| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Department of Land Resources and Environmental Sciences, Montana State University-Bozeman, P.O. Box 173120
Bozeman, MT 59717-3120
215 South 7th Street, Livingston, MT 59047
* Corresponding author (paulhook{at}earthlink.net)
Received for publication March 22, 2002.
 |
ABSTRACT |
|---|
 TOPNOTES ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
The ability of riparian buffers to trap nonpoint-source pollution from forests, croplands, and pastures has been demonstrated extensively in humid regions (Wenger, 1999; Correll, 2000; Dosskey, 2001), but in the interior western USA, relevant buffer research is very limited except in relation to sediment from logging roads (Belt et al., 1992; Dosskey, 1998). Buffer design practices vary widely, but guidelines typically specify minimum buffer widths and limit management activities (Belt et al., 1992; Wenger, 1999; Fischer et al., 2000). Factors such as slope, soil type, vegetation type, and ground cover may be taken into account, but this complicates implementation (Fischer et al., 2000). Despite widespread use of buffers, there is relatively little quantitative information that relates design requirements to site conditions (Gilliam, 1994; Daniels and Gilliam, 1996; Correll, 2000).
Effects of grazing on riparian and aquatic ecology are well researched in the western USA (Platts, 1991; Mosley et al., 1997; Clary, 1999), but few studies have evaluated the effectiveness of herbaceous buffers in protecting water quality on the region's rangelands (Pearce et al., 1998a,b) or interspersed, semiarid croplands (Fasching and Bauder, 2001). Factors known to control buffer performance in other regions probably apply in rangelands, but additional factors related to grazing may also be important. Differences in vegetation are likely to be important because of large variations in site water balance. Streamside vegetation on rangelands varies from dense wetland sedges to sparse bunchgrasses or shrubs. Furthermore, grazing temporarily reduces plant biomass and can cause long-term changes in vegetation. Riparian grazing guidelines recommend leaving a certain amount of plant material, often specified in terms of stubble height, to promote desired ecological conditions and channel characteristics, limit erosion, and promote sediment deposition (Clary and Webster, 1989; Clary and Leininger, 2000). Utilization guidelines have shown success in improving riparian and aquatic ecological conditions (Clary, 1999; Clary and Leininger, 2000), but research has raised questions about the use of stubble height to assess capacity for sediment retention (Abt et al., 1994; Clary et al., 1996; Pearce et al., 1998a, b).
This study was conducted to improve understanding of factors that affect sediment retention in rangeland riparian areas. Small-plot runoff simulations were used to evaluate the influence of the following factors on sediment retention: (i) buffer width; (ii) slope; (iii) vegetation characteristics including vegetation type, biomass, and ground cover; and (iv) stubble height. Results are relevant both to situations where riparian areas with intact vegetation lie downslope from more disturbed areas that are obvious potential sources of sediment, such as croplands, surface mining areas, roads, or areas with concentrated livestock use, and to rangeland pastures where the distinction between sediment source area and buffer area is not as clear. The term riparian buffer refers here to any riparian area managed for water quality and other conservation benefits (Dosskey, 1998) and is not restricted to ungrazed sites (Mosley et al., 1997).
|
METHODS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Paired runoff plots were established in 1997–1998 in locations that represented typical upland, transitional, and wetland sites and spanned 2 to 20% slopes; the experiments reported here used four upland, five transitional, and four wetland plot-pairs. Each pair included two parallel plots 2 m wide by 6 m long, separated by 1 m, and aligned to follow the slope. Lawn edging was installed on the top and side edges of plots. Plots were fenced to exclude livestock, deer, and elk. Following setup and characterization, plots were ungrazed and undisturbed for at least one fall through spring period.
