Some Uses of Channel Bed Sediment Concentration Data

• Determine the spatial distribution of trace metals– Identify point and non-point sources of pollution– Assessment of the rates and patterns of contaminant dispersal

• First approximation of potential ecological and human health effects (and regional water quality surveys)

• Monitoring of potential impacts of waste waters from industrial or municipal sites

• Geochemical exploration (surveys)

Downstream Trends in Channel Bed

• Where point sources are present the concentrations generally decline from the point of input.

From Salomons& Forstner, 1984

Concentrations of Cu and Ni in the <63 um fraction of channel bed sediments from the Po River, Italy. Samples were collected in the summer (grey bars) and winter (black bars). Acronyms along x-axis represent successive downstream sampling sites. Note minimal variations in concentration between seasons.

Viganò, L., and 14 others, 2003. Quality assessment of bed sediments of the Po River (Italy). Water Research, 37:501-518. (from 2 of 8 graphs from figure 3, page 507)

Downstream Trends in Channel Bed

• Where point sources are present the concentrations generally decline from the point of input.

From Salomons& Forstner, 1984

Characteristics of channel deposits (adapted from Knighton 1998; Church and Jones 1982; Hoey 1992)

Scale Characteristic

Transitory deposits Micro-forms Meso-forms

Bedload temporarily at rest

Coherent structures such as ripples with λ ranging from 10-2 to 100 mFeatures with λ from 100 to 102 m; includes dunes, pebble clusters and transverse ribs

Alluvial bars Macro-forms

Mega-forms

From by lag deposition of coarse-grained sedimentStructures with λ from 101 to 103 m such as riffles, point bars, alternate bars, and mid-channel barsStructures with λ > 103 m such as sedimentation zones

Braided Island Braided Anastomosing

Meandering Braided

Meandering

Meandering

Laterally Inactive Channels Laterally Active Channels

Straight-Island Form

Straight-Ridge Form

Sinuous-Stable

Braided

AnabranchingRivers

Meandering

Single-ChannelRivers

Single-ChannelRivers

AnabranchingRivers

(B)

(A) Bedload Mixed Load Suspended LoadStraight Straight Straight

Figure from Huggett, J.R., 2003. Fundamental of Geomorphology, Routledge Fundamentals in Physical Geography, Routledge, London, fig. 7.7, p. 185.

Ele

vatio

n

Distance Downstream

Riffle

Riffle

Riffle

Pool

Pool

Mean Water Surface Slope

Point Bars Pools

Riffle

Point BarPoint Bar

Point Bar

Point Bar

Point Bar

Pools

Riffle

(A)

(B)

(C)

Meandering Channel

Straight Channel

Longitudinal Profile

From Markham, A.J. and Thorne, C.R., 1992. Geomorphology of gravel-bed river bends. In: P. Billi, R.D. Hey, C.R. Thorne, and P. Tacconi, (eds.), Dynamics of Gravel-bed Rivers, pp. 433-456, New York, Jonh Wiley and Sons, Ldt., figure 22.2, p. 436.

Secondary Flow Directions

Point Bar

Super ElevatedWater Surface

From Thompson, A., 1986. Secondary Flows and the Pool-Riffle, Earth Surface Processes and Landforms, 11:631-641., Figure 4, p. 636.

Reading, H.G., 1978. Sedimentary Environments and Facies, Blackwell Publications,New York, Fig. 3.26, page 34. Company may have been purchased by Elsevier?

Figure 20. Laremie River, Wyoming (photo by J.R. Balsley); obtained from USGS Photo Library

Knighton, D., 1998. Fluvial Form and Processes: A New Perspective, Arnold, London.Fig. 5.23, p. 233.

Grain-Size & Compositional Variations

• Ladd et al. 1998– Examined trace metal concentrations in 7

morphological units in Soda Butte Creek, Montana

– (lateral scour pools, eddy drop zones, glides, low gradient riffles, high gradient riffles, attached bars, and detached bars)

– Highest concentrations in eddy drop zones and attached lateral bars with largest amount of fine sediment

•Slingerland and Smith (1986) define a placer as “a deposit of residual or detrital mineral grains in which a valuable mineral has been concentrated by a mechanical agent,”

•A contaminant placer is defined here as a concentration of metal enriched particles by the hydraulic action of the river. Where they occur, trace metal concentrations will be locally elevated in comparison to other areas (Miller & Orbock Miller, 2007)

Density-Dependent Variations

Guilbert, J.M. and Park, C.F., Jr., 1986. The Geology of Ore Deposits. New York, W.H. Freemand and Company, figures 16-1, p. 746 and 16-4b p.749.

Bateman, A.M., 1950. Economic mineral deposits, 2nd edition. New York, Wiley and Sons.

Reservoir

0

1

1B2

2B

3 4

5

67

7B

7C 7D9 10

11

1213 14

15

16 1718

Gagingstation

GagingStation

MineralCanyon

Dayton

CarsonCity

VirginiaCity

TableMtn.

Canyon

(Brunswick)

FortChurchill

395

0

0 1 2 3 4

1 2 3 4 5 Miles

Km

95

Six Mile

Canyon Fan

Gold Canyon

Six M

ile Canyon

Fork

Reno Fallon

CarsonCity

Carson River

Watershed Boundary

Carson Lake

Carson Playa Stillwater

WildlifeRefuge

Lahontan Reservoir

LakePyramid

SIERRA NEVADA

NevadaCalifornia

ForkE

ast

Truckee R

.

Car

son

R.

Tru

ckee

R

.Wes

tF

ork

C

arso

n R. .

