Terrestrial concentration ratio database: Analyses by wildlife groups

Terrestrial concentration ratio database:
Analyses by wildlife groups and RAPs
Wood MD, Beresford NA, Howard BJ,
Copplestone D & Yankovich TL
Concentration ratios
Wildlife Transfer Database (WTD)
Wildlife Transfer Database (WTD)
Wildlife Transfer Database (WTD)
Ecosystem sub-category
Definition
Terrestrial
Generic ecosystem including data from all
terrestrial ecosystem types (excludes
terrestrial areas of estuarine systems)
Terrestrial - semi-natural
grassland
Includes: mountain and upland grasslands,
heath and shrub lands, and some Arctic
ecosystems
Terrestrial - forest
Land with tree crown cover of more than 10
% over an area of more than 0.5 ha and with
trees, which are able to reach a minimum in
situ height of 5 m at maturity
Terrestrial - agricultural
grassland
Terrestrial - coastal sand dunes
Managed grasslands
Terrestrial - wetland
Marsh, fen, peatland (excludes estuarine
saltmarshes)
Coastal sand dunes (excludes marine
organisms)
Differences between sub-groups
The Plan
• Use Wildlife Transfer Database
• Derive CRwo-media for sub-categories
– e.g. broad soil types for human foodchain
• Test statistical significance of sub-category CRwo-media
values
• Easy?........NO!
Wildlife Transfer Database (WTD)
Wildlife Transfer Database (WTD)
• WTD entries include both individual and summarised values
(n, arithmetic mean, SD)
• Some values n>1, arithmetic mean, but no SD
The solution
• Ideal – return to original source data, but >520 sources (19502010)
• For each summarised data line, treat as n = 1 (IAEA human
foodstuff approach)
– Loses information on within study variation and omits the weighting
of larger studies
• For each summarised data line
– Assume lognormal
– Generate distribution
– Sample n times
(random percentiles)
The solution
• Lognormal, so lnx follows normal distribution
• Derive arithmetic mean and SD of lnx
ln x
1 2
 ln  x   ln x
2
 ln x
  x2 
 ln 1  2 
 x 
• Generate n random percentiles, sample lnx at these percentiles, reverse
transform (explnx = x)
– Uses Z tables (full details in paper)
• But.......
The solution
• What about n>1, arithmetic mean, but no SD?
• CV calculated for each study reporting both mean and SD
i
CV 
i
• Mean CV for used to estimate missing arithmetic SDs
 i  CV  i
• Enables us to derive a full set if individual summary values (Reconstructed
Database or RDB)
A calculator to make things easier
Available at: https://wiki.ceh.ac.uk/x/PgC6Cw
The solution
n

 xi
i 1
N
g  n xi xi 1  xn
n

 x   
i 1
2
i
N


 g  exp 




(ln ) 

