Toxicity of Metal-Contaminated Sediments to Benthic Invertebrates
Department of the InteriorU. S. Geological Survey
John Besser, Bill Brumbaugh, and Chris Ingersoll
Columbia Environmental Research Center, Columbia, Missouri
TESNAR Workshop: ‘Understanding the Impacts of Mining in the Western Lake Superior Region’
September 12-14, 2011 Bad River Convention Center, Odanah, Wisconsin
1
Mining activities produce metal-contaminated sediments
• Metals enter aquatic ecosystems from mining, ore processing, and smelting.
• At neutral pH, metals tend to move from water to sediment:
– settling of particulates (e.g. mine wastes);
– precipitation of insoluble metal species;
– sorption of metals on sediment particles.
High concentrations of metals in bed sediments can lead to toxic effects on benthic organisms.
2
Applications of sediment toxicity testing
• Ecological risk assessment (e.g., Superfund)
• Document ecological injury (e.g., NRDAR)
• Pre- and post-remediation assessment
• Effluent monitoring/Toxicity Identification Evaluation
• Characterize waste or dredged material
• Establish or validate sediment quality guidelines
3
CERC mining-related sediment studies
4
Upper Columbia River (WA)
Clark Fork River (MT)
Whiskeytown NRA (CA)
San Carlos Reservoir (AZ)
Upper Animas River (CO)
Tri-State (MO/KS/OK)
Old Lead Belt (MO)
Viburnum Trend (MO)
Palmerton smelter (PA)
Vermont Copper Belt
Types of sediment test methods
• Whole-sediment toxicity testing
– Simulate natural water+sediment exposure
• Pore-water toxicity testing
– Isolate water exposure route
• Elutriate testing (sediment-water suspension)
– Effects of dredging or resuspension
• Sediment extracts or leachates
– Source identification; prioritize cleanups
5
Whole-sediment testing
6
• Goal: simulate surficial sediments and overlying water
– Allow development of limited depth gradient (3-4 cm)
– Realistic role of overlying water (water quality, replacement rate)
Whole-sediment toxicity tests
• Direct measure of effects on benthic organisms
• Support cause-effect findings
• Wide applicability
• Limited special equipment is required
• Rapid and inexpensive
• Legal and scientific precedents
• Integrates interactions of contaminant mixtures
• Amenable to field validation
7
Pore-water testing
8
• Goal: isolate aqueous exposure route
– Use standard aquatic test organisms
• Advantages:
– Simplicity and sensitivity of test methods
– Compare aqueous vs. solid-phase exposure
• Disadvantages
– Difficulty of pore-water collection
– Artifacts of testing with water-column organisms
Daphnid
Fathead Minnow
Comparison of exposure routesPalmerton smelter, PA (Besser et al. 2009)
Tested surface water, pore water, and sediment with Hyalella
Toxicity in surface water and pore water from same three sites
Limited toxicity of whole sediment (one site)
Consistent with metal inputs from groundwater seepage
Fine sediments scarce in contaminated stream reach
9
Surface water
BC1
(R)
AC8
(R)
AC6
AC5
AC2/
3AC1
LR4
(R)
LR3.
5LR
3LR
2
LR2
(D)
Su
rviv
al (o
f 10)
0
2
4
6
8
10
* * *
Green = referencePink = smelter
Yellow = downstream
Pore water
BC1
(R)
AC8
(R)
AC6
AC5
AC5
(D)
AC2/
3AC1
LR5
(R)
LR3.
5LR
30
2
4
6
8
10
*
*
*
BC1
(R)
AC8
(R)
AC6
AC5
AC2/
3AC1
LR5
(R)
LR4
(R)
LR3.