Plot elevations were surveyed on a 1-m grid that extended 1 m beyond plot edges. Average slopes were estimated for the entire 6-m length of each plot, for the lower 2 m, and for the final 1 m using elevations of points on plot edges and midlines. Six soil samples (15 cm deep) were collected outside each pair of plots, composited, and analyzed for coarse fragments (>2 mm) and texture of fines (hydrometer method; Gee and Bauder, 1986). Soil organic matter was estimated by loss on ignition (360°C for 2 h). Plant biomass was sampled in the growing season before the experiment by clipping vegetation at the ground surface in six quadrats per plot; quadrats were 20 x 50 cm in wetland and transitional plots, where vegetation was very dense, and 50 x 50 cm in upland plots, where vegetation was sparse. Surface cover was estimated at 10 locations per plot using a point frame with ten 3-mm-diameter steel rods that were pushed toward the ground until they contacted (i) bare soil; (ii) plant litter, lichen, or moss; or (iii) basal parts of vascular plants (Bonham, 1989). Microtopography was sampled at the same locations using a 1-m-long array of 100 vertical, 3-mm-diameter rods that passed through a horizontal bar and rested on the ground surface (McCool et al., 1981). The profile of the top of the rods was photographed against a white backdrop with black horizontal lines 1 cm apart and used to estimate the variation in surface elevation within each 1-m profile and, from these observations, describe qualitative differences in microtopography among vegetation types.
Immediately before runoff simulations, one plot of each pair was clipped to an approximately 2- to 5-cm stubble height ("severe clipping") and the other to an approximately 10- to 15-cm stubble height ("moderate clipping") using a gas-powered weed trimmer. These two clipping treatments assessed the consequences of aggressive and moderate defoliation and represented contrasting conditions that are common on grazed rangelands. The severe clipping treatment was judged to be the most aggressive vegetation removal possible without significant soil disturbance, and it left shorter stubble than recommended in riparian grazing guidelines (Clary, 1995). The moderate clipping treatment was representative of stubble heights commonly recommended for riparian grazing (Clary and Webster, 1989; Clary, 1995; Clary and Leininger, 2000). Plots were raked lightly to remove most loose detritus without disturbing the soil surface. Initial installation of lawn edging predated simulations by 8 to 24 mo to allow soil and vegetation time to recover from disturbance. To further contain runoff, 30-cm-tall sheet-metal strips were inserted inside the edging on plots' long sides and sealed with bentonite. Significant leakage of runoff from the edging was not observed.
Runoff Simulations
Runoff simulations employed a run-on distribution device, a hopper for adding sediment to run-on, and a metal apron and portable flume to collect and measure runoff. Water was pumped from Warm Springs Creek to a portable tank trailer, where another pump delivered water to the run-on distribution device, minimizing the effects of elevation and distance on pumping and allowing close control of runoff rates. A manifold fed water into the bottom of a 2-m-long trough divided into eight compartments; overlying water stilled the incoming water. Water flowed across the trough's downhill lip onto a 45° apron that delivered run-on to the plot as gentle, nonerosive flow that was spread evenly across the width of the plot as it left the apron. A hopper with a 2-m-long, 0.5-cm-wide slot was used to add dry, sieved sediment to the incoming flow.
Runoff from the lower end of the plot was collected on a sheet-metal apron with its upslope lip folded down at 90°, inserted into a slot in the soil, and sealed with bentonite. Water flowed via a central gutter into a trapezoidal flume. The flume was leveled, and the slope of the central gutter was adjusted to the same slope at each plot. The flume's stilling well was equipped with a clear tube and scale to observe stage (water level) manually and with a float-and-pulley water-level recorder to record stage automatically. To prevent backwash from affecting flume readings, a sump was dug below the flume outfall and emptied continuously by a pump.
Simulations were conducted in June through August 1999. Steady state runoff was established by increasing the rate of water delivery to plots until the target runoff rate, estimated by flume stage, was maintained for at least several minutes. Elapsed time between first runoff and initial addition of sediment at the top of each plot was 13 to 35 min (mean = 22 min). Flume stage (cm) was calibrated against runoff rate (L/s) by measuring the time taken to fill a 12 L bucket from the flume outfall three times during each simulation. Stage was monitored throughout simulations and the rate of water delivery to plots was adjusted to maintain a constant runoff rate if necessary. Use of steady, equal rates of runoff across simulations was intended to minimize confounding due to potential differences in infiltration and runoff rate among sites; this approach focused on evaluating effects of site factors on sediment retention under uniform runoff.
