| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Iowa Dep. of Natural Resources, Geological Survey Bureau, 109 Trowbridge Hall, Iowa City, IA 52242-1319
* Corresponding author (kschilling{at}igsb.uiowa.edu)
Received for publication June 25, 2001.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
Abbreviations: SQW2, Squaw Creek watershed outlet • USGS, United States Geological Survey • WNT1, upper portion of the Walnut Creek watershed • WNT2, lower portion of the Walnut Creek watershed
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
In Iowa, the most common nonpoint-source pollutant is NO3–N, and, on an annual basis, it is primarily delivered to streams through baseflow ground water discharge and tile drainage (Hallberg, 1987). Transport of NO3–N has been shown to vary markedly with season, with stormflow contributions increasing during late winter and spring (Alberts et al., 1978; Owens et al., 1991; Jaynes et al., 1999; Pionke et al., 1999). Variations in streamflow composition and pollutant loading rates may also occur due to geologic controls on ground water discharge (Schnabel et al., 1993) and land use differences (Owens et al., 1991). Seasonal agricultural activity or inputs from storm events further accentuate intermittent loading of nonpoint-source pollutants (Carpenter et al., 1998).
Streamflow and concentration data presented herein describe variations in baseflow and stormflow transport of NO3–N, Cl, and SO4 in two rural watersheds in central Iowa (Fig. 1) . Water quality conditions in both watersheds have been monitored as part of the Walnut Creek Watershed Monitoring Project, a project designed to evaluate changes in chemical transport resulting from conversion of large tracts of land from row crop agriculture to native prairie in a treatment watershed (Walnut Creek) compared with a highly agricultural control watershed (Squaw Creek) (Schilling and Thompson, 1999, 2000). The project was established in 1995 in relation to the watershed restoration activities implemented by the U.S. Fish and Wildlife Service at the Neal Smith National Wildlife Refuge.
|
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
The study area is in a humid, continental region with average annual precipitation of 850 mm. Highest monthly rainfall totals typically occur in May and June, although large storms occurring throughout the summer can lead to rapid rises in discharge. High streamflows are also associated with major snowmelt events that often occur in February. Three United States Geological Survey (USGS) stream gauges used in this study are in the upper (WNT1) and lower (WNT2) portions of the Walnut Creek watershed and at the Squaw Creek watershed outlet (SQW2) (Fig. 1). Discharge in both watersheds tends to be flashy, displaying rapid responses to precipitation. For example, stream discharge at WNT2 has ranged from a high of 50085 m3/h to a low of 6.8 m3/h (Schilling, 2000).
In 1992, land use in the Walnut and Squaw creek watersheds consisted of 70% row crop agriculture and 27% grass (Schilling and Thompson, 1999). From 1992 to 2000, 876 ha or 16.8% of the Walnut Creek watershed was converted from row crop agriculture to native prairie (Schilling, 2001). During the same time period, land use in the Squaw Creek watershed remained essentially unchanged. Most restoration occurred during the period 1993 to 1998 when approximately 140 ha/yr were converted from row crop agriculture to prairie. Prairie restoration has ranged from less than 60 to 210 ha/yr. Although some prairie plantings were initially targeted along stream corridors for purposes of improving stream water quality, the majority of planting sites were sited piecemeal in the watershed based on political, ecological, and operational needs of the refuge. In addition to the land conversions, 294 ha of refuge-owned lands in the Walnut Creek watershed (5.6% of the watershed) has remained in row crop production during the restoration period but is farmed on a cash-rent basis by local farmers. In these areas, improved agricultural management practices are mandatory. No fall application of fertilizer is allowed and a maximum of 112 kg/ha of nitrogen is allowed on conventional corn (Zea mays L.) rotation. A typical application rate of nitrogen fertilizer in Squaw Creek watershed is approximately 168 kg/ha (Schilling and Thompson, 1999).
Combining the prairie planting areas and restricted application areas, land use changes have been implemented on 22.4% of the Walnut Creek watershed, nearly all of which is in the core of the watershed between the upstream and downstream gauging stations (Fig. 1). The upper portion of the Walnut Creek watershed, above the upstream gauge (WNT1; Fig. 1) is predominantly row crop land use, averaging about 75 to 80% in any given year. Including other U.S. Fish and Wildlife Service–owned land in the watershed (cool season grass or wooded areas), USFWS controls 1759 ha, or 33.7%, of the Walnut Creek watershed above the WNT2 gauging station. From 1992 to 1997, nitrogen applications in the Walnut Creek watershed were reduced by an estimated 18.1% (Schilling and Thompson, 1999).