5LR
30
2
4
6
8
10
*
Sediment
Sediment vs. Pore-water testsViburnum Trend MO (Besser et al 2008a)
Whole-sediment tests with Hyalella (left) identified several toxic sites
Pore-water tests with Ceriodaphnia (right) were more sensitive, but had variable survival in reference sites (green)
Limited tolerance for PW constituents (e.g. ammonia)
10
Hyalella Ceriodaphnia
MT
01
MT
02
MT
03
MT
04
MT
06
MT
07
MT
10
MT
12
MT
14
MT
20
CL
1C
L2
CL
3C
L4
CL
5Am
ph
ipo
d s
urv
iva
l (%
)
0
20
40
60
80
100
*
*
MT14
MT12
MT3
MT4
MT2
MT1
MT10
MT7
MT6
MT20
CL3
CL2
CL5
CL4
CL1
Ce
rio
da
ph
nia
su
rviv
al (%
)
0
20
40
60
80
100
Characteristics of sediment test organisms
• Sensitivity to toxicants (metals)
• Availability / Ease of culture
• Life cycle / Potential endpoints
• Taxonomic group
• Distribution and abundance
• Ecological importance
11
Standard sediment test organisms
12
Amphipod (Hyalella) Midge (Chironomus) Oligochaete (Lumbriculus)
Mussel (Lampsilis)
Alternative test organismsMayfly (Hexagenia)
Sensitivity of benthic taxa to metalsNi-spiked sediment (Besser, unpublished data)
13
•Differences among species:HA=Hyalella (amphipod)
GP=Gammarus (amphipod)
HS=Hexagenia (mayfly)
CD, CR=Chironomus (midge)
TT=Tubifex (oligochaete)
LV=Lumbriculus (oligochaete)
LS=Lampsilis (mussel)
•Sediment differences Metal bioavailability
Sediment Nickel
Toxicitty threshold (EC-20, g/g)
100 1000 10000
Cu
mu
lati
ve p
rop
ort
ion
0.0
0.2
0.4
0.6
0.8
1.0
GP*
HA
HS
LV*
CD*
CR*
LS*
TT*
HA
GP
HS
TT
CD
CR
LV
LS*
SR sediment
WB sediment
Differences in sensitivityBig River, Missouri (Besser et al. 2010)
Toxicity to mussels was more closely associated with
sediment metals. 14
LampsilisHyalella
Amphipod growth
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 17 18
Len
gth
(m
m)
2
3
4
5
Mussel length
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 17 18
Len
gth
(m
m)
1.0
1.5
2.0
2.5
Sediment metals
1 2 3 4 5 6 7 9 10 11 12 13 14 15 17 18
Su
m (
PE
Q)
0
10
20
30 Pb
Zn
Cd
Test endpoints
• Survival
– Severe effect; acute or chronic test
• Growth (length or weight)
– Often more sensitive than survival
• Biomass production
– Sensitive; integrates effects on survival and growth
• Reproduction
– Sensitive but variable; long/complex test methods;
• Bioaccumulation
– Document bioavailability; characterize dietary exposure of fishes
15
Hyalella survival and reproductionViburnum Trend, MO (Besser et al. 2008a)
Survival was high in reference sediments (green); few toxic sites
Reproduction was sensitive, but varied among reference sites
Influence of nutrients, organic matter, etc.
16
MT
01
MT
02
MT
03
MT
04
MT
06
MT
07
MT
10
MT
12
MT
14
MT
20
CL
1C
L2
CL
3C
L4
CL
5Am
ph
ipo
d s
urv
iva
l (%
)
0
20
40
60
80
100
*
*
MT
01
MT
02
MT
03
MT
04
MT
06
MT
07
MT
10
MT
12
MT
14
MT
20
CL
1C
L2
CL
3C
L4
CL
5
Yo
un
g p
er
fem
ale
0
2
4
6
8
10
12
14
**
*
Interpretation of toxicity data
• Control sediments – define test performance
– Quality assurance for studies with field-collected sediment
– Treatment comparisons in experimental studies
• Reference sediments – define ‘baseline’ conditions
– Single site for simple study area (e.g. upstream/downstream)
– Multiple sites (‘reference envelope’) to represent broader area
• Concentration-response relationship
– Experimental studies (e.g., spiking) or field data with gradient of metal concentrations
– Estimate toxicity value (e.g., LC50, EC20)
17
Comparisons to reference site(s)(Seal et al. 2010; Besser et al. 2010)
Ely Mine, VT: upstream reference sites to match each stream segment
Big River, MO: multiple reference sites (both upstream and regional) – Wide range of sediment type from headwaters to mouth
18
Ely Mine, Vermont
EB1
EB2
EB3
EB4
SB1
SB3
SB4
SB5a
SB5b
OR1
OR3
Mid
ge g
row
th, m
g
0.0
0.5
1.0
1.5
2.0Reference
Mining
*
* ** *
Old Lead Belt, Missouri
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 17 18 19 20 21
Mu
ssel le
ng
th (
mm
)
1.0
1.5
2.0
2.5
Reference
Toxic
Non-toxic
Laboratory-Field Comparisons
• Establish cause-effect relationships
– Community data can be influenced by historic impacts (e.g. species loss) and habitat alteration
– Lab tests use taxa of interest, minimize influence of habitat
• Estimate site-specific toxicity thresholds
– Use of local species or surrogate
– Simulate ambient water quality
19
Laboratory vs. field responses(Besser et al. 2010; Seal et al 2010, in press)
Missouri: reduced mussel growth predicts community impacts
Vermont streams: gradient of amphipod survival vs. benthos taxa richness
Acid sites (red): low taxa richness, but sediment not toxic
