home  publications                                                                                   Kontakt   Gästebuch  Links  Flohmarkt
 
 

Final covering of the Ronneburg uranium mining site

Uwe Hoepfner, Halle/Saale

Proceedings of the 11th International Conference on Environmental Remediation and Radioactive Waste Management ICEM2007
September 2-6, 2007, Oud Sint-Jan Hospital Conference Center, Bruges, Belgium    Poster pdf (0,6 MB)    Paper pdf (0,3 MB)
Abstract   
The rehabilitation of WISMUT’s former Ronneburg uranium mining site involves backfilling of waste rock to the Lichtenberg open pit. The relocation project comprises about 110 million m3 of sulphide-bearing and ARD-generating waste rock which makes it the most important and most cost-intensive single surface restoration project conducted by WISMUT at the Ronneburg site. The backfilled waste rock has to be covered on an area of about 220 ha to control water infiltration and gas diffusion. Design planning for the final cover placement which began in 2004 had to be based on a comprehensive cost-benefit analysis as well on field tests of alternative cover options which are in compliance with legal requirements. An intensive testing program concerning the vadose zone of soil covers has therefore been started in 2000. The paper presents an overview of the monitoring program and the results of the vadose zone measurements. The water and gas balances of soil covers have to be predicted for extended evaluation periods. Therefore water balance simulations of single layer covers (storage and evaporation concept) taking current and future soil and climate conditions into account are performed with the HYDRUS_2D code.

Introduction

During the „Cold War“ years until 1990, uranium production by WISMUT totalling some 231,000 tons of uranium had made the German Democratic Republic the world’s third largest uranium producer (behind the U.S. and Canada) and provided a large portion of uranium supply in the former COMECON economic area, along with the USSR and CSSR [1]. Production was suspended for economic reasons in late 1990; since then, the federally-owned WISMUT company is cleaning up the uranium post-mining legacies. The closeout and cleanup project in Saxony and Thuringia involving costs in the order of an estimated total of € 6.2 bn is one of the largest-scale environmental restoration world-wide [2, 3], the reclaimed mineland near Ronneburg a clearly visible symbol of the ongoing remedial effort [4]. Primary emphasis of the cleanup project is on a) flooding of the underground mines at the Aue, Königstein, and Ronneburg sites, b) reclamation of the waste rock piles at the Aue, Königstein, and Ronneburg sites, as well as the backfilling of a mined-out open pit mine near Ronneburg, and c) reclamation of the ponds containing uranium mill tailings near Seelingstädt and Crossen.

Generation of acid seepage combined with high contaminant levels is one of the environmental problems encountered at former mining sites. Such acid mine drainage is due to pyrite weathering which produces acid in the presence of oxygen and causes the leaching of heavy metals, salts, and radionuclides from waste rock or mill tailings [5–7]. Generation of acid mine drainage can be diminished by stimulating neutralisation reactions on the one hand, and by suppressing oxygen entry, on the other. A compacted and permanently humid soil cover constitutes an efficient means to reduce oxygen diffusion. For this reason, one of the approaches to remediation calls for the capping of such sites with earthen covers combining hydraulic performance and diffusion control (dry barrier concept, [7]). As a consequence thereof, surface capping systems have to be designed for large areas at former mining sites.

Periods of time in the range of 200 to 1,000 years considered by regulatory bodies in evaluating various cover options are very long and exceed by several orders of magnitude both the periods of observation available during the design or licensing phases and the periods for which there is substantiated knowledge of how the surface covers will perform. As a consequence, the use of prognoses methods and tools becomes a crucial element in the selection process, on the one hand, and, on the other, concepts with long-term performance predictions become highly appealing when selecting an option. To such situations may apply the concept of natural analogues as designed for nuclear waste repositories [8].

In this case, natural analogues for final cover designs are sites of natural vegetation with elevated evaporation levels to reduce seepage rates (ET-Cover, Storage and Release Concept, [9, 10]) as well as grounds with high subsoil compaction which would reduce gas diffusion by their high degree of water saturation. Prognoses of water and gas balances of final covers designed in that way have to take long-term aging of soil substrates (root penetration, soil fabric) as well as the long-term modification of meteorological boundary conditions into account.

The relocation of about 110 million m3 of sulphide-bearing waste rock into the Lichtenberg open pit mine is the most important and cost-intensive single surface restoration project at the former Ronneburg uranium mining site. Mine wastes from the Absetzerhalde and the Nordhalde dumps were placed into different zones of the open pit depending on the materials' acid generating potential. Under that scheme, waste material having the highest acid generation potential is to be placed in the deepest zone of the pit below the anticipated flood level within the anoxic zone. Material with low acid generating potential is to be placed in zone above the flood level (zone of reduced O2) while the near surface and therefore a relatively oxygen-rich zone (thickness approx. 10 m) is to be filled with acid consuming waste rock.