After the target runoff rate was achieved, but before sediment was added, at least one 1-L water sample was collected from the flume outfall to establish a baseline sediment concentration. Then, 36 L (approximately 50 kg) of dry, sieved sediment was added to the hopper and allowed to fall into run-on over 1 to 4 min (mean = 2.5 min). The sediment, a byproduct of sand and gravel mining, was a uniform mixture consisting of approximately equal proportions of silt, very fine sand, and fine sand, and a smaller proportion of clay (Table 1). Runoff was sampled using 1-L bottles immediately after adding sediment and at 15-s to >5-min intervals thereafter; during each simulation, 11 to 27 samples (mean = 17) were collected over 7 to 29 min (mean = 14 min). Sampling frequency and duration was adjusted to represent the passage of the sediment pulse accurately. Sampling was most frequent early in simulations when sediment concentrations were changing rapidly. Sediment concentration was measured in the field with Imhoff cones (Sojka et al., 1992) to calibrate visual judgments of changes in concentrations and their return toward baseline levels, which were used to adjust sampling frequency and decide when to terminate sampling. Sediment concentrations were remeasured in the laboratory by filtering each 1-L sample through a preweighed filter, drying overnight at 105°C, and reweighing; these measurements were used in data analyses.
|
Total sediment yield for each simulation was calculated as the sum of flow-weighted sediment outputs for each sampling period. The period represented by each sample was considered the time between midpoints between successive sampling times. For each sampling period, runoff rate was estimated from the calibrated flume stage relationship. Sediment mass was calculated as the product of sediment concentration, runoff rate, and duration for each time period. Sediment mass was not corrected for background concentrations, which averaged 3% of mean concentrations during each simulation. Sediment retention was calculated as the difference between mass of sediment applied (50 kg) and total mass measured in runoff and was expressed as a percentage of the sediment applied [% retention = 100 x (50 - sediment yield)/50].
Data Analysis
Data from 13 plot-pairs and 78 simulations (two stubble heights and three buffer widths in each plot-pair) were used in statistical analyses. Analyses of variance were performed with the SAS GLM procedure (SAS Institute, 2001); effects were considered significant at p 
0.05. Slope, runoff, and sediment retention were analyzed using a split-split-plot design, with vegetation type, clipping treatment, and buffer width as the whole-plot, sub-plot, and sub-sub-plot factors, respectively. Each plot-pair was treated as a statistical replicate or "whole-plot." F tests for vegetation type effects used the mean squares for plot-pair (nested within vegetation type) as the error term, while F tests for clipping treatment and clipping treatment x vegetation type interactions used the mean squares for the interaction between clipping treatment and plot-pair, and F tests for width and its interactions used the residual error. Slope was included as a covariate in the analysis of sediment retention because many studies have shown that increasing slope can reduce buffer performance (Phillips, 1989; Tollner et al., 1976; Munoz-Carpena et al., 1999; Ghadiri et al., 2000). Because residuals were nonnormally distributed with heterogeneous variances, sediment retention data were rank-transformed; the statistical results reported are for rank-transformed data, but results for untransformed data were nearly identical. Plot characteristics that were measured separately in each plot (biomass, plant basal cover, litter cover, and bare soil area) were analyzed using a split-plot design, with vegetation type as the whole-plot factor and clipping treatment as the sub-plot factor. F tests for vegetation type effects used the mean squares for plot (nested within vegetation type) as the error term, while F tests for clipping treatment and clipping treatment by vegetation type interactions used the residual error. Soil characteristics, which were measured for one composite sample per plot-pair, were analyzed by one-way analysis of variance (ANOVA). Means separations for main effects of vegetation type and buffer width were performed with Tukey's test. Where significant interactions between factors were found (e.g., between vegetation type and buffer width for sediment data), effects of one factor were tested within each level of the other factors using the SLICE option of the SAS LSMEANS statement, followed by pairwise comparisons of least-square means. Controls of sediment retention were also analyzed using forward stepwise regression analyses with buffer width, stubble height, vegetation type (upland = 1, transitional = 2, wetland = 3), vegetation characteristics, slope, and runoff as independent variables and either raw or exponentially transformed sediment retention as the independent variable.
|
RESULTS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
Runoff and Sediment
Flume stage and runoff rates were consistent across simulations. Calibrated runoff rate averaged 2.4 L/s and did not vary significantly with vegetation type, clipping treatment, or buffer width (F2,10 = 1.12, p = 0.36; F1,10 = 0.05, p = 0.82; and F2,40 = 0.83, p = 0.44, respectively). For a 2-m-wide plot, this would be equivalent to an average velocity of 6 cm/s at a depth of 2 cm. The calibrated rate of water delivery to plots averaged 3.3 L/s.
