CONCLUSIONS A special system of multilevel monitoring wells, called hydro-geochemical cross-section, was constructed for the collection of data and for the detailed hydrogeochemical study. Wells were constructed for specific purposes, mainly for study of hydrogeochemical transport processes in the riverside zone of the Danube, changes in oxidation-reduction state of Danube water recharging the thick gravel aquifer, processes of oxidation of organic matter, etc. The system of wells consists of 11 multilevel hydro-gechemical wells. Individual wells were situated along the forecasted ground water flow corresponding to the situation after putting the hydropower structures into operation. The distance of individual wells from the Danube and reservoir and the depth of their filters is defined according to the geological profile and presupposed course of geochemical processes. Configuration of wells and their filters - horizontal and vertical distribution of filters - enables one to follow development of the ground water quality downwards the aquifer. A filter part of piezometer (observation well with short filter) is 0.5 m long which enables taking representative focused water samples even in the case of expressive vertical variability of chemical composition of ground water. Inside the 50 m deep bore(holes, 6 and in 70 m deep bore-holes 8 observation levels were built in, respectively.
Suggested method of two pumps (submersible and suction) overcome negative influence of stagnant water in a well already after short period of pumping. Pumping by suction pump, localised close under water level in a well ensures vertical up-flow of previously stagnant water in the casing and an inflow of fresh ground water via the well filter into the well, toward the sampling submersible pump and measuring electrodes. Sampling is carried out by submersible pump, which do not cause such marked changes in the composition of gases in water as suction pump does. Building up of hydro-geochemical cross-section at Kalinkovo enables to obtain ground water samples of high quality for detailed study of hydro-geochemical processes in the infiltrated Danube water. The hydro-geochemical cross-section is an addition to the up to the present time monitored wells chosen for the monitoring of the impact of water engineering structures Gabcikovo. The advantage of the described method of sampling is evident mainly at the elder large diameter monitoring wells, which are typical for the area of the Danube Lowland.
Water Research Institute (VUVH) samples ground water from multilevel observation wells localised in the riverside zone of the Danube.
Two municipal waterworks companies, namely Bratislava Waterworks and Canalisation (VaK) Western Slovakia Waterworks and Canalisation (ZsVaK) perform regular watching of ground water quality in the wells used for municipal drinking water supply. At monitoring wells localised around the important waterworks well fields ground water quality and water level regime is monitored. Exploited wells are usually of large diameter. They are installed with long filter and equipped with built-in permanent pumps. Wells are pumped continually at higher discharge (e.g. 50 l/s). On one hand, this causes water mixing between different depths of aquifer, on the other hand it enables one to characterise ground water quality from broader area of the production wells. Sanitary protection of the wells and the methods used for sampling are unified. We regard the data obtained from the wells used for municipal water supply to be the most suitable and reliable for the interpretation of ground water quality, especially in regional scale.
Placement of the hydrogeochemical profile near Kalinkovo was chosen on the basis of 3D model of ground water flow [8, 10]. It is characteristic for the flood(plain zone of the Danube upstream Kalinkovo that water from the Danube recharges ground water of Zitny ostrov island and at the same time, both, river bed and inundation area remain rather narrow even after putting the Hydropower structures Gabcikovo into the operation.
The hydro(geochemical profile consists of 11 multi-level hydrogeological wells, which serve for ground water sampling. Individual wells are situated along the general course of the ground water flow. The distance of individual wells from the Danube and the depth of their filter parts is defined according to the geological conditions and presupposed course of geochemical processes in the infiltrating ground water. Horizontal and vertical distribution of filters enables one to follow development of the ground water quality along the flow lines in various depths. (see Fig. 2, 3).
Covering loam layer reaches thickness in the range 1-2 m. Aquifer consists of highly permeable and poorly graded sandy gravel with irregular lenses of sands. The permeability of gravely sediments is characterised by the permeability coefficient the value of which is regularly larger then 1 x 10-2 m/s. The thickness of sandy-gravel aquifer increases with the distance from the Danube. The underlying aquitard is composed of less permeable fine- to medium-grained sands, sandstone and clays of Neogene age (Fig. 3).
The boreholes were drilled in the period from October 1992 to February 1993 using rotary-percussion drilling rig with introductory diameter of 630 and 920 mm. Foremost 8 boreholes of the profile were drilled to the depth of 50 m which enables to characterise virtually the whole cross-section of sandy-gravel aquifer. The others 3 wells were drilled to the depth of 70 m. They did not reach the Neogene basement.
During the drilling works the drilling mud was not used, not to influence the future water samples by long-term ion-exchange with the clay in the drilling mud [1].
Along with the drilling, the samples of sediments for grain-size analyses were taken from each meter of the depth and from each lithological change. From the obtained grain(size curves, the conductivity coefficients were calculated, using method of Carman-Kozeny [in 9]. In this manner particular types and permeability of sediments were identified in more details. These data were utilised as input parameters in the model of ground water flow.
The soil samples for chemical analyses were taken at individual filter levels. Samples were analysed for content of carbonates, grain(size composition, COD(Mn), COD(Cr), BOD(after 1, 5, 10, 28 days), TOC, content of organic carbon "glucose-extractable", content of humic substances, total content of Fe and Mn and the content of Fe and Mn in the form of oxides [11]. The samples were transported to chemical laboratory within the day of sampling.
Definite well casing and screen is made from polyethylene pipes with inner diameter of 90 mm. Material of pipes is even at 15-20% perforation sufficiently resistant to ground pressure in the depth of 70 m. Inner diameter of pipes was designed to comply the used method of sampling which requires to drop both the submersion pump and the measurement equipment onto the filter level.
