USGS Banner

Title: Assessment of Changing Land-Use Practices on Basin Sediment Yields and Provenance in Western North Carolina Using Multivariate Fingerprinting Techniques

Focus Categories: SED, G&G, SW

Keywords: Sedimentation, Geomorphology, Water Quality, Reservoir Siltation

Duration: From March 1, 2000 to February 28, 2001

Federal Funds $39,903

Non-Federal Funds $11,856

Principal Investigators: Jerry Miller, Lawrence Kolenbrander, Steven Yurkovich, and Mark Lord, Department of Geosciences and Natural Resources Management, Western Carolina University, Cullowhee, NC 28723

Congressional District No.: 11

Research Needs: Federal, State, and Local agencies have all argued that erosion and sedimentation resulting from anthropogenic activities represent a serious, if not the most significant, threat to aquatic ecosystems within the mountainous terrains of western North Carolina. Given the projections for unprecedented growth and development in the region during the next decade, the environmental impacts associated with enhanced sediment loads are likely to increase. It is, therefore, imperative that effective sediment control management practices are developed and implemented. The design of these management practices must be based on a firm understanding of the erosional and depositional processes that function within these high gradient environments, and the factors that control the magnitudes of erosion and sedimentation in any given area. Unfortunately, few detailed studies of sediment transport dynamics on hillslopes or within the axial channels of these mountainous terrains have been conducted. In fact, the topic of highest priority for the WRRI Advisory Committee for Fiscal Year 2001 is research into the development and performance of sediment control measures that may be applied to land-disturbing activities in the mountains (e.g., road construction and housing development). Particularly lacking is a coherent, quantitative understanding of the relative contributions of sediment to the region’s streams and rivers from differing forms of land-use practices. Clearly, it is necessary to determine where a majority of the sediment is coming from before it can be controlled using structural or bioengineering approaches. Also missing are insights into the natural rates of sedimentation that characterized the area prior to watershed development, and the variations in the rates of upland erosion that are associated with annual fluctuations in hydrologic conditions.

Expected Results: The proposed investigation will combine studies of reservoir bed sediments with multivariate fingerprinting and mixing model techniques to determine basin sediment yields and provenance within watersheds characterized by differing magnitudes of land-use. While the combined procedures have been used in other environments (particularly within the U.K.), they have not, to our knowledge, been applied in the southern Appalachians. Thus, this investigation will represent the first attempt to utilize these emerging analytical techniques in the mountainous environments of western North Carolina. The results obtained from this approach will provide valuable insights into the natural rates of upland erosion prior to watershed development. These data are essential to (1) determine the magnitude of human impacts on sediment yields, (2) assess the effectiveness of developed management practices, and (3) establish targets for the amount of sediment that can reasonably be removed from the aquatic system. In addition, the analysis will provide a preliminary, quantitative understanding of the relative contributions of sediment from delineated land-cover types in developed watersheds, and the changes in source areas through time that result from watershed development. Ultimately, these data may be used to focus limited financial resources on controlling sediment migration from the most important contributors of debris to the region’s streams, rivers, and reservoirs.

Nature, Scope, and Objectives of Research

It is not uncommon for the sediment trapping efficiency of reservoirs to exceed 90 %. Most of the captured sediments are deposited on the reservoir bed where they record temporal variations in sediment loads derived from upstream areas that may result from natural disturbances (e.g., the erosional impacts of extreme hydrologic events), or from anthropogenic disturbances (e.g., those associated with watershed development and alterations in land-use practices). It follows, then, that the analysis of these deposits, when linked to a understanding of sediment provenance, may provide valuable insights into the historical changes in the nature and magnitude of erosional and depositional processes operating within the region. The primary objectives of this investigation are to utilize reservoir and sediment provenance data to:

(1) document the natural rates of sedimentation in “undisturbed” watersheds by determining the thickness and age of the bed materials within reservoirs fed by basins where land-disturbing activities are limited, and by analyzing reservoir materials deposited prior to significant development in the basins where human impacts have occurred;

(2) quantify the relative contributions of sediment, at any given time, from specific land-cover types (e.g., pastures, housing developments, roads, etc.), and obtain insights into the most important contributors of debris to the region’s streams, rivers, and reservoirs; and

(3) determine the human impacts on basin sediment yields and provenance during the past several decades by comparing data obtained for disturbed and undisturbed watersheds, and by relating the temporal variations in sedimentation rates and provenance to an understanding of watershed development.

Inherent in this investigation is an assessment of whether physical and geochemical fingerprinting, when used in combination with sediment mixing models, can be effectively applied in western North Carolina to determine the relative contributions of debris from distinct land-use categories. The study will also represent one of the first attempts to quantitatively investigate the impacts of land-use alterations on reservoir sedimentation in the region, and the first investigation aimed at quantifying the relative contribution of sediment from the different source areas in the area.

Three reservoirs were chosen for this study based upon the control they afford in assessing the sedimentation record of western North Carolina. The reservoirs include Upper Sapphire Lake, Fairfield Lake, and Lower Sapphire Lake, all of which are located in the Blue Ridge Province of the southern Appalachians in Jackson County, North Carolina (Fig. 1). These reservoirs are positioned in the northern headwaters of the Savannah River system and were constructed in the late 1800s or early 1900s. Thus, they may contain sediment records of up to 109 years in length (Table 1). The local base level is about 940 m with mountain peaks reaching just over 1400 m. The mean annual temperature in Jackson County is about 13°C, but is considerably cooler at high elevations. The mean annual precipitation is also highly variable, but the region in the vicinity of the reservoirs (south of Cashiers) receives about 2500 mm/yr (Fig. 1).

The geology of the region, like much of the Blue Ridge Province of western North Carolina, is dominated by varieties of high-grade metamorphic gneisses, locally ranging from muscovite-rich granodioritic gneiss to hornblende gneiss (McKniff, 1967). The rock formations strike approximately southwest-northeast and two-four major rock assemblages are present in each of the drainage basins (Table 1). Soils in the region predominantly consist of Inceptisols on steep slopes and Ultisols on gentle slopes.

The study area is located in the high growth, affluent area of Cashiers. Large, summer-season homes, private communities, and golf courses blanket much of the region. Upper Sapphire lake drains a densely forested, nearly undisturbed region (Fig. 2A). The Fairfield Lake watershed is mottled with new homes and a few resorts, although most of the land remains forested (Fig. 2B). The Lower Sapphire Lake watershed is intensely developed with a variety of land uses, including several golf courses (Fig. 2C). Most of the reservoirs in this region were developed about 100 years ago and served as a centerpiece for secluded resorts. During the past century, the region has experienced a large increase in population and shifted away from quiet resorts to an affluent second home destination. The fact that distinct land-use changes have occurred in the area, and that the type and magnitude of land-use activities differ between the three reservoir/catchment systems to be examined, suggests that this area should provide an excellent setting to accomplish the above objectives.