Time series of sediment concentrations (data not shown) reflected the pulsed nature of the simulations. Sediment started to flow off plots soon after addition (usually within 15–120 s), increased in concentration rapidly, peaked within a few minutes, then declined toward background levels in 5 to 10 min in most cases. For the 6- and 2-m buffer width simulations, baseline concentrations before sediment additions averaged from <0.005 to 0.06 g/L depending on vegetation type. Baseline concentrations were no higher before 2- than 6-m simulations in each plot, indicating that sediment applied in the 6-m simulations was not carried over and collected during the 2-m simulations. Baseline concentrations in 1-m buffer width simulations were slightly elevated (0.45, 0.17, and 0.06 g/L for upland, transitional, and wetland plots, respectively) indicating that in some plots a small amount of sediment applied during 2-m simulations was carried over and collected during the subsequent 1-m simulations; because these values were low compared with concentrations measured during 1-m simulations, which peaked at an average of 21 to 63 mg/L depending on vegetation type, the estimated amount of sediment carried over was quantitatively insignificant (mean of 0.16 kg or 3% of total).
Mean sediment retention ranged from 63 to 99% of applied sediment (Fig. 1)
. For individual simulations, retention ranged from 38 to 99.5%. Sediment retention was affected significantly by buffer width (F2,40 = 218, p < 0.0001) and vegetation type (F2,10 = 6.63, p = 0.015), and the effect of vegetation type depended on buffer width (interaction F4,40 = 3.05, p = 0.028). There were large, statistically significant improvements in sediment retention with each increase in width regardless of vegetation type (F2,40 
43, p < 0.0001 for width tested within each vegetation type). Mean sediment retention averaged 83, 94, and 99% for 1-, 2-, and 6-m buffer widths, and these differences corresponded to large differences in sediment mass in runoff (8.3, 3.2, and 0.6 kg, respectively). Effects of vegetation type were smaller than those of buffer width, and they diminished at larger buffer widths. For 1- and 2-m simulations, sediment retention was significantly less in upland plots than either transitional or wetland plots (F2,40 = 10.7 and 11.6, p = 0.0002 and 0.0001, respectively). For 6-m simulations, sediment retention did not vary significantly among vegetation types (F2,40 = 2.40, p = 0.10). Clipping treatment (F1,10 = 2.60, p = 0.15) and its interactions with all other factors (F 0.78, p 0.46) did not affect sediment retention significantly. Slope, included as a covariate, did not affect sediment retention in this analysis (F1,40 = 1.02, p = 0.32). Because flow rates were equal, statistical results for average sediment concentration were essentially a mirror image of those for sediment retention and are not reported. Mean, flow-weighted sediment concentrations corresponding to results shown in Fig. 1 ranged from 0.2 to 9.9 g/L.
|
0.89), which were due mainly to the gross differences among vegetation types summarized above. When different buffer widths were analyzed separately, sediment retention was explained best by biomass and slope (1- and 2-m buffer widths) or slope alone (6-m buffer width); these models explained about half the variation in sediment retention in 1- and 2-m buffers but only a small percentage of variation in 6-m buffers. Neither runoff rate nor stubble height was a significant factor in any regression analysis, confirming that minor variations in runoff among simulations did not affect inferences about controls of sediment retention and that height of residual vegetation did not affect sediment retention even after accounting for other factors. |
DISCUSSION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
2 m wide will be vulnerable to much higher rates of sediment delivery. Results of this study and previous buffer research are consistent, suggesting that until further research is completed on semiarid rangelands, general recommendations can be extended to other locations with caution. Sediment retention rates in individual runoff simulations were consistent with the 40 to 100% range reported for most studies of herbaceous filter strips and forested buffers in humid regions (Gilliam, 1994; Dosskey, 2001).
Effects of Buffer Width, Slope, and Physical Conditions
Within the range of conditions evaluated, buffer width had the greatest effect on sediment retention. Averaged across other factors, decreasing buffer width from 6 to 1 m reduced average sediment retention from 99 to 83%, which corresponded to 13 times more sediment in runoff. In contrast, moving from wetlands to uplands or increasing slope from 2 to 20% each reduced average sediment retention from 96 to 91%, which corresponded to just 2.5 times more sediment. For situations where easy-to-administer, rule-based guidelines are needed, specifying an adequate buffer width is likely to be the simplest way to assure effective buffers for nonpoint-source sediment. Based on study results, 6 m is suggested as a starting point for designing rangeland buffers for sediment retention, consistent with cropland filter strip guidelines (USDA Natural Resources Conservation Service, 2000). Under the conditions tested, even the least favorable sites studied (uplands with 10–20% slope and 90–265 g/m2 biomass) achieved 95% sediment retention at a 6-m width, suggesting this width may be effective across diverse rangeland riparian sites.