Following completion of the backfill that will be up to 60 m above the initial ground level, a dry cover will be placed on top of the backfilled mine wastes. The area to be covered amounts to about 220 ha. Design planning for the final cover placement which began in 2004 had to be based on a comprehensive cost-benefit analysis as well on field tests and modeling studies of alternative cover options which are in compliance with legal requirements. An intensive testing program concerning the vadose zone of soil covers had therefore been started in 2000.

Methology

A number of test plots, which represents different cover concepts like Sealing Layer, ET-Cover and Capillary Barrier were used to study the water and gas balances for their respective performance over a period of several years [10-11]. During the investigation period (2000–2005), parameters established from test cover results comprised flow rates, soil moisture and soil moisture tension as well as site meteorological conditions.

Four test plots were constructed during spring 2000 on an already contoured area of the Lichtenberg open pit backfill. The test plots have surfaces of 50 x 60 m each, the average slope inclination is 10 %. Three different soil covers were constructed using local soil material. The fourth test plot is a reference plot without any soil cover. The lowermost layer of all the cover profiles consists of 0,6 m of old cover material of the relocated Absetzerhalde and Nordhalde waste rock dumps (ZAN-material). It was separated during the relocation process and is planned to be reused for the final cover construction of the backfilledLichtenberg open pit. The vertical profiles of the plots are:
All the test plots were instrumented with lysimeters and nests of vadose zone instruments. To measure the percolation into the waste rock, lysimeters with an area of 50 m2 are located in the waste rock material 0.6 m below the cover. Surface runoff and interflow above the sealing layer, on top of the ZAN-layer and above the waste rock are measured for the entire field. The measuring devices are tipping buckets and flumes. Suspended sediment is collected in runoff traps. Field-testing of the vadose zone includes soil water content, soil suction and temperature. Water content is measured using time domain reflectometry probes. Soil suction is detected with tensiometers. To detect soil suction up to the wilting point, so-called equitensiometers are used. Pore water can be sampled with suction cups. A meteorological monitoring station is located in the testing area, to measure all relevant meteorological parameters. In addition, a number of natural sites and waste piles covered for many years with vegetation from natural seeding were analysed for root penetration and site hydrology (interception, stem flow).

Calibrated with the results obtained from field investigations, the HYDRUS_2D code [12] was used to establish the present-day prognosis of the cover systems’ performance from a water balance impact prospective. Based on water saturation in soil substrate, oxygen flux was calculated using a stationary diffusion approach. In the second modeling step, long-term water balance prognoses were developed based on the field validated modeling approaches. Pedotransfer functions were used to take soil development scenarios into account. Conditions for future climate change were simulated by downscaling the IPCC-A1 scenario for Eastern Thuringia. To calculate forest stand scenarios, the HYDRUS_2D code was extended to include a combined interception-transpiration model which allows to calculate transpiration in periods of wet leaf surfaces (by means of the GASH interception model) and in dry periods with the pine-parameterised PENMAN-MONTEITH approach (for details see [10]).

Results

Measurement results of water balance and of hypodermic flow (lateral drainage) in particular yield valuable insights as to single and double-layer cover hydraulics. Contrary to the working hypothesis – the assumption that significant lateral drainage would only occur within a cover comprising a sealing layer – tension measurement results showed at an early stage that perched water tables develop over long periods of time within single layer systems as well and may cause corresponding interflow down the test slopes. For instance, a surprisingly high interflow was observed in the 1 m thick single layer cover and on the waste pile surface, respectively (21 % of precipitation).

This flow is caused on the one hand by the deeper section of the recultivation layer, which was higher compacted, and by the waste rock material itself, on the other. Comparative tests to determine the permeability coefficient of the waste rock material reveal that very low permeabilities of 3–5 x 10-9 m s-1 have to be assumed with regard to this substrate which not only does have a strong sealing effect but also a considerably lower total pore volume. Interflow also occurs in the upper level of the recultivation layer where the emergence of a topsoil structure shifts the runoff-interflow ratio at the ground surface in favour of the interflow.

The HYDRUS_2D code was successfully calibrated and validated on the basis of data obtained from the test cover plots. This allowed simulation of the flow during each of the three-month calibration and validation periods which yielded results in good agreement with the measured discharge hydrograph, with cumulative deviations of < 10 % for total flow in all of the three covers investigated. Average absolute deviations of daily flows during the respective calibration and validation period  vary between < 0.01 and 0.5 mm d-1.  Measured and simulated interflow within a 1 m thick single layer cover is shown in Fig. 1 – which clearly demonstrates good simulation of the process dynamics of rapid interflow as well as good agreement with the cumulative height of discharge.