Horizontal Line

Table 1

Major characteristics of reservoirs chosen for study.


Major Stream

Nature of Land-Cover

Primary Land Use or Land Cover


Drainage Area

sq. km

Basin Relief


Basin Slope m/km

Lake Area

sq. km

Date of Dam Closure

Upper Sapphire Lake

Nix Creek



Sgn, mgs, hgn






Fairfield Lake

Trays Island Creek

single land use

home sites

Cgn, hgn






Lower Sapphire Lake

Horsepasture River

mixed land use

town, golf courses, home sites

Sgn, Cgn, hgn, mgs






Sgn: Sapphire Gneiss, Cgn: Cashiers Gneiss, hgn: hornblende gneiss and schist, mgs: micaceous gneiss and schist (from McKniff, J. M., 1967)

Figure 1: Locations of study reservoirs and their drainage basins. From Fontana Lake, 1:100,000 series topographic map, 1983, U.S. Geological Survey.

Figure 1: Locations of study reservoirs and their drainage basins. From Fontana Lake, 1:100,000 series topographic map, 1983, U.S. Geological Survey.

Figure 2: Vertical air photos showing relative land use patterns in each study basin.  A) Upper Sapphire Lake Area: forest, B) Fairfield Lake Area: home sites, C) Lower Sapphire Lake: golf courses and home sites.  Photos taken on 25 March 1995 by the National Aerial Photography Program.  North at top; photo width is about 800m.


Methods, Procedures, and Facilities

The objectives put forth in Section 12 will be accomplished by systematically completing a series of research tasks, and integrating the results of each. A summation of the specific tasks, in the order in which they will be addressed, are presented below.

(1) The collection, sampling, and dating of bed sediments from the Upper Sapphire, Fairfield, and Lower Sapphire Reservoirs. The sediments will be obtained using a Livingston coring device and dated using 210Pb and 137Cs. This will allow for a determination of sediment age as a function of depth below the reservoir bed, and will lead to an analysis of the changes in sediment loading rates through time (including those that may be related to changes in land-use);

(2) Characterization of the physical, mineralogical, and geochemical nature of the sediments found within the reservoir beds, and the vertical variations in these properties that result from changes in sediment provenance;

(3) An analysis of the physical, mineralogical, and geochemical characteristics of sediments found within delineated source areas (land-cover categories), and a determination of the parameters that most effectively distinguish the sediments associated with each land-cover category using a multivariate statistical approach (e.g., cluster analysis and/or discriminant analysis);

(4) The development of sediment mixing models that allow reservoir sediments deposited during discrete time periods to be related to the delineated sediment source areas;

(5) Validation of mixing model estimates by (a) comparing the actual sediment properties with those predicted by the models to assess “goodness-of-fit”, (b) applying the models to experimental mixtures of source area sediments created in the laboratory, and (c) comparing variations in source area contributions, as predicted using the mixing models, to an understanding of watershed development; and

(6) An examination of the variations in basin sediment yields and the relative contributions of debris from each delineated land-cover category to the reservoirs through time. Emphasis will be placed on identifying the most important sources of sediments to the region’s rivers and lakes.

The procedures that will be used to complete these tasks are discussed in more detail below.

Extraction and Collection of Reservoir Sediments

The use of lake and reservoir sediments in the reconstruction of watershed scale erosional processes over decadal time scales is well founded, and numerous studies have analyzed limnic sediments with the explicit goal of deciphering the rates and variations in basin sediment yields (Dearing and Foster, 1993). During this investigation, sediment yields from three catchments will be documented by collecting, dating, and analyzing cores extracted from the Upper Sapphire, Fairfield, and Lower Sapphire Reservoirs. Reservoir siltation is a significant problem in western North Carolina. For example, in 1993 the Southeast United Methodist Assembly paid over $500,000 to dredge portions of Lake Junaluska (Gibson, 1998; also see Yurkovich, 1984 for an outline of the sediment history of the Lake Junaluska basin). Thus, it is likely that the reservoirs to be examined in this study, which were developed in the late 1800s and early 1900s, will contain an excellent sedimentologic record.

We will collect a minimum of three cores from each reservoir using a Livingston Piston Coring device. The coring sites will be located in areas where the preservation of long-term sediment records is most likely to occur. The Livingston Coring device allows the sediments to be extruded into 2” ID plastic tubes that can be cut and capped to limit disaggregation of the sediments during transport. The see-through nature of the plastic tubes will allow a preliminary examination of the sediments in the field, and care will be taken to ensure that the cores extend through the entire thickness of the reservoir (lacustrine) deposits. The sedimentology and stratigraphy of the cores will be described in detail in the Sedimentology/Soils Laboratory at Western Carolina University. Core descriptions will subsequently be compared to determine the spatial variations in deposition within a given reservoir, particularly with regards to unit thickness. The most stratigraphically complete core from each reservoir will then be sampled for isotopic dating using 137Cs and 210Pb. The initial sampling interval will be based on the total thickness of the deposits, relative to the age of the reservoir; attempts will be made for each slice (sampling interval) to correspond to 1-5 year increments, although the sampling increment will not exceed 1-5 cm (in practice, 1 cm is the thinnest slice that can be obtained unless the cores are frozen).

The use of 137Cs and 210Pb to estimate the rates of sediment accumulation in lakes and reservoirs has become a well-established procedure (Wise, 1980; Davis et al., 1984; McCull et al., 1984; Melieres et al., , 1988; Pouchet et al., 1989). 210Pb is produced from the decay of 222Rn, the latter of which is emitted from soils and sediment to the atmosphere. Most 210Pb is subsequently returned to the Earth’s surface within a few days where it may become incorporated and strongly attached to sediments in lakes and reservoirs (Dickin, 1997). The atmospheric addition of 210Pb to any given sediment package stops once they are buried. As a result, the rate of sediment accumulation can be determined by documenting the 210Pb activity in surface materials, determining the decrease in activity with depth, and removing the 210Pb in sediments that is produced by the decay of 226Ra, a non-atmospheric source referred to as supported 210Pb. The supported 210Pb activity is typically assumed to be that value at which the total activity becomes a constant with depth and, thus, can easily be determined for most sites. With a half-life of 22.26 years, 210Pb is ideally suited for the dating of sediments on a decadal scale over time-frames of 100-150 years.

