Content uploaded by Diederik Schowanek
Author content
All content in this area was uploaded by Diederik Schowanek on Mar 14, 2014
Content may be subject to copyright.
GEO-REFERENCED REGIONAL EXPOSURE ASSESSMENT TOOL FOR EUROPEAN
RIVERS (GREAT-ER): A CASE STUDY FOR THE RUPEL BASIN (B)
F. Verdonck*, G. Boeije*, D. Schowanek** & P.A. Vanrolleghem*
* BIOMATH Department, University of Gent, Coupure Links 653, 9000 Gent, Belgium
** Procter & Gamble, Eurocor, Temselaan 100, 1853 Strombeek-Bever, Belgium
ABSTRACT
In the GREAT-ER project (Geography-referenced Regional Exposure Assessment Tool for
European Rivers), an accurate aquatic exposure prediction tool for chemicals was developed and
validated for use within environmental risk assessment schemes.
In this paper, the application of the GREAT-ER methodology to the Rupel basin in Belgium is
presented. Spatial concentration patterns of the anionic surfactant Linear Alkylbenzene Sulphonate
(LAS) were predicted for this basin. Different scenarios were simulated. LAS removal in waste
water treatment plants is high (98-99.9%). However, since only 30-40 % of the wastewater in the
Rupel basin is treated, predicted environmental concentrations (PECs) are high. For some
subbasins, calculated PEC-values were positioned against biological and physico-chemical water
quality measurements.
This case study showed the feasibility of implementing a new large catchment on a short term and
indicated which efforts are required to apply GREAT-ER on a large, ultimately pan-European
scale. It was found that it is possible to simplify the river network without considerable loss of
accuracy. This allows to gain time and reduce effort.
INTRODUCTION
The goal of comprehensive risk assessment is to estimate the likelihood and the extent of adverse
effects occurring to man, animals or ecological systems due to possible exposure(s) to substances.
The assessment of whether a substance presents a risk to organisms in the environment is based on
the comparison of a predicted environmental concentration (PEC) with a predicted no effect
concentration (PNEC) to ecosystems (Feijtel et al., 1997).
As the current generic ‘unit world’ techniques to assess regional exposure do not account for spatial
and temporal variability and do not offer realistic predictions of actual concentrations, they are
merely applicable on a screening level. The objective of the GREAT-ER project (Geography-
referenced Regional Exposure Assessment Tool for European Rivers) was to develop an accurate
tool to predict chemical exposure in the aquatic environment (Feijtel et al., 1997).
In GREAT-ER, a new database, model and software system were worked out to calculate the
distribution of PEC’s of 'down-the-drain' chemicals in European surface waters, on a river and
catchment area scale. The system uses a Geographic Information System (GIS) for data storage and
visualisation, combined with simple mathematical models for the prediction of chemical fate. At
present, the system contains information for four catchments in Yorkshire (UK), one in Italy, and
two in Germany. GREAT-ER 1.0 has been validated by comparing simulations with the results of
an extensive monitoring campaign for the detergent ingredients Linear Alkylbenzene Sulphonate
(LAS) and boron (Feijtel et al., 1997; Schowanek et al., in press).
1
The output of GREAT-ER 1.0 is three-fold (Schowanek et al., submitted):
1. a colour-coded GIS map with the distribution of a chemical’s PEC in the river basin.
2. a profile of the chemical concentration as a function of the distance for a selected branch of the
river.
3. aggregated PEC’s (i.e. PEC
initial
and PEC
catchment
) to integrate the results for an entire catchment
(Boeije et al., in press).
In this paper, the application of the GREAT-ER
methodology to the Rupel basin is presented. The
Rupel is a tributary to the river Schelde (Figure
1). Compared to the other pilot study catchments,
the Rupel catchment is significantly larger, with
more than 1800 discharges and a catchment area
of about 7000 km
2
. Spatial concentration patterns
of the anionic surfactant LAS were predicted for
this basin.
Figure 1. The Rupel basin in Belgium
METHODOLOGY
A large amount of input data is needed to run the GREAT-ER model. Information about treated and
untreated waste water emissions, hydrological data and chemical market data are needed. These
data were kindly provided by several organisations, i.e. AMINAL ('Administratie Milieu-, Natuur-,
Land- en Waterbeheer', Brussels), AQUAFIN (Aartselaar), AWZ ('Administratie Waterwegen en
Zeewezen', Brussels), MRW ('Ministère de la Région Wallonne', Namur), VMM ('Vlaamse
MilieuMaatschappij', Aalst), and ECETOC (European Centre for Ecotoxicolgy and Toxicology of
Chemicals, Brussels).
Flows, flow velocities and depths are needed as hydrological input data. To this end, an empirical
hydrological model was developed. Since the study area is very large and complex, it was not
possible to apply a deterministic model which needs a lot of input parameters. The applicability of a
power function relating flow to the sum of the lengths of all upstream rivers was demonstrated
(Verdonck, 1999). A correlation coefficient of 0.94 was found.
A limited validation exercise was made. It appeared that the ratio of the modelled to measured flow
varied from 1 to 3 (Verdonck, 1999). Because GREAT-ER's aim is to predict chemical
concentrations with a accuracy factor 3 to 5 (ECETOC, 1999), the accuracy of this hydrological
model was considered acceptable.
RESULTS & DISCUSSION
LAS removal in waste water treatment plants (WWTP’s) is in the range of 98-99.9% (Schowanek et
al., submitted). However, since only 30-40 % of the wastewater in the Rupel basin is currently
treated, predicted environmental concentrations are rather high. An example of a concentration
profile is shown in Figure 2 for the ‘Kleine Nete’ river.