In all three test cover plots, the hydraulic permeability of the waste rock material as established by calibration falls within the highly impermeable range. Thus, the HYDRUS_2D calibration confirms and corroborates the results derived from field measurements on the hydraulic effect of the waste rock material acting as a sealing boundary layer underneath the soil cover. Permeability coefficients of the recultivation layers as established during the calibration, differ to a higher degree from field and laboratory results. Permeabilities by a factor of 2–200 higher than indicated by laboratory studies performed in the year 2000 had to be assumed, in particular for some decimetres at the top of the recultivation layer. Thus, the assumed fabric development of the topsoil are confirmed by the HYDRUS_2D calibration. Moreover, an anisotropic permeability distribution had to be assumed for the recultivation layer of all of the three test covers, with an increase by a factor of 1.5–100 for horizontal hydraulic permeabilities, in contrast to vertical permeabilities perpendicular to the slope. Also, somewhat steeper soil-water characteristic curves were established for the recultivation layer substrate.

The complete water balance during the investigation period as well as that of the long-term meteorological series (1970–1999) was calculated using the validated HYDRUS_2D model for single and double layer covers for the Lichtenberg test plots. Actual evapotranspiration in the long-term average is 532–539 mm a-1 or 72.0–72.9 % of precipitation, respectively, for all of the three covers, with an average annual precipitation rate of 739 mm a-1 and an average potential evaporation of 658 mm  a-1(according to PENMAN-MONTEITH, grass vegetation). As a consequence, total flows have similar levels at all of the three test plots and average around 196–250 mm a-1. Due to differences in the substrate properties and in layering, flow conditions are manifestly differentiated resulting in unequal surface runoff rates and in particular with regard to the ratio of hypodermic flow versus deep vertical infiltration. The ratio of permeability coefficients between the draining and sealing layers (i.e. sealing layer in the two-layer system and waste rock material in the single-layer system) acts as the controlling and driving parameter for deeper flows in the cover.

The processes of cover system aging, of vegetation evolution of a coniferous forest as well as of future climate change accounted for in the scenario simulations yield modified evapotranspiration conditions on the one hand and modified total flows and flow rates, on the other. The long-term prediction of the water balance is exemplified by conditions beneath a pine forest under anticipated future climatic conditions between the year 2036 and 2055 (Fig. 2). Changing hydrometeorological boundary conditions, along with an increasingly negative climatic water balance, result in declining flows and increasing soil dryness. Total flows will decline during the period 2016–2035 by a factor of 2 down to levels of 102–125 mm a-1 in comparison with average conditions during the 1970–1999 period, and continue to decline during 2036–2055 by a factor of 3 down to 73–84 mm a-1. Individual flow rates will shift in comparison with recent conditions as the drop in interflow will be disproportionally high.

On balance, this demonstrates the capabilities of a numerical two-dimensional simulation model in reproducing the complex hydraulic conditions of surface covers in good agreement with field test results as well as the feasibility of a secure and substantiated water balance prediction. Based on the prediction of hydraulic conditions within the cover and on saturation conditions in particular,   a well-founded prediction of oxygen diffusion is possible which, in turn, is the basis for predicting the generation of acid mine drainage at rehabilitated mining sites.

Conclusions

Based on the test plot and simulation results, a simple one-layer cover combined cover of cohesive soil material from on-site excavation overlain by a 0.4 m thick recultivation layer to restore natural soil functions for re-vegetation was derived and submitted for approval. Together with hydraulic measures, this approach is to meet any requirement in terms of radiology, water protection, stability, erosion protection, and reuse. Following weighing of cover-related short and long term costs it was found that higher construction costs for the various cover options will not be offset by cost savings in the long term water treatment [13]. As a consequence of pit flooding, even a costly multiple cover including an impermeable layer as it is represented at the sealing layer test plot will only insignificantly impact on concentrations and loads at ground water spill over points.

Despite the universal spread and use of standard modeling tools like HYDRUS, simulating the water balances of soils, both natural and man-made, still remains afflicted with numerous uncertainties. This makes model calibration and validation mandatory. Against the background of high sensitivities of the relevant input parameters, which are also time variables in the light of prediction periods in the order of centuries or even millennia, the quality of model predictions has to bee critical considered. Even the prediction of a validated model only amounts to "history matching" [14].

