MAGPlan Clean groundwater for Stuttgart

Final Report WP1 (short version) Stefan Spitzberg BoSS Consult GmbH 10.02.2012

MAGPlan Clean groundwater for Stuttgart

Content 1 

Motivation ................................................................................................................................2 

2 

Executed work steps ..............................................................................................................3 

2.1  2.2  2.2.1  2.2.2  2.2.3  2.2.4  2.2.5  2.3  2.3.1  2.3.2 

Installation of groundwater monitoring wells .............................................................................3  Hydrogeological model..............................................................................................................6  Geological sequence and hydrostratigraphy.............................................................................6  Gypsum leaching and tectonics ................................................................................................9  Groundwater flow ....................................................................................................................11  Hydraulic permeabilities ..........................................................................................................13  Additional field investigations ..................................................................................................15  Numerical flow model ..............................................................................................................19  Development of the model ......................................................................................................19  Calibration of flow....................................................................................................................22 

3 

Conclusion .............................................................................................................................26 

Tables Table 1: Faults with hydraulic blockage effect (no flow boundary)........................................................ 10  Table 2: Variety of hydraulic conductivities ........................................................................................... 13  Table 3: Hydrogeological units and assignment to model layers .......................................................... 20  Table 4: Measured and calculated spring discharges ........................................................................... 25  Table 5: Calculated water balance of Muschelkalk ............................................................................... 25 

Figures Fig. 1: Project area MAGPlan (blue line), Stuttgart city area (red line) ................................................... 3  Fig. 2: Stuttgart city map with wells MAG 1 bis MAG 6 ........................................................................... 4  Fig. 3: Drilling core of MAG 2 showing the transition zone from Gypsum to Lower Keuper at 56,4 m below ground level................................................................................................................................... 5  Fig. 4: Drilling equipment at drilling point MAG 1 .................................................................................... 6  Fig. 5: Geological subsurface map of the MAGPlan area (light green-framed) ...................................... 7  Fig. 6: Hydrogeological structuring of aquifers ........................................................................................ 8  Fig. 7: Schematic diagram of gypsum leaching with different rock overlying of Gypsum Keuper .......... 9  Fig. 8: Gypsum leaching within Grundgipsschichten in the MAGPlan area (light blue-framed) ........... 10  Fig. 9: Fence-chart showing groundwater flow and genesis of spring water within project area .......... 11  Fig. 10: Groundwater flow in the surrounding of Karlshöhe within the Lower Keuper .......................... 12  Fig. 11: Distribution of hydraulic conductivities in Lower Keuper (section) ........................................... 14  Fig. 12: Histograms of Upper Muschelkalk (Trigonodusdolomit) .......................................................... 15  Fig. 13: Groundwater flow in the Upper Muschelkalk aquifer on 04.07.2011 ....................................... 16  Fig. 14: Cone of depression at the end of immission pumping test in P172 (data shown in cm) ......... 17  Fig. 15: Hydraulic analysis of P172 (high-speed pressure measurement) ........................................... 18  Fig. 16: Hydraulic analysis of a hydro-seismogram in monitoring well BK17.1/4.................................. 18  Fig. 17: Horizontal extension of model area in the different hydrogeological units ............................... 21  Fig. 18: Comparison between calculated and measured groundwater levels of Muschelkalk.............. 23  Fig. 19: Calculated distribution of piezometer heads [m+NN] within upper part of Muschelkalk .......... 24 1

MAGPlan Clean groundwater for Stuttgart

1

Motivation

The Department for Environmental Protection of the City of Stuttgart, in cooperation with the State Institution for Environment, Measurements and Nature Conservation in Karlsruhe (LUBW) apply the MAGPlan “Management plan to prevent threats from point sources on the good chemical status of groundwater in urban areas“. The MAGPlan project is funded by the European Commission under the program LIFE+ 2008, running from 2010 to 2014. The project was implemented due to the long-term and wide-spread use of volatile chlorinated hydrocarbons (CHC) in the inner city of Stuttgart which resulted in significant contamination of several Keuper aquifers. Furthermore, the underlying karst aquifer (Upper Muschelkalk), being the source of the Bad Cannstatt and Berg mineral springs with a capacity of 500 l/s, is also affected. The MAGPlan project comprises an area of 26 km² of inner city of Stuttgart, which is situated in a valley directly upstream from the mineral springs of Stuttgart (see fig. 1). Against this background, the focus is placed on investigations of the hydraulic characteristics of affected aquifers in the Nesenbach Valley. The investigations include descriptions of the hydrostratigraphy and aquifer geometry, groundwater flow, hydraulic conductivity and storage properties as well as vertical interactions between various groundwater horizons. Findings are presented as a hydrogeological concept model. As the next step, a numerical groundwater flow model will be developed which, based on the hydrogeological model, enables a precise quantitative description of the groundwater flow and of the groundwater budget within the project area. The numerical flow model provides the basis for a following transport model, which in turn, is expected to detect migration paths in the valley of Stuttgart and identify main input sources. Data collection and modeling for this work package (WP1) has already been carried out, which included drilling for groundwater monitoring wells, the hydrogeological model and the numerical groundwater flow model, processing status dated from 31.01.2012. Those institutions involved in the process are:  management drilling works:

