Abstract

In Wallonia, the ecotoxicological tests have been chosen as complementary descriptive data of chemical and physical parameters for the monitoring of surface water quality. As they are able to take into account interactions between chemicals and the environment, they allow developing an integrated approach of risk assessment. ArcInfo environmental databases and functionalities are tools used in this way.

INTRODUCTION

ISSeP is commissioned by the Walloon Government of Belgium to manage several monitoring networks of the quality of the environment (air, water, waste deposits, Figure 1). A physical and chemical monitoring program of the rivers has been developed, according to Belgian norms as well as to European directives related to the quality of surface water. Since 1994, ISSeP has to perform the sampling and the analysis of about 90 parameters on 140 points sampled every 4 weeks, to manage the data and to present the results. Even if no regulatory constrain still exist on ecotoxicity, a complementary program involving toxicity tests measurements has been early implemented in order to describe the biological adverse effects of toxic pollution.
This paper will focus on the valorisation of the ecotoxicological information stored in the ArcInfo environmental networks database as a useful parameter for risk assessment.

Figure 1
The environmental networks managed by ISSeP (1998)

THE ECOTOXICOLOGICAL DATA

Ecotoxicology is "the science of toxic substances in the environment and their impact on living organisms" (Jorgensen, 1990). In our approach, the ecotoxicological monitoring consists in measuring in lab the effects of surface water samples on organisms, using microbiotests or conventional tests and to interpret the results in terms of water pollution.
Since ecosystems involve complex interactions, it is difficult to explain the response of biological systems to chemicals or samples. Assessment is further complicated by the ability of the organisms to adapt, by the diversity of species and the differences in responses among the biological components. The ecotoxicological measurement gives however a reliable response of organisms to sample exposure which takes into account interactions between (not always analysed) chemicals as well as their bioavailability in complex environments.
The vulnerability of aquatic ecosystems depends on several factors including physical and chemical properties that can have an important effect on the biological activity, but also on concentrations of toxicants entering the ecosystem, duration and type of exposure, properties of the ecosystem itself and its location in a potential toxic area. Given the functionalities required for the data management, modelling, overlay and presentation both in space and time, those factors lead to "think GIS" for this multidisciplinary approach.
The monitoring of the surface water toxicity in Wallonia has been an evolving exercise depending mainly on the technical developments of biotests (Van der Wielen & Halleux, 1998). Different strategies may be used according to the scale of studies. At a small scale, Microtox acute toxicity testing with the luminescent bacteria Vibrio fischeri is performed, which an interesting cost-effective and reliable alternative to conventional normalised tests and allows to point out "black points" (Van der Wielen, 1995). At a larger scale (river section, small basin), a battery of tests including algae, crustacean and fishes is used in order to cover a wider spectrum of ecological functions through organisms of different sensitivities.
Statistical analysis and mathematical modelling are used to quantify the biological effects in terms of effect concentration, to validate the results, to point out unexpected values and to analyse distributions. Procedures are picked up from external software (estimation of end-points) or developed within the database system (validation procedures, input helps, on-line warnings, and reference values). The final results are to be validated by the ecotoxicologist (application of quality criteria, comparison to reference values) and finally stored in the ArcInfo environmental database of ISSeP.

ECOTOXICITY AT A SMALL SCALE: THE MEUSE AND SCHELDT BASINS

Since 1994, about 1500 acute toxicity data have been recorded on 70 given sampling points; only 9% are toxic. A classical thematic mapping seems to be convenient for having a quick overview of the water quality at this scale. As the sampling points are fixed and the data density is too low for performing studies in space such as drawing toxicity profiles along rivers or evaluating detailed impact of industries, the valorisation of the results seems to be restricted to the study the intensity and frequency of the toxic pollution and the interpretation of the results according to chemistry of the water samples. The best data model seems to be of network type (line and point topology) with dedicated attribute tables even if the dynamic segmentation model with toxicity results given as events is an interesting one (Halleux, 1996).
The paper maps have to summarise the information according to well-defined criteria answering, one by one, the end-user requirements. Figure 2 shows an example of the maximum toxicity registered in 1997 coupled with the frequency of toxic pollution.
Dynamic edition of results can be performed through applications developed using ArcInfo, ArcView or MapObjects. It appears however that when the end-user questions or achievements are not precisely defined or when data are to be used in a research or a step-by-step context, the more flexible and useful editing way is the ArcView toolbox. The software functionalities such as joining ecotoxicity tables with chemical results, drawing results against time or chemicals, performing statistics and showing different thematic maps by opening several windows, ... are of particular interest.
The projects can be defined using ArcExplorer or ArcView. ArcExplorer is available on Esri URL without charge and is widely used (more than one million downloads until now); ArcView is preferred when additional features are to be built or when data processes are needed.