Consequently, with ever-increasing modeling complexity, growing importance is to be attached to field observations regarding the water balance, either as lysimeter studies or in other forms of measurements of soil moisture or flows in the field logging data from water balance processes under changing soil, climatic, and land use conditions for long periods of time.



Hydrus 2D calibration: Interflow

Figure 1:     Measured and simulated lateral interflow at the base of the recultivation layer, thin 1 m single layer cover (Lichtenberg test plot near Ronneburg/Thuringia)


Water balance evapotranspiration cover

Figure 2:     Predicted water balance, single layer cover. Climate change scenario, pine forest: Cumulative precipitation, cumulative potential and actual Evapotranspiration, cumulative flows (surface runnoff, lateral interflow, deep percolation into waste rock)
References
[1]  Lange, G.; Mühlstedt, P.; Freyhoff, G. & Schröder, B. [1991]: Der Uranerzbergbau in Thüringen und Sachsen – ein geologisch-bergmännischer Überblick. Erzmetall, 44, 162–171.
[2] Mager, D. [1996]: Das Sanierungsprojekt WISMUT. Geowissenschaften,  11,  443–447.
[3] Hagen, M.; Jakubick, A. & Paul, M. [2003]: Status and results of the Wismut environmental remendiation project. In: Jakubick, A.; Mager, D. & Metzler, D. (Ed.): UMREG 2003. Uranium Mine Remediation Exchange Group Meeting, Oxford, 24.9.2003, 1–12.
[4] Hoepfner, U. & Paul, M. [2003]: Die Berg­baufolgelandschaft bei Ronneburg. In: Liedtke, H.; Mäusbacher, R. & Schmidt, K.-H. (Ed.): Relief, Boden und Wasser. Nationalatlas Bundesrepublik Deutschland. Institut für Länderkunde, Leipzig, 52–53.
[5] Rogowski, A.; Pionke, H. & Broyan, J. [1977]: Modeling the impact of strip mining and reclamation processes on quality and quantity of water in mined areas: A review. Journal of Environmental Quality,  6,  237–244.
[6] Harries, J. [1987]: A review of acid drainage from waste rock dumps at mine sites (Australia and Scandinavia). Australian Nuclear Science and Technology Organization. Rep. ANSTO/E692.
[7] Tremblay, G. & Hogan, C. [2001]: MEND Manual. Vol. 4 – Prevention and control. Mine Environment Neutral Drainage (MEND) Program, 2001.
[8] Waugh, W.; Petersen, K.; Link, S.; Bjornstad, B. & Gee, G. [1994]: Natural analogs of the long-term performance of engineered cover. In: Gee, W. & Wing, N. (Ed.): In-situ remediation: Scientific basis for current and future technologie. Proc. thirty-third Hanford symp. on health and the environment, Pasco, 7-11.11.94, 379–409.
[9] Hoepfner, U. & Zepp, H. [2005]: Das Konzept der ET-Caps – Eine Alternative zur Oberflächenabdichtung? Mitt. Deutsche Bodenkundliche Gesellschaft, 107, 481–482.
[10] Hoepfner, U. [2006]: Das Speicher- und Verdunstungskonzept alternativer Oberflächenabdeckun­gen: Messung und Simulation des Wasser- und Gashaushaltes und sein Einsatz bei der Sanierung der Wismut-Altlasten. Dissertation, Bochum.
[11] Hoepfner, U.; Jahn, S.; Neubert, G. & Paul, M. [2003]: Lysimeteruntersuchungen zum Wasser- und Stofftransport bei der Haldensanierung des Uranerzbergbaus in Thüringen. In: BAL (Ed.): Bericht über die 10. Gumpensteiner Lysimetertagung. Bundesanstalt für alpenländische Landwirtschaft Gumpenstein, 29-30.4.2003. Irdning, 179–180.
[12] Šimůnek, J.; Šejna, M. & van Genuchten, M. [1999]: The HYDRUS-2D software package for simulating the two-dimensional movement of water, heat, and multiple solutes in variably-saturated media. U.S. Salinity Lab., Agric. Res. Service, Riverside, California.
[13] Paul, M.; Kahnt, R.; Eckart, M.; Jahn,  & Baacke, D. [2003]: Cover design of a backfilled open pit based on a systems approach for a uranium mining site In: Farell, T. & Taylor, G. (Ed.): Proc. Sixt Intern. Conf. Acid Rock Drainage (ICARD 2003), Cairns, North Queensland, Australia, 4-7.7.2003, 351–36.
[14] Konikow, L. & Bredehoeft, J. [1992]: Ground-water models cannot be validated. Advances in Water Resources,  15,  75–83.

    


Bluecounter Website StatisticsBluecounter Website Statistics                Made with Nvu