Klinger and Partner GmbH

 hydrogeological model:

Department of Environmental Protection for the City of Stuttgart & BoSS Consult GmbH

 groundwater flow model:

Prof. Kobus and Partner GmbH

2

MAGPlan Clean groundwater for Stuttgart

general groundwater flow direction Fig. 1: Project area MAGPlan (blue line), Stuttgart city area (red line). Reference: Amt für Umweltschutz, Stuttgart

2

Executed work steps

2.1

Installation of groundwater monitoring wells

The first step, involved installing six groundwater monitoring wells within the MAGPlan project area during the period May to July 2011. The locations of the drilling positions MAG 1 to MAG 6 are shown in fig. 2. The drilling points were specifically placed in zones with lack of knowledge about the level of groundwater contamination. To that end, four of six wells were placed in the southern city and the surrounding Karlshöhe area (see fig. 2).

3

MAGPlan Clean groundwater for Stuttgart

Fig. 2: Stuttgart city map with wells MAG 1 bis MAG 6 Reference: Amt für Umweltschutz, Stuttgart

All six drilling points tap the groundwater of the Lingula-Dolomite strata, belonging to the upper Lower Keuper. In previous groundwater investigations in the Stuttgart city area, the Lingula-Dolomite aquifer has been shown to be a preferred spreading path for contaminations within the Gypsum and Lower Keuper system. Figure 3 shows the drilling core from MAG 2 exemplarily to illustrate the transition zone from Gypsum to Lower Keuper. Due to the scarcely depleted Gypsum layer in this core, an effective barrier against the inflow of contaminants from the Gypsum layer into the Lower Keuper is built. The lower picture shows the drilling equipment at drilling point MAG 1 located beside the SWR (broadcast station in Stuttgart). Already during the drilling process, short-term pumping tests and depths-defined samplings were carried out in higher located aquifers.

4

MAGPlan Clean groundwater for Stuttgart

Fig. 3: Drilling core of MAG 2 showing the transition zone from Gypsum to Lower Keuper at 56,4 m below ground level. Reference: Klinger und Partner GmbH

After the installation of groundwater monitoring wells, four-day aquifer-tests were run for each well in order to determine hydraulic properties and to sample the groundwater. It was noted, that the water capacities varied strongly, minimal 0,025 l/s (MAG 4) to maximal 2,3 l/s (MAG 1), within the Lingula-Dolomite aquifer. However, this is, attributed to the different thicknesses of overlying rock (72 m at MAG 4 and 28 m at MAG 1) resulting in varying degrees of weathering. 5

MAGPlan Clean groundwater for Stuttgart

Fig. 4: Drilling equipment at drilling point MAG 1, Reference: Klinger und Partner GmbH

2.2

Hydrogeological model

2.2.1 Geological sequence and hydrostratigraphy The geology of the MagPlan project area is mostly characterized by the Keuper sequence. Under the subsurface, Keuper strata from Rhaetian to Sandstone and Gypsum Keuper to Lower Keuper occur under a cover of quaternary rock (see fig. 5 geological subsurface map). Whereas younger strata occurs predominantly along the plateaus and the valley slopes, older strata (Grundgipsschichten, Lower Keuper) were incised only in the valley basin due to the erosive effect of the streams Nesenbach and Neckar. The bedrock of the valley basin is covered with fluviatile gravel and sand sediments, whereas the bedrock on valley slopes is covered with residual loess and cover sediments called “Fließerden”.

