposed by Durand et al. (1999).

Hysteresis Study
A quantitative technique to study the hysteresis behavior of dissolved nutrient export was developed by splitting the event hydrograph into the rising limb and falling limb (also called recession limb). For both limbs independently, nutrient and SS loads were plotted against discharge as the independent variable and a correlation was computed. If both correlation coefficients were above an arbitrarily fixed value of 0.90, a regression equation was computed for the rising and falling limbs. The area under the curve for the two regression equations was estimated by integration using minimum and maximum discharges observed in the event as the lower and higher limits. Finally, a hysteretic index (H) was computed using the ratio of these two areas as expressed by:

Figure 3 illustrates the calculation of an H index for SRP for one rain event (3 June). A hysteretic index around 1 indicates weak hysteresis behavior, greater than 1, clockwise hysteresis, and less than 1, counterclockwise hysteresis. The use of this index is possible when both regression lines fit the data well (r > 0.90). We chose to not apply the H index if regression lines crossed each other (figure-eight loop) because this is a special situation combining both clockwise and counterclockwise hysteresis. We believe that using a single value of H in that situation would confuse that special case with those of either clockwise or counterclockwise hysteresis.



View larger version (27K):
[in this window]
[in a new window]
  		Fig. 3. Example of a hysteresis calculation based on the 3 June event for Field 1.

 
Statistics
The properties of the two fields were compared using two-tailed t tests. Seven correlations were performed between (i) w and Q/R, (ii) discharge and C/SS ratio, (iii) H and SS concentrations, (iv) H and mean rainfall intensity, (v) H and maximum rainfall intensity, (vi) H and w, and (vii) H and soil water-extractable nutrients. Because time is considered to be measured without any errors, linear regressions between nutrient concentrations and time were performed. Because SS concentrations and discharge are measured with errors, major axis analyses were used between SS and nutrient concentrations and between discharge and C/SS ratio. To compare nutrients, homogeneity tests of slopes were performed between time and dissolved nutrients on a load basis, between SS concentrations and dissolved nutrient also on a load basis, and finally between discharge and C/SS ratio. Comparison of dates for each water-extractable nutrient was done using a randomized complete block design with more than one replicate using fields as blocks. All analyses, except major axis analyses, were performed using CoStat 6.003 (CoHort Software, 1998). For all statistical analyses, a probability value of <=0.05 was used to determine significance, with a probability value of <=0.01 considered highly significant.


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Soils Properties
Soils in both fields are clay textured (Table 2). Water-extractable pH in all samples was neutral, while the CaCl2 solution lowered pH by about 1.7 units. Both fields contained similar amounts of organic carbon and oxalate-extractable Al and Fe. The cation exchange complex was dominated by Ca and Mg with small amounts of K and Na. Statistical analysis (t tests) showed that none of these properties varied between the two fields (P <= 0.05). Therefore, it was hypothesized that runoff from these fields contained similar nutrient concentrations.


View this table:
[in this window]
[in a new window]
  		Table 2. Selected soil physical and chemical properties at the study site.{dagger}

 
In 2001, four rain events produced surface runoff for which three of them occurred on moist soil conditions (w > 30%) and the remaining rain event occurred under drier antecedent conditions (w = 10–14%) (Table 3). Antecedent soil water was significantly correlated with Q/R (r = 0.87; P <= 0.05), given w between 13 and 47% and runoff ratios between 2 and 47% (Table 3). Before all rain events, NO3–N was the most abundant water-extractable anion in soil, followed by SO4–S and SRP. Relative abundance of cations was less clear. Across the rain events, concentrations of Na and Ca were higher than K and Mg (Table 3). For all cations and anions, randomized complete block design analyses showed no significant differences in amounts extracted with water among rain events. Soil characteristics of Field 1 were also used for Field 2 for the 3 June event because the rain commenced before sampling of Field 2 was completed.