g 
i 1

N



n
xi
2
What did we find?
Element
Wildlife
group
Wildlife subcategory
N
Arithmetic
mean±SD
Geometric
mean(SD)a
Cs
Arthropod
Carnivorous
15
(2.4±4.7)E-1
4.1E-2(6.7)*
Detritivorousd 56
(1.1±2.7)E-1
3.1E-2(4.1)*
Herbivorousd
5
(3.7±2.8)E-2
2.9E-2(2.0)*
Carnivorous
11
(1.5±1.8)E-1
6.0E-2(5.5)#
Herbivorous
57
1.0±1.5
3.2E-1(4.8)*
Omnivorous
79
(6.1±19)E-1
1.8E-1(4.3)*,#
Carnivorous
231
(5.4±19)E-1
1.4E-1(4.1)^
Herbivorous
1879
3.8±8.4
1.8(3.5)#
Omnivorous
333
3.3±6.0
1.0(6.2)+
17±16
11(3.1)*
Cs
Cs
Bird
Mammal
Rangifer spp.c 916
What did we find?
Element
Wildlife
group
Wildlife subcategory
N
Arithmetic
mean±SD
Geometric
mean(SD)a
Cs
Arthropod
Carnivorous
15
(2.4±4.7)E-1
4.1E-2(6.7)*
Detritivorousd 56
(1.1±2.7)E-1
Generic
ok
3.1E-2(4.1)*
Cs
Cs
Bird
Mammal
Herbivorousd
5
(3.7±2.8)E-2
2.9E-2(2.0)*
Carnivorous
11
(1.5±1.8)E-1
6.0E-2(5.5)#
Herbivorous
57
1.0±1.5
3.2E-1(4.8)*
Omnivorous
79
(6.1±19)E-1
1.8E-1(4.3)*,#
Carnivorous
231
(5.4±19)E-1
1.4E-1(4.1)^
Herbivorous
1879
3.8±8.4
1.8(3.5)#
Omnivorous
333
3.3±6.0
1.0(6.2)+
17±16
11(3.1)*
Rangifer spp.c 916
What did we find?
Element
Wildlife
group
Wildlife subcategory
N
Arithmetic
mean±SD
Geometric
mean(SD)a
Cs
Arthropod
Carnivorous
15
(2.4±4.7)E-1
4.1E-2(6.7)*
Detritivorousd 56
(1.1±2.7)E-1
3.1E-2(4.1)*
Herbivorousd
5
(3.7±2.8)E-2
2.9E-2(2.0)*
Carnivorous
11
(1.5±1.8)E-1
6.0E-2(5.5)#
Herbivorous
57
1.0±1.5
3.2E-1(4.8)*
Omnivorous
79
(6.1±19)E-1
1.8E-1(4.3)*,#
Carnivorous
231
(5.4±19)E-1
1.4E-1(4.1)^
Herbivorous
1879 Supports
3.8±8.4
1.8(3.5)#
Omnivorous
333exclusion
3.3±6.0
1.0(6.2)+
Cs
Cs
Bird
Mammal
Rangifer spp.c 916
17±16
WTD
in
11(3.1)*
What did we find?
Element
Wildlife
group
Wildlife subcategory
N
Arithmetic
mean±SD
Geometric
mean(SD)a
Cs
Arthropod
Carnivorous
15
(2.4±4.7)E-1
4.1E-2(6.7)*
Detritivorousd 56
(1.1±2.7)E-1
3.1E-2(4.1)*
Herbivorousd
5
(3.7±2.8)E-2
2.9E-2(2.0)*
Carnivorous
11
(1.5±1.8)E-1
6.0E-2(5.5)#
Herbivorous
57
1.0±1.5
Statistical
79
(6.1±19)E-1
differences,
231
(5.4±19)E-1
but order?
3.2E-1(4.8)*
1879
3.8±8.4
1.8(3.5)#
333
3.3±6.0
1.0(6.2)+
17±16
11(3.1)*
Cs
Bird
Omnivorous
Cs
Mammal
Carnivorous
Herbivorous
Omnivorous
Rangifer spp.c 916
1.8E-1(4.3)*,#
1.4E-1(4.1)^
So what next?
• Wildlife sub-categories (more analysis being done)
– Available data limit our ability to divide CRwo-media values
• What about main wildlife groups?
– Are they really different?
WARNING
What follows are initial results
Results - Am
6.00E-01
A
CRwo-soil (GM)
5.00E-01
4.00E-01
3.00E-01
2.00E-01
1.00E-01
A
A
A
A
A
AB
0.00E+00
Wildlife Group
AB
AB
AB
B
Results - Am
6.00E-01
• Lichen n=3
• Mollusc, amphibian, most of