6

MAGPlan Clean groundwater for Stuttgart

Fig. 5: Geological subsurface map of the MAGPlan area (light green-framed) Reference: Amt für Umweltschutz, Stuttgart

The vertical structuring within the project area is oriented to the hydrostratigraphic circumstances (see fig 6 hydrogeological structuring of aquifer formations). Figure 6 shows relevant aquifers framed in red. The first aquifer is composed of quaternary unconsolidated sediments, which are common in the Nesenbach and Neckar Valley (porous aquifers, e.g. river gravel of river Neckar). Underneath in the bedrock, there are seven more aquifers, which are considered as fractured aquifers assigned to Gypsum Keuper (Mittlerer Gipshorizont, Dunkelrote Mergel, Bochinger Horizont, Grundgipsschichten, Grenzdolomit), Lower Keuper and Upper Muschelkalk. A part of the fractured aquifers also possess karst characteristics which appear if for example Upper Muschelkalk or Grundgipsschichten are affected by intense gypsum leaching. 7

MAGPlan Clean groundwater for Stuttgart

Im MAGPlan-Projektgebiet relevante Grundwasserstockwerke Fig. 6: Hydrogeological structuring of aquifers, Reference: Amt für Umweltschutz, Stuttgart 8

MAGPlan Clean groundwater for Stuttgart

2.2.2 Gypsum leaching and tectonics An important role for the aquifer formation and the horizontal structuring within the model area plays gypsum leaching and tectonic structures. In figure 7 the general genesis of different aquifer types in Gypsum Keuper is presented as a function of the degree of leaching.

Fig. 7: Schematic diagram of gypsum leaching with different rock overlying of Gypsum Keuper Reference: Amt für Umweltschutz, Stuttgart

According to figure 7, non-leached zones can be found under thick overlying rock at valley slopes. Generally, these zones are not water-bearing. Whereas, the valley basins are completly leached and moderately water-bearing. Between both zones, there is a narrow band within the Gypsum Keuper which is partly leached and often acts as an abundant aquifer. Figure 8 shows very well the distribution of different facies types of Grundgipsschichten. Also, the aquifer geometry of the remaining Gypsum Keuper aquifers and of Grenzdolomit aquifers are considerably determined by the process of gypsum leaching. However, faults within deep aquifers of Lower Keuper or Upper Muschelkalk can have substantial hydraulic influence (see table 1 and chapter 2.2.3).

9

MAGPlan Clean groundwater for Stuttgart

Fig. 8: Gypsum leaching within Grundgipsschichten in the MAGPlan area (light blue-framed) Reference: Amt für Umweltschutz, Stuttgart Table 1: Faults with hydraulic blockage effect (no flow boundary)

mo: Upper Muschelkalk in total and Upper Dolomite of Middle Muschelkalk, moδ: Trigonodusdolomit, ku: Lower Keuper, km1: Gypsum Keuper 10

MAGPlan Clean groundwater for Stuttgart

2.2.3 Groundwater flow The eight mentioned groundwater reservoirs can be summarised into two aquifer units:  Upper aquifer system: Quaternary, Mittlerer Gipshorizont, Dunkelrote Mergel, Bochinger Horizont, Grundgipsschichten, Grenzdolomit  Lower aquifer system: Lower Keuper, Upper Muschelkalk The first system is mainly structured by the relief-depending process of gypsum leeching and so is only hydraulically effective in the valley of Stuttgart and the basin of Cannstatt. In the quaternary sediments, the groundwater flow is restricted to the Neckar Valley and narrower ranges of the Nesenbach Valley. The groundwater flow of the upper aquifer system is orientated with the topography and in the Nesenbach Valley is generally directed towards north-east (see fig. 1).

Fig. 9: Fence-chart showing groundwater flow and genesis of spring water within project area Reference: Amt für Umweltschutz, Stuttgart 11

MAGPlan Clean groundwater for Stuttgart

Principally, the surface influence decreases with increasing depth. Underneath the Grundgipsschichten and especially underneath the Grüne Mergel-formation (overlain by Grenzdolomit) the regional-acting lower aquifer system becomes more important. From the Lower Keuper and especially from the Upper Muschelkalk the spacious groundwater inflow changed, compared to the Quaternary groundwater inflow, coming from west and south. As figure 9 illustrates, the majority of the mineral water originates from a southwesterly karst aquifer inflow within the Upper Muschelkalk, flowing along the Nesenbach Valley. The detailed map at figure 10 shows the groundwater flow within the Lower Keuper around the Karlshöhe area. Often, the hydraulic system is complex, presented here by taking the example of the Karlshöhe-fault, which has a separating effect on hydraulics. Moreover, potential jumps of about 10 m appear around the faults, which often serve as evidence for high heterogeneity and high permeability contrasts. The tongue-shaped trend of the 240 m-isohypse could trace a zone parallel to the valley of higher transmissivity, whereas the 250 m-isohypse around Karlshöhe is caused by higher thickness of overlying rock and thus having a lower permeability. It should be noted, that the construction of water table contours in deeper aquifers becomes more unreliable due to generally less deep drillings and lower density of geological data.

