DEVELOPMENT OF AN EUROPEAN
QUANTITATIVE EUTROPHICATION RISK
ASSESSMENT OF POLYPHOSPHATES IN
DETERGENTS
Barbara M. de Madariaga;
M. José Ramos,
José V. Tarazona
BACKGROUND
CEEP, within the voluntary initiative HERA, presented
a risk assessment report on polyphosphates in
detergents
The risk estimation was based on a toxicity
assessment following the TGD; the RAR stated that it
was not possible to estimate the eutrophication risk.
The CSTEE considered that the environmental risk
of polyphosphates should be related to its
contribution to the eutrophication risk and that the
available information should be sufficient for
conducting such assessment.
This study is a follow up of this consideration, and has
been funded by CEEP and conducted by Green Planet
(a technological base spin-off company) and INIA ( a
Spanish public research institute)
1

---

THE STUDY WORK PLAN
Green Planet and INIA developed the initial proposal
The proposal was presented for discussion at an ad
hoc international expert workshop (Nov 2005)
The proposal was adapted to consider the experts'
opinions and has been used for a quantitative
eutrophication risk estimation
The draft report was distributed for comments and
peer review by the experts
The final report was submitted
Additional scenarios have been considered
RISK ASSESSMENT
CONCEPTUAL MODEL
EXPOSURE ASSESSMENT
EFFECT ASSESSMENT
RISK CHARACTERIZATION
RISK COMMUNICATION
MATHEMATICAL IMPLEMENTATION
RESULTS FOR THE PAN-EUROPEAN
ASSESSMENT
RESULTS FOR NATIONAL SCENARIOS
2

---

CONCEPTUAL MODEL
REGIONAL (LARGE RIVER BASINS)
BASIN RISK = RISK FOR SENSITIVE
AREAS
POTENTIAL RISK (= PEC/PNEC IN TGD)
RISK OF PHOSPHATES IN DETERGENTS
CURRENT USE PATTERNS
INDICATED BY CEEP/AISE
EXPOSURE
NEEDS:
ANNUAL AVERAGE TOTAL PHOSPHOROUS
CONCENTRATION
CONTRIBUTION OF THE SOURCE TO BE
EVALUATED (E.G. DETERGENTS) AS
CONCENTRATION OR PERCENTAGE
OPTIONS
GENERIC MODEL FOR LARGE RIVER BASINS
SPECIFIC MODELS
MONITORING DATA
3

---

THE SIMPLIFIED MODEL
PAN-EUROPEAN ASSESSMENT
GENERIC (NOT GIS) MODEL
BASED ON AVERAGE EXPORT COEFFICIENTS
APPLICABLE TO LARGE RIVER BASINS
DISCRIMINATE THE CONTRIBUTIONS FROM
DETERGENTS, OTHER POINT SOURCES AND
DIFFUSE SOURCES
VALIDATED FOR THE DANUBE RIVER BASIN
CAN BE REPLACED BY SITE-SPECIFIC
MODELS AND/OR MONITORING DATA
VALIDATION
Table 1. Export coefficients selected for the simplified model and reported range in the
literature.
Land use
Units
Coefficient
Range
References
Arable Land
kg ha-1 year-1
0.66
0.02 - 123
Lasevils and Berrux, 2000.
Pasture
kg ha-1 year-1
0.4
0.002 ­ 5.8
Hilton et al., 2002
Forest
kg ha-1 year-1
0.02
0.01 ­ 0.51
Hanrahan et al., 2001
Other
kg ha-1 year-1
0.2
0.02 - 3
De Wit and Bendoricchio,
2001

10000
9000
8000
)
/s
7000
3
y = 0,0064x
m
6000
(
R2 = 0,8141
w
o
5000
r
Fl
4000
ve
3000
Ri
2000
1000
0
0
500000
1000000
1500000
Catchment Area (km2)
4