The filter and casing are made from polyethylene which corresponds to the trends of geochemical studies. Iron and manganese are central parameters of the studies of oxidation-reduction processes in the riverside zone of the Danube. Steel building material was not used mainly due to its low resistance to chemical and electrochemical corrosion. In the riparian zone of the Danube the factors of corrosive conditions are the following: dissolved oxygen at a concentration greater than 2 mg/l, high velocity of ground water flow and the content of total dissolved solids in the ground water. Mainly the corrosion on the metallic casing surface, but also bi-metallic corrosion at casing joints, which can cause development of holes in the casing must be taken into consideration. In the beginning, leaks are expressed only as leakage of water, followed by inflow of sand into the well, and finally by reduction of well space or total destruction of well. The product of a corrosion are iron and manganese, when galvanised steel is used, also Zn and Cd. At oxidising conditions these metals are found in the form of hydrous metal oxides covering walls of casing, or they accumulate in the well bottom. Under reducting conditions high level of dissolved Fe and Mn can be expected [6]. Moreover, the surfaces on which corrosion occurs also present very active adsorption surfaces for the trace organic and inorganic substances and potential sites for a variety of chemical reactions, such as formation of organic-metallic complexes are [5, 6].
Polyethylene well-casing material with low content of plasticizers is resistant to the electro-chemical corrosion and is both chemically and mechanically sufficiently resistant. In the given method of sampling, adsorption and leaching of some microcomponents (e.g. Sn and Sb from thermoplastic stabilisers utilised in thermoplastic production) are negligible even when organic microcontaminants are studied. Before installation, all filters and casings, used for monitoring wells, were washed with fresh water [3, 6].
A filter part of piezometer is 0.5 m long which enables taking representative depth specific water samples even in the case of major vertical variability of chemical composition of ground water. Inside the 50 m deep bore(holes, 6 observation levels were built in (filters with slotted perforation), 70 m deep bore(holes have 8 levels (the deepest filter have drilled perforation). Configuration of filters enables detail study of hydro-geochemical processes. Under a filter, a bottom cap of 0.5 m length is placed. Individual parts of casing are connected by threaded joints without glue so as to prevent contamination of the water samples.
As a filter pack and seal between filters the excavated and sieved sandy(gravel material were used. As a rule this material was placed into the same depth from which it was taken. For a gravel pack the coarse-grain fraction of sieved sandy-gravel material was used. The seal between individual screen parts was filled with fine-grained fraction of excavated material. This fraction is of relatively low permeability and at the same time it is immobile in the aquifer. There are neither negative cation-exchange processes nor increasing of pH in water samples, which is the case e.g. of bentonit seal [4].
The fact that the excavated material was used restricted possibility of additional contamination in the process of the gravel-pack production and/or transport to the site and it also eliminated input of foreign rock materials into the well. It is known that in the case of using industrially produced gravel pack a chemical analysis of filter pack is usually recommended [3].
After final installation all multiple piezometers were cleaned and developed by using gradually increased pumping until suggested maximal discharge (1 l/s) or maximal drawdown (usually < 1 m) was reached. During sampling, neither this maximal discharge, nor drawdown were passed over.
Proposed method of two pumps (submersible and suction) overcome negative influence of stagnant water in a well already after short period of pumping. Pumping by suction pump, localised close (1 - 3 m) under water level in a well ensures vertical up-flow of stagnant water in the casing and inflow of fresh water into the well, to the sampling submersible pump and measuring electrodes. Sampling is carried out by submersible pump, which do not cause such marked changes in the composition of gases in water as suction pump does. The method is simple, effective, and applicable wherever inner diameter of a casing enables to drop a submersible pump into the filter (Fig. 4.).
The volume of pre-pumped water by suction pump is determined so as to ensure water refresh in the filter and its gravel pack. After pumping out of necessary volume of water by suction pump, the additive submersible sampling pump GRUNDFOS MP-1, placed in the filter depth, is given to parallel operation. The yield of the submersible pump should be less then that of the suction pump (approx. 20 - 50 %). In comparison with standard methods of sampling the period of pre-pumping is considerably shortened. The extent of pumping has to ensure only washing of narrow Teflon up-flow pipes (when up-flow pipe is 50 m long, has inner diameter of 2 cm, and pumping capacity is 0.1 l/s, it takes 10 minutes to flush it once). Simultaneously it can be stated that if the samples are taken directly from the well filter, any interaction of water sample with perforated polyethylene material of the screen is excluded.
On the basis of the data on discharge and drawdown obtained during development of piezometers, optimal capacity for both a submersible and a suction pump was defined for each individual object. During sampling constant discharge of pumps is kept within all the period of pumping. Exact records on the operation of pumps, discharge and water levels help not only at interpretation of chemical analyses, but also in the identification of changes in the technical state of piezometers.
A measurement of electrical conductivity, temperature, pH, dissolved oxygen, redox potential is performed in situ during the pre-pumping of object by suction pump. The measurement electrodes are localised on the upper edge of filter. Alkalinity is determined in the field immediately after taking out a sample. Water samples are filtered under pressure through the 0.4 m filter and transported to the laboratory at the same day.
Our experience confirm the necessity of field in situ measurement of at least water temperature, redox potential, content of dissolved oxygen directly in the well at the filter depth. Important is also the use of water sampling by submersible pumps instead of using suction pumps.
