USGS North Carolina WSC - Projects - Sediment Collection Surry County

The design of this study follows that of a before-after control impact paired (BACIP) experiment. Two impacted sites were selected, one in which BMPs were implemented (Bull Creek), and a second site, in which channel stabilization, a type of in-stream restoration, was performed in addition to BMP implementation (Pauls Creek). A third site, in which neither BMP implementation nor in-stream restoration was planned was selected as a control site (Hogan Creek). Samples were collected before and after BMP implementation and in-stream restoration.

The paired basin study design requires that the basins to be compared are similar except for the variables of interest (in this case, the stream improvement activity). Land-use data compiled for 2002, by the U.S. Department of Agriculture, for the Bull Creek and Hogan Creek basins were similar. Pasture and hay production were the dominant land uses in these basins and represented slightly more than half of the total land area. Much of the land in the Pauls Creek drainage basin is forested and the northern part of the basin is in the Blue Ridge physiographic province.

Stream gages were established at each of the sites and stage and streamflow data were obtained from 2003-2007. Suspended-sediment samples were collected on a biweekly basis. Water-quality samples were collected on a quarterly basis from 2003-2007. Habitat assessments were made annually from 2004-2006. Benthic invertebrate samples were obtained annually from 2004-2007.

Assessing the effects of agricultural BMP implementation and stream channel stabilization on water quality, stream habitat, and biota was complicated by climatic conditions and changes in land use that occurred during and prior to the study. Complications that occurred during the study included road construction in the Pauls Creek basin; withdrawals from an upstream impoundment and various land use changes in the Bull Creek basin; and logging and BMP implementation in the Hogan Creek basin (the control site). Drought conditions and corresponding declines in streamflow contributed to trends in water quality observed at the study sites. Complicating factors occurring prior to the study included implementation of BMPs in all basins, stream channel modifications in the Pauls Creek basin, and the construction of an impoundment upstream from the Bull Creek site.

Similarities in size and land use characteristics of the Bull and Hogan Creek basins indicate that the Hogan Creek site is an appropriate control site for the Bull Creek site. However, differences in size and land use characteristics indicate that the Hogan Creek site may not be an appropriate control site for the Pauls Creek site.

Records of stage were obtained from the gaging stations established at each of the study sites. The gaging station at the Hogan Creek site was established on August 11, 2004; the gaging station at the Bull Creek site was established on June 1, 2004; and the gaging station at the Pauls Creek site was established on August 26, 2004. Streamflow measurements were made at each of the three sites to establish stage-discharge relations to enable calculation of discharge from the stage data. Because of dry conditions during much of the study, most of the discharge measurements were made at low stage. The relation between stage and flow at mid-to-high ranges was predicted from a one-dimensional step-backwater hydraulic model using the U.S. Army Corps of Engineers Hydrologic Engineering Centers River Analysis System (HEC-RAS) software.

Daily mean streamflow data are available for water years 2004-2007 for Hogan Creek, Bull Creek, and Pauls Creek. Streamflow conditions in the study area were similar to or less than long-term annual means during the study. Prolonged drought conditions impacted the ability to obtain sediment samples and discharge measurements during high streamflow conditions.

Physical characteristics of water, including water temperature, pH and dissolved- oxygen concentration show generally similar data distributions. Differences in water temperature, dissolved-oxygen concentration, pH among the basins were not statistically significant. Pauls Creek had statistically significant higher specific conductance and lower alkalinity than the other sites. These differences are likely a function of the different geologic setting and larger drainage area of the Pauls Creek basin in comparison to the Bull and Hogan Creek basins.

Total ammonia plus organic nitrogen, dissolved ammonia, and dissolved nitrite plus nitrate, and total nitrogen all showed statistically significantly higher concentrations at Bull Creek than the other sites. Concentrations of nitrogen constituents in samples from Bull Creek were high in comparison to these reported values for forested and rural basins in North Carolina and the southeast. Forested riparian buffers, which are effective in decreasing stream concentrations of nitrogen, are absent or minimal along much of the main stem of Bull Creek in the vicinity of the study site. In contrast, riparian zones are largely forested in the vicinity of the Pauls Creek and Hogan Creek study sites. Potential sources of nitrogen in Bull Creek include a campground, livestock, and poultry.