Figure 2
Map of surface waters toxicity (1997)

ECOTOXICITY AT A LARGE SCALE: THE WARCHE RIVER IN MALMEDY

In 1994, the Warche river showed toxic events that seemed to be inconsistent with the "good quality" of the river given by its chemical quality index (DGRNE & ISSeP, 1995) and macroinvertebrates biotic index (Vanden Bossche, 1994). A detailed study has been implemented in order to understand this paradox and to propose reliable measurement points. Three waste water discharges were identified as main sources of water pollution (2 are from paper factories, 1 from a municipal water treatment plant) and characterised using a battery of tests in order to identify causal sources, to estimate the risk and to choose the most appropriate bioassay(s) for effluent monitoring and further ambient monitoring.
In this stepwise approach, two toxicity surveys took place for the three effluents. The battery of tests used is representative of the different trophic levels (producer, decomposer, consumer) and includes acute (short-term) toxicity tests as well as chronic (long term) ones (Figure 3). Results of acute tests are expressed in EC50 with its confidence limits and stored in Toxicity Units.

Figure 3
Characteristics of the battery of tests

The bar chart of the acute toxicity (Figure 4), drawn from the point attribute tables, allow classifying the effluents: in decreasing order of toxicity, Paper Factory 1, Paper Factory 2, Municipal water treatment Plant. The toxicity of the last industry during Survey1 is still surprising but can be explained by the chromium remaining after treatment of effluents from a tannery closed before Survey2. The effluents of the paper factories are toxic to highly toxic.

Figure 4
Results of the battery of tests

The analysis of toxicity against chemistry allows to classify the biological systems into two classes: the first one includes only Daphnia magna, related to aluminium; all the other tests are grouped in the second one. The results indicate that Raphidocelis subcapitata is the most sensitive test in this case study. Brachyonus calyciflorus (Roxtoxkit) and Vibrio fischeri (Microtox) are also good indicators of the effluent toxicity. The chemistry of the effluents is characteristic of the industrial activity: richness in ammoniac nitrogen, phosphorus and boron for the municipal plant, high content in organic nitrogen and aluminium for the paper factories.
Simulations can be performed in order to estimate the safety factors for the acute exposure of organisms. Both effluent and river flows can be managed as impedance on river sections. As the effluent toxicity may vary widely according to the paper production, the exercise is based on the data of effluent toxicity taken from this profile study and the statistics of the river and effluent flows. Results (Figure 5) show high potential risks of river pollution whose extension has to be estimated through the analysis of the river profile.

Figure 5
Potential toxicity of the Warche river

Only the Microtox and Rotoxkit systems were used on 16 samples taken the same day along the river (Figure 6); the conventional algae test is not used because of the lab capacity limits.

Figure 6
Profile sampling of the Warche river

Using point events of Dynamic Segmentation allows drawing a cross-section of the river toxicity. The toxic plume is well described by the Microtox response. Crossing the ecotoxicological, physical and chemical data shows that water conductivity is a good indicator of the discharge in the river (Figure 7). The location of the chemical network sampling point (far from PF1, upstream PF2) has to be revised for providing a more reliable information about the water quality.

Figure 7
Toxicity of the Warche river

The toxicity information is be finally overlaid with thematic maps, like topography, land use, ecosystems or critical loads maps (Eloy, 1998), in order to integrate other useful parameters for estimating the impact of effluent toxicity on the vulnerability of aquatic systems. A more precise estimation of the risk requires however a more accurate description of the effluent variability and river flows.

DIFFUSION OF RESULTS

Until now, toxicity results are sent quarterly as numerical reports to the Authority. The annual report, formatted as HTML-files, covers the interpretation, the detailed studies and the new researches. Developments are performed in order to provide on-line results with comments and maps on the Web using MapObjects and ArcView IMS.

CONCLUSION

Ecotoxicology provides a reliable and efficient support for water pollution risk assessment at different scales: global monitoring of water quality, measurement and control of effluents toxicity, study of spatial spread-out of toxic plume, estimation of the impact on ecosystems. Within ArcInfo, a computer-aided decision support has been developed on the basis of GIS functionalities and ecotoxicologist expertise. Further improvements will integrate a more systematic use of this application.

REFERENCES

• DGRNE, ISSeP (1995). "Reseau de mesure de la qualite des eaux de surface en Wallonie - Annee 1994". 235p.
• Eloy S. (1998), "Modelling, mapping and managing critical loads for forest ecosystems using GIS: approach of Wallonia (Belgium) to the study of long-range transboundary air pollution effects on ecosystems in Europe", Proc. Ecosystem Vulnerability Symposium, Seattle, 8/98, SETAC, to be published.
• Halleux I. (1996), "What's the best data model for managing the quality of surface water?", 11th EUC, London, 10/96, Esri-UK.
• Jorgensen S.E. (1990), "Modelling in Ecotoxicology", Elsevier, New York.
• Vanden Bossche J.P. (1994). "Carte de la qualite biologique des eaux de surface. 1986-1994". MRW, DGRNE, Namur, Belgium.
• Van der Wielen C. (1995), "Toxicite aigue des eaux de surface en Region Wallonne par Photobacterium phosphoreum (MicrotoxÆ) - Annee 1994", DGRNE - ISSeP, 33p.
• Van der Wielen C., Halleux I. (1998), "Toxicity monitoring of the rivers Scheldt and Meuse in Wallonia, Belgium", Proc. Symposium on New Microbiotests, Brno, 6/98, SETAC, to be published.