View this table:
[in this window]
[in a new window]
  		Table 3. Characteristics of the soil samples collected before the four rain events for each plot.{dagger}

 
Nutrient Concentrations in Surface Runoff
Figure 4 illustrates the hyetograph and resulting hydrographs for both fields for the four rain events that produced surface runoff in 2001. In general, hydrographs were single peaked except the 4 July event for which only the first 60 min were considered for analysis because of the presence of a second peak around 70 min into the event. For the 3 June and 10 July events, both fields had similar hydrographs. For the 11 July event, Field 1 had a lower a peak discharge than Field 2. For the 4 July event both hydrographs had similar concentration curves but Field 1 decreased more rapidly during the recession. Moreover, only a few samples were collected during the concentration curves of the events of 10 and 11 July; as a result calculation of the H index could be less reliable. During the 3 June rainfall, dissolved nutrient concentrations were related linearly to the logarithm of time (Fig. 5) . However, these trends were not observed for all events, that is, slope values were not significant (P <= 0.05) for all dissolved nutrients (Table 4). Generally, the slopes of lines relating anion concentrations in runoff to log of time were greatest for NO3, followed by SO4 and SRP indicating that the concentration of NO3 in runoff water increased more rapidly than SO4 or SRP. During runoff events, the concentration of Ca in runoff water increased more quickly than Mg and Na concentrations, but K concentrations were generally not related to time. These trends of increased dissolved nutrients within an event are in contradiction with those observed in box studies for SRP and Br, in which decreased concentrations over time were attributed to exhaustion of solute sources (Sharpley, 1980; Ahuja et al., 1983).



View larger version (35K):
[in this window]
[in a new window]
  		Fig. 4. Hydrograph and hyetograph of each rain event. Lines with the open and closed circles are for Fields 1 and 2, respectively. The circles represent the sampling time. The lines were smoothed for a better visual presentation.

 


View larger version (25K):
[in this window]
[in a new window]
  		Fig. 5. Dissolved cation concentrations and anion concentrations, plotted against the logarithm of time during the 3 June runoff event for Field 1.

 

View this table:
[in this window]
[in a new window]
  		Table 4. Slope values of the regression between log of time and dissolved nutrients in runoff water.

 
Correlation coefficients between concentrations of dissolved nutrients and suspended sediments with runoff discharge at time of sampling (Table 5) showed significant negative correlations (P <= 0.05) for anions, which were observed in five cases for NO3–N and four for SRP and SO4–S. For cations, negative correlations were found for five cases for Ca and Mg, four cases for Na, and two cases for K. Negative correlations were observed between dissolved nutrients and runoff discharge, which might indicate a dilution effect. However, a pure dilution effect would show nutrient concentrations to decrease in synchrony with discharge increases, and that loads of nutrients remain relatively stable during the runoff event. Therefore, any hysteretic behavior would be minimal (H = approximately 1) if dilution dominates change in solute concentration during runoff. At these sites, however, for both anions and cations, most H values were less than one, meaning that loads were usually higher for a given discharge during the recession of the hydrograph (Table 6). These hysteresis indices were not significantly related with rainfall characteristics such as storm amounts and storm duration. Because the temporal increase of nutrient concentrations during an event is neither explained by chemical exhaustion, dilution effects, nor rainfall characteristics, we suggest two alternative hypotheses: the first applies if suspended sediments influence dissolved nutrient concentrations whereas the second relies on antecedent soil moisture conditions.


View this table:
[in this window]
[in a new window]
  		Table 5. Correlation coefficients between runoff discharge at time of sampling and suspended sediments (SS) and between runoff discharge at time of sampling and dissolved nutrients.