Fig. 10: Groundwater flow in the surrounding of Karlshöhe within the Lower Keuper Reference: Amt für Umweltschutz, Stuttgart 12

MAGPlan Clean groundwater for Stuttgart

2.2.4 Hydraulic permeabilities In the first step, all existing analysis of pumping tests and hydraulic tests from the project area were compiled, verified and if necessary re-evaluated. Alltogether, more than 1600 pieces of experimental data were obtained. About half of the data was taken from the city project Stuttgart-21. Table 2 gives a summary of the aquiferdepending data density and the variety of hydraulic conductivities. Table 2: Variety of hydraulic conductivities

Commonly, there are considerably more data available for the four upper aquifers compared to the lower aquifers. The most incomplete data record exists for Grundgipsschichten, because it is rarely tapped separately. Most often, the basal Grundgipsschichten are combined with Grenzdolomit within one screen section and are, in such a case, hydraulically assigned to Grenzdolomit. The permeabilities vary considerably, partly even more than about ten orders of magnitude. The permeability is dependant on the position of the measuring point: Usually, high values can be measured in the valley with minor rock overlay, whereas the lowest permeabilities occur in great depth and only slightly in non-leached zones. In the second step, the spatial distribution of permeabilities was mapped for every single aquifer within the project area. Additionally, for aquifers of Lower Keuper and Muschelkalk, being supra-regional important, overview plans were made. With regard to the later following groundwater model, the Muschelkalk was divided into three sublayers:  Trigonodusdolomit  Nodosus- and Trochitenschichten above Haßmersheimer Schichten  Trochitenschichten below Haßmersheimer Schichten, and Upper Dolomite of Middle Muschelkalk 13

MAGPlan Clean groundwater for Stuttgart

This division enables the classification of permeability contrasts within the Muschelkalk. Compared to the underlying limestone, higher permeabilities usually occur in Trigonodusdolomit. The Haßmersheimer Schichten-formation forms a sealing layer in the lower part of the Upper Muschelkalk. The sulfate rock of Middle Muschelkalk represents the basis of the observed aquifer system. Figure 11 shows a section of a permeability distribution map of Lower Keuper.

Durchlässigkeit [m/s]

Fig. 11: Distribution of hydraulic conductivities in Lower Keuper (section) Reference: BoSS Consult, Stuttgart

Usually, the highest conductivities within Gypsum and Lower Keuper can be found along the axis of the streams, Nesenbach and Neckar. At high-lying locations, especially in non-leached sulfate rock, low permeabilities of less than 1E-06 m/s prevail. Remarkably, permeability zones on the north-west edge of the Nesenbach Valley often move tongue-like back and forward in west-east direction. It can be seen as a reflection of hydraulic influences of local fault zones. The hydraulic permeabilities of the Upper Muschelkalk, determined from the pumping well data, are often considerably lower than the permeabilities received from the surrounding reactions in observation wells. This karst-characteristic phenomenon is 14

MAGPlan Clean groundwater for Stuttgart

also demonstrated in both histograms for Trigonodusdolomit in figure 12. Therefore the mean permeability, which was calculated of values from pumping wells, representing the local conditions around the well, was found to be with about 4E-04 m/s approx. 6 times lower compared to the regional mean of approx. 2E-03 m/s, calculated from values of the monitoring wells.

Fig. 12: Histograms of Upper Muschelkalk (Trigonodusdolomit) Reference: BoSS Consult, Stuttgart

The regional permeabilities, determined from the monitoring wells, were taken as a basis for the maps of the spatial distribution of permeabilities in the Upper Muschelkalk, because the values of the monitoring wells represent the karst aquifer better for the here used scale of several kilometer. 2.2.5 Additional field investigations In addition to the evaluation of existing data, new field investigations of Upper Muschelkalk were implemented. The Upper Muschelkalk plays an important role as a supply and storage reservoir for the mineral springs of Stuttgart. The key to understanding the input and transport of contaminants to the mineral springs is the knowledge of the hydraulic interaction between the higher groundwater horizons; Gypsum and Lower Keuper and the Upper Muschelkalk. Unfortunalely, because of depth and karst formations, there are only sparse data for the Upper Muschelkalk which can be used for the transfer to a broader area. Therefore, it was appropriate to implement specific investigations of the Upper Muschelkalk at regional and local scale to enlarge the current database and so improve understanding of the processes systems at work.