Two scenarios were simulated: the current situation and a hypothetical situation without any
WWTP’s. The effect of the WWTP’s at Dessel and Herentals is clearly shown. The untreated
discharge at Geel causes an increase of the LAS-concentration in the river. The decreasing parts of
the concentration profile are due to dilution, confluences with other rivers and biodegradation.
2
0
0,2
0,4
0,6
0,8
1
1,2
0 5000 10000 15000 20000 25000 30000 35000 40000 45000 50000
Distance (m)
PEC (mg/l)
no purification
current situation
WWTP DESSEL
7300 I.E.
Discharge GEEL
12680 I.E.
WWTP Herentals
14700 I.E.
Confluence
with ’de Aa’
Figure 2: Concentration profile of LAS in the ’Kleine Nete’
Different ‘what if‘-scenarios could be simulated (Verdonck, 1999):
• What if all discharges are treated and what if none are?
• What if a trickling filter is replaced by an activated sludge system?
• How does building a new WWTP at a specific location affect the PEC in the rivers?
• How do in-stream-removal and in-sewer-removal affect the PEC in the rivers?
Results show that WWTP’s induce a significant though local improvement on LAS-concentrations
in the rivers. An activated sludge system has a better influence on the PEC compared to a trickling
filter. LAS-concentrations are rather high in the Zenne basin and the downstream part of the Dijle
basin, intermediate in the Demer basin and relatively low in the Nete basin and the upstream part of
the Dijle basin.
The need to select many or few rivers in a river network (i.e. the level of geographical detail) was
investigated. It was found that it is possible to simplify the river network without considerable loss
of accuracy. Figure 3 shows 4 possible levels of geographical detail (selection only of stretches with
flow rate Q larger than a certain value) and their effect on an aggregated concentration of the
Zuunbeek basin, a subbasin of the Rupelbasin.
0
1
2
3
4
5
6
7
11,52 3
PEC (mg/l)
Level 1:
Q > 0.33 m³/s
Level 2 :
Q > 0.065 m³/s
Level 3:
Q > 0.0065 m³/s
Level 4:
All rivers
Figure 3: Aggregated PEC
catchment,volume
for different levels of geographic detail
for the Zuunbeek basin, a subbasin of the Rupel basin (Q = flow in m
3
/s)
3
If the level of detail increases (for the same emissions of LAS), the total length and residence time
in the river will increase (more distance travelled by the chemical in the modelled basin). Hence, the
impact of biodegradation will increase, and as a consequence the predicted concentration will
decrease. However, the concentration at level 3 and 4 were found to be practically the same. Hence,
simplifying the network to level 3 appeared possible and allows to reduce the required effort in
terms of data collection (less geographical detail) and calculations (fewer stretches to evaluate).
This is an important conclusion in view of the pan-European application of GREAT-ER.
CONCLUSIONS
• The final deliverable of the first stage of the GREAT-ER project is a CD-ROM which contains
the exposure assessment software, the validation data for boron and for LAS in six pilot study
areas, and consumption data for these substances. The results illustrate that GREAT-ER can
deliver accurate simulations of chemical concentration in a river basin, provided reliable
datasets are used.
• A first version of the GREAT-ER implementation for the Rupel basin was developed. The
GREAT-ER simulations for the Rupel are realistic, but further refinement of the data such as
inhabitants, in-stream removal, in-sewer removal and flows can increase the predictive power.
• This case study also showed the feasibility of implementing a new catchment and indicated
which efforts are required to apply GREAT-ER on a large, ultimately pan-European scale.
Problems concerning data collection may possibly occur. Fortunately, this was not the case in
Belgium. Integrating the data in the GIS was the most time consuming work package. Data
incompatibility and inconsistencies could induce some problems, but generally it appeared
feasible to implement a large catchment in a short period of time using (semi-) automatic
procedures.
ACKNOWLEDGEMENTS
The authors thank the European Centre for Ecotoxicology and Toxicology of Chemicals (ECETOC), the Environmental
Risk Assessment Steering Management committee (ERASM) of the Association Internationale de la Savonnerie, de la
Détergence et des Produits d`Entretien (A.I.S.E.) and the Comité Européen de Agents de Surface et Intermédiares
Organiques (CESIO) and the UK Environment Agency for their financial and management support.
REFERENCES
Boeije G, Wagner J-O, Koormann F, Vanrolleghem PA, Schowanek D & Feijtel TCJ (in press) New PEC definitions
for river basins applicable to GIS-based environmental exposure assessment. Contribution to GREAT-ER #8.
Chemosphere
Ecetoc (1999) Personal communication
Feijtel T, Boeije G, Matthies M, Young A, Morris G, Candolfi C, Hansen B, Fox K, Holt M, Koch V, Schröder R,
Cassini G, Schowanek D, Rosenblom J & Niessen H (1997) Development of a geography-referenced regional exposure
assessment tool for European rivers - GREAT-ER. Contribution to GREAT-ER #1. Chemosphere, 11:2351-2373
Schowanek D, Fox K, Holt M, Schroeder, FR, Koch V, Cassani G, Matthies M, Boeije G, Vanrolleghem P, Young A,
Morris G, Gandolfi C & Feijtel TCJ (submitted) GREAT-ER: a new tool for management and risk assessment of
chemicals in river basins. Contribution to GREAT-ER #10. Submitted to IWA Paris 2000 Congress.
Verdonck F (1999) Toepassing van een geografisch gerefereerd regionaal blootstellingsmodel voor chemicaliën op het
Rupelstroombekken. Master's thesis, Universiteit Gent, Faculteit Landbouwkundige en Toegepaste Biologische
Wetenschappen
