Aquifers modelling

Analytical models using classical reservoir engineering techniques such as material balance, aquifer modelling and displacement calculations can be used in combination with field and laboratory data to estimate recovery factors for specific situations. These methods are most applicable when there is limited data, time and resources, and would be sufficient for most exploration and early appraisal decisions. However, when the development planning stage is reached, it is becoming common practice to build a reservoir simulation model, which allows more sensitivities to be considered in a shorter time frame. The typical sorts of questions addressed by reservoir simulations are listed in Section 8.5. 

Obs. Walls = Well 1 Aquifer Model Leaky Solution... 

Fig. 21.4. Transport of benzene within an aerobic aquifer, modeled in two dimensions. Contaminated water containing 1 mg kg-1 benzene leaks into the aquifer over the course of two years, at the point indicated. As in the previous model (Fig. 21.3), the benzene is retarded by sorption to organic matter in the aquifer and attenuates due to sorption, biodegradation, and dispersive mixing. Plots were rendered using the matlab software.

Not all phases are present or were probably present in the aquifer. Model solutions that require phases unlikely to have ever been present in the aquifer are not realistic. Careful petrographic observation and mineralogy studies are the most direct evidence for mass transfer reactions, but are not commonly employed in modeling studies. 

WiNTERHALDER, J. R. Hann, J. H. 1991. Wanaea/ Cossack dynamic aquifer model multidisciplinary approach. In Asia-Pacific Conference, Proceedings. Society of Petroleum Engineers, 615-628. 

McNab, W.W.J., and T.N. Narasimhan. 1995. Reactive transport of petroleum hydrocarbon constituents in shallow aquifer Modeling geochemical interactions between organic and inorganic species. Water Resour. Res. 31 2027-2033. 

Reservoir pressure is measured in selected wells using either permanent or nonpermanent bottom hole pressure gauges or wireline tools in new wells (RFT, MDT, see Section 5.3.5) to determine the profile of the pressure depletion in the reservoir. The pressures indicate the continuity of the reservoir, and the connectivity of sand layers and are used in material balance calculations and in the reservoir simulation model to confirm the volume of the fluids in the reservoir and the natural influx of water from the aquifer. The following example shows an RFT pressure plot from a development well in a field which has been producing for some time. 

Detailed sampling can include, but is not limited to, the installation of monitoring well networks. After the wells have been installed, aquifer tests are typically performed. Once the aquifer tests are performed and the aquifer characteristics are determined, time series sampling for a given contaminant, or a surrogate, is undertaken. The combined results of these efforts provide the basis for development of a treatment strategy. Modeling can be used as part of this effort to help determine the best technical and most cost-effective techniques to be used at a site. 

The modeling of a groundwater chemical pollution problem may be one-, two-, or tlu-cc-dimcnsional. The proper approach is dependent on the problem context. For c.xamplc, tlie vertical migration of a chemical from a surface source to the water table is generally treated as a one-dimensional problem. Within an aquifer, this type of analysis may be valid if the chemical nipidly penetrates the aquifer so that concentrations are uniform vertically and laterally. This is likely to be the case when the vertical and latcrtil dimensions of the aquifer arc small relative to the longitudinal scale of the problem or when the source fully penetrates the aquifer and forms a strip source. 

Groundwater pollution problems, however, are modeled using a two dimensional analysis. A typical aquifer has area dimensions that are much hirger than the vertical dimension. Therefore, chemictils dissolved in the groundwater aclticvc ertical uniformity a short distance from the source and the chemical plume will move in the lateral and longitudinal directions. 

The solubility of methyl parathion is not sufficient to pose a problem in runoff water as determined by an empirical model of Wauchope and Leonard (1980). Some recent monitoring data, however, indicate that methyl parathion has been detected in surface waters (Senseman et al. 1997). In a study to determine the residue levels of pesticides in shallow groundwater of the United States, water samples from 1,012 wells and 22 springs were analyzed for methyl parathion. No methyl parathion was detected in any of the water samples (Kolpin et al. 1998). In a study of water from near-surface aquifers in the Midwest, no methyl parathion was detected in any of the water samples from 94 wells that were analyzed for pesticide levels (Kolpin et al. 1995). Leaching to groundwater does not appear to be a significant fate process. 

Metwally M E-S, NL Wolfe (1990) Hydrolysis of chlorostilbene oxide. II. Modelling of hydrolysis in aquifer samples and in sediment-water systems. Environ Toxicol Chem 9 963-973. 

Note that the models aim to define general descriptive bulk parameters for an aquifer. Values for bulk adsorption or weathering will reflect the weighted range for a variety of phases present in complex, natural systems, and provide an overall measure of the behavior of the aquifer. In contrast, individual phases can be examined in the laboratory to determine the specific processes involved. 

Modeling of the transport of the long-lived nuclides, especially U, require knowledge of the input at the water table as a boundary condition for aquifer profiles. There are few studies of the characteristics of radionuclides in vadose zone waters or at the water table. Significant inputs are likely to occur to the aquifer due to elevated rates of weathering in soils, and this is likely to be dependent upon climatic parameters and has varied with time. Soils may also be a source of colloids and so provide an important control on colloidal transport near recharge regions. 

Conventionally, the sample is initially saturated with one fluid phase, perhaps including the other phase at the irreducible saturation. The second fluid phase is injected at a constant flow rate. The pressure drop and cumulative production are measured. A relatively high flow velocity is used to try to negate capillary pressure effects, so as to simplify the associated estimation problem. However, as relative permeability functions depend on capillary number, these functions should be determined under the conditions characteristic of reservoir or aquifer conditions [33]. Under these conditions, capillary pressure effects are important, and should be included within the mathematical model of the experiment used to obtain property estimates. 

Based on the results a conceptual model is created and the hydraulic properties of the aquifer and its surrounding layers are derived from the pumping test. 

The groundwater model simulation indicates that the shallow aquifer could be restored to 5 pg/L (MCL) in 25 to 40 yr with soil remediation. Without soil remediation, between 60 and 100 yr would be required. 

Groundwater models and other analytic techniques are available to assist in proper pump siting, choosing pump capacities, and calculating the movement of the contaminant plume. The characteristics of the aquifer, the flow of groundwater, and the size of the plume should be known. 

In most instances, hydrodynamic dispersion is not great enough to require detailed consideration in hydrogeologic modeling for fate assessment of deep-well-injected wastes. Flowever, regional variations (such as the presence of an USDW in the same aquifer as the injection zone, as is the case in parts of Florida) should be evaluated before a decision is made to exclude it. 

If osmotic effects are possible, several other effects would need to be considered in a geochemical-fate assessment, depending on whether the solute concentration is increased or decreased. If solute concentrations are increased, pressures associated with injection would increase beyond those predicted without osmotic effects. Also, the movement of ions to the injection zone from the aquifer with lower salinity (above the clay confining layer) would increase the salinity above those levels predicted by simple mixing of the reservoir fluid and the injected wastes. This action could affect the results of any geochemical modeling. 

---