annelid and most arachnid all from
sand dunes with marine Am source
(small sample numbers)
A
CRwo-soil (GM)
5.00E-01
4.00E-01
3.00E-01
2.00E-01
1.00E-01
A
A
A
A
A
AB
0.00E+00
Wildlife Group
AB
AB
AB
B
Results -Cd
2.00E+01
A
1.80E+01
CRwo-soil (GM)
1.60E+01
1.40E+01
1.20E+01
1.00E+01
8.00E+00
6.00E+00
4.00E+00
2.00E+00
B
C
D
E
DE
F
0.00E+00
Wildlife Group
F
G
G
G
Results -Cd
2.00E+01
• Arachnid (n=30, 1 st.), Annelid (n=452, 5 st.),
Arthropod (n=684, 7 st.), mollusc (n=34, 1 st.)
• Annelid & Arthropod
(justifiable CR groups)
• Amphibian (n=5, 1 st.),
Birds (n=5, 1 st.), Mammals (n=430, 5 st.)
• Arguably vertebrate CR appropriate
A
1.80E+01
CRwo-soil (GM)
1.60E+01
1.40E+01
1.20E+01
1.00E+01
8.00E+00
6.00E+00
4.00E+00
2.00E+00
B
C
D
E
DE
F
0.00E+00
Wildlife Group
F
G
G
G
Results - Cs
2.50E+00
A
CRwo-soil (GM)
2.00E+00
1.50E+00
1.00E+00
5.00E-01
A
AB
B
B
B
B
B
0.00E+00
Wildlife Group
C
CD
CD
CD
D
Results - Cs
2.50E+00
• Mammal (n=3421 of which 1/3 is Rangifer)
A
CRwo-soil (GM)
2.00E+00
1.50E+00
1.00E+00
5.00E-01
A
AB
B
B
B
B
B
0.00E+00
Wildlife Group
C
CD
CD
CD
D
Results - Cs
1.20E+01
• Mammal (n=3421 of which 1/3 is Rangifer)
• Amphibian (n=139, 5 st.), Bird (n=227, 10 st.),
Reptile (n=141, 6 st.) – but all may include
riparian
A
CRwo-soil (GM)
1.00E+01
8.00E+00
6.00E+00
4.00E+00
2.00E+00
B
B
BC
C
C
C
C
C
0.00E+00
Wildlife Group
D
DE DE DE
E
Results - Cs
1.20E+01
• Mammal (n=3421 of which 1/3 is Rangifer)
• Amphibian (n=139, 5 st.), Bird (n=227, 10 st.),
Reptile (n=141, 6 st.) – but all may include
riparian
• Vertebrates vs invertebrates
A
CRwo-soil (GM)
1.00E+01
8.00E+00
6.00E+00
4.00E+00
2.00E+00
B
B
BC
C
C
C
C
C
0.00E+00
Wildlife Group
D
DE DE DE
E
Results - Pb
1.20E+00
A
CRwo-soil (GM)
1.00E+00
8.00E-01
6.00E-01
4.00E-01
2.00E-01
B
BC
BC
BC
0.00E+00
Wildlife Group
ABCD
CD
D
Results - Pb
1.20E+00
A
CRwo-soil (GM)
1.00E+00
8.00E-01
6.00E-01
•
•
•
•
Lichen & bryophyte (n=349, 10 st.)
Amphibian (n=24, 2 st., one of which all are juveniles)
Mammal (n=800, >20 st.)
Arachnid (n=2), Arthropod (n=563, 5 st.), Annelid
(n=701, 7st)
• Grasses & herbs (n=347, 13 st.)
4.00E-01
2.00E-01
B
BC
BC
BC
0.00E+00
Wildlife Group
ABCD
CD
D
Results - Ra
4.00E-01
A
CRwo-soil (GM)
3.50E-01
3.00E-01
2.50E-01
2.00E-01
1.50E-01
A
• Arthropod (n=27, 3 st. – 0.05-0.09 & 68 from same st.), Lichen & Bryophtes
(n=252, 11 st.), Shrub (n=527, 9 st. –
mainly Canada)
• Grasses & herbs (n=478, >20 st.),
mollusc (n=10, 1 st.), bird (n=48, 2 st.)
A B • Mammal (n=84, 13 st.)
• Tree (n=32, 2 st.)
1.00E-01
5.00E-02
C
BCD
0.00E+00
Wildlife Group
CD
D
E
Results - Ra
4.00E-01
A
CRwo-soil (GM)
3.50E-01
3.00E-01
2.50E-01