Data Collection
Surface water samples have been collected at upstream and downstream locations in Walnut and Squaw Creek watersheds on a weekly to bimonthly basis since 1995 as part of an ongoing monitoring project (Schilling and Thompson, 2000). Water samples were analyzed for NO3–N, Cl, and SO4 by the University of Iowa Hygienic Laboratory using standard methods. Summary concentration data for NO3–N, Cl, and SO4 for Water Years 1996 to 2000 are presented in Table 1
. For the five-year period, 86 to 90 water samples were collected at the stream gauging sites and analyzed for NO3–N, Cl, and SO4.
|
Chemical Load Estimation
Hydrograph separation into baseflow and runoff components was performed on streamflow data collected at the three USGS gauging sites using an automated method developed by Sloto and Crouse (1996). A local-minimum method was used, which essentially connects the lowest points on the hydrograph and provides estimates of daily baseflow discharge between local minimums by linear interpolation (Sloto and Crouse, 1996) (Fig. 2)
. Daily runoff discharge was determined at each stream gauge site by subtracting daily baseflow discharge from daily streamflow discharge. For purposes of this study, runoff was defined as contributions from overland flow and subsurface stormflow (or interflow) (Freeze and Cherry, 1979).
|
Daily chemical load data were tabulated and summarized by month and water year. Load data were normalized on a unit area basis by dividing the total annual load at each gauging site by the watershed area above the gauge. In the case of Walnut Creek watershed, the load per unit area between the two gauge sites was determined by subtracting the load estimated at WNT1 from WNT2. Flow-weighted concentrations were calculated by dividing the daily constituent load by daily discharge. Statistical comparisons among data sets were performed using the MINITAB Release 13 statistical software package (MINITAB, 2000).
The accuracy of the modeling approach was evaluated by comparing actual concentration data with estimated flow-weighted concentrations generated by the ESTIMATOR model for the entire data record and for the baseflow period only (Fig. 3) . In general there was good correlation between the modeled concentrations and measured values. Pearson correlation coefficients ranged between 0.77 and 0.94 for NO3–N concentrations, 0.36 and 0.73 for Cl concentrations, and 0.58 and 0.89 for SO4 concentrations. All but baseflow concentrations of Cl at WNT1 were significant at P < 0.001 (Cl at WNT1 was significant at P = 0.06). Correlation coefficients were typically higher for baseflow periods than for the entire data record, probably due to less variability in discharge associated with the baseflow data set.
|
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
|
|
|
|
|
Average NO3–N concentrations are approximately 1 mg/L lower in Walnut Creek compared with Squaw Creek (Table 1) and recent statistical analysis has indicated that a statistically significant change in NO3–N is occurring in the treatment watershed (Walnut) while adjusting for the control watershed (Squaw) (P < 0.05; Schilling, 2001). Schilling (2001) reported that NO3–N concentrations have decreased by 0.0028 mg/(L wk) over the course of the sampling period (326 wk), equivalent to 0.91 mg/L since 1995. While not statistically significant, downstream concentrations in Walnut Creek are also decreasing over time while considering the upstream control at WNT1 (P = 0.19; Schilling, 2001).
Concentrations of Cl and SO4 average approximately 3 mg/L less at WNT2 compared with SQW2 (Table 1). The Cl concentrations at WNT1 are comparable with SQW2 but SO4 concentrations are substantially lower than either WNT2 or SQW2 (Table 1).
It is noteworthy that the summary statistics describing the subset of water quality data describing baseflow concentrations were remarkably similar to statistics describing the entire water quality record (Table 1). Considering fluctuations of concentrations that occur throughout the year, the comparable data sets suggest that the samples collected during baseflow periods were scattered throughout the year and not focused during any one season. Minor concentration differences between Cl and SO4 data sets at SQW2 (Table 1) may be similar to observations of increasing dissolved solids concentrations found in other southern Iowa watersheds during particularly low baseflow periods, which has been linked to increasing contributions from bedrock sources (Horick and Steinhilber, 1973).