15

MAGPlan Clean groundwater for Stuttgart

In July 2011 there was a two-week immission pumping test carried out, exemplarily for the monitoring well P172. Water was pumped at a rate of 5 l/s to examine the regional structure of the contaminant plume, the hydraulic influence of a directly adjacent fault zone as well as possibly existing anisotropies and hydraulic boundaries. Figure 13 shows the local groundwater flow directions before the immission pumping test was carried out. The fault zone crossing the Nesenbach Valley causes a step change in piezometer head of about 2 m and thus, works as a hydraulic barrier. However, the contour lines of piezometer head converge in the inflow of P172, which seem to be situated in a permeable corridor. Figure 14 illustrates the cone of depression at the end of the immission pumping test, which confirms the general barrier effect of the fault zone. South of the fault zone, all observation wells reacted spontaneously within a range of decimeter, to the north they reacted not at all or greatly delayed within a range of centimeter.

Fig. 13: Groundwater flow in the Upper Muschelkalk aquifer on 04.07.2011 Reference: BoSS Consult, Stuttgart

In Figure 13 the catchment area of P172 to the end of the immission pumping test is shown (green-framed). A capture width of about 190 m was achieved. The level of volatile chlorinated hydrocarbons concentration (CHC) stayed constant during the test between 10 - 12 µg/l, which is a sign of a relatively broad and homogenous CHC-plume. The parallel measured amounts of sulphur hexafluoride (SF6) showed 16

MAGPlan Clean groundwater for Stuttgart

only a small age shift within the pumped groundwater of initially 25 years mean residence time to 28 years at the end of the test.

Fig. 14: Cone of depression at the end of immission pumping test in P172 (data shown in cm) Reference: BoSS Consult, Stuttgart

Besides the immission pumping test in the monitoring well P172, some unconventional and innovative measuring techniques were applied. One of them was a special high-speed pressure transducer, able to measure 100 values per second. This technique was carried out to detect the earliest flow regimes within a karst aquifer as well as anthropogenic-induced pressure signals in order for more accurate hydraulic analysis. Figure 15 shows the comparison of the hydraulic analysis of the undercritical damped oscillation response after the immission pumping test (left) and the oscillation of confined groundwater table, induced by compression of pore space due to the load of running subways (right) for P172. Another method employed is the long-term-use of a data logger with continuous highfrequency recording of water level (constantly 1 sec). This measuring technique is applied at the monitoring well BK17.1/4 for 2 years to detect uncontrolled natural signals (e.g. earthquakes). Figure 16 shows the hydraulic analysis of the earthquake from 23.10.2011 in Turkey. For this purpose, a sequence in the recorded hydroseismogram was chosen, where the enforced oscillation moved to a free damped oscillation of the pressure table (see rectangle in figure 16). Both approaches have shown to be promising methods. One advantage is, that the local surrounding of the investigated wells can be hydraulically characterized without 17

MAGPlan Clean groundwater for Stuttgart

pumping and thus without the occasional need to clean contaminated pumped water. In further course, a methodological verification of the introduced methods has to follow.

K = 1,8E-03 m/s

Fig. 15: Hydraulic analysis of P172 (high-speed pressure measurement) Reference: BoSS Consult, Stuttgart

K = 6,5E-03 m/s

Fig. 16: Hydraulic analysis of a hydro-seismogram in monitoring well BK17.1/4 Reference: BoSS Consult, Stuttgart

