Transcript Dissolved Metal Concentration

1
Incorporation of bioavailability
Patrick Van Sprang – ARCHE
OECD Workshop on Metals Specificities in Environmental
Hazard Assessment, Paris, 7-8 september 2011
Introduction
• Metals are found in different forms in the environment
2
• These are referred to as metal “species”
• Changing in the environment is called “ (chemical) speciation” or
“transformation”
• Kinetics and chemical speciation under environmentally relevant
conditions crucial for PNEC derivation & read-across
• Important point: Not all metal species are toxic
Metals exist in the environment…
Particulate Metal
– adsorbed to suspended solids
or mineral surfaces)
 each of these processes may reduce metal bioavailability/toxicity
(POC1
Dissolved Metal Complexes
– with inorganic ligands (OH-, CO32-, HCO3-,..)
– with dissolved NOM2 (measured as DOC3: humic and fulvic acids)
 each of these processes may reduce metal bioavailability/toxicity
Dissolved Free Metal
– Free-ion forms tend to bind to biological ligands, e.g.physiologically
active sites at the gill
 these species mainly causes metal toxicity
1POC:
Particulate Organic Carbon
2NOM: Natural Organic Matter
3DOC: Dissolved Organic Carbon
3
Metal toxicity can be expressed as…
Total Metal Concentration
4
For terrestrial and sediment systems, the concentration of a metal that is
determined after destruction of the mineral matrix. For aqueous systems: the
total amount of metal present, including the fraction sorbed to particles and to
dissolved organic matter and the fraction in the mineral matrix;
= particulate (sorbed + precipitated) + dissolved (inorganic complexes + organic
complexes + free ionic forms)
Dissolved Metal Concentration*
most often, the dissolved fraction in ecotoxicity tests refers to the fraction that
passes through a filter of 0.45 µm.
= inorganic complexes + organic complexes + free ionic forms
* It should be noted, however, that this definition may not necessarily refer to the metals
in solution. In the range of 0.01- 0.45 µm colloid inert particles containing metal ions that
remain suspended, may still exist;
Metal toxicity can be expressed as…
Bioavailable Metal Concentration
- the degree to which a metal species is available to interact with
the biotic ligand (e.g. physiologically active sites at the gill) to exert
its effect.
= free ionic forms (mainly)
- Free Ion Activity Model (FIAM): assumes that the free
metal ion activity reflects the chemical reactivity and
toxicity of the metal
- Biotic Ligand* Model (BLM): assumes that both the free
metal ion activity and the interaction of the available
cationic forms with the organism reflect the toxicity.
* A "biotic ligand" is a biochemical receptor that is metal-binding and
is treated similarly to other ligands in the exposure water, except that
it is on the organism. An example of a biotic ligand is a fish gill.
5
Why incorporate bioavailability in
CSR of metals?
6
• NOEC/EC10 in laboratory test media which often maximizes
bioavailability (e.g. low DOC in water; low OC in soil) may not reflect ‘real
environment’ (rivers may have different DOC, pH) !
• Database often contains NOEC/EC10 obtained in test media with widely
varying chemistry (= very different bioavailability) which toxicity values to
select (species mean ?, lowest NOEC/EC10 ?) ?
• Generic/uncorrected SSD does not represent ‘intrinsic sensitivity’ alone
but rather a mix of ‘intrinsic sensitivity’ + bioavailability effects toxicity
values should therefore be normalized towards similar physico-chemical
conditions !
Why inorporate bioavailability in
CSR of metals?
Incorporation
bioavailability
Concentration (µg/l)
Generic PNEC
Cumulative Distribution
Function (%)
Cumulative Distribution
Function (%)
• Bioavailability models ‘remove’ the variability in sensitivity due to
differences in physico-chemistry
Concentration (µg/l)
Normalized PNEC
7
Does bioavailability matter in EU waters ?