Chemical Loads
Total export of NO3–N, Cl, and SO4 from Walnut Creek (WNT2) and Squaw Creek (SQW2) watersheds was similar, ranging from 23.9 to 28.5 kg/ha of NO3–N, 31.7 to 38.8 kg/ha Cl, and 59.4 to 66.8 kg/ha SO4 (Table 2). Annual losses of NO3–N were greatest in 1998 when they ranged from 43.6 to 56.3 kg/ha. Maximum monthly export loads of NO3–N and Cl exceeded 20 Mg in most years, with load peaks typically occurring in February and May–June of any given year (Fig. 5). Peak loads for all constituents occurred in February and May 1996, May 1997, and June 1998 (Fig. 5). Although total export loads of NO3–N, Cl, and SO4 were lower in Walnut Creek (WNT2) compared with Squaw Creek (SQW2), paired t tests were not statistically significant (P > 0.1).
Differences in NO3–N and Cl losses between Walnut and Squaw Creek watersheds can be traced to substantially lower losses emanating from the lower portion of Walnut Creek watershed (WNT2 - WNT1) (Table 2). Nitrate and Cl losses from upper Walnut Creek were similar or slightly higher than Squaw Creek, but losses from lower Walnut Creek were substantially less (Table 2). Differences in chemical loss rates among watershed areas appear to be more evident in baseflow than stormflow. Average baseflow losses of NO3–N and Cl in lower Walnut Creek watershed were 40 to 50% less than upstream Walnut Creek and Squaw Creek, whereas average stormflow losses were approximately 30 to 40% lower for NO3–N and 5 to 10% lower for Cl (Table 2). Paired t tests indicated that total and baseflow losses of NO3–N and Cl from lower Walnut Creek were significantly lower than either upstream Walnut Creek (WNT1) or Squaw Creek (SQW2) (P < 0.05).
In both watersheds, export of NO3–N, Cl, and SO4 occurred primarily with baseflow, ranging from 61 to 68% at the watershed outlets (WNT2, SQW2; Table 3). Differences in baseflow percentages of chemical loads were noted between upper and lower Walnut Creek watershed. Baseflow percentages at WNT1 were similar to those at SQW2 but those in the lower portion of Walnut Creek were less. In all cases, the percentage of chemical load exported with baseflow was higher than the baseflow percentage of streamflow alone, suggesting that a higher proportion of chemical transport occurred through baseflow discharge than stormflow.
Annually, baseflow export of NO3–N, Cl, and SO4 varied from 40 to 80% of the total export load and tended to follow patterns dictated by discharge (Fig. 4 and 5). Within any given year, chemical losses from stormflow comprised a much larger percentage during February (Fig. 6). Contributions from baseflow inputs tended to increase through the latter part of the year so that by December inputs of NO3–N, Cl, and SO4 to streams were derived from more than 90% baseflow (Fig. 6).
Losses of NO3–N and Cl during baseflow and stormflow were variable among years. At the Walnut Creek outlet (WNT2), baseflow NO3–N losses ranged from 7 to 28 kg/ha and stormflow losses ranged from 3 to 16 kg/ha (Table 2). In Squaw Creek, combined NO3–N losses from baseflow and stormflow ranged from 13 to 56 kg/ha. Baseflow and runoff losses were equally variable among years within Walnut Creek watershed, with baseflow NO3–N losses ranging from <5 kg/ha in lower Walnut Creek to 43 kg/ha in upper Walnut Creek and runoff losses ranging from 2 to 20 kg/ha, respectively. Regardless of watershed area, losses of NO3–N, Cl, and SO4 were all greatest in years with higher precipitation and streamflows.
Few consistent differences are noted in SO4 loads among watershed areas (Table 2). Schilling and Wolter (2000) reported that SO4 concentrations in surface water did not relate significantly to agricultural land use in the Walnut Creek watershed, so SO4 provided a marker for tracking baseflow inputs independent of land use. Thus, the lack of consistent trends in SO4 losses among watershed areas suggests that differences in NO3–N and Cl may be related to land use changes in Walnut Creek watershed.
