Decision-support modelling
The purpose of this study is to approach the issue of spreading of a benzene plume in an unconfined anaerobic sulphate-rich sand aquifer. In order to do this, we perform both analytical and synthetic modelling.
By identifying the main influencing factors, how they interrelate, i.e. the effects that they have on each other, and how they affect the system, it is possible to achieve a better understanding of the system as a whole and approach the issue we would like to address.In this case, five priority parameters have been identify:
P1: Sulphate concentration in groundwater.
P2: Groundwater flow variations (direction and velocity).
P3: Water table fluctuations.
P4: Rate of benzene biodegradation.
P5: Benzene depletion from NAPL.
From the analytical modelling step, we obtained the results plotted in Fig. 1. The cause-effect plot provides a clearer visualization of the parameters interrelationship. The variable interaction intensity is greater as one moves up along the diagonal; thus, if a parameter is plotted within the lower left corner it means that it is both less responsive (low total impact of others) and influential (low total impact on others). Fig. 2 shows the strenght of the interaction between the factors.
Fig. 1. Cause-effect plot. WT: water table. GW: groundwater: SO4 cc.: sulphate concentration in groundwater. Both Depletion NAPL and Biodegradation refer to benzene. The average point corresponds to (5.8, 5.8).
Fig. 2. Influence diagram. The thickness of the arrows depicts the strength of the interaction between the parameters.
Performing a synthetical modelling, the parameters influence is assessed in relation to the issue we are trying to address. In order to do so, we perform a Multi-Criteria Evaluation (MCE). The values of the variables are standardized by means of a utility function (built upon empirical data mainly) for each of the parameters. The results of the MCE are display in Fig. 3.
The efficiency of both models on addressing the issue of spreading of a benzene plume depends largely on how they were built, i.e. on each of the steps that were followed and the quality and quantity of data input. At the same time, the synthetic model value is largely determined on the analytical model, as is based on it.
From Fig. 1 is clearly visible that the variable water table fluctuations have the highest total impact (and does not response to any other parameter) while benzene biodegradation proves to be the most responsive one. However, these results have to been considered carefully and revised when possible. Though every model has its limitations, these can be significantly reduced if, for instance, a multi-disciplinary group of experts works on it and there is a fair amount of empirical data against which to test the model results.
Total Utility for scenario 1 is as twice (actually, a bit over the double) as larger as the one for scenario 2 ; i.e. scenario's 1 conditions would imply a plume size the double than the conditions for scenario 2.
The applications and limitations of both the analytical and synthetic models depend largely on a variety of aspects. In this particular case, since we are dealing with the spreading of an aromatic organic compound in an aquifer, the complexity of the system itself accounts for a vast part of the uncertainties.
The simplifications and assumptions made, the lack of enough knowledge and/or experience on the subject and the scarcity of data impact on the modelling results. In spite of this, both models aid in getting a better understanding of the system as a whole and of the interrelation of its main factors and provide a scale of importance of these as to the spreading of the contamination. The MCE in particular is a great tool since it enables the evaluation of diverse scenarios and makes it easier to visualize the risks.
References
Appleyard, S. (2003): Groundwater Quality in the Perth Region. (4), 103-113.
Davis, G.B., Barber, C., Power, T.R., Thierrin, J., Patterson, J.M., Rayner, J.L. and Wu, Q. (1999): The variability and intrinsic remediation of a BTEX plume in anaerobic sulphate-rich groundwater. , 265-290.
Stevens, R. (2014): Environmental Geology Course Compendium. Department of Earth Sciences, University of Gothenburg.
Yang, Y.J., Spencer, R.D., Mersmann M.A and Gates T.D. (1995): Ground-Water Contaminant Plume Differentiation and Source Determination Using BTEX Concentration Ratios. (6), 927-935.
Fig. 3. Numerical results for the MCE of the two scenarios compared. uS1: utility for scenario 1; uS2: utility for scenario 2.
2-