Acute effects (LC50 in µg/l) of copper to Daphnia magna,
tested in 11 different EU surface waters
48h-EC50 in µg Cu/L
De Schamphelaere et al., 2002
400
648
300
200
100
0
01 02 03 05 06 07 07 08 09 10 11
Sampling location
Factor 30 difference in acute Cu-toxicity across EU surface
waters !!
8
Does bioavailability matters in EU soils ?
Chronic effects of nickel (NOEC/EC10 in mg/kg) to soil
organisms/processes tested in 16 different EU surface soils
Soil
Houthalen
Zegveld
Montpellier
Rhydtalog
Jyndevad
Kövlinge II
Aluminusa
Borris
Woburn
Ter Munck
Souli **
Marknesse
Brécy
Cordoba 2
Cordoba 1
Guadalajara
pH (CaCl2)
3.6
4.1
4.1
4.2
4.5
5.1
5.6
5.6
6.1
6.7
7.0
7.6
7.5
7.6
7.6
7.7
Nitrific.
Glucose
196
38
68
59
39
62
89
104
97
253
66
156
196
61
33
37
191
16
555
97
425
148
457
712
190
54
Maize
Barley
Tomato
87
24
226
71
91
26
109
185
110
416
103
283
233
504
396
105
192
24
226
27
48
14
27
47
54
136
103
283
233
504
72
105
192
603
46
297
54
169
84
801
712
68
411
9
Eisenia Folsomia
55
966
78
179
31
282
309
31
303
169
378
299
609
514
195
312
32
619
113
510
87
22
103
183
884
298
559
583
941
875
79
542
Factor between 10-45 difference in chronic Ni-toxicity across
EU soils !!
Approaches for bioavailability ?
TOTAL METAL LEVELS (MONITORING DATA)
10
WATER
SOIL
SEDIMENT
Total
Me-concentration
Total
Me-concentration
Total
Me-concentration
PHYSICO-CHEMICAL SPECIATION MODELLING
KD, SS
Dissolved
Me-fraction
Ca, pH, DOC,…
(speciation model)
Porewater or free
ion model
SEM, AVS
SEM – AVS
Me-fraction
Free ionic
Me-fraction
Toxicity-based models (Biotic Ligand Model, Regression Models,…)
Bioavailable Metal Fraction
BIOAVAILABILITY ASSESSMENT MODELLING
Biogeochemical Region X1
Biogeochemical Region X2
Biogeochemical Region Xn
1. Transformation from total to soluble fraction approach
- Assumption: Dissolved metal concentrations more closely approximate the
biologically available fraction than does total metal concentrations
Total metal concentrations
= dissolved metal
concentrations
(e.g. Ni, Cu, Zn)
≠ dissolved metal concentrations (e.g. Pb)
Conversion needed
No conversion needed
11
- Speciation model (e.g. Minteq: requires
phys-chem characterisation of medium
(pH, Hardness,..))
- Analytical measurements testing
(filtration, then e.g. ICPMS/AAS)
2. Use of speciation models - approach
- Assumption: chemical species (mainly free metal ion activity) is able to explain 12
the observed toxicity FIAM
Dissolved metal concentrations
- Speciation model (e.g. Minteq/WHAM: requires physchem characterisation of medium (pH, Hardness,..))
- Analytical measurements testing (filtration, then e.g.
Ion Selective Electrode (ISE), Donnan membrane
technique (DMT))
3. Toxicity related bioavailability models:
approach
- Assumption: chemical species (mainly free metal ion activity) and the interaction13of
the available cationic forms with the organism reflect the toxicity  BLM
- Chronic BLM’s have been developed & validated for several metals (e.g. Ni, Zn,
Co, Cu in freshwater)/ Acute BLM’s also exist for other metals (e.g. Ag, Cd)
Dissolved metal concentrations
- FIAM - speciation model (e.g. Minteq/WHAM)
- Gill Surface Interaction Model (e.g. CHESS)
The BLM requires a description of water chemical parameters
that can influence metal toxicity:
- pH
- DOC (a convenient measure of NOM)
- Major ions: Calcium, Magnesium
- Others: e.g. Sodium (Cu)
Ca2+
Na+
Mg2+
Log KCaBL
Water
Log KMgBL
Log KNaBL
Me-DOC
Organism
H+
‘biotic ligand’ e.g.
gill, cell surface
14
Log KHBL
pH
Log KHBL
Me2+
[Me] on ‘biotic
ligand’
pH
MeCO3
Toxic effect
MeOH+
Speciation (WHAM) Competition (log K’s)
Intrinsic sensitivity
3. Toxicity related bioavailability models
BLM: development (1)
) (µM)
Acute
70
60
50
40
30
20
10
0
2
2+
R = 0.9559
21d EC50 (Ni
48h EC50 (Ni
2+
) (µM)
15
2
R = 0.9672
0.0
0.5
1.0
1.5
2.0
1.2
Chronic
1.0
2
0.8
R = 0.8086
0.6
0.4
2
R = 0.9918
0.2
0.0
0.0
Cation activity (mM)
0.5
1.0
Cation activity (mM)
Ca
Mg