Lake Tahoe

Lahontan

Eureka Mill, Brunswick Canyon

0

2

4

6

8

10

12

Me

rc

ury

Co

nc

en

tra

tio

ns

(u

g/g

)

0 20 40 60 80 100

Distance Downstream from 395 (km)

Pool Riffle Point BarChannel

Hg Concentrations In Channel Bed

LahontonReservoir

GoldCanyon

SixmileCanyon

MineralCanyon

Carson City

Variations Dependent on Time and Frequency of Inundation

• Examples– Queens Creek, Arizona– Rio Pilcomayo, Bolivia

Graf, W.L., Clark, S.L., Kammerer, M.T., Lehman, T., Randall, K., Tempe, R., and Schroeder, A., 1991. Geomorphology of heavy metals in the sediments of Queen Creek, Arizona, USA. Catena, 18:567-582, figures 2, p. 572 and 6, p. 578.

StudyArea

“Modern Mine”

Pb & Zn ConcentrateBall Mill

Floatation Process

Sampling Site RP-1~1.5 miles from Mills

Floatation Mill, Potosi

20 miles from Mills

You want me to live where?

Effluent

20 miles from Mills

~60 miles from Mills

Yikes!

High-WaterChannel Deposits

Low-Water Channel Deposits

Rio Pilcomayo, southern Bolvia near Uyuni. Photo taken in July during the dry season.

Implications to Sampling

• Local variations – referred to as small scale or field variance (Birch et al. 2001)– Can be on the order of 10 to 25 % relative standard

deviation and may be significantly greater than analytical variation (error)

– May hinder ability to decipher differences in contaminant levels between sample sites

• Reconnaissance level surveys and sample stratification by morphological units ?

• Sampling of specific units only?

– Composite sampling to minimize within unit variations

Changes in Sediment Composition Can:

• Influence the spatial and temporal concentration patterns observed in aquatic systems

• Hinder the determination of localized inputs of trace metals from either natural sources (e.g., ore bodies) or anthropogenic sources (e.g., mining operations or industrial complexes).

• Changes in grain-size have a particularly significant influence on metal concentrations.

Types of Mathematical Manipulations Commonly Applied to Bulk Metal Data

After Horowitz, 1991

• Corrections for Grain-size differences

• Normalization to a single grain-size range

• Carbonate content corrections

• Recalculation of concentration data on a carbonate-free basis

• Normalization to a conservative elemental

• Use of multiple Normalizations

Methods of Handling the Grain Size Effect

• Analysis of a specific grain-size fraction which is considered to be the chemical active phase– Does not provide for an understanding of the actual

concentrations that exist in the bulk sample– Inhibits the calculation of total trace metal transport rates

• Normalize the metal concentration data obtained for the bulk (< 2mm or sand) sized fraction using some form of mathematical equation and grain size data obtained from a separate sample– Provides actual concentration found in bulk sample– Poorly documents the concentrations that would actually be

measured in the finer-grain size fractions

Designation of Chemical Active Sediment Phase

• Numerous size fractions have been used as the chemical active phase including <2 µm, <16 µm, <20 µm, <63 µm, <70 µm, <155 µm, <200 µm (Horowitz, 1991)

• Argument for using < 63 µm fraction– It can be extracted from the bulk sample via sieving, a

process which does not alter trace metal chemistry

– It is the particle size most commonly carried in suspension by rivers and streams and may therefore be the most readily distributed through the aquatic environment

Grain Size Normalization

Normalized Concentration

= (DF *Bulk Metal Concentration)

Where,

DF = Dilution Factor

= 100/(100 - % of sediment > size range of Interest)

Concentration vs. Quantity of Fine Sediment

Sizes Frequently Used

•2 μm

•16 μm

•62.5 μm

•63 μm

•70 μm

•125 μm

•200 μm

Data from deGroot et al., 1982

Data from Horowitz and Elrik, 1988

Differences between Measured and Normalized Values

• Selected chemical active phase (grain size fraction) may not contain all of the trace metals

• Differences in concentration are not solely due to grain size variations

• Data contain analytical errors associated with grain size or geochemical analyses

Fractional Contributions of Selected Metals in Suspended Sediments

ConcentrationPercent Contribution

Constituent(mg/kg)

<63 μmfraction

>63 μmfraction

Total Sample

<63 μmfraction

>63 μmfraction

Arkansas River (sampled 5/11/87)a,b

Mn 1100 600 800 50 50

Cu 51 22 33 58 42

Zn 325 110 190 63 37

Pb 52 25 35 54 46

Cr 56 44 49 43 57

Ni 32 16 22 55 45

Co 15 11 12.5 45 55

Cowlitz River (sampled 4/20/87)a,c

Mn 650 670 660 40 60

Cu 63 33 46 57 43

Zn 62 68 59 42 58

Pb 12 10 10.8 45 55

Cr 35 19 25 56 44

Ni 25 16 19 53 47

Co 14 14 14 41 59

aThe represents the mean of the initial and final composite samples obtained at these sampling sites. b<63 μm fraction equaled 37 %, >63 μm fraction equaled 63 %, c <63 μm fraction equaled 41 %, >63 μm equaled 59 %.

(modified from Horowitz et al., 1990)

Carbonate Correction

• Assumes: Carbonate does not contain substantial quantities of trace metals and, thus, acts as a diluent. May not be true of Cd and Pb.

• Generally applied to streams in calcareous terrains, particularly those in areas with karst.

Where,

DF = Dilution Factor

= 100/(100 - % of carbonate in sample)

Normalized Concentration (DF *Bulk Metal Concentration)=

Conservative Element Corrections

• Assumes that some elements have had a uniform flux from crustal rocks. Thus, normalization to these elements provides a measure (or level) of dilution that has occurred.

• Elements most commonly used are Al, Ti, and to a lesser extent, Cs and Li.

• Normalized value = (Concentration of Trace Metal) (Concentration of conservative element)

Note: this generates a ratio, not a concentration as did the previous procedures