137Cs is an artificial radionucelide that was produced by the atmospheric testing of nuclear weapons in the 1950s and 1960s. Like 210Pb, the 137Cs that is deposited from the atmosphere is rapidly and strongly bound to fine-grained particles. Thus, the 137Cs content of reservoir and lake sediments is related to the rate of fallout at the time of their deposition. The first appearance of 137Cs can be dated to the early 1950s, while the peak fallout levels occurred in 1963 (Walling and He, 1993). Determination of the depth at which the peak 137Cs content exists in reservoir materials, which is assumed to correspond to 1963, can be used to obtain rates of sediment accumulation over the past 30-40 years (Ritchie et al., 1973; Robbins and Edgington, 1975; Krishnaswami and Lal, 1978; Walling and He, 1993).

The rates of sediment accumulation can be transformed into basin sediment yields by (1) using the dated cores to estimate the volume of sediment deposited within the reservoir during a specific time-interval (usually 1-year), (2) converting the volume estimates to units of weight, and (3) by dividing the estimated weight of sediment deposited in the reservoir per year by the upstream basin area. The conversion between units of volume and weight requires an understanding of the bulk densities of the sediments, and the changes in bulk densities that may occur as a result of compaction.

Accurate dating of the sediments within the cores will allow for a determination of the average sedimentation rates within the three reservoirs. Depending on the dating resolution of the obtained cores, it may also be possible to assess variations in the rates of accumulation, and basin sediment yields, since dam construction. These changes can subsequently be related to an understanding of land-use alterations within the catchments (determined from the mapping of land-cover types on sequential aerial photographs and from the analysis of historical records), and will provide important insights regarding the impact of development on stream and reservoir siltation. Insights into the magnitude of sedimentation resulting from land-disturbance will also be obtained by comparing the rates of sediment accumulation documented for the undeveloped Upper Sapphire reservoir system with the highly developed Fairfield and Lower Sapphire catchments.

Once the age of the cores have been determined, they will be sampled for geochemical, mineralogical, and physical analysis. The sampling interval will be adjusted so that the data are obtained from sediments deposited over a 2 to 5 year period. This sampling resolution has been chosen because data obtained over time periods shorter than 1-2 years may reflect materials derived from only a portion of the basin during an extreme hydrologic event (i.e., a flood), rather than the dominant sources of sediment to the reservoirs measured over longer-time intervals characterized by the more typical runoff regimes (Passmore and Macklin, 1994). The upper limit reflects an attempt to maximize our understanding of the alterations in basin source areas through time. The results of the geochemical, mineralogical, and physical analyses will be used in conjunction with sediment mixing models to determine the primary sources of sediment to the reservoir for each of the sampled increments of known age (as described in more detail below).

Identification, Characterization, and Differentiation of Contemporary Sediment Source Areas

Identification and Mapping of Land-Cover Categories Through Time: Potential sediment source areas will be differentiated and mapped from aerial photography using the North Carolina system for the mapping of land use and land cover (NCCGIA, 1994). This procedure is patterned after the U.S. Geological Survey land use and land cover classification system (Anderson,, 1976), but has developed four levels of classification (in comparison to three used by the U.S. Geological Survey). A level III classification is recommended by Anderson, et al. (1976) for use with remote sensing data ranging in scale from 1:20,000 to 1:80,000 (1976). The North Carolina system allows mapping of both land use and land cover to distinguish between regions that have been significantly impacted by human activities and those characterized by limited disturbances and natural vegetative and aquatic areas.

Mapping of Land use Changes through Time: Land use and land cover changes during the past several decades will be documented by applying the above classification system to aerial photography obtained at different times, a process that will allow the magnitude of watershed development to be determined during at least 2 distinct time-periods (before the oldest photo set, and between the two photo sets). Relatively current photography (1995) of the area is available from the U.S. Geological Survey. Aerial photography of earlier date has been produced by the USDA, Forest Service in conjunction with forest inventory projects (1976 and earlier); by the USDA, Natural Resources Conservation Service in conjunction with cooperative soils mapping and district conservation work planning (1980 and earlier); and by Jackson County, NC in conjunction with tax mapping (1960). These sets of aerial photographs will be reviewed to select the oldest and most recent images of useable scale, thereby creating the longest time interval between photographic sets possible, with the inclusion of one or more sets intermediate in time. The date of the intermediate photos will be selected to correspond with significant changes in sedimentation rates determined from the sediment cores, if possible.

The land use and land cover classification system will be applied to each set of aerial photography and summaries of the total land areas for each of the various classes will be developed for the associated time period. This will allow a comparison of changes in the number and types of classes and their respective areas through time. In addition to analyzing photographic sets, town governments and historical societies will be contacted to determine other significant changes in land use. Each of the land use and land cover classes will be related to the underlying bedrock composition as discussed below.

13.2c Delineation and Mapping of Basin Bedrock Geology: Geologic mapping, at a scale of one inch to 2000 feet, of the Cashiers - Highlands area was completed by McKniff in 1967. Later work by Hadley and Nelson (1971) and by Brown et al. (1985) further elucidated the local geology and has resulted in a basic understanding of the bedrock geology of the study area. The Nix Creek watershed that supplies the Upper Sapphire Lake (reservoir) is underlain by biotite gneisses and amphibolites of the Tullulah Falls Formation which are interbedded with meta-diorites. The remainder of the Lower Sapphire watershed, including the rocks underlying the Horsepasture River, is predominantly composed of Devonian aged quartz diorites or granodiorites (locally called Whiteside Granite) which are themselves intrusive into the Tullulah Falls rocks. Unfortunately, McKniff (1967), Hadley and Nelson (1971) and Brown et al. (1985) use different rock descriptions and nomenclature which will require some reconnaissance mapping to sort through this information.

An initial step in the project will be to compile the existing geologic bedrock maps into a single usable base map using a scale of 1:24000 or larger. This map (covering the project's watersheds) will include primarily bedrock geology and all known mineral deposit locations. Distribution of alluvium and colluvium may also be incorporated if it can be extracted from the Soil Survey of Jackson County, NC (Sherrill, 1997) or other sources.

Multivariate Differentiation of Sediment Source Areas: The geochemical fingerprinting approach requires a chemical, mineralogical, and physical characterization of the surface sediments found within each of the delineated land-cover categories. Because the composition of the local bedrock geology is likely to influence the nature of these surface materials, it is not only necessary to stratify the collected data by land-cover type, but by the lithology of the underlying rock units. Thus, the land use and land cover data, obtained from 1995 or more recent aerial photographs, will be overlain on the geologic base maps outlined in Section 13.2c. These composite maps will be used to guide the sampling process.