18

MAGPlan Clean groundwater for Stuttgart

2.3

Numerical flow model

The groundwater flow model provides the basis to simulate the transport of volatile chlorinated hydrocarbons within the MAGPlan project area. The groundwater flow model allows calculating the flow conditions in three dimensions. Thus, it provides a consistent interpretation of the groundwater flow situation for a following transport simulation. The development of the model was carried out according to the hydrogeological system model (cf. chapter 2.2). It contains all relevant information for the geological-hydrogeological structure and for the flow conditions which are proven by numerous groundwater observation wells and partly by long time series of groundwater levels and spring discharges. Subsequent to the buildup of the model, the groundwater model has undergone a calibration process whereby variation of hydraulic conductivities and leakage coefficients the best possible consistency between measured and calculated piezometer heads and spring discharges was achieved. 2.3.1 Development of the model Vertical structuring and buildup The groundwater model involves relevant hydrogeological units of Gypsum Keuper, Lower Keuper and Upper Muschelkalk as well as the Quaternary aquifer of valley sediments in the Nesenbach Valley of Stuttgart. Considering the low conductive aquitards, table 3 gives a detailed overview about aquifers of relevant hydrogeological units. The numerical concept demands a layer-based model structure why consequently different hydrological units were assigned with at least one model layer. The Quaternary was divided into two model layers to enable a separation between the higher permeable debris layer at the basis and the overlying, usually lower permeable parts (cover sediments, alluvial loam). Likewise, the comparatively thick Grundgipsschichten were divided into two model layers. The reason for that is firstly, to regard the mainly vertical acting contaminant transport in parts of complete gypsum leaching and secondly, to enable a vertical differentiation of partly leached sections. The unit of Upper Muschelkalk was divided into four model layers to separate the hydraulic properties of Trigonodusdolomit from the other layers of Upper Muschelkalk with Haßmersheimer Layers. Layers of Gypsum Keuper are only hydraulically active in sections with gypsum leaching. That is why only these ones were implemented in the model, whereas non-leached sections remain inactive in the model.

19

MAGPlan Clean groundwater for Stuttgart

Table 3: Hydrogeological units and assignment to model layers Hydrogeological unit Quartär Mittlerer Gipshorizont with Bleiglanzbank Dunkelrote Mergel Bochinger Horizont Grundgipsschichten Grenzdolomit Grüne Mergel Middle part of Lower Keuper Estherienschichten Trigonodusdolomit Lower part of Upper Muschelkalk until Haßmersheimer Schichten Haßmersheimer Schichten Upper Muschelkalk below Haßmersheimer Schichten, and Upper Dolomite of Middle Muschelkalk

stratigraphy Q

hydraulic function GWL

thickness [m] variable

Model layer

km1

GWL

37

3 and 4

km1 km1 km1 ku ku ku ku mo

GWL GWL GWL/GWG GWL GWG GWL GWG GWL

16 5 variable 2 4 10 6 10

5 and 6 7 8 and 9 10 11 12 13 14

mo

GWL

60

15

mo

GWG

8

16

mo/mm

GWL

20

17

1 and 2

Q: Quartär, km1: Gypsum Keuper, ku: Lower Keuper, mo/mm: Upper/Middle Muschelkalk, GWL: aquifer, GWG: aquitard

The local position of model layers (horizontal layering) is orientated to the layer boundary between Gypsum and Lower Keuper as well as to the thickness of Grundgipsschichten which depends on the leaching degree. Because of the complete gypsum leaching in the central Nesenbach Valley, the Grundgipsschichten are transfered into an aquifer with low conductivity, which represents an important separation function. Currently, gypsum leaching takes place in valley slopes, which is why higher permeabilities are present and the layer thicknesses increase until the non-leached area. Remaining hard rock aquifers were assumed to have constant layer thicknesses. There are variable thicknesses of individual hydrological units below the Quaternary basis, which are, due to erosion, missing in certain areas of central Nesenbach Valley. Horizontal discretisation For the simulation of groundwater flow, the finite-difference-program MODFLOW, developed by United States Geological Survey, is used. The model area is subdivided horizontally in a row- and column-orientated model grid. For the central part of the Nesenbach Valley, a model discretisation of 10 m x 10 m is applied to be able to simulate the transport of the CHC-plume. For the marginal areas, model cells with 50 m cell lengths are applied. In total, the discretisation results in 6.450.764 active model cells within the model area.

20

MAGPlan Clean groundwater for Stuttgart

Model boundaries and boundary conditions Within the horizontal direction, the model was defined according to natural boundary conditions. The Muschelkalk aquifer together with Lower Keuper aquifer show, in accordance with figure 17, the greatest extension. They cover the Nesenbach Valley, the western main inflow, the southern inflow along the Neckar Valley and the spring rising zone in Bad Cannstatt and Berg. The model area is bounded in the south-west by the Fildergraben main fault, which is assumed to be impermeable. As a result, the groundwater inflow into the model area from the supra-regional Muschelkalk aquifer is subdivided into two components: first, a western inflow coming from recharge areas, and second, a southern inflow from Swabian Alb foreland.

