Tab. 1 Mole % CO2 and 13C values of CO2 at baseline and at the final monitoring event for the eight observation wells

During CO2 storage operations in mature oilfields or saline aquifers it is desirable to trace the movement of injected CO2 for verification and safety purposes. We demonstrate the successful use of carbon isotope abundance ratios for tracing the movement of CO2 injected at the Cardium CO2 Storage Monitoring project in Alberta between 2005 and 2007. Injected CO2 had a d13C value of -4.6±1.1 per mil that was more than 10 per mil higher than the carbon isotope ratios of casing gas CO2 prior to CO2 injection with average d13C values ranging from -15.9 to -23.5 per mil. After commencement of CO2 injection, d13C values of casing gas CO2 increased in all observation wells towards those of the injected CO2 consistent with a two-source end-member mixing model. At four wells located in a NE-SW trend with respect to the injection wells, breakthrough of injected CO2 was registered chemically (>50 mol % CO2) and isotopically 1-6 months after commencement of CO2 injection resulting in cumulative CO2 fluxes exceeding 100000 m**3 during the observation period. At four other wells, casing gas CO2 contents remained below 5 mol % resulting in low cumulative CO2 fluxes (<2000 m**3) throughout the entire observation period, but carbon isotope ratios indicated contributions between <30 and 80% of injected CO2. Therefore, we conclude that monitoring the movement of CO2 in the injection reservoir with geochemical and isotopic techniques is an effective approach to determine plume expansion and to identify potential preferential flow paths provided that the isotopic composition of injected CO2 is constant and distinct from that of baseline CO2.

(Table 5) Experimental hydrodynamic parameters for different soil types obtained in the Lüchow-Dannenberg district, Lower Saxony

In three typical sandy soils of Northern Germany the mobility of radioactive fission products of technetium, iodine, ruthenium and zirconium have been investigated in dependence of the hydrodynamic and physico-chemical soil properties. The laboratory experiments, which simulated fall-out events, used soil columns (1 m length, 30 cm diameter) taken as undisturbed as possible. By measurements of the breakthrough curves in the percolate and of the depth distribution of radionuclides in the soil columns after 6 months the average transport velocity could be determined. These values could be compared with the average water velocity measured by 3H tagging. Three qualitative mobility relations were observed: Ranker: Tc > Ru > I > Zr; Podsol: Tc > Ru > I > Zr; Brown forest soil: Tc = Ru > I > Zr. Relations between some physico-chemical soil properties and the retardation of radionuclides due to adsorption could be observed (eg. retardation of iodine and technetium by organic substances). The average retardation factors of the radionuclides and the hydrodynamic soil parameters are used in a model which gives a quantitative assessment of the hazard of groundwater contamination by a fall-out event in areas covered with comparable soils.

Results of Closed-Flow Column Experiments (zip-archive 2.3 MB)

The identification of transport parameters by inverse modeling often suffers from equifinality or parameter correlation when models are fitted to observations of the solute breakthrough in column outflow experiments. This parameters uncertainty can be approached by the application of multiple experimental designs such as column experiments in open-flow mode and the recently proposed closed-flow mode. Latter are characterized by the recirculation of the column effluent into the solution supply vessel that feeds the inflow. Depending on the experimental conditions, the solute concentration in the solution supply vessel and the effluent follows a damped sinusoidal oscillation. As a result, the closed-flow experiment provides additional observables in the breakthrough curve. The evaluation of these emergent features allows intrinsic control over boundary conditions and impacts the uncertainty of parameters in inverse modeling. We present a comprehensive sensitivity analysis to illustrate the potential application of closed-flow experiments. We show that the sensitivity with respect to the apparent dispersion can be controlled by the experimenter leading to a decrease in parameter uncertainty as compared to classical experiments by an order of magnitude for optimal settings. With these finding we are also able to reduce the equifinality found for situations, where rate-limited interactions impede a proper determination of the apparent dispersion and rate coefficients. Furthermore, we show the expected breakthrough curve for equilibrium and kinetic sorption, the latter showing strong similarities to the behavior found for completely mixed batch reactor experiments. This renders the closed-flow mode a useful complementary approach to classical column experiments.