Croatian geological survay1
The groundwater pumping sites Zapadno polje and Luke, where groundwater from Orljava alluvial aquifer is abstracted, are the main sources of public water supply of PoÂžega and surrounding settlements during dry seasons. Zapadno polje has been operating since middle 1960s, while Luke was established in late 1990s. Both pumping sites are situated near the Orljava river - Luke on the left and Zapadno polje on the right bank of the river, approximately 750 m downstream from Luke. Groundwater quality was never an issue at Zapadno polje and, on the contrary, demanganisation device has been operating at Luke since the establishment of the pumping site because of the high content of natural manganese in groundwater. However, in 2009 high manganese concentration was detected for the first time at the catchment area of the pumping site Zapadno polje as well. Hence, comprehensive hydrogeological research were carried out in order to prevent spreading of manganese towards abstraction wells where the groundwater quality still complies with the provisions of the Regulation on health and safety of drinking water.
METHODS AND MATERIALS
3-D numerical groundwater flow model was developed in order to test hydrogeological conceptual model, evaluate flow net, and groundwater flow velocities and different scenarios which would help preventing spreading of manganese across the catchment area of the pumping site. MODFLOW code (Harbaugh et al, 2000) was applied which is based on the finite difference method. For the purpose of isolating the catchment areas of individual wells, particle tracking model was developed using MODPATH (Pollock, 1994). All necessary data for model developing and subsequent calibration was acquired and interpreted, including data on lithological features of the aquifer and covering aquitard across the model domain, boundary conditions (precipitations, stages of Orljava river, abstraction quantities at wells on daily basis) and monitoring of groundwater levels at catchment area of the pumping site. Transient simulation of groundwater flow for the period of 6 months, devided in stress periods was performed for the calibration purposed. Once the model was calibrated, it was used to evaluate different approaches to prevent further groundwater quality deterioration. In the first approach a grout curtain construction, which would isolate the area where high concentration of manganese had been detected and thus prevent manganese spreading, was evaluated, whereas in the second approach the effects of different groundwater abstraction rates at individual wells were tested.
RESULTS AND DISCUSSION
During calibration process hydraulic conductivity of the aquifer was modified in order to obtain best possible fit between observed and simulated groundwater heads. Except the conductance values of the layer dividing river boundary and model flow domain, boundary conditions were not changed at all. The best calibration result was obtained for the range of hydraulic conductivities of aquifer between 20 and 430 m/day. NRMS error was below 4,5% and for the most stress periods was below 3%, which entirely fulfilled calibration targets. Furthermore, mass balance error was 0,01% for most stress periods. Simulation of different scenarios of groundwater source management showed that manganese movement towards abstraction wells with good groundwater quality could be prevented with grout curtain construction and at least minimised with adjusted groundwater abstraction regime. In the case of physical barrier, the contaminated area is completely isolated, but, at the same time, geometry of the pumping site catchment area and boundary conditions is changed leading to reduced quantity of bank filtration from the Orljava river and, consequently, decreased groundwater levels and reduced pumping site abstraction rates. On the other hand, with adjusted groundwater abstraction regime hydraulic barrier was established. In this case, groundwater from contaminated catchment area ends up in 3 abstraction wells with already poor groundwater quality. Groundwater abstracted by these 3 wells should be delivered to drinking water supply system only after the treatment at existing demanganisation devise at nearby Luke pumping site. As a consequence of hydraulic barrier, the catchment areas of remaining wells are changed, occupying zones with, based on current knowledge, satisfactory groundwater quality. However, it has to be emphasised that, because of many different factors influencing behaviour of manganese in aquifers, the possibility of presence of high manganese content in groundwater even in newly formed catchment areas cannot be completely excluded.
Numerical modelling of groundwater flow and particle tracking in saturated zone of the aquifer were once again proved as convenient tools for testing and evaluation of different scenarios for solving specific groundwater management problems. Both evaluated scenarios, aiming at preventing the spread of manganese towards abstraction wells, have advantages and shortcomings. In the case of physical barrier, contaminated zone of the catchment area is completely isolated but, consequently, groundwater replenishment from river bank filtration is reduced. In this regard, adjusted abstraction regime at pumping site appears to be better solution, but then groundwater abstracted by wells which form hydraulic barrier must be treated at demanganisation devise before it is delivered to public water supply system. Harbaugh, A.W., Banta, E.R., Hill, M.C., McDonald, M.G. (2000) MODFLOW-2000, The US Geological Survey modular ground-water model - user guide to modularization concepts and the ground-water flow process. US Geological Survey Open-File Report 2000-92, 121 pp.
Pollock, D.W. (1994) User's guide for MODPATH/MODPATH-PLOT, Version 3; a particle tracking post-processing package for MODFLOW, the US Geological Survey finite-difference ground-water flow model. US Geological Survey Open-File Report 94-46, 248 pp.