---

DANUBE RIVER BASIN
Figure from Behrendt, H., Huber, P., Kornmilch, M, Opitz, D., Schmoll, O.,
Scholz, 500
G. & Uebe, R. 2000. Nutrient Emissions into river basins of Germany.
UBA-Texte 23/00, 266 pp
)
L
400
ug/
(
e
g
300
r
a
e
v
l

a
200
annua
100
P
0
)
)
l
ein
ge)
al
tó
st
age
ge
zán
sziget
isto
Reni
che
vera
olfsth
ver
era
n-Pr
Jo
Jesenice
n-W
isza
(av
Da
Dan-
Inn (a
a (a
Hercegs
va-
Dan-
Da
rav
Sa
Sava
Mo
Tisa-T
Dan-
MONITORING DATA 2001
MONITORING DATA 2002
MODEL ESTIMATIONS
POINT EMISSION SOURCES
HUMAN METABOLISM
1.5 gP/person and day
DETERGENTS
EU average 0.36 gP/person and day
Maximum (Hungary): 0.84 gP/person and day
STP/WWTP REDUCTION
20%
60%
5

---

EFFECTS ASSESSMENT
CRITERIA ADAPTED FROM WATER
FRAMEWORK DIRECTIVE
APPLIED TO A 303 FIELD CASES
DATABASE
ESTIMATES RELATIONSHIPS BETWEEN
PHOSPHOROUS CONCENTRATION AND
EUTROPHICATION POTENTIAL
EFFECT ASSESSMENT FOR PHOSPHATES
DOSE/RESPONSE RELATIONSHIPS
ADVERSE CONSEQUENCES AS DEFINED BY
THE WFD
RESPONSE DEPENDS ON A LARGE VARIETY OF
VARIABLES
EVEN FOR THE SAME ECOSYSTEMS AND UNDER
CONTROLLED CONDITIONS
ALTERNATIVE
FIELD OBSERVATIONS COVERING THE NATURAL
VARIABILITY
PROBABILITY ESTIMATIONS FOR EFFECTS
6

---

A significant undesirable disturbance is a direct or indirect anthropogenic
impact on an aquatic ecosystem that appreciably degrades the health or
threatens the sustainable human use of that ecosystem
Table 1: Significant undesirable disturbances that may result from accelerated growth of
phytoplankton, macroalgae, phytobenthos, macrophytes or angiosperms
(a)Causes the condition of other elements of aquatic flora in the ecosystem to be moderate or
worse
(a)Causes the condition of benthic invertebrate fauna to be moderate or worse
(a)Causes the condition of fish fauna to be moderate or worse
(a)Compromises the achievement of the objectives of a Protected Area for economically
significant species
(a)Compromises the achievement of objectives for a Natura Protected Area
(a)Compromises the achievement of objectives for a Drinking Water Protected Area
(a)Causes a change that is harmful to human health (e.g. shellfish poisoning)
(a)Causes a significant impairment of, or interference with, amenities and other legitimate uses of
the environment
(a)Causes significant damage to material property
The condition of phytoplankton, phytobenthos, macrophytes, macroalgae or
angiosperms would not be consistent with good ecological status where, as a
result of anthropogenic nutrient enrichment, changes in the balance of taxa had
occurred that are likely to adversely affect the functioning of the ecosystem
Table 2: Examples of ecologically significant undesirable changes
to the balance of taxa
(a) An entire functional group of taxa, or a keystone taxon, normally
present at reference conditions is absent;
(a) A nutrient-tolerant functional group of taxa not present under
reference conditions is no longer rare
(a) A substantial change in the balance of functional groups of taxa has
occurred;
(a) A group of taxa, or a taxon, of significant conservation importance
normally present at reference conditions is missing
7