In addition to stratifying the data according to geological and land-cover constraints, it is necessary to minimize sampling biases to the extent possible, given site access and another sampling difficulties. In this case, sampling sites will be located on the basis of random sampling techniques (e.g., those presented by Cochran, 1963) in which a numbered grid will be overlain on the maps produced in Section 13.2c. Some biases may be included in the procedure to ensure that a minimum of 5 samples is obtained from each land-cover type, underlain by each of the major delineated geologic units (i.e., the Tullulah Falls Formation and Whiteside Granite). The actual number of samples collected from each type of land-cover/bedrock source will be adjusted (weighted) on the basis of the area that the units comprise within the watershed.

Collins et al. (1997b) argued that sediment provenance studies should concentrate on surface materials. We will follow their suggestion by collecting materials within 5 cm of the surface. Composite sediment samples will also be collected from the outer 5-10 cm of floodplain deposits exposed in the channel banks at a minimum of 10 locations along each of the axial stream systems. It is expected that a total of 150-200 samples will be collected and analyzed from the upland areas of the three catchments.

All samples will be labeled in the field, double-bagged, and transported to Sediment/Soils Laboratory for analysis. Recent investigations by Passmore and Macklin (1994) and Collins et al. (1997a, 1997b, 1998) have argued that the differentiation of source area sediments is most successful when a wide range of fingerprinting properties are analyzed. During this investigation, we will examine numerous elements that exhibit different geochemical behaviors including trace metals (e.g., Fe, Mn, Al), heavy metals (e.g., Cu, Zn, Pb, Cr, Ni, Co), and base metals (e.g., Na, Mg, Ca, K). In addition, the materials will be characterized for grain mineralogy and grain-size. The analytical procedures used to carry out these analyses are outlined in more detail below.

There have been a wide variety of multivariate techniques used to differentiate sediments from different land-cover categories, including cluster analysis, principal component analysis, and discriminant analysis. Regardless of the exact approach that is invoked, these analyses generally begin by determining the parameters that are most effective in differentiating between delineated source area types. Collins et al. (1997b) started with a relatively simple, bivariate analysis, in which each parameter was examined independently to determine if it could statistically distinguish between disturbed source areas and channel bank materials. The parameters that proved successful were utilized in subsequent, multivariate analyses. We will use a similar approach here.

It addition to determining the best parameters to use, it may be necessary to combine various land-cover types that possess highly similar physical and geochemical signatures. In the southern Appalachians, this may involve the union of road cuts and housing excavations (as opposed to pastures, clear-cut terrains, golf courses, or forested lands). We will begin our multivariate treatment by using cluster analysis to determine major groupings of land-cover categories. Subsequently, the significance of these groupings will be analyzed using multivariate ANOVA and/or discriminate analysis. The end product of these statistical treatments will be a list of land-cover types (sediment source areas) that can be separated from each other nearly 100 percent of the time.

Nature and Development of Sediment Mixing Models

The complex processes involved in the erosion, transport, and deposition of sediment ultimately result in a deposit that represents a mixture of material derived from multiple source areas within the watershed. If the physical and geochemical properties of the source area sediments are conserved during the transportational and depositional process, then it should theoretically be possible to determine the relative contributions that each source contributed to the resulting mixture. Total conservation of parameter values is rarely achieved in nature. Nonetheless, some properties are generally conserved. This prompted Yu and Oldfield (1989) to develop an empirically based, sediment mixing model that was capable of determining the relative contributions of material from different source areas to the Rhode River estuary on the western shore of Chesapeake Bay. The model proposed by Yu and Oldfield (1989) has since been modified and improved to determine the provenance of suspended sediments (Collins et al., 1997b, 1998), floodplain deposits (Collins et al., 1997a), and reservoir bed materials (Yu and Oldfield, 1993).

We will modify the basic sediment mixing model first proposed by Yu and Oldfield (1989) and later modified by Collins et al. (1997b). Constraints on the mixing model require that (1) each source type contributes some sediment to the mixture, and thus the proportions derived from individual source areas (Ps) must be non-negative (0 < Ps < 1), and (2) the contributions from all of the source areas must equal unity, i.e.:


       å Ps = 1.


In addition, some differences (error) between the values of the measured parameters in the source area and the mixture must be allowed. For any individual parameter, the error can be determined as follows:


              ej = bj - å xi aij


where bj (j = 1, 2, 3…..n) are the measurements on “n” independent parameters within the sediment mixture, aij (i = 1, 2, 3….m) is the measurement on the corresponding parameter within the ith source area, xij is the proportion of the ith source component in the sediment mixture, and ej is the residual error. When the number of measured parameters equals or is greater than the number of sources areas (n>m), as in the case of this study, the system of equations is over-determined, and a “solution” must be obtained using an iterative computational method that minimizes an objective function, thereby obtaining a best fit solution to the entire data set (Yu and Oldfield, 1989). There are a number of objective functions that may be utilized, but in previous studies, it has taken the form of the sum of the relative errors (Yu and Oldfield, 1989; Collins et al., 1997b), where:


           f(xi) = å ej/bj.


In essence, these sediment mixing models require that the parameters that most effectively separate the land-cover types are measured in both the source area sediments and the mixture (deposits). Subsequently, the relative contributions from each sediment source are adjusted using an iterative computer technique until the differences (error) between the measured parameter values in the source materials and the deposits are minimized for all of the utilized properties. In this study, one or more sediment mixing model(s) will be developed that relates sediments of known age contained within our study reservoirs to the defined land-use/geology categories.

Recent studies by Collins and his colleagues have argued that various correction factors need to be incorporated into the mixing models to account for difference in parameter values that are likely to result during erosion, transport and deposition. Of most importance is the inclusion of a particle size correction factor. Particle size has a significant influence on sediment geochemistry and mineralogy (Horowitz and Elrick, 1987; Salomons and Forstner, 1984; Ackermann, 1980; Forstner, 1982; Whitney, 1975). Thus, small differences in particle size distribution that occur by hydraulic sorting or selective deposition, may result in large differences in parameter concentrations. As a result, many geochemical and mineralogical properties cannot be directly compared unless a correction factor is utilized in the sediment mixing models (Collins et al., 1997b), or the analyses are carried out using only the fine-grained (<63 mm) sediment fraction (Yu and Oldfield, 1989). For the proposed study, we will examine the relationships between geochemical/mineralogical properties and the grain-size and organic matter content of the deposits using correlation analysis to determine if correction factors are warranted.