Fig. 17: Horizontal extension of model area in the different hydrogeological units Reference: Kobus und Partner, Stuttgart

Lower Keuper and Muschelkalk aquifer have the same extension at the south and at the south-west. At the north-west, the Lower Keuper is bounded by a watershed. The model layers of Gypsum Keuper are orientated to existing gypsum-leached zones. 21

MAGPlan Clean groundwater for Stuttgart

Completely unleached areas are located outside of the model area. The Quaternary aquifer of the Nesenbach and Neckar Valley is represented according to its extension, whereas in the Nesenbach Valley only some quaternary trenchs are water-bearing. The rising mineral springs, originating from Muschelkalk, flow out in the Quaternary aquifer too, as far as they are not tapped by wells. Moreover, known groundwater extractions within the city area were considered, and the river Neckar was simulated as a leakage boundary condition along the Neckar Valley. The local groundwater recharge is derived from a hydrologic model, which supplies the boundary inflows from valley slops in the Nesenbach Valley too. These flow into the layers of Gypsum Keuper close by the leaching front. The artesian confined mineral springs were implemented in the model as so-called drainageboundary condition. The diffuse discharges of mineral water into river Neckar were considered as local anomalies with rises directly out of the Muschelkalk. Furthermore, local enhanced vertical permeabilities of Lower Keuper and Gypsum Keuper result in the rise of mineral water into the Quaternary of the Neckar Valley. 2.3.2 Calibration of flow In the context of the development of the hydrogeological model, all available pumping tests of respectively hydrogeological units were evaluated and differentiated maps of horizontal permeabilities within aquifers evolved. These permeability distributions were transferred to the model and varied within the known spectrum from the hydrogeological model. Thus, the horizontal permeability structure has been further refined. Because the hydrogeological units of the rock aquifer structure consist of different geological sublayers, which cannot be implemented in detail in the groundwater model, the vertical permeability of the hydrogeological units in the groundwater model is considered to be independent from the horizontal permeability. For that reason, the vertical permeability is an important factor in the flow calibration process. Within the model area, numerous fault zones exist, which especially play an important role for the vertical exchange between aquifers. In certain areas, the hydraulic effect of faults can be identified by groundwater levels (see fig. 13). In other areas and with consideration of findings of table 1, a hydraulic characterization concerning the vertical connection or the hydraulic separation in horizontal or vertical direction was made within the framework of the model calibration. Due to a possible crucial impact of vertical aquifer interactions to the transport of water ingredients, they have to be verified in further transport model of CHC and if necessary adjusted. The calibration of flow was performed for steady-state-conditions according to the measured piezometer heads of the reference campaign from 28./29.10.2010. At this time piezometer heads exceeded a bit the long-term value for mean water level. This finding was especially taken into account for the definition of constant heads at the model boundary of the Muschelkalk aquifer, which has been deduced from adjacent measuring points. 22

MAGPlan Clean groundwater for Stuttgart

Piezometer heads The information about groundwater levels of 864 observation wells were used as benchmark. By variation of horizontal and vertical conductivities, the best possible adjustment of calculated and measured groundwater levels was achieved. The comparison of calculated and measured piezometer heads for the Muschelkalk aquifer is shown in figure 18.

260

255

250

berechnet [m]

245

240

235

230

225

220 220

225

230

235

240 gemessen [m]

245

250

255

260

Fig. 18: Comparison between calculated and measured groundwater levels of Muschelkalk Reference: Kobus und Partner, Stuttgart

Piezometer heads within the upper part of the Muschelkalk aquifer, calculated with the model, are shown in figure 19. There is a divided inflow from western and southern direction, which joins together in the Nesenbach Valley and runs valleyparallel to the spa- and mineral springs.

23

MAGPlan Clean groundwater for Stuttgart

Fig. 19: Calculated distribution of piezometer heads [m+NN] within upper part of Muschelkalk Reference: Kobus und Partner, Stuttgart

Spring discharges Apart from groundwater levels, which are most often taken from reference campaigns at certain dates, the measured discharges of spa- and mineral springs were compared with the calculated discharge amounts during the calibration process. The discharge of mineral springs is determined by the difference of hydraulic potential between groundwater level and drainage height as well as by the leakage factor of the springs. The leakage factor was achieved during calibration. Calculated and measured spring discharges are compiled in table 4.