 

View this table:
[in this window]
[in a new window]
  		Table 6. Hysteresis indices (H) for each nutrient and suspended sediment (SS) for the four rain events.{dagger}

 
Suspended Sediments Hypothesis
Cation and anion concentrations declined with suspended sediment concentrations in runoff (Fig. 6) indicating that an increase in suspended sediments leads to a decrease in dissolved nutrients. This suggests that suspended sediments might sorb dissolved nutrients during a rain event. However, for all events, slope values of the major axis between dissolved nutrients and log of SS were not always significantly different from zero (P <= 0.05; see Table 7), meaning that desorption or resorption of dissolved nutrients were negligible for some events. The ranking of slope values for the major axis analysis between suspended sediments and cation concentrations was Ca < Mg < Na < K, meaning that Ca would have been more sensitive to resorption if the concentration of suspended sediments increased whereas K would have been the least sensitive. For anions having a slope value different from zero, NO3–N had the steepest slopes followed by SO4–S and SRP, meaning that more NO3–N was sorbed when suspended sediment concentrations increased in surface runoff. This apparent higher sorption of NO3–N to suspended sediments than SRP and SO4–S could be explained by the presence of easily transferable NO3–N from soils compared with SRP and SO4–S as suggested by their water-extractable values. Indeed, the steeper slope for NO3–N is not due to a higher sorption tendency but rather to the large amount of NO3–N in runoff water.



View larger version (25K):
[in this window]
[in a new window]
  		Fig. 6. Relationships between concentration of dissolved cations and anions and logarithmic values of suspended sediments in runoff during the first runoff event for Field 1.

 

View this table:
[in this window]
[in a new window]
  		Table 7. Slopes of the major axis between logarithm of sediment concentration and dissolved nutrient of runoff for both fields.

 
The resorption phenomenon could explain contradictory trends for dissolved concentrations observed with time here and in previous studies from Sharpley (1980) and Ahuja et al. (1983) done at different time and space scales of measurement. The latter studies were conducted on 1-m-long boxes that may not have been long enough to allow dissolved chemicals to fully resorb on suspended sediments whereas in this study the time differences between peaks of high rainfall intensity and peaks of runoff discharge suggested that runoff water flowed approximately 20 min on the soil surface before sampling, thus giving enough time for resorption to occur and allowing enough time for the solution to reach equilibrium. Increased nutrient concentrations during a rain event may result from larger suspended sediment transfers during the early part of a rain event (H > 1; Table 6). However, it is also possible that the equilibrium was not completely achieved at the time of sampling, and, because samples were not filtered immediately in the field, the present results are not representative of field conditions due to sorption–desorption reactions occurring in the sampling bottle. Even in the presence of this artifact, we do not believe that the shift in relative concentrations between dissolved versus adsorbed nutrients occurring in the sample bottle would be great enough to change the ranking among cations and anions.

To depict the effects of hydrology on possible resorption of dissolved nutrients by suspended sediments, the C/SS ratio of Durand et al. (1999) was used in relation with runoff discharge. Figure 7 illustrates this relationship and shows that interactions between dissolved chemicals and suspended sediments were low (high C/SS ratio) at low runoff discharge and high (low C/SS ratio) at high runoff discharge. With the exception of one event for all nutrients and for Na during the first rain event in Field 1, these trends were always significant at P <= 0.05 (Table 8). For anions, C/SS ratios of NO3–N were the most sensitive to runoff discharge (indicated by a steeper slope) followed by SO4–S and SRP. For cations, C/SS ratios of Ca were the most sensitive to runoff discharge followed by Mg, Na, and K. Relationships between C/SS ratio and runoff discharge can be explained by more sediments having been present in runoff water at high flow rates as observed by positive correlations between these two parameters at P <= 0.05 (Table 5), therefore allowing more resorption of dissolved chemicals on these sediments and decreasing the C/SS ratio.



View larger version (25K):
[in this window]
[in a new window]
  		Fig. 7. Runoff concentrations of dissolved cations and anions over suspended sediments during a hydrological event for the first runoff event for Field 1. Note the different vertical scales for both graphs.

 

View this table:
[in this window]
[in a new window]
  		Table 8. Slopes of major axis between logarithm of discharge and ratio of concentration of dissolved chemicals to concentration of suspended sediments (C/SS ratio) for all rain events for both fields.

