7.2 Chemical Reactions in Variable-density Flow Systems
Geochemical reactions and groundwater flow and transport processes are closely linked. On the one hand, flow and transport redistribute solutes across the subsurface, thereby affecting the delivery of reactants for, and removal of reaction products of chemical reactions. On the other hand, chemical reactions may alter the fluid composition in such a profound way, that its density changes (e.g., as shown by the deviation of some data points from the theoretical relationship in Figure 4 of Section 2), and thereby, the flow field changes. Mixing by dispersion and diffusion is a key driver for geochemical reactions. The greater contact area between fluids of different density during convective fingering enhances mixing and thus promotes reactions.
Some of the early classical hydrochemical studies in a variable-density groundwater setting focused on the dissolution of calcite in freshwater-saltwater mixing zones in coastal areas (Sanford and Konikow, 1989; Wigley and Plummer, 1976). Ore formation controlled by thermal convection has been studied in basin-scale geology (Person et al., 1996; Raffensperger and Garven, 1995). The mobilization of arsenic in groundwater storage and recovery in brackish aquifers has received attention (Wallis et al., 2011). Precipitation of minerals, such as carbonates, has been extensively studied in geothermal problems due to its relevance for aquifer thermal energy storage (Brons et al., 1991).
The equation of state as given by Equation 35 is valid for a single species. For a multispecies solution, density is no longer a function of the concentration of a single solute but now depends on the concentrations of all solutes present. In that case, the density can be calculated using a relationship of the type presented in Equation 38.
 |
(38) |
where:
| ρ0 | = | fluid density when Ci = Ci,0 (kg m−3) |
| NS | = | total number of solutes that contribute to the density |
 |
= | slope of the density-concentration relationship for species i |
| Ci | = | species concentration (kg m−3) |
| Ci,0 | = | reference concentration of species i (kg m−3) |
Models based on this type of relationship have been used previously in studies by Zhang and Schwartz (1995) and Mao and others (2006). It is also possible to compute fluid density using a thermodynamic framework (Monnin, 1994, 1989), which was implemented in the studies by Freedman and Ibaraki (2002), Hamann and others (2015), and Post and Prommer (2007).
To assess and quantify the potential effects of reactive transport in a variable-density flow problem, Post and Prommer (2007) converted the Elder problem into a reactive multispecies transport problem to examine the importance of density changes induced by cation exchange and calcite equilibrium reactions on free convection processes. They found that chemical reactions were important only at low density contrasts. This is because the change in density induced by chemical reactions was only significant for low density contrasts, and negligible for high density contrasts.
Bauer-Gottwein and others (2007) investigated how variable-density flow, evapotranspiration and geochemical reactions work together to control the hydrology, and solute transport processes underneath salt pans on the islands in the Okavango Delta. In these systems evapoconcentration triggers (i) mineral precipitation and (ii) convective, density-driven downward transport of the salt into deeper aquifer systems. Their modeling results showed that the time to onset of density driven flow can increase with mineral precipitation and carbon dioxide degassing. Hamann and others (2015) studied a playa system to gain insights into the feedback between density‐driven flow and the spatiotemporal patterns of precipitating evaporites and brine evolution. By comparing their results to nonreactive models, they found that the nonreactive simulations could overestimate the solute mass in the aquifer below the playa by up to 20 percent compared to the reactive multicomponent transport simulations. Both studies highlight the important role those chemical reactions may have on the flow field in systems where solute concentrations are so high that salts will precipitate from the solution.
Free convection may also be expected to exert a strong control on tracer concentrations that are used to infer groundwater ages. As an example, Figure 33 shows the results of a simulation of a semi-confined aquifer that was originally fresh and suddenly became inundated by seawater. An age stratification was present before the inundation, with groundwater age increasing with depth, as reflected by decreasing 14C concentration of the groundwater. Some of the original stratification can still be recognized after 1500 years but it is also apparent that free convection has completely overhauled the pre-existing tracer concentration patterns that were present in the aquifer. Young, high-salinity water has sunk to the bottom of the aquifer, and old, fresher water ascends between the descending salt plumes. In this process, mixing of water of various ages is taking place. Consequently, where convective flows are important, the interpretation of tracer concentration patterns will be infinitely more complex than in settings with forced convection. Moreover, the mixing of the fluids is likely to trigger hydrochemical reactions that could have their bearing on the tracers of interest as well. For 14C it is well known that the interaction between the groundwater and carbon sources in the aquifer will bias the inferred age. Therefore, failing to recognize the role of convective flow may lead to serious misinterpretation of age tracer data.
Figure 33 – Numerical model results of a freshwater aquifer after 1500 years since it became inundated by seawater. Top: Chloride distribution. Bottom: 14C concentration.
Free convection is also important in the storage of CO2 in underground reservoirs. The injected CO2 tends to be lighter than the resident groundwater and floats to the top of the target rock layer and accumulates below the confining unit that acts as a seal of the reservoir. The CO2 dissolves in the brine, creating a fluid that is one to two percent denser as a result, and therefore tends to sink (Hidalgo and Carrera, 2009). Because the CO2 lowers the pH of the groundwater, carbonate minerals like calcite and dolomite can dissolve, which has the potential to enhance porosity and permeability (Islam et al., 2016; Sainz-Garcia et al., 2017).
Finally, free convection also has relevance for the emission of greenhouse gases from peat layers. Rappoldt and others (2003) found that nocturnal cooling of a water-saturated peat moss layer resulted in downward flow of the relatively cold surface water, and upward flow of the warmer water from below. This results in a periodic flow and mixing regime that has consequences for CO2 and methane transport in the upper layers of water-saturated peat.