De Schamphelaere & Janssen, 2002
1.5
3. Toxicity related bioavailability models
BLM: validation (1)

De Schamphelaere et al., 2002
Sampling of waters
16
Factor 2
BLM
Chemical analyses
(pH, DOC, Ca, Na,…)
Test
Adding ≠ concentrations
Determine toxicity
3. Toxicity related bioavailability models
BLM: validation (2)
..for invertebrates and fish
17
10000
predicted EC50 (µg/L)

Daphnia - acute - Cu
Daphnia - chronic -Cu
Daphnia -acute -Zn
1000
Daphnia - chronic -Zn
Daphnia - acute - Ni
100
Field cladocerans acute - Cu
Rainbow trout chronic - Zn
10
10
100
1000
10000
observed EC50 (µg/L)

Factor 10 to 30 variability in toxicity…

reduced to factor 2 in > 90% of the cases
3. Toxicity related bioavailability models
BLM: validation (3)
..for algae
predicted EC50 (µg Cu/L)

18
10000
1000
100
P. subcapitata
Chlorella sp.
C. reinhardtii
P. subcapitata (NOEC, field)
10
10
100
1000
10000
observed EC50 (µg Cu/L)

Factor 10 to 30 variability in toxicity…

reduced to factor 2 in > 90% of the cases
3. Toxicity related bioavailability models
BLM: similar response across metals ? (1)
19
Algae
Invertebrates
invertebrates fish
Fish
algae
NOEC (µg/l Zn)
NOEC (µg/l Zn)
Invertebrates, fish
& algae
NOEC (µg/l Zn)
NOEC (µg/l Zn)
- Zn
pH
DOC
pH
Invertebrates, fish
& algae
Ca > Mg
Hardness
DOC
Invertebrates,
fish, algae
pH
NOEC (µg/l Ni)
Invertebrates, fish
& algae
NOEC (µg/l Ni)
NOEC (µg/l Ni)
- Ni
Invertebrates, fish
& algae
Mg > Ca
Hardness
3. Toxicity related bioavailability models
BLM: similar response across metals ? (2)
- Cu
DOC
invertebrates fish
algae
Hardness
does not
significantly
affect chronic
toxicity
NOEC (µg/l Cu)
Invertebrates, fish
& algae
NOEC (µg/l Cu)
NOEC (µg/l Cu)
20
pH
pH
- Toxicity response = f(organism; phys-chem parameter)
- Toxicity response
= pH: similar for algae; different for
invertebrates & fish
= DOC: similar for all organisms
= H: ± similar for all organism (> Ca for Zn;
>
Mg for Ni; less significant for Cu)
3. Toxicity related bioavailability models
BLM: applicability domain
21
• BLMs developed & validated within 90th % of phys.-chem from EU
waters and should therefore only be applied within such boundaries !!
• Specific conditions outside boundaries need special attention (e.g.
model extrapolation, additional specific testing….
3. Toxicity related bioavailability models
BLM: extrapolation across other species ? (1)
• BLM developed for limited number of species:
– P. subcapitata (green alga)
– D. magna/C. dubia (cladoceran, invetebrate),
– O. mykiss/P. promelas (fish)
• Ecotoxicity database contains NOEC/EC10 for other
species/taxonomic groups (e.g. molluscs, rotifers, insects)
• Given that individual development for all existing aquatic species is
impossible, can a BLM developed for one species be used for
another species?...
• Extrapolation assumes similar mechanism of actions (e.g. similar
stability constants between the cations (Ca, Mg, H) and the biotic
ligands, similar site of action)
22
3. Toxicity related bioavailability models
BLM: extrapolation across other species ? (2)
BLM model fish (rainbow trout)