Model Verification

As noted by Yu and Oldfield (1993), objective independent tests to assess the performance of sediment mixing models are currently not available. Nevertheless, the utilization of the models requires some form of verification to insure that the results are credible. For this study, we will use three approaches. First, we will examine how well the model fits the data (i.e., the goodness-of-fit) by comparing the measured parameter values in the mixture with the predicted values. The assessment of these relative errors should provide insights as to whether the mixing model results in an acceptable prediction of the fingerprinting properties. Second, we will create mixtures of sediments in the laboratory using different proportions of material from the various land-cover categories. The model(s) will be applied to these samples to assess their ability to correctly determine the relative amounts of material from each source area contained in the mixture. This procedure will provide an understanding of the conditions under which spurious matches can occur, if any. Finally, we will document temporal changes in the relative contributions of sediment from the identified source areas by applying the model to materials collected at different depths below the reservoir’s bed (as described in more detail below). These data will be used to assess whether the temporal changes in the relative contributions from the various land-cover categories, as predicted by the sediment mixing model, are consistent with changes in land-use practices as determined from sequential aerial photographic mapping. In essence, this process will assess whether the model results are intuitively correct (there is some danger of circular reasoning in this latter approach, as these relations will also be used to assess the impacts of land-disturbing activities on sediment source areas; nonetheless, this qualitative method will provide useful insights into the application of sediment mixing models in this area).

Interpreting Temporal Changes in Sediment Yields and Sources

The application of the sediment mixing models to the core samples will allow a chronology of the changes in the relative contributions of sediment to the reservoirs from the different source areas to be determined. Thus, we will be able to assess temporal changes in both basin sediment yields (by knowing deposit thicknesses and age) and sediment provenance (via the results of the mixing models). In combination, the data will provide a wealth of information on the natural variations in sediment provenance and loads in the area, and the impacts of human activities on basin sediment yields and sources.

Natural variations in sediment loads and sources will primarily be assessed by examining the modeling results from the Upper Sapphire basin which has been subjected to only limited amounts of land-disturbing activities. Data concerning natural changes in source area contributions and yields may also be obtained from Fairfield and Lower Sapphire Reservoirs by examining deposits that pre-date significant development in the region. The impacts of human activities on sediment yields and sediment sources will primarily be obtained by examining the results from more recent deposits within the Fairfield and Lower Sapphire Reservoirs, and comparing the results to the natural sediment yields and sources in the area. Of primary concern will be the identification of the land-cover types that have contributed the most sediment to the reservoirs during the past several decades. This analysis will be aided by the fact that the Fairfield and Lower Sapphire catchments have had different development histories and exhibit different land-cover types. Thus, it should be possible for the impacts of differing land-use practices to be isolated and documented. If it is possible to quantify the nature and timing of sediment control practices in the area, a comparison of natural and human influenced sediment yields and provenance may also provide insights into the effectiveness of previously utilized sediment management programs within the Fairfield and Sapphire catchments.

Analytical Procedures

Geochemical Analysis: The reservoir and source area sediments will be analyzed for a variety of constituents, including trace, heavy, and base metals. These analyses will be carried out within the geochemical laboratories at Western Carolina University. Fe, Mn, and Al extraction will involve both pyrophosphate-dithionite (Bascomb, 1968) and oxalate (Deb, 1950) methods, whereas heavy metals will be digested in hot aqua regia. Base cation extraction will be conducted using acid ammonium acetate (Qui and Zhu, 1993). Concentrations in the extracts will primarily be determined using an Inductively Coupled Plasma Atomic Emission Spectrometer (ICP-AES)(note that some modification of extraction procedures may be used in the study to obtain the best results using an ICP-AES). Quality of the data will be assured through analysis and monitoring of blanks, replicates, controls, and standard reference materials (e.g., USGS, NIST, and in-house and international SRMs).

Isotopic Analysis of 210Pb and 137Cs. The analysis of 210Pb and 137Cs is a relatively expensive procedure. During this investigation, we will reduce the costs of analysis by working closely with Dr. Paul Lechler at the Nevada Bureau of Mines and Geology. Dr. Lechler is currently in the process of developing a new method to analyze for these isotopes using an ICP-MS. To insure quality control, a limited number of sample splits may be analyzed at another laboratory, such as Teledyne/Brown Engineering, that has been accurately conducting these analyses for some time.

Mineralogical Analysis and Relations to Bedrock Geology: Many provenance studies have been done on fine fluvial sediments including analysis using mineralogy, color, magnetics, radionuclides, palynology, heavy minerals, and chemistry. This portion of the investigation will focus primarily on the mineralogy of sediment and underlying bedrock using procedures similar to those of Fan (1976) and Phillips (1992).

he mineralogical analysis of the samples is a relatively time-consuming process, and it will not be possible to analyze approximately 150-200 source areas samples and about 150 additional samples from the reservoirs given the time constraints placed on this investigation. Thus, the mineralogical analysis during this study will focus on the identification of mineralogical signatures that can be used qualitatively (i.e., outside of the multivariate statistical treatments) to delineate major source area contributions. While these data cannot be used in the more rigorous statistical treatments, the possibility that unique mineralogical signatures may allow for highly specific determination of sediment provenance makes this analysis an important component. Moreover, these data may provide valuable insights into the nature of the sedimentological and mineralogical changes that occur during the erosion and transport of the source area sediments to the reservoirs. Therefore, a subset of the samples collected from the sediment source areas and the reservoir deposits will be characterized for grain mineralogy. In addition, the underlying bedrock will be collected from those sites, or as close to them as rock exposure allows.

Analyses of the source area sediments should show the types of minerals present from each formation, and microscopic analysis will provide descriptive mineral characteristics of some minerals (e.g., fractured quartz) that may help establish provenance. At each site, a thin section of the underlying bedrock will be prepared and analyzed using standard petrographic techniques. A minimum of 1000-1500 counts will be run per thin section to produce the necessary reliability to accurately determine the volume percent of mineral constituents in the rock. The mineralogy of the bedrock can then be compared to the mineralogy of the sediment.

Mineralogical data, including the existence of heavy minerals, will also be obtained from sediments collected at selected point bars and channel beds along the course of the streams. If heavy minerals are observed in low concentrations they will be separated from the sediment by various techniques (e.g., heavy liquids, magnetic separation, etc.) to determine their type and concentration. If they are relatively abundant in the sediment, they will be characterized by optical techniques.