Where flexible guidelines for site-specific buffer design are appropriate, narrower widths than 6 m may be adequate on level or gently sloping areas with dense vegetation, such as moist floodplain sites. The regression model indicated that under favorable site conditions (wetlands with 2% slope and 1000 g/m2 biomass), a 2-m buffer would achieve 96% retention. Numerous studies have shown that most sediment is retained in the first 1 to 3 m of filter strips (Tollner et al., 1976; Daniels and Gilliam, 1996; Robinson et al., 1996; Pearce et al., 1998b). Incremental improvements in sediment retention are relatively small beyond 7.5-m widths (Schmitt et al., 1999). However, because the sediment used in this study was relatively coarse (predominantly fine sand and silt), the retention rates reported should be considered optimistic and do not support use of very narrow (e.g., 1–2 m) buffers for general control of nonpoint-source pollutants. Buffers of a given width are less effective for retention of relatively mobile contaminants such as suspended clay, bacteria, or dissolved nutrients and herbicides than for coarse sediment (Schmitt et al., 1999). With the sediment, runoff regime, and plot conditions used in this study, most of the coarser sediment (fine to coarse sand) was unlikely to leave any of the plots, and differences in buffer width and plot characteristics probably affected the amount of clay, silt, and very fine sand transported.
Sediment transport and deposition in vegetated filter strips are controlled mainly by amount and velocity of runoff (Tollner et al., 1976; Correll, 2000). Infiltration enhances pollutant retention by reducing runoff volume, and decreasing slope enhances performance by reducing energy for sediment transport (Daniels and Gilliam, 1996; Robinson et al., 1996). In this study, uniform runoff rates were imposed, eliminating potential effects of site characteristics on infiltration and runoff, but differences in infiltration among riparian vegetation and site types are likely to be important under natural runoff conditions (Frasier et al., 1998). The modest range of slopes evaluated in this study (2–20%) limited its ability to characterize effects of slope on sediment retention. Steeper slopes are common on rangelands and are likely to result in lower sediment retention than reported here.
Effects of Buffer Vegetation and Surface Conditions
Results indicated that the dense vegetation of moist and wet riparian sites generally retained sediment effectively, whereas lower sediment retention was associated with sparse vegetation. With few exceptions, lower than average sediment retention for each buffer width occurred in upland plots or transitional plots with below-average biomass. Biomass, basal cover, and bare area explained sediment retention reasonably well and were strongly correlated with each other.
Pearce et al. (1998a) also reported high sediment retention (>90% in 10- and 2-m plots) in dense sedge and grass-sedge riparian communities in northern Colorado rangeland. Differences in sediment retention in that study were explained partially by variation in plant density, cover, and growth form, bare soil area, surface roughness, slope, and runoff, but dominant controls were not apparent. Sediment retention in simulated and real filter strip vegetation increases with stem density (Tollner et al., 1976; Munoz-Carpena et al., 1999; Ghadiri et al., 2000).
The higher risk of sediment transport through upland vegetation appeared to reflect fewer barriers to flow and presence of relatively direct, open flow paths in some plots. In rangelands, sparse vegetation like this is common along stream reaches where shallow ground water or seasonal flooding do not supplement rainfall. Upland range studies show that erosion and sediment transport are sensitive to surface cover, especially when very sparse. Linse et al. (2001) reported that variation in erosion from upland plots correlated well with cover when less than 30%, but above 30% cover, erosion rates were low and unrelated to cover. They interpreted these differences as reflecting better development of microchannel networks at low cover.