Flow-Weighted Concentrations
Flow-weighted concentrations followed a similar pattern exhibited by chemical mass losses (Table 4)
. Flow-weighted concentrations of NO3–N and Cl in baseflow and stormflow were higher in Squaw Creek and upstream Walnut Creek than lower Walnut Creek. Average flow-weighted concentrations of NO3–N were >10 mg/L in Squaw Creek and upper Walnut Creek but were 6.6 mg/L in lower Walnut Creek (Table 4). Similarly, concentrations of Cl were >14 mg/L in Squaw Creek and upper Walnut Creek but 9.2 mg/L in lower Walnut Creek. Sulfate concentrations were slightly higher at SQW2 compared with WNT2, and concentrations were higher in lower Walnut Creek compared with upper Walnut Creek (Table 4). Discharge of ground water from Pennsylvanian bedrock in lower reaches of Walnut Creek is believed to contribute to SO4 differences within the Walnut Creek watershed (Schilling and Wolter, 2001).
|
In general, average flow-weighted concentrations compared favorably with the average of all analyses measured during the same five-year period (Water Years 1996–2000) (Table 1).
Relations to Land Use
This study suggests that chemical loading rates are different in lower Walnut Creek watershed compared with upper Walnut Creek and Squaw Creek. Hydrologic changes and reduced chemical loads are consistent with the hypothesized effects of prairie restoration and nutrient management activities occurring in lower Walnut Creek watershed. However, attempting to attribute these differences solely to land use changes is problematic for several reasons. First, flow and chemical loads have varied according to precipitation. Areas that received greater rainfall produced more streamflow and subsequently higher chemical loads, so differences in loading patterns may, in some cases, be an artifact of variable rainfall input. Jaynes et al. (1999) noted that short-term studies of watersheds (four years in the case of Jaynes et al., 1999) are not capable of accurately assessing agricultural chemical losses without sufficient accounting for yearly rainfall differences. In this five-year study, rainfall has varied widely both spatially between Walnut and Squaw Creek watersheds and temporally among water years. Second, restoration activities have not been implemented on a uniform basis across the watershed area. Prairie restoration sites are located throughout lower Walnut Creek watershed and annual plantings have ranged from 60 ha/yr to more than 200 ha/yr. Thus, observations of hydrologic or load changes that may be directly attributable to prairie restoration are confounded when integrated across the landscape and over many time scales (Schilling and Thompson, 1999). Third, flow separations and chemical load data described in this study were estimates derived from actual streamflow and chemical sampling data. Attempting to develop a derivative relationship, such as rate of change, based on estimated data would, in all likelihood, exceed the accuracy of the estimated data. Using statistical methods to detect gradual changes over time is probably best left to actual data rather than derivative data. In this case, actual concentration data have shown a gradual decrease in NO3–N concentrations in Walnut Creek compared with Squaw Creek (Schilling, 2001).
Nevertheless, based on a five-year average of accumulated data, there is evidence to suggest that prairie restoration and land management changes may be affecting watershed hydrology and chemical loading rates. In terms of hydrologic differences, average annual baseflow from lower Walnut Creek containing most of the prairie restoration areas was lower than average baseflow from upper Walnut Creek or Squaw Creek watersheds. However, inputs from subsurface tile drainage in the watersheds confound interpretation of baseflow changes. Although ground water discharge from drainage tiles contributes to both baseflow and stormflow portions of the hydrograph, it is especially noticeable during dry periods when the tile water discharge increases sustainable baseflow. While tile water inputs were not quantified in upper Walnut Creek and Squaw Creek watersheds, a stream survey of lower Walnut Creek revealed 52 tiles entering the stream channel between the two gauges, 19 of which were flowing sufficiently for sampling in May 1999 (Schilling and Wolter, 2001). Flow from the tiles accounted for only 4% of the total flow from the watershed and 7% of the export loads of NO3–N, Cl, and SO4. Tiles discharging from restored prairie areas along the main stem of Walnut Creek showed low concentrations of NO3–N (<1 mg/L) and Cl (<3 mg/L) (Schilling and Wolter, 2001). Tile water discharge and solute concentrations are believed to be greater in upper Walnut Creek and Squaw Creek watersheds where land use is dominated by row crop production. Many first-order streams in both watersheds originate as tile flow from row crop fields.
Unlike baseflow discharge, mean annual stormflow from lower Walnut Creek was higher than the other watershed areas. Although contrary to expectations, some increased stormflow can probably be traced to differences in precipitation among the watershed areas. Despite sharing a common basin divide to minimize differences, variability in precipitation between lower Walnut Creek and upper Walnut Creek and Squaw Creek watersheds was often quite large, with lower Walnut Creek receiving an average of more than 100 mm of precipitation over the five-year period. Greater rainfall in lower Walnut Creek would contribute to more runoff and may have masked possible reductions in stormflow discharge due to prairie restoration. However, three rain gauges in these 5000-ha watersheds are probably insufficient to fully characterize the variability in precipitation occurring in the area. Clearly, additional rainfall monitoring will be needed if hydrologic changes are to be fully elucidated at the watershed scale.