24

MAGPlan Clean groundwater for Stuttgart

Table 4: Measured and calculated spring discharges name Auquelle Brunnen im Maurischen Garten Leuzequelle Inselquelle Berger Westquelle Berg, Südquelle 2 Berg, Urquelle Berg, Mittelquelle Berg, Nordquelle Berg, Ostquelle Summe Berger Quellen Wilhelmsbrunnen 1 Mombachquelle Kellerbrunnen neu

date 28.10.2009 28.10.2009 28.10.2009 28.10.2009 28.10.2009 28.10.2009 28.10.2009 29.10.2009 28.10.2009 29.10.2009 29.10.2009 11.11.2009 seit 2010 seit 2011

measured discharge [l/s] 22,48 14,03 34,97 37,38 4,18 1,50 27,30 9,30 6,67 4,55 53,50* 8,17 41,00 7,30

calculated discharge [l/s] 23,80 14,36 35,40 38,69 4,85 1,59 28,17 10,38 7,61 4,96 57,57 8,81 41,38 7,29

* discrete measured value

Water balance of Muschelkalk aquifer With help of the model, the water balance of Muschelkalk aquifer can be compiled, which is shown in table 5. The main water inflow of the Muschelkalk aquifer into the model area, about 750 l/s, is coming from the western model boundary. The main outflow components of Muschelkalk aquifer are the spa- and mineral springs, and the rise into the river Neckar via the known anomalies. Table 5: Calculated water balance of Muschelkalk Component of balance

inflow [l/s]

outflow [l/s]

Western inflow Southern inflow („Mittel- und Hochscholle“) groundwater recharge of Muschelkalk and total exchange with the overburden outflow Muschelkalk mineral springs river Neckar with anomalies

750 146

13

438 228 309

Total

975

975

66

25

MAGPlan Clean groundwater for Stuttgart

3

Conclusion

With the aim of the hydraulic characterization of the aquifer system of the Nesenbach Valley in Stuttgart, both, classical methods (data evaluation, drillings, pumping tests, reference measurements at certain dates) and innovative methods (high-speed pressure measurements, anthropogenic tracers) were applied. The findings are documented in a hydrogeological model, which is ongoing updated, and will be incorporated into a numerical groundwater flow model. It links all information within time and space with the aim to simulate the hydraulic interactions between involved groundwater horizons, which helps to understand contamination migration paths and relevant transport mechanisms. The high degree of complexity of the regarded aquifer system (fractured and karst aquifer properties, hydraulically relevant tectonics), the size of the project area and the fact, that for the deepest and most important aquifer the fewest dataset is available, imply a challenge for the modeling. The transfer of local findings on regional level is especially subjected to uncertainties in the karst aquifer of the Upper Muschelkalk. Also, precise small-scale data of the contaminant source and contaminant migrating are needed. Hence, it is important that hydrogeological model, numerical flow model and transport model remain in a mutual dynamic optimization process to identify and adjust discrepancies and deficits. Overall, the wealth of available data required a considerable effort for their acquisition and processing. The graphic presentation of geological and hydrogeological information of the project area enabled, for the first time, a closed spatial picture of the whole inner city of Stuttgart. Indeed, additional implemented drillings and pumping tests closed local lacks of knowledge, but also raised new issues. This concern, for example, the relevance and effect of small-scaled hydraulic heterogeneities on groundwater flow and contaminant transport. Thus, the fractured and karst characteristics of the regarded aquifer system in the Nesenbach Valley set natural limits for the investigation of the process understanding. It is important to apply integral methods, which representatively describe a greater volume. Against this background, for instance, a combined tracer test in the system “Lower Keuper - Upper Muschelkalk” would be desirable to understand the flow and contaminant migrations paths at key points in the project area.

26

BoSS Consult GmbH Stefan Spitzberg Lotterbergstr. 16 70499 Stuttgart Telefon +49 711 601 606 811 Fax +49 711 601 606 810 Email [email protected] www.bosscon.de

Die Autoren tragen die alleinige Verantwortung für den Inhalt dieser Veröffentlichung, die nicht die Meinung der LIFE-Programmbehörde wiedergibt. Die LIFE-Programmbehörde haftet nicht für die etwaige Benutzung der in der Veröffentlichung gemachten Angaben. The paper in hand reflects the author‘s views and the LIFE Programme is not liable for any use that may be made of the information contained therein.

Mit Unterstützung des Finanzierungsinstruments LIFE der Europäischen Gemeinschaft