Principal Investigator

At a Glance

In earlier work, the Celia group studied CO2 injection into depleted shale-gas systems and concluded that it was not feasible for most situations. That modeling work has now been extended to study the fate of fracking fluids in shale-gas systems. Modeling results indicate that the large amount of fracking fluids left underground is unlikely to pose any significant environmental risk.


Research Highlight

Earlier Modeling Work: CO2 Sequestration in Shales

In earlier work on shale gas systems, the Celia group developed a model for CO2 injection into depleted shales and analyzed the feasibility of large-scale CO2 sequestration in shales. The model involved two-component (methane and CO2) single-phase (gas) flow in the shale rock, and included competitive sorption as well as pressure-dependent nonlinearities. While being a reduced-order model, the results obtained matched gas production data very well, which gave confidence that the two-component model was reasonable.

The overall finding was that large-scale injection of CO2 is not feasible because of the excessive number of wells required to inject significant quantities of CO2. For example, injection of emissions from just the four largest coal-fired power plants in southwest Pennsylvania requires many thousands of wells in the Marcellus Formation. Rather than inject captured CO2 into depleted shales, it makes more sense to build a larger pipeline to transport the CO2 to a more suitable injection location like the Illinois Basin.

New Modeling Work: Fate of Fracking Fluid

The earlier modeling work has been extended to focus on the fracturing (“fracking”) fluid. One of the persistent questions about hydraulic fracturing in shale systems is the ultimate fate of the injected fracking fluid. In many shale systems, most of the injected fracking fluid does not return to the surface. This has led to concern that the injected fluid might be available to migrate out of the formation and leak into other formations, and potentially into groundwater zones.

The Celia group extended their earlier modeling work and developed a two-phase model (aqueous-based liquid fracking fluid and resident natural gas) to simulate the time period immediately following the fracking, including the shut-in period, and the subsequent longer period associated with production of gas (Figure 2.1.1). This modeling is based on an open-source, full reservoir simulator (MRST), and includes all relevant fluid flow processes, including capillary hysteresis. Detailed water injection and production data were made available for several wells in the Horn River Formation in British Columbia, and those data were used to test the model. Results showed very good matches to the amount of water remaining underground as well as the timing of the water that does flow back to the surface. The two-phase model also produces excellent matches with gas production data. These comparisons between model and data give confidence that the model produces useful results.

Figure 2.1.1. Illustration of horizontal shale gas well (top) with the model representation below it. The model representation includes the dark gray inner region representing the propped fracture, which contributes to fracking fluid imbibition and subsequent gas flow, and the light gray unpropped region that opens during fracking fluid injection but closes afterward. The dashed orange box shows the numerical model domain. Red arrows denote gas flow and blue arrows denote fracking fluid flow. Not to scale2.

In terms of the fate of the fracking fluid that remains underground, simulation results show that very strong capillary imbibition into the shale rock matrix leads to large amounts of fracking fluid being imbibed into the rock (Figure 2.1.2). Capillary hysteresis enhances the strong underlying capillarity in the rock and leads to almost all of the imbibed liquid remaining in the rock matrix during the gas production process. This result appears to be robust across a range of parameters including different representations of the capillary pressure functions.

Figure 2.1.2. Water volume in the subsurface as a percentage of total volume injected. The graph shows total water (Sum) and the distribution of that total between the fractures and the matrix. The time period (2) is the shut-in period, and the later time (3) is the gas production period. Notice that even after gas production begins, water continues to imbibe from the fractures into the matrix2.

These overall results show that the large volume of fracking fluid that remains underground is imbibed strongly into the host rock and remains there in the long term, therefore posing very little risk in terms of migration out of the shale production zone.


1 Edwards, R.W.J. and M.A. Celia, 2018. Shale Gas Well, Hydraulic Fracturing, and Formation Data to Support Modeling of Gas and Water Flow in Shale Formations. Water Resources Research, in revision.

2 Edwards, R.W.J., F. Doster, M.A. Celia, and K.W. Bandilla, 2017. Numerical Modeling of Gas and Water Flow in Shale Gas Formations with a Focus on the Fate of Hydraulic Fracturing Fluids. Env. Sci. and Tech. 51(23): 13779-13787. doi: 10.1021/acs.est.7b03270.