2.00E-01
1.50E-01
A
• Arthropod (n=27, 3 st. – 0.05-0.09 & 68 from same st.), Lichen & Bryophtes
(n=252, 11 st.), Shrub (n=527, 9 st. –
mainly Canada)
• Grasses & herbs (n=478, >20 st.),
mollusc (n=10, 1 st.), bird (n=48, 2 st.)
A B • Mammal (n=84, 13 st.)
• Tree (n=32, 2 st.)
1.00E-01
5.00E-02
C
BCD
0.00E+00
Wildlife Group
CD
D
E
Key findings from group comparisons
• Many group specific CRs not statistically different
– Often low n and few studies
– Large variability
• Removal of Rangifer from mammal is appropriate
• Screening assessments
– Higher level CR groupings (e.g. invertebrate,
vertebrate)
ICRP RAPs
RAP v Wildlife Group
1.00E+01
1.00E+00
1.00E-01
1.00E-02
1.00E-03
1.00E-04
Solid fill =
no significant difference
Am – Deer
(n=13,
st.), Not deer
(n=14, 2 st.)
RAP
v 1Wildlife
Group
1.00E+01
1.00E+00
1.00E-01
1.00E-02
1.00E-03
1.00E-04
Solid fill =
no significant difference
Pu – DeerRAP
(n=21,v3Wildlife
st.), Not deer
(n=50, 5 st.)
Group
1.00E+01
1.00E+00
1.00E-01
1.00E-02
1.00E-03
1.00E-04
Solid fill =
no significant difference
Cd – DeerRAP
(n=13,v3Wildlife
st. E-3 to E-2),
Group
Not deer (n=20, 1 st. 5 – 8)
1.00E+01
1.00E+00
1.00E-01
1.00E-02
1.00E-03
1.00E-04
Solid fill =
no significant difference
Pb – DeerRAP
(n=12,v3Wildlife
st.), Not deer
(n=92, 13 st.)
Group
1.00E+01
1.00E+00
1.00E-01
1.00E-02
1.00E-03
1.00E-04
Solid fill =
no significant difference
Sr – Deer RAP
(n=72,v6 Wildlife
st.), Not deer
(n=53, 5 st.)
Group
1.00E+01
1.00E+00
1.00E-01
1.00E-02
1.00E-03
1.00E-04
Solid fill =
no significant difference
Sr – Rat (n=46,
Not rat (n=170,
RAP 4vst.),
Wildlife
Group5 st.)
1.00E+01
1.00E+00
1.00E-01
1.00E-02
1.00E-03
1.00E-04
Solid fill =
no significant difference
Cs - RAP v Wildlife Group
CRwo-soil (GM)
Solid fill =
no significant difference
1
0.1
Cs – DeerCs
(n=- 1762,
st.), Not deer
(n=137, 9 st.)
RAP 10
v Wildlife
Group
CRwo-soil (GM)
Solid fill =
no significant difference
1
0.1
Cs – Rat Cs
(n=82,
6 st.),vNot
rat (n=270,
6 st.)
- RAP
Wildlife
Group
CRwo-soil (GM)
Solid fill =
no significant difference
1
0.1
Cs – Duck
4 st.),
Not duck (n=112,
4 st.)
Cs(n=40,
- RAP
v Wildlife
Group
CRwo-soil (GM)
Solid fill =
no significant difference
1
0.1
Is it justified to use RAP-specific CRs?
Is it justified to use RAP-specific CRs?
• Probably not……
Is it justified to use RAP-specific CRs?
• Probably not……
– Based on RAP analysis
Is it justified to use RAP-specific CRs?
• Probably not……
– Based on RAP analysis
– Based on wildlife group analysis
Is it justified to use RAP-specific CRs?
• Probably not……
– Based on RAP analysis
– Based on wildlife group analysis
Questions…