20
Old Lead Belt, Missouri
Mussel length (mm)
1.2 1.4 1.6 1.8 2.0 2.2 2.4
Mu
ssel ta
xa
0
10
20
30
Non-toxic
Toxic
Vermont Copper Belt
Amphipod survival (%)
0 20 40 60 80 100
Nu
mb
er
of
taxa
0
10
20
30
40
50
E1
O1
O3
S1
S3
S4
S5 S5
E2
E3
C10
C10A
C10B
C10C
C10D
P4P4A
P4C
P4E
P5
P5AP5
P6
Metal bioavailability in sediment
• Estimate available metal fractions
– Selective extractions (e.g., Luoma 1989, Tessier et al. 1984)
• Characterize major metal-binding phases
– Acid-volatile sulfide and total organic carbon (Ankley et al 1996; USEPA 2005)
– AVS strongly limits metal solubility: Ag, Cu, Pb, Cd, Zn, Ni
– TOC has weaker binding but high capacity; more stable
Allows estimation of pore-water metals (highly bioavailable)
21
Metal fractions and bioavailabilityLake Roosevelt, WA (Besser et al. 2008b; Paulson and Cox 2007)
Upstream site (LR7) was most toxic and had greatest total metals
Downstream toxic sites (LR3, LR2) had much lower total metals
Metals are in easily-extractable fractions (F1 and F2) 22
Midge Growth
REF LR7 LR6 LR5 LR4 LR3 LR2 LR1
Dry
we
igh
t (m
g)
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
*
*
*
*
Sediment Pb fractions
LR7 LR6 LR5 LR4 LR3 LR2 LR1 SA8
Pe
rce
nt
of
tota
l
0
20
40
60
80
100
F4 (Residual)
F3 (Sulfide/OM)
F2 (Oxide)
F1 (Sorbed)
Total Sediment Metals
REF LR7 LR6 LR5 LR4 LR3 LR2 LR1
Su
m-P
EQ
0
20
40
60
80
100
Zn
Pb
Cu
Cd
Metal bioavailability in pore water
• Measure dissolved metal concentrations
– Field: Push-point (large volume) or airstone (small volume)
– Lab: Centrifuge or pressure (large volume)
– Lab or Field: Peeper (small volume)
• Free or labile metal fraction
– Specialized samplers (e.g., DGT)
– Geochemical modeling
– Biotic ligand models (BLM): model metal binding to site of uptake
23
Pore-water sampling methods
24
Peeper
Centrifuge
DGT (Zhang et al. 1995)
4 cm2-4 cm sediment
0-2 cm sediment
0-2 cm water
4 cm2-4 cm sediment
0-2 cm sediment
0-2 cm water
Push-point
BLM models for pore water?(Copper toxicity and DOC; Wang et al 2009)
BLM predicts acute copper toxicity across wide range of water quality
Few BLM studies with pore-water
25
50 100 150 200 250 3000
100
200
300
400
500
Water hardness (mg/L as CaCO3)
0 100 200 300 400 500 6000
100
200
300
400
500
Me
asu
red
Cu
EC
50
(µ
g/L
)
BLM predicted Cu EC50 (µg/L)
Pond
Eagle Bluffs
Ditch #6
Luther Marsh
Humic Acid
Variable Hardness
Reference tests
Sediment quality guidelines(to protect benthic organisms)
1. Empirical SQGs (MacDonald et al. 2000)
Based on frequency of toxicity in large datasets (sediments with multiple toxicants)
– Probable Effect Concentration (PEC) is concentration associated with increased frequency of toxicity
– PEC-Quotient = sediment metal concentration / PEC
• Can sum PEC-Quotients to characterize metal mixtures
26
Application of PEC Quotients(Besser et al. 2010; MacDonald et al 2009)
Big River, MO (left): mussel toxicity at Zn-PEQ >1.0
Tri-States (right): amphipod toxicity at Sum-PEQ near 1027
Amphipod survivalMussel shell length (mm)
Zinc PEQ
0 1 2 3 4 51.2
1.4
1.6
1.8
2.0
2.2
2.4
Reference
Toxic
Non-toxic
Sediment quality guidelines (continued)
2. Equilibrium Sediment Benchmarks (ESBs; USEPA 2005)
Assumes pore water is primary exposure route
– Normalize metals to acid-volatile sulfide (AVS):
• No toxicity if simultaneously-extracted metals (SEM) > AVS
• (SEM = sum of Ag, Cu, Pb, Cd, Zn, and Ni)
– Then normalize to TOC: (SEM‐AVS)/TOC
• Range of uncertain toxicity = 130 to 3000 umol/g (USEPA 2005)
28
29
AVS normalization of nickel toxicityNi-spiked sediments (Besser, unpublished data)
Wide range of toxicity (expressed as total Ni) among eight sediments
Normalizing to [SEM-AVS] reduces variation among sediments
[SEM Nickel - AVS]
SEM(Ni)-AVS (umol/g)
-40 -20 0 20 40 60 80 100S
urv
ival (o
f 10)
0
2
4
6
8
10
Total Nickel
TR-Nickel ( g/g)
0 2000 4000 6000 8000
Su
rviv
al (o
f 1
0)
0
2
4
6
8
10
30
Application of sediment ESB(Tri-State Mining District; MacDonald et al 2009)
Hyalella survival corresponds to [(SEM-AVS)/TOC]:
– Narrower range of uncertainty (low AVS, low TOC)
Amphipod toxicity -- Tristate 2007
SEM-AVS/foc, umol/g
10 100 1000 10000
Am
ph
ipo
d s
urv
iva
l (%
)
0
20
40
60
80
100
Logistic regression:
(r2 = 0.65)
EC20 ~1200
Take-home points
• Sediment toxicity testing has many applications.
• Whole-sediment tests are realistic and broadly applicable.
• Test with multiple species and endpoints.
• Select of appropriate reference site(s).
• Validate toxicity vs. evidence of community impacts.
• Characterize controls on metal bioavailability.
31