Physical Analysis: The grain-size distribution of each of the sediment samples will be determined in the Sedimentology/Soils Laboratory using wet-sieving and pipetting techniques following the procedures modified from Singer and Janitzky (1986). The organic matter content of the samples will be estimated by loss-on-ignition (550°C, 4hrs.).


Sedimentology/Soils Laboratory: The Sedimentology/Soils Laboratory has recently been renovated in the Department of Geosciences and Natural Resources Management (GNRM) at Western Carolina University. It consists of two separate facilities, one of which includes a core and sample storage room, work areas to describe, subsample, and photograph cores, and a facility for the preparation of samples for geochemical, mineralogical, and physical analysis. The second lab provides the equipment necessary to characterize the sedimentology and weathering characteristics of the collected materials including grain-size distribution, organic matter content, carbonate content, electrical conductivity, and pH. Binocular and petrographic microscopes are available in the petrology lab for analysis of grain mineralogy.

Computer Facilities. The Department of Geosciences and Natural Resources Management is well equipped with both PC and Macintosh desktop Computers. The statistical and numerical modeling studies to be conducted during this investigation will be carried out on a Gateway, Pentium III, 550 Mhz processor with 256 MB SDRAM. Other hardware devices include digitizing tablets, HP scanners, flatbed scanners, a Nikon slide scanner, large format plotters, Polaroid 7000 film recorder, and color laserjet printers. The statistical analysis will be performed using SYSTAT. Other available software includes Mircrosoft and Corel office suites, Adobe Photoshop, Micrografix Designer, ARC-VIEW, HEC-RAS, and SWAT.

Geochemical Facilities: The geochemistry labs at WCU are equipped with the basic tools necessary to conduct the geochemical analysis required for this project. Specific analytical equipment includes a Thermo Jarrell Ash AA with CTF-188 Graphite Furnace, a Thermo Jarrell Unicam 869 Atomic Absorption spectrometer with a Hg analyzer, and a Thermo Jarrell Ash Iris ICP.

Related Research

The use of sedimentologic records from lakes and reservoirs to calculate basin sediment yields has received considerable attention since the late 1960s. Early studies primarily relied on repeat surveys of bottom topography to calculate volumes of sediment that were deposited in the reservoir through time, and generally disregarded processes such as compaction in the calculations of basin sediment yields. More recent studies, however, have relied heavily on the sampling, dating, and analysis of one or more cores extracted from the beds of lakes and reservoirs (Dearing and Foster, 1993). When linked to an understanding of the changes in land-use practices within the watershed, these studies have been able to describe major changes in sedimentation through time that result from human disturbances within the watershed. A major problem with these earlier investigations, however, is that they were unable to identify the major areas contributing sediment to the lakes and reservoirs; rather, they simple illustrated that sediment yields had increased, and the most likely source was disturbed lands.

To circumvent this problem, numerous approaches have been used to quantify the source(s) of suspended channel sediments including (1) the delineation and mapping of barren lands, or regions of obvious sheet, rill, gully, and bank erosion, and (2) the use of direct monitoring techniques including suspended sediment sampling, erosion pins, cross-channel surveys, and runoff troughs (Imeson, 1974; Peart and Walling, 1986). These methods are plagued by operational (sampling) difficulties and generally provide a data set that is both temporally and spatially limited (Peart and Walling, 1986). Some studies have relied on the use of soil loss equations to estimate the relative sediment yields from subbasins of a watershed. This approach raises questions as to the reliability of the utilized equations, and the uncertainties associated with calculating the sediment delivery ratios required to convert the estimates of upland erosion to one of downstream sediment yield (Peart and Walling, 1986).

As a result of the problems outlined above, many investigators have abandoned these traditional approaches and turned to physical and geochemical fingerprinting techniques to determine the provenance of both suspended sediments and historical deposits within floodplains and terraces at the basin mouth. Since the late 1970s, a variety of parameters have been utilized in these investigations including grain size (Fan, 1976; Knox, 1987; Sutherland, 1991), grain mineralogy (Fan, 1976), mineral magnetics (Oldfield et al., 1979; 1985; Walling et al., 1979), radionuclides, (Peart and Walling, 1986), and heavy metal pollutants (Knox, 1987, 1989; Lewin and Wolfenden, 1978; Macklin, 1985, 1988, 1996; Passmore and Macklin, 1994).

Perhaps one of the more significant conclusions reached by these earlier investigations is that erroneous sediment-source area associations may occur when only a single fingerprinting parameter is used. As a result, more recent investigations have relied on multiple parameters and the utilization of multivariate statistical techniques to manipulate the data (e.g., Passmore and Macklin, 1994). The application of more rigorous statistical methods has been paralleled by the development of various sediment mixing models which allow a more accurate assessment of the relative contribution of material derived from each of the delineated source areas (Collins et al., 1997b, 1998; Peart and Walling, 1986). In light of the above, it is now generally accepted that fingerprinting techniques represent a reliable means of assessing sediment-source area relations and the transport and storage of sedimentary particles over historic time frames. Moreover, the application of these methods in several geological settings has shown that the technique can be extremely useful in determining the impact of land-use changes on hillslope sediment delivery processes on a basin-wide scale (Collins et al., 1997b).

The work proposed here represents (1) one of the first detailed investigations of reservoir sedimentation and the changes in sedimentation relative to land-use impacts in the mountainous terrains of western North Carolina, (2) the first attempt to quantitatively determine the relative contributions of sediment from different source areas in the region, and (3) the first use of geochemical fingerprinting in the area. In terms of the “larger” implications of the research, it will improve on the nature of the sediment mixing models, particularly with regards to the nature of the correction factors and the means of differentiating sediment source area.

Progress Review: Not Applicable

Schedule of Tasks

The investigation put forth here will be conducted between July 1, 2000 and June 30, 2001. The starting and completion dates of each of the research tasks are shown on Figure 3.

Figure 3.