23
Read across?
BLM model algae (Raphidocelis)
Read across?
BLM model water flea (Daphnids)
Read across?
23
3. Toxicity related bioavailability models
BLM: extrapolation across other species ? (3)
How ?: perform ‘spot checking’ of the BLMs for species for
which no validation had been undertaken.
Literature toxicity data (e.g. Cu)
Toxicity testing (e.g. Ni)
- Insect: Chironomus tentans
- Rotifer: Brachionus calyciflorus
- Molluscs: Lymnaea stagnalis
- Higher plant: Lemna minor
BLM predictions were within a factor ± 3
24
3. Toxicity related bioavailability models
BLM: Implementation in risk assessment (1)
25
3 BLM’s (alga, invertebrate,
fish) available ?
No
Partial BLM normalization
allowed
Yes
‘Spot checks’ available for at
least 3 other species ?
Yes
Full BLM normalization
across all species
No
Partial BLM
normalization or BioF
approach
3. Toxicity related bioavailability models
BLM: Implementation in risk assessment (2)
Partial BLM normalization or
BioF approach (e.g. Zn RA)
Full BLM normalization across
all species (e.g. Ni, Cu-RA)
- Use D. magna/C. dubia BLM to normalise
all other invertebrates (e.g. molluscs,
rotifers,..)
- Use O. mykiss/P. promelas to normalize all
fish/amphibians
- Use R. subcapitata to normalize all other
algae
Eco-region HC5
DOC = 2,5 3,2 3,7 2,8 2,5 8,0
pH = 7,7 8,1 8,2 7,8 6,7 7,6
100
90
12,0 mg/l
6,9
Scenario ditchThe Netherlands
26
1. Normalise the NOEC for the BLM species
towards site specific conditions (NOECx) and
towards EU reference water chemistry
conditions (NOECref)
2. Calculate the bioavailability factors (BioF) for
the BLM species
BioFwater, X 
NOECref
NOECx
Scenario lake Monate Italy
80
Scenario river Rhine The Netherlands
3. Select the highest BioF for the BLM species
Scenario river Otter United Kingdom
70
percentile
Scenario river Teme United Kingdom
60
Scenario acid lake Sweden
Scenario river Ebro Spain
4.- Calculate the bioavailable PEC concentration
50
40
PECbioavailable=PEC * BioFwater,X
30
20
10
0
1
10
9,9 µg/l 14,7 µg/l
100
25,4 µg/l
10,9/11,8 µg/l 17,5 µg/l
57,3 µg/l
NOEC (µg/l)
1000
10000
- Or calculate the bioavailable PNEC
concentration
PNECbioavailable=PNECgeneric /BioFwater,X
4. Bioavailability models in marine waters
- Coastal/open sea waters are characterised by…
- high pH (typically between 7.8–8.3), high salinity (35‰), high ionic27
strength.
- DOC levels may vary considerably between marine waterbodies
- Freshwater and marine organisms face very different iono- and
osmo-regulatory issues related to living in either a very dilute or
concentrated salt environment.
freshwater BLMs can NOT directly be used for marine
environments
- Me-DOC binding freshwater different then marine waters
= Speciation modelling to be modified with the ionic strength
DOC normalization if applicable = bioavailability correction = not
species-specific
27
4. Bioavailability models in marine waters
Bioavailability correction (DOC): derivation of normalized PNEC
value – e.g. Cu
Toxicity - DOC regressions for
6 marine species
28
Model accuracy - Bioavailability
prediction within a factor of 2