Despite mixed results in terms of flow, losses of NO3–N and Cl were substantially less in lower Walnut Creek watershed, which contained the prairie restoration sites. Losses of NO3–N and Cl from lower Walnut Creek were approximately one-half the mass lost from upper Walnut Creek and Squaw Creek watershed areas, which were dominated by row crop fields. Reduction of NO3–N and Cl from baseflow inputs was greater than for stormflow, but it was noteworthy that stormflow loads of NO3–N and Cl from lower Walnut Creek were also less than Squaw Creek or upper Walnut Creek. This was evident in the flow-weighted concentration data that showed average stormflow concentrations of NO3–N and Cl approximately 4 to 5 mg/L less in runoff from lower Walnut Creek. Thus, greater stormflow discharge from lower Walnut Creek was not manifested in correspondingly greater chemical loads.
While NO3–N losses were clearly less in the lower Walnut Creek watershed compared with Squaw Creek and upper Walnut Creek, NO3–N losses remained considerably higher than those reported for pristine tallgrass prairie watersheds (Dodds et al., 1996). Observations of surface water flowing from four watersheds on Konza Prairie Research Natural Area in Kansas indicated annual NO3–N losses <0.2 kg/ha. Previous research at Walnut Creek watershed, consisting of a single sampling event completed during a baseflow period in May 1999 showed that some small (<100 ha) interior watersheds containing nearly 100% restored prairie had NO3–N and Cl in surface water <1 mg/L and <3 mg/L, respectively (Schilling and Wolter, 2001). Extrapolating from a single sampling event, annual NO3–N losses for restored prairie watersheds in Walnut Creek watershed were <1 to 2 kg/ha.
Average annual losses of NO3–N from Squaw Creek and upper Walnut Creek (28–34 kg/ha) are similar to those reported for a different Walnut Creek in Story County, Iowa (Jaynes et al., 1999). Although wide variability in annual losses was observed, average NO3–N loss from Walnut Creek (Story) was 29 kg/ha for the period 1992 to 1995 (Jaynes et al., 1999). Comparable NO3–N losses are noteworthy considering that Walnut Creek in Story County is located in a heavily agricultural and tile-drained watershed developed on poorly drained, recent glacial materials (Wisconsinan Des Moines Lobe). In contrast, Walnut and Squaw Creek watersheds in Jasper County are located on older glacial sediments with well-developed drainage. However, land use in Squaw Creek and Walnut Creek (Story) is similar [73 and 83% row crop agriculture, respectively; Walnut (Story) data from Hatfield et al., 1999]. Perhaps nitrogen export from the two watersheds is comparable despite major differences in landscape morphology because of greater reliance on artificial tile drainage in the Walnut Creek (Story) watershed in lieu of steeper slopes and increased contributions from stormflow in the Jasper County watersheds. Although the manner of NO3–N delivery from cropped fields to streams may be different in the two landscape settings, NO3–N losses in Iowa streams are invariably linked to row crop agriculture (Schilling and Libra, 2000).
Results from this study demonstrate the utility of partitioning loads into baseflow and stormflow components to identify sources of pollutant loading to streams and assist investigation of land use effects on water quality. In the case of Walnut Creek and Squaw Creek watersheds, transport of NO3–N, Cl, and SO4 occurred primarily with baseflow, but loading rates varied considerably among watershed areas. Implementing land use or land management changes that serve to reduce baseflow loading rates to streams may be an appropriate best management practice in some cases to reduce pollutant concentrations on a watershed scale.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  K. E. Schilling and J. Spooner Effects of Watershed-Scale Land Use Change on Stream Nitrate Concentrations J. Environ. Qual., October 27, 2006; 35(6): 2132 - 2145. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. C. Scharf, N. R. Kitchen, K. A. Sudduth, J. G. Davis, V. C. Hubbard, and J. A. Lory Field-Scale Variability in Optimal Nitrogen Fertilizer Rate for Corn Agron. J., March 1, 2005; 97(2): 452 - 461. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 |
(average estimated values ± one standard error).