Schedule of the Initiation and Completion of Research Tasks

















1. Collection of sediment cores


2. Compilation of bedrock geology maps

3. Mapping of contemporary land-cover types

4. Integration of land-cover and bedrock geology maps

5. Description and sampling of cores

6. Collection of source area samples

7. Geochemical analysis of samples

8. Mineralogical analysis of samples

9. Mapping of temporal changes in land-cover types

10. Multivariate analysis of source area sediments

11. Development and verification of mixing models

12. Determination of temporal variations in basin sediment yields and provenance

13. Assessment of the links between land-cover variations and changes in basin sediment yields and provenance through time


See Attachments A and B at the back of the proposal

Investigator Responsibilities and Qualifications:

This study will be conducted by an interdisciplinary team consisting of a hydrologist (Lord), a geomophologist (Miller), a mineralogist/petrologist (Yurkovich), and a natural resource manager (Kolenbrander) from the Department of Geosciences and Natural Resources Management, at Western Carolina University. Drs. Lord and Miller will oversee the collection, description, and sampling of sediment cores from the Upper Sapphire, Fairfield, and Lower Sapphire Reservoirs. Dr. Miller will oversee the geochemical and physical analysis of the collected samples, and the development of the sediment mixing models. Dr. Yurkovich will be responsible for quantifying grain mineralogy of the deposits, and for compiling data on the underlying bedrock geology of each of the study’s catchment areas. Watershed development, including the delineation and mapping of land-cover types, will be coordinated by Dr. Kolenbrander. Dr. Miller will be responsible for the overall coordination of the project. For investigator qualifications, see attached Vitae.

Student Training Potential

The use of undergraduate and graduate students in research projects is a particularly important means of providing them with a working knowledge of laboratory and field procedures, the use of “high-tech” equipment, the process of hypotheses development and testing, the methods utilized in data manipulation, and the requirements for effectively presenting the study results. In this investigation, three students will actively participate in various aspects of the work. A master’s student in the Department of Chemistry will be trained to prepare and conduct the required geochemical analysis. This will include training of sound laboratory procedures related to sediment digestion, and the use of a recently purchased ICP-AES. It is expected that the student will work approximately 20 hours per week on the project, and will conduct a thesis that is closely related to the investigation. In addition, two undergraduate students will be hired as research assistants; one student will work closely with Dr. Kolenbrander to delineate and map land-cover types on aerial photographs of the area. The second student will be involved in the mineralogical analysis of the collected samples. It is hoped that all three students will be able to participate in the collection, description, and sampling of the sediment cores. Moreover, each of the students will be expected to participate in the manipulation, interpretation, and presentation of the collected data.

Information Transfer

Due the rather large amount of the work that has been outlined above, the tasks will not include the publication of the study results; rather, they will focus on the collection and manipulation of the required data. The publication of the study’s results will occur in subsequent years, and will be complimented by additional data that we expect to collect following June 30, 2001. At that time, we expect to present the materials at several professional meetings and in refereed journals. If the fingerprinting/sediment mixing model approach proves successful, we will consider the development of a workshop on their use, both within the State, and, perhaps, at the regional Geological Society of America meeting.

Submittal of Proposal to Other Agencies

We are planning on submitting a related proposal to the USDA –CSREES-NRICGP Watershed Processes/Water Resources Grants program. These proposals are due on November 15, 1999, and if successful, the project will begin in June, 2000. The USDA proposal will focus on different catchments and reservoirs which are dominated by differing land-use practices (primarily logging). Moreover, it will likely include additional components, related to the impacts of sediment on reservoir/stream biota, which are not included here.

Notice of Research Project

The associated form is attached.


Ackermann, F., 1980. A procedure for correcting for grain size effect in heavy metal analyses of estuarine and coastal sediments: Environ. Technol. Lett., 1: 518-527.

Anderson, J.R., Hardy, E.E., Roach, J.T., and Witmer, R.E., 1976. A Land Use and Land Cover Classification System for use with Remote Sensor Data. Geological Survey Professional Paper 964. Govt. Printing Office, Washington, DC 28pp.

Bascomb, C.L., 1968. Distribution of pyrophosphate-extractable iron and organic carbon in soils of various groups. Journal of Soil Science, v. 19, p. 251-268.

Brown, P.M. et. al., 1985, Geologic Map of North Carolina, NC Dept. of Environment, Health and Natural Resources, Geology Division.

Cochran, W.G., 1963. Sampling Techniques. 2nd edition, New York: John Wiley and Sons.

Collins, A.L., Walling, D.E., and Leeks, G.J.L., 1998. Use of composite fingerprints to determine the provenance of the contemporary suspended sediment load transported by rivers. Earth Surface Processes and Landforms, v. 23, p. 31-52.

Collins, A.L., Walling, D.E., and Leeks, G.J.L., 1997a. Use of the geochemical record preserved in floodplain deposits to reconstruct recent changes in river basin sediment sources. Geomorphology, v. 19, p. 151-167.

Collins, A.L., Walling, D.E., and Leeks, G.J.L., 1997b. Source type ascription for fluvial suspended sediment based on a quantitative composite fingerprinting technique. Catena, v. 29, p. 1-27.

Davis, R.B., Hess, C.T., Norton, S.A., Hanson, D.W., Hoogland, K.D., and Anderson, D.S., 1984. 137Cs and 210Pb dating of sediments from soft-water lakes in New England (U.S.A.) and Scandinavia, a failure of 137Cs dating. Chem. Geol., 44: 151-185.

Dearing, J.A. and Foster, I.D.L., 1993. Lake sediments and geomorphological processes: some thoughts. In: J. McManus and R.W. Duck (Editors), Geomorphology and Sedimentology of Lakes and Reservoirs, pp. 5-14, New York, John Wiley and Sons.

Deb, B.D., 1950. The estimation of free iron oxides in soils and clays and their removal. Journal of Soil Science, v. 1, p. 212-220.

Dickin, A.P., 1997. Radiogenic Isotope Geology, Cambridge, Cambridge University Press, 490p.

Fan, P-F., 1976, Recent silts in the Santa Clara river drainage basin, southern California: a mineralogical investigation of their origin and evolution, J. Sediment. Petrol., v46, 802-812.

Forstner, U., 1982. Cumulative phases for heavy metals in limnic systems: Hydrobiologia, v. 91, p. 299-313.

Gibson, P., 1998. Mountain Streams: our State’s watershed. Sediments, v. 5, no. 4, p. 1-3.

Hadley, J. B. and A. E. Nelson, 1971, Geologic Map of the Knoxville Quadrangle, North Carolina, Tennessee, and South Carolina, U. S. Geol. Survey, Misc. Geol. Inv. Map I-654.

Horowitz, A.J. and Elrick, K.A., 1988. Interpretation of bed sediment trace metal data: methods of dealing with the grain size effect. In: J.J. Lichtenberg, J.A. Winter, C.I. Weber, and L. Fradkin, (Editors), Chemical and Biological Characterization of Sludges, Sediments, Dredge Spoils, and Drilling Muds. American Society for Testing and Materials, pp. 114-128.

Imeson, A.C., 1974. The origin of sediment in a moorland catchment with special reference to the role of vegetation. In: Fluvial Processes in Instrumented Watersheds, Inst. Brit. Geogr. Spec. Pub. No. 6, p. 69-72.