Dissolved orthophosphate showed statistically significantly higher concentrations at Hogan Creek than the other sites. Dissolved orthophosphate and total phosphorous concentrations for the study basins were generally near detection limits. Total phosphorus concentrations in samples from the study sites were generally within the ranges reported for Piedmont forested basins, with the exception of one high (0.47 mg/L) sample from Hogan Creek.

Suspended-sediment loads and yields were calculated for the three basins. Strong seasonal variations in suspended-sediment load were evident for the study sites with considerable basin-to-basin and year-to-year variation. Sediment yield in 2007 at the Pauls Creek site, following completion of channel stabilization activities, was less than yields calculated for the other sites. This reduced yield may have been due in part to stabilization of the stream channel upstream from the Pauls Creek site or to effects of drought on streamflow. Suspended-sediment loads for the study sites fall within the range of 60th to 90th percentiles when compared to 20 other basins across the Southeast. Suspended-sediment load equations were based primarily on data for low streamflow conditions and mid- to high-range samples were not collected before and after stream improvements were made. Thus, the accuracy of load estimates is limited when extrapolated to mid- to high- streamflow conditions.

Analysis of variance was used to test differences in suspended-sediment concentration and sediment discharge between sites (stream improvement and control) and differences before and after the improvements. No statistically significant (p=0.05) differences were found for suspended-sediment concentrations and sediment discharge at Bull Creek (the BMP site) when compared to those at Hogan Creek (the control site). Comparison of Pauls Creek with the control site indicated statistically significant decreases in suspended-sediment concentrations and sediment discharge before and after channel stabilization at Pauls Creek. These findings, coupled with the reduced sediment yields estimated for Pauls Creek following channel stabilization, provide evidence of improved stream quality at the Pauls Creek site. Other studies have shown rapid improvements in stream quality following in-stream restoration.

Bull and Hogan Creeks showed similar habitat characteristics. Habitat characteristics at the Pauls Creek reach differed from the two smaller basins in that lacked pools in the study reach, and had more run areas than riffle areas following channel stabilization. The Bull Creek reach showed little change in habitat characteristics during the study period. Habitat characteristics at the Pauls Creek site showed the greatest variation over time because of the stream channel modifications that were part of the channel stabilization effort. The duration of this study likely was inadequate for detection of changes in habitat characteristics and observed changes in habitat could not be definitively associated with the basin treatments. However, the habitat measurements may prove useful in a future reassessment of the study sites.

Statistical analysis showed that invertebrate assemblages in the streams differed from one another and changed significantly from 2004 to 2007. Based on the North Carolina Index of Biotic Integrity (NCIBI) Bull Creek had the poorest conditions and also showed the greatest improvement in invertebrate communities during the study with increased diversity and taxa richness. The assemblage change trajectory of Bull Creek differed from that of Hogan and Pauls Creeks, in which decreases in richness and diversity were observed. The improvement at Bull Creek suggests that implementation of BMPs had a positive effect on this stream whereas channel stabilization and BMP implementation in the Pauls Creek did not have an observable improvement. Changes in Pauls Creek tend to mirror those of the control stream (Hogan Creek) suggesting that the in-stream restoration and BMP implementation in Pauls Creek have not resulted in significant change. This might be anticipated given the excellent condition (based on the NCIBI) of Pauls Creek at the start of the study. However, attributing observed changes in benthic invertebrate assemblages to stream-improvement activities is questionable given the drought conditions, uncontrolled disturbances, and the small portion of the basins modified by BMPs and in-stream restoration. The duration of this study probably was insufficient to detect of changes associated with best management practices and stream channel restoration.

Several lessons were learned from this study regarding the BACIP design, site selection and characterization, and data collection. First and foremost was the recognition that experimental designs are difficult to implement for basin-level studies. Efforts to assess the effectiveness of agricultural BMPs and in-stream restoration were complicated by unplanned land-use activities that occurred during this study as well as historical land-use activities. Differentiating between the effects of these unplanned activities and the planned BMP implementation and in-stream restoration activities on stream quality was not feasible using the BACIP design. Perhaps this design would be more appropriate for assessing smaller stream basins or basins where land-use activities are less diverse and more regulated, such as public lands.