 
Soil Water Hypothesis
If nutrient concentrations in runoff were controlled by release from the soil surface, an increase in "mobile" ions such as Na, NO3–N, and SO4–S during an event could be due to easier water exchange between "old" soil water and event rainwater as conditions became wetter during an event, leading to H < 1. In the case of adsorbed ions that are more sensitive to diffusion mechanisms such as P and K (Havlin et al., 1999), a concentration increase during an event could be explained by soil disaggregation. Indeed, Linquist et al. (1997) studied the effects of soil aggregation on P release in an Ultisol and showed that its rate of extraction was inversely related to aggregate size due to longer diffusion paths found in large aggregates. This phenomenon was also observed for K by Horn and Taubner (1989). In our study, shrinkage of these clay-textured soils under dry conditions may have increased aggregate stability. Stable aggregates will resist breakdown by either rainfall or runoff. We can hypothesize that the initially soluble nutrient concentration in runoff as a result of the movement of the "old" soil water into runoff is not appreciable on a dry field. As a result, nutrient concentrations in runoff water will be controlled by desorption from soil aggregates having low diffusion rates. However, as the fields become wetter during the rain event, soil aggregates become progressively easier to break down and therefore P and K releases will be enhanced (Sharpley et al., 1981). Consequently, higher loads of nutrients are measured during the falling limb of the hydrograph (H < 1). Ahuja et al. (1983) reported that aggregates enhance release of solute into runoff in the early part of the hydrograph, but their study was performed on prewetted soil boxes. This wetting makes aggregates less stable and more vulnerable to raindrop impact (Hillel, 1998) and, in our study, soils had varying initial conditions. However, only hysteresis indices of SO4–S had a significant relationship with soil water, providing support for this hypothesis.


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
The results of this study provide insight regarding the transfer of dissolved chemicals and suspended sediments at the field scale from soil to surface water under natural rainfall conditions. The descending rank of nutrient sensitivity to hydrological parameters (slopes of C/SS ratios with discharge) was Ca, Mg, K, and Na for cations and NO3–N, SO4–S, and SRP for anions. To explain this ordering, two different hypotheses are suggested in which the chemistry of runoff water is either controlled mainly by suspended sediments or by soil wetness. Results also suggested that hysteresis behavior of nutrient concentrations in runoff water increased with antecedent soil water. This phenomenon should be accounted for in studies looking at relationships between dissolved nutrient load in runoff and soil test nutrients, such as in the case of phosphorus. For hysteresis behavior of suspended sediments in runoff water, no significant relationships with either soil water content or rainfall characteristics were observed. However, we do not suggest a universal rejection of the hypothesis that soil water affects hysteresis behavior of suspended sediments and dissolved nutrients in runoff water even if results in the present study seem to reject it. Use of rain simulators on large areas in future research could help to more thoroughly examine this hypothesis.


   ACKNOWLEDGMENTS
 
We thank Romuald Langlois, Gilles Tremblay, and all the field staff of the research station for their incredible help. We thank the late Dr. Régis Simard for his advice. We also thank Dr. William Hendershot and Dr. Chandra Madramootoo for equipment loan. Finally we thank Peter Enright for technical field support and Hélène Lalande for technical laboratory support. This work was supported in part by FCAR and NSERC.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 


Related articles in JEQ:

This Issue in Journal of Environmental Quality

JEQ 2003 32: 1931-1938. [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 Google Scholar
Google Scholar
Right arrow Articles by Langlois, J. L.
Right arrow Articles by Mehuys, G. R.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Langlois, J. L.
Right arrow Articles by Mehuys, G. R.
Agricola
Right arrow Articles by Langlois, J. L.
Right arrow Articles by Mehuys, G. R.
Related Collections
Right arrow Water Quality
Right arrow Watershed and Landscape Processes
Right arrow Surface Hydrology
Right arrow Nitrogen
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