Knox, J.C., 1987. Historical valley floor sedimentation in the upper Mississippi valley. Ann. Assoc. Am. Geogr, 77: 224-244.

Knox, J.C., 1989. Rates of floodplain overbank vertical accretion. In : J. Hagedor (Editor), Floodplain Evolution. Abstr. Int. Geomorphology Floodplain Symp., Gottingen, September 9-12.

Krishnaswami, S. and Lal, D., 1978. Radionuclide limnochronology. In: Lakes: Chemistry, Geology, Physics, A. Lerman, (Ed.), Springer Verlag, p. 153-177.

Lewin, J. and Wolfenden, P.J., 1978. The assessment of sediment source: a field experiment. Earth Surf. Processes, 3: 171-178.

Macklin, M.G., 1985. Floodplain sedimentation in the upper Axe valley, Mendip, England. Trans. Inst. Brit. Geogr., N.S., 10: 235-244.

Macklin, M.G., 1988. A fluvial geomorphological based evaluation of contamination of the Tyne basin, north-east England by sediment-borne heavy metals, Unpublished report to the Natural Environmental Research Council, 29p.

Macklin, M.G., 1996. Fluxes and storage of sediment-associated heavy metals in floodplain systems: assessment and river basin management issues at a time of rapid environmental change. In: M.G. Anderson, D.E. Walling, and P. Bates, (Editors), Floodplain Processes, John Wiley and Sons, Chichester.

McCall, P.L., Robbins, J.A., and Matisoff, G., 1984. 137Cs and 210Pb transport and geochronologies in urbanized reservoirs with rapidly increasing sedimentation rates. Chem. Geol., 44: 33-65.

McKniff, J. M., 1967, Geology of the Highlands-Cashiers area, North Carolina, South Carolina, and Georgia, Ph.D. Thesis, Rice University, Houston, TX., 100 p.

Melieres, M.A., Pourchet, M., Bouchez, R., and Piboule, M., 1988. Chernobyl 134Cs, 137Cs, and 210Pb in high mountain lake sediment: measurements and modeling of mixing process. Jour. of Geophysical Res., 93: 7055-7061.

NCCGIA (N.C. Center for Geographic Information & Analysis), 1994. A Standard Classification System for the Mapping of Land Use and Land Cover. Raleigh, N.C. 59 p.

Oldfield, F., Maher, B.A., Donoghue, J., and Pierce, J., 1985. Particle-size related, mineral magnetic source sediment linkages in the Rhode River catchment, Maryland, USA. J. Geol. Soc. London, 142: 1035-1046.

Oldfield, R., Rummery, T.A. Thomas, R. and Walling, D., 1979. Identification of suspended sediment sources by means of magnetic measurements: some preliminary results. Water Resour. Res. 15: 211-218.

Passmore, D.G. and Macklin, M.G., 1994. Provenance of fine-grained alluvium and late Holocene land-use change in the Tyne basin, northern England. Geomorphology, 9: 127-142.

Peart, M.R. and Walling, D.E., 1986. Fingerprinting sediment source: the example of a drainage basin in Devon, U.K. In: R.F. Hadley, (Editor), Drainage Basin Sediment Delivery, IAHS-AISH Publication 159.

Qui, X.C. and Zhu, Y.Q., 1993. Rapid analysis of cation exchange properties in acidic soils. Journal of Soil Science, v. 155, p. 301-308.

Phillips, J. D., 1992, Delivery of upper basin sediment to the lower Neuse River, North Carolina, USA., Earth Surf. Process. Landforms: v. 17, 699-709.

Pouchet, M., Pinglot, J.F., and Melieres, A., 1989. Cesium 137 and lead 210 in alpine lake sediments: measurements and modeling of mixing processes. J. Geophysical Res., 94: 12761-12770.

Ritchie, J.C., McHenry, J.R.,, and Gill, A.C., 1973. Dating recent reservoir sediments. Limnology and Oceanography, v. 18, p. 254-263.

Robbins, J.A. and Edgington, 1975. Coupled lakes model for estimating the long-term reponse of the Great Lakes to Time-dependent loadings of particle-associated contaminants. U.S. National Oceanic and Atmospheric Administrations, 20pp.

Salomons, W. and Forstner, U., 1984. Metals in the Hydrocycle. Springer-Verlag, Berlin.

Sherrill, M. L., 1997, Soil Survey of Jackson County, North Carolina, USDA, National Conservation Service, 322p.

Singer, M.J. and Janitzky, P., 1986. Field and laboratory procedures used in a soil chronosequence study. U.S. Geological Survey Bulletin 1648, 49p.

Sutherland, R.A., 1991. Selective erosion and sediment source identification, Baringo District, Kenya. Z. Geomorphol., 35: 293-304.

Walling, D.E. and He, Q., 1993. Towards improved interpretation of 137Cs profiles in Lake Sediments. In: J. McManus and R.W. Duck (Editors), Geomorphology and Sedimentology of Lakes and Reservoirs, pp. 31-53, New York, John Wiley and Sons.

Walling, D.E., Peart, M.R., Oldfield, F., and Thompson, R., 1979. Suspended sediment sources identified by magnetic measurements. Nature, 281: 110-113.

Wise, S.M., 1980. Caesium-137 and lead-210: a review of the techniques and some applications in geomorphology. In: R.A. Cullingford, D.A. Davidson, and J. Lewin, (Editors), Timescale in Geomorphology, pp. 109-127.

Wolfenden, P.J. and Lewin, J., 1978. Distribution of metal pollutants in floodplain sediments. Catena, 4: 309-317.

Whitney, P.R., 1975. Relationship of manganese-iron oxides and associated heavy metals to grain size in stream sediments: Journal of Geochemical Exploration, v. 4, p. 251-263.

Yu, L. and Oldfield, F., 1989. A multivariate mixing model for identifying sediment source from magnetic measurements. Quaternary Research, v. 32, p. 168-181.

Yu, L. and Oldfield, F., 1993. Quantitative sediment source ascription using magnetic measurements in a reservoir-catchment system near Nijar, S.E. Spain. Earth Surface Processes and Landforms, v. 18, p. 441-454.

Yurkovich, S.P., 1984. Non-point sources of pollution – Lake Junaluska Watershed. Technical Report prepared for the Center for Improvement of Mountain Living, and the Lake Juneluska Assmebly, 36 p.

U.S. Department of the Interior, U.S. Geological Survey
Maintained by: John Schefter
Last Updated: Wednesday November 2, 2005 12:44 PM
Privacy Statement || Disclaimer
|| Accessibility