---

Characteristics
Descriptors
Units and endpoints
Geographical
European Ecological Region
name
identification
River Basin
name
Waterbody Name
name
Morphological and
Waterbody Type
name
physico-chemical
Area
ha
description
Mean Depth
m
Depth Classification
Deep/Shallow
Conductivity
µS/cm
Temperature
ºC
Dissolved Oxygen
mg/L
Secchi disk
m
pH
-
TP & TN annual average conc.
µg/L
Ecological variables
Trophic Status
OECD (1982)
Most relevant
Dominant Species
Number of species and structure
Ecosystem structure
(per taxa group)
Effect endpoints
Chlorophyll a
µg/L
Algal blooms
yes / no
Shifts in Species Composition, Abundance, Structure:
yes / no
Phytoplankton, Invertebrates, Other aquatic flora, Other
Relevant changes
fauna
Relevant changes
Sediment organic matter
yes / no
Change in water quality
yes / no
Oxygenation conditions at hypolimnion
Oxygenated, hypoxia, anoxia
Other specific local effects
yes / no
Eutrophication
Rationale
Direct & indirect effects
Assessment
Ecologically Relevant Effects (ERE)
yes / no
ERE - semi quantitative discrimination
from -3 to +3
Data Validation
Trend in the semi-quantitative classification
MorphoEdaphic Index (MEI based on conductivity) following Vighi, and Chiaudani, 1985.
8

---

Ov erlay Chart
TP Conc Affected
1,000
Mediter anean Affected
,750
A+C Eur. Deep Affected
,500
conditional probability p(TP | G-)
A+C Eur. Shal ow Af ected
,250
Al sites Affected
,000
0,00
250,00
500,00
750,00
1.000,00
Ov erlay Chart
TP Conc Non-Affected
1,000
Mediter anean Non-Af ected
,750
conditional probability p(TP | G+)
A+C Eur. Deep Non-Affected
,500
A+C Eur. Shal ow Non-Affected
,250
Al sites Non-affected
,000
0,00
200,00
400,00
600,00
800,00
FROM FIELD DATA TO RISK
CHARACTERIZATION
· Conditional probabilities p(TP | G+) and p(TP | G-) are used to
define the eutrophication risk as
Relative (0-100%) conditional probability of a water body to be
in less than good status given a certain TP concentration
· p(G- | TP) corrected by maximum value of p(G-)
· Defined in the range:
From 1- p(TP | G+) to p(TP | G-)
· With a most likely value of
mlp(G- | TP) = p(TP | G-)mlp(G-) / p(TP)
9

---

RISK CHARACTERIZATION
ATLANTIC SHALLOW
100
75
i
sk %
r
n
i
o
50
i
cat
h
p
t
r
o
25
u
E
MEDITERRANEAN
0
1
10
100
100 1000
Total P concentration ugP/l
75
i
sk %
r
n
t
i
o
50
ca
phi
t
r
o
25
u
E
0
1
10
100
1000
Total P concentration ugP/l
MATHEMATIC IMPLEMENTATION
INPUTS
Units
Figures
Scenario
MEDITERRANEAN
Effect assessment distribution
2
PopulationDensity
person/ha
1,17
CatchmentArea
ha
10000000
RiverFlow
m3/s
640
LanduseArableLand
%
26
LandusePasture
%
26
LanduseForest
%
38
LanduseOther
%
10
ArableLand coefficient
kg/ha/year
0,66
Pasture coefficient
kg/ha/year
0,4
Forest coefficient
kg/ha/year
0,02
Other uses coefficient
kg/ha/year
0,2
P emission from Population
g/person/day
1,5
P emission from Detergents
g/person/day
0,36
Current P reduction at STP
%
20
Sites with non-good status
%
33
10

---

RESULTS
MEDITERRANEAN
EUTROPHICATION RISK ESTIMATIONS
PREDICTED EXPOSURE LEVELS
Units
Units
1-p(TP | G+)
p(TP | G-) mlp(G- | TP)
Units
TP total concentration
465,1
µg P/l
100
%
TOTAL RISK
93,6
80,5
86,1
%
TP conc. from Detergents
60,9
µg P/l
13,1
%
Risk without Detergents
92,0
76,0
82,4
%
TP conc. from Other Point sources
253,9
µg P/l
54,6
%
Risk without Point sources
81,0
43,0
52,7
%
TP conc. from Diffuse sources
150,2
µg P/l
32,3
%
Risk without Diffuse sources
89,2
67,5
75,5
%
100
90
80
K
I
S
70
R
60
I
ON
T
A
50
I
C
H
40
OP
R
T
30
U
E
20
10
0
TOTAL RISK
Risk without Detergents
Risk without Point
Risk without Diffuse
sources
sources
The line represents the range
Most likely value
RESULTS:
CONTRIBUTION OF DETERGENTS
TO THE OVERALL RISK
GENERIC EUROPEAN
SCENARIOS
11