Several observations suggest that practical field indicators of a site's potential for sediment retention should emphasize major differences in vegetation that can be evaluated visually rather than accurate, quantitative measurements. First, regression models with plant cover, bare area, vegetation type, and biomass fit data about equally well; in other words, even describing the amount of vegetation as low, medium, or high predicted sediment retention as well as time-consuming biomass measurements. Second, gross differences in flow paths between sparse and dense vegetation appeared to influence sediment retention. Though many people have observed that spatial patterns of plant cover and microtopography affect runoff patterns and sediment transport in rangelands, such effects have been difficult to quantify and predict, and it may only be practical to identify major differences in cover and flow patterns that affect erosion and sediment movement (Linse et al., 2001). In fact, presence of significant plant cover may be most important. In a study of grassed filter strip performance in south-central Montana cropland, presence of vegetation was more important than the species planted (Fasching and Bauder, 2001). Finally, factors that are not easily quantified may be important or random variability may be high. In the present study, much of the variation in sediment retention within vegetation types was not correlated with vegetation amounts and surface characteristics. Observations suggested that litter dams, surface depressions, and open flow paths sometimes affected sediment retention, as observed by others (Pearce et al., 1998a; Ghadiri et al., 2000; Linse et al., 2001).
Role of Stubble Height and Grazing
Results of this study and previous research do not support the use of stubble height as a predictor of buffer capacity for retention of sediment in overland flow. Pearce et al. (1998a)(b) applied simulated rainfall, run-on, and sediment to 10- and 2-m-long plots that were clipped to the ground surface, clipped to 10 cm, or left unclipped. Stubble height generally did not affect runoff (Frasier et al., 1998) or sediment retention except over distances less than 2 m (Pearce et al., 1998a,b). In deeper flows representative of concentrated runoff or overbank flooding, short stubble enhanced sediment deposition most, but taller vegetation favored subsequent retention (Abt et al., 1994; Clary et al., 1996).
Although one of the goals of riparian grazing guidelines has been to reduce nonpoint-source pollution (Clary and Webster, 1989), publications that recommend stubble height criteria generally do not argue that they predict pollutant retention. Instead, stubble height can serve as an indirect indicator of trampling, soil compaction, streambank damage, and shrub browsing, as well as a direct measure of herbaceous plant defoliation (Clary and Webster, 1989; Clary, 1995; Clary and Leininger, 2000). This approach reflects the historical focus of rangeland riparian research, particularly in the U.S. Pacific Northwest, on effects of grazing on vegetation, channel characteristics, aquatic habitat, and fisheries, in contrast to the midwestern and eastern USA, where water quality has been emphasized (Platts, 1991; Dosskey, 1998). Research has demonstrated improvements in riparian vegetation and aquatic habitat under moderate or light cattle grazing (35–50 or 20–25% forage utilization, 10 or 14 cm stubble height, respectively) (Clary, 1999). The limited research now available on sediment retention in rangeland riparian areas shows that where herbaceous vegetation is dense, commonly recommended stubble height criteria in the 5- to 15-cm range (Clary, 1995; Clary and Leininger, 2000) will probably maintain vegetation capable of trapping sediment in either shallow overland runoff (Pearce et al., 1998a, b; this study) or overbank flow (Abt et al., 1994; Clary et al., 1996).
Other research shows that grazed riparian buffers can protect water quality but that intensive grazing may sometimes reduce their effectiveness. A grazed buffer in Denmark retained all sediment and particulate P in overland flow (Kronvang et al., 2000). In contrast, sediment retention was inadequate before elimination of intensive grazing from a riparian filter strip in a steep New Zealand pasture (Cooper et al., 1992). In two Montana studies, grazing did not cause consistent increases in sediment, nitrate, or phosphate; coliform counts were elevated but usually did not exceed water quality standards (Finck et al., 2000; C. Marlow, personal communication, 2001). Streamside grazing and associated trampling affect hydrologic processes much more than clipping (Flenniken et al., 2001) and contribute directly to nutrient and bacterial contamination (Trlica et al., 2000).
|
ACKNOWLEDGMENTS |
|---|
|
NOTES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
REFERENCES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Related articles in JEQ:
This article has been cited by other articles:
|
|
 |
|
  A. D. Ziegler, J. Negishi, R. C. Sidle, P. Preechapanya, R. A. Sutherland, T. W. Giambelluca, and S. Jaiaree Reduction of stream sediment concentration by a riparian buffer: filtering of road runoff in disturbed headwater basins of montane mainland southeast Asia. J. Environ. Qual., January 1, 2006; 35(1): 151 - 162. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Journal of Natural Resources and Life Sciences Education |
Vadose Zone Journal | ||||
| Soil Science Society of America Journal | Journal of Plant Registrations | The Plant Genome | |||