Site selection is also an important component of water-quality studies. The BACIP design requires that the sites be similar in all aspects except the impact, which in this study, was to be either BMP implementation or in-stream restoration. Careful characterization of the study sites prior to selection is essential. To characterize the sites, the physical setting and land-use activities of the drainage basins should be evaluated prior to selection. Physical characteristics should include basin size, topography, and geologic setting. In addition to field reconnaissance, land-use characteristics should be evaluated using GIS applications, aerial photographs, and information from local, state, and federal governmental agencies. Locations of features such as confined animal feeding operations and impoundments, which can directly impact water-quality conditions, can be identified with aerial photographs. Additional spatial information related to land-use characteristics such as population, land cover, agriculture, and is available from various sources. Because historical characteristics of a site also contribute to its present characteristics, an effort should be made to understand the past land use activities in the stream basin. Past land-use activities affect present day stream conditions. Aerial photographs obtained by the U.S. Department of Agriculture for most of the United States since the 1930s are one potential source of historical information. Historical maps, including topographic maps, can also provide information about past land use, such as the presence of impoundments or mills. The history of stream-improvement activities in the drainage basins of potential sites should also be obtained so that impacts from prior activities do not confound analysis of data.

The nature of the impacts should also be evaluated. In the case of agricultural BMPs, many are not designed to directly impact stream quality and some types are designed to impact only certain aspects of stream quality. For example, the construction of the cement-lined chicken composter in the Bull Creek basin would likely have minimal or no impact on stream sediment and a only a slight impact on stream nutrient concentrations considering the distance from the stream. Multiple BMPs implemented over a several- year period make determination of before and after periods difficult, and for short term studies, much of the study can fall into a “during” impact category and should not be included in the “before” and “after” comparison.

In addition to selection of impacted sites, the BACIP design requires selection of appropriate control sites. To identify control sites that are comparable to the site being impacted by a planned stream-improvement activity, the sites should be of similar size and physical setting and have similar past and present land use characteristics. In this study, based on size and land use, the Hogan Creek appears to be a suitable control site for comparison with the Bull Creek site; however, it does not appear to be suitable for the Pauls Creek study site. Control sites should not have stream-improvement activities in their drainage basins during or before the study. However, identifying similar sites is challenging and a great deal of effort is needed to identify candidate sites, especially control sites.

Data should be collected over a long enough period to show effects of the stream-improvement activity and to encompass variation associated with climatic conditions. The length of time over which data should be collected and frequency of collection varies with the type of data. For example, it is likely that suspended-sediment and nutrient concentrations would change more rapidly in response to livestock exclusion than would habitat or invertebrate assemblages. Several years may be needed for changes in habitat characteristics to occur following livestock exclusion as vegetation is established and stream banks are stabilized. Normal seasonal variations in water-quality and inter-annual variation in benthic invertebrate community structure may mask response to stream improvement efforts if data are not collected over a long enough period to accommodate this variability. In theory, the BACIP design should eliminate the variability associated with climatic conditions provided that the sites are in climatically similar locations and observations are made over a long enough period to allow for response to the impact and climatic variation.

The BACIP design is advantageous because it allows assessment of the effects of impacts with few sampling sites, requiring basically an impacted site and a control site. However, other study designs for evaluating impacts of stream-improvement activities, such as comparison of locations upstream and downstream from the impact locations, should be considered if land-use activities within the study site drainage basins cannot be controlled. Sampling upstream and downstream from stream-improvement areas should be considered to lessen the effects of unplanned land use activities and climatic variability as well as eliminate the need for control sites. However, comparison of upstream and downstream effects is best suited to areas where the impact covers a small portion of the drainage basin or is near the stream. Regardless of design, basing site selection on careful evaluation of drainage-basin characteristics is essential for a successful assessment of the effects of stream-improvement activities on stream quality.

To circumvent unnecessary complexity in data analysis such as that caused by unanticipated land use activities within the study basins, the suggestions for future studies include:

For future studies, to determine the amount of improvement that results from the implementation of BMPs, sampling locations upstream and downstream from the BMP area should be selected with samples routinely taken from each. Comparisons of upstream and downstream samples could then be analyzed and the effects of the BMP more readily identified. Likewise, any influences on the water quality of the stream in the stream basin upstream from the upstream sampling point would already be quantified thereby simplifying the analysis and enabling a determination of the effectiveness of the BMP in reducing suspended sediment concentrations.

Project Overview

USGS IN YOUR STATE

Sediment Collection Surry County

Project Summary