---

Scenario
Detergent
Difference between total risk and risk without
TP conc.
contribution
detergents
Ecoregion&type
Class
Upper bound
Lower bound
%
µg/l
mlp(G-|TP)
1-p(TP|G+)
P(TP|G-)
1a
13.1
465
Mediterranean
1.6
4.5
3.7
1b
13.1
465
At-N&C shal ow
0.2
1.2
0.5
1c
26
546
Mediterranean
3.4
8.1
7.6
1d
26
546
At-N&C shal ow
0.4
2.3
1
2a
13.1
232
Mediterranean
1.6
4.7
4.4
2b
13.1
232
At-N&C shal ow
0.4
2.8
1.1
2c
26
273
Mediterranean
3.4
10.3
9.3
2d
26
273
At-N&C shal ow
0.8
5.4
2
3a
8
255
Mediterranean
0.9
2.8
2.5
3b
8
255
At-N&C shal ow
0.2
1.4
0.6
3c
16.8
282
Mediterranean
2
6.3
5.5
3d
16.8
282
At-N&C shal ow
0.5
2.9
1.1
4a
9.6
212
Mediterranean
1.1
3.3
3.2
4b
9.6
212
At-N&C shal ow
0.4
2.1
0.8
4c
19.8
239
Mediterranean
2.5
7.4
6.9
4d
19.8
239
At-N&C shal ow
0.7
4.4
1.6
5a
9.9
154
Mediterranean
1.1
3
3.2
5b
9.9
154
At-N&C shal ow
0.4
3.3
1.4
5c
20.4
174
Mediterranean
2.5
6.8
7.2
5d
20.4
174
At-N&C shal ow
0.8
6.7
2.7
Table ES.2.. Median and arithmetic mean values obtained for the different generic scenarios.
Detergent
Difference between total risk and risk
TP conc.
contribution
without detergents
Parameter
Upper bound Lower bound
mlp(G-
% µg/l
1-p(TP|G+)
P(TP|G-)
|TP)
All scenarios
Median
15 247
0.85
3.85
2.6
Arith mean
16 283
1.24
4.48
3.31
Mediterranean scenarios
Median 15 247 1.80
5.50
4.95
Arith mean
16
283
2.01
5.72
5.35
Atlantic-N&Central shallow scenarios
Median 15 247 0.40
2.85
1.10
Arith mean
16
283
0.48
3.25
1.28

12

---

PROBABILISTIC IMPLEMENTATION
MONTE CARLO ANALYSIS
Overlay Chart
Cumulative Comparison
1,000
TOTAL RISK min
,750
Risk without detergents min
,500
TOTAL RISK max
,250
Risk without detergents max
,000
0,0
25,0
50,0
75,0
100,0
Overlay Chart
Cumulative Comparison
1,000
,750
Risk without detergents mlp(G-|TP)
,500
,250
TOTAL RISK mlp(G-| TP)
,000
0,0
25,0
50,0
75,0
100,0
APPLICATION TO
SPECIFIC/NATIONAL
SCENARIOS
CALIBRATION OF EXPOSURE LEVELS
COMPARISON WITH OBSERVED EFFECTS
13

---

SPANISH SCENARIOS
Upper bound
Lower bound
TP conc.
mlp(G-|TP)
Example
1-p(TP|G+)
P(TP|G-)
(µgP/l)
(%)
(%)
(%)
1a­Tajo-Trillo
36
42.5
16.1
12.1
2a­Tajo-Aranjuez
98
76.2
31.9
39.82
3a­Tajo-Polan
1370
100
94.6
100
4a­Tajo-Alcantara
295
88.5
65.1
73.6
5a-Ebro-Miranda
36
42.4
16.1
12.1
6a-Ebro-Mendavia
166
82.09
46.01
55.8
7a-Ebro-Zaragoza
173
82.5
47.2
57.1
8a-Ebro-Tortosa
129
79.2
38.7
47.9
14

---

Example
Catchment/Station
Detergent
Difference between total risk and risk
TP conc.
contribution
without detergents
Upper bound
Lower bound
%
µg/l
mlp(G-|TP)
1-p(TP|G+)
P(TP|G-)
1a
Tajo - Trillo
8.4
36
6
0.2
1.1
1b
Tajo - Trillo
5
36
3.4
0.1
0.6
2a
Tajo - Aranjuez
6.7
98
0.8
1.6
1.9
2b
Tajo - Aranjuez
3.8
98
0.4
0.9
1.1
3a
Tajo - Polan
13.9
1370
0
1
0
3b
Tajo - Polan
9.3
1370
0
0.6
0
4a
Tajo - Alcantara
18.3
295
2.2
7
6
4b
Tajo - Alcantara
13.7
295
1.6
5.2
4.4
5a
Ebro - Miranda
4.7
36
3.2
0.1
0.6
5b
Ebro - Miranda
2.6
36
1.7
0
0.3
6a
Ebro - Mendavia
11.4
166
1.4
3.6
3.9
6b
Ebro - Mendavia
7.2
166
0.8
2.2
2.4
7a
Ebro - Zaragoza
11
173
1.3
3.5
3.7
7b
Ebro - Zaragoza
6.9
173
0.8
2.2
2.3
8a
Ebro - Tortosa
9.4
129
1.1
2.6
3.01
8b
Ebro - Tortosa
5.7
129
0.6
1.5
1.8
APPLICABILITY TO THE DANUBE
RIVER BASIN
100
90
80
K
S
70
PRELIMINARY ESTIMATIONS
N RI
60
I
O
50
CAT
HI
P
BASED ON UBA 2003
40
TRO
30
U
DETERGENTS CONTRIBUTION
E
20
24% OF POINT SOURCES
10
MONITORING DATA
0
TOTAL RISK
Risk without Detergents
Risk without Point
Risk without Diffuse
90 and 140 ug P/l
sources
sources
The line represents the range
Most likely value
100
90
80
K
I
S
70
R
N
60
I
O
T
CA
50
HI
P
40
O
R
T
30
U
E
20
10
0
TOTAL RISK
Risk without Detergents
Risk without Point
Risk without Diffuse
sources
sources
The line represents the range
Most likely value
15

---

CONCLUSIONS
THE STUDY HAS DEVELOPED A MODEL FOR A
QUANTITATIVE ASSESSMENT OF THE
EUTROPHICATION RISK ASSOCIATED TO
PHOSPHOROUS EMISSIONS/CONCENTRATIONS
THE MODEL CONSTITUTES A TOOL FOR ASSESSING
THE CONTRIBUTION OF DIFFERENT SOURCES TO THE
EUTROPHICATION RISK
THE PAN-EUROPEAN ASSESSMENT FOR THE
CURRENT SITUATION SUGGESTS THAT
POLYPHOSPHATES IN DETERGENTS INCREASE THE
EUTROPHICATION RISK BY:
2-8% UNDER MEDITERRANEAN CONDITIONS
0.4-2% UNDER ATLANTIC CONDITIONS
CONCLUSIONS cont.
THE MODEL CAN BE ADAPTED TO SPECIFIC
RIVER BASINS, THUS:
IF YOU HAVE PREDICTIONS OR
MONITORING DATA FOR PHOSPHOROUS
CONCENTRATIONS.....
....THE MODEL BECOMES THE TOOL FOR
MOVING FROM P CONCENTRATIONS TO
EUTROPHICATION RISKS.
16

---

THANK YOU
17

---