By Larry Gavin
July 17, 2025

On March 17, 2025, Charles F. Harvey, a Professor of Civil and Environmental Engineering at Massachusetts Institute of Technology, published a report titled, “Limitations of science and engineering relevant to protecting drinking water: The Mahomet Aquifer as a case study.”

The Mahomet Aquifer stretches across central Illinois and is designated as a “sole-source aquifer.

Dr. Harvey’s  report explains why the “best science” and the “best engineering” cannot guarantee the containment of carbon dioxide stored in geologic formations deep underground.  It concludes, “Our descendants will drink water and use water for agriculture. Geologic carbon dioxide storage is an experimental project. Allowing private interests to risk the safety of public resources is not a risk that can be justified by our current state-of-the art science or engineering.” Report, p. 9

While the report uses the Mahomet Aquifer as a case study, the report says, “conclusions provided here are applicable to any jurisdiction considering the regulation or injection of CO2, near important water sources, especially but not limited to single-source aquifers.” Report, p.1

Dr. Harvey holds a Ph.D. from Stanford University in geological and environmental science and is a founding member of the Science Roundtable on Carbon Capture and Storage. The roundtable is a group of scientists and engineers formed to provide evidence-based assessments of carbon capture and storage. It is not funded by oil companies. 

Susan D. Hovorka, a senior research scientist at the Bureau of Economic Geology, Jackson School of Geosciences, at the University of Texas at Ausin, challenges Harvey’s conclusions. Hovorka holds a Ph.D. in geology from the University of Texas at Austin.

Although asked, the EPA declined to comment on Harvey’s report.

The EPA Class VI permits for WCS’s CO2 project

Dr.  Harvey’s report was published two days after Region 5 of the U.S. Environmental Protection Agency (EPA) finalized two  Class VI permits issued to Wabash Carbon Services, LLC (WCS, a subsidiary of Wabash Valley Resources)  to construct and operate two carbon dioxide injection wells, one in Vermillion County and another in Vigo County, Indiana. The permits allow WCS to inject 1.67 million tons of liquified carbon dioxide into the ground each year for 12 years, a total of 20 million tons. The carbon dioxide will be permanently stored in rock formations about 5,000 feet below ground.

 WCS projects that the carbon dioxide will migrate underground horizontally in about a two-mile radius from each well’s bottom point. The carbon dioxide will also migrate vertically. This migration area is referred to as the “pore area.” The extent of the carbon dioxide migration is called the “plume.” The EPA says the geology of the pore area for WCS’s project includes “native formation fluids (mostly brine),” or in other words, saline water.

One important question is what happens after all this carbon dioxide is pumped into and stored underground. As one condition of the permits, WCS must monitor what’s happening in the pore areas during the 12-year injection period, and WCS must also monitor what’s happening in the pore areas for 10 years after the injections cease. EPA Region 5 decided to allow a 10-year post-injection monitoring period, rather than the 50-year default period provided in the EPA’s regulations.

The permits were initially given to WCS on January 19, 2024, but four farmers appealed. On March 3, 2025, an EPA Appeals Board concluded that Region 5 failed to state the basis for its decision to use a 10-year post-injection monitoring period, rather than a 50-year default period, and that it failed to identify the crucial facts supporting that decision in the context of ten elements and seven other criteria that had to be considered and documented. So, the Appeals Board remanded the case and ordered EPA Region 5 to provide the required information. (EPA Appeals Board decision, pp.145-46).

On May 15, 2025, Region 5 prepared a 37-page memorandum providing the information ordered by the Appeals Board. But the region decided not to provide a new public hearing or comment period.

Dr. Harvey’s report

Dr. Harvey’s report says, “Injection of CO2 into brine is so rare that it must be considered experimental.” He says that as of 2024, worldwide, there existed only 12 operational carbon capture and storage (CCS) facilities injecting CO2 into deep saline aquifers (ie. underground layers of porous rock saturated with brine), half of which began operating in 2021 or later. “Any injection of CO2 into brine must therefore be considered experimental.” Report, p.4

The report also says, “The few existing geologic carbon storage sites are plagued with serious, unforeseen problems.” In the U.S., Archer Daniels Midland’s geologic carbon sequestration project, “stands alone as the only project in Decatur, Illinois, to inject significant amounts of CO2 not used to advance oil recovery from a reservoir.  After violating its Class VI permit, ADM is not currently injecting CO2 at its Decatur facility,” says the report, p. 4

“Globally,” the report says, “many marquee geologic carbon storage projects have either been shut down or substantially reduced because of unexpected problems with injection of CO2. … It is important to note that lessons learned from oil and gas extraction do not transfer to the use of injecting supercritical CO2 in brine.” Report, p.4.

In his report, Dr. Harvey makes four broad conclusions, and supports each with specific findings:

·      “Our best science cannot guarantee geologic containment of CO2”;

·      “Our best engineering design cannot guarantee our infrastructures will durably, continuously, and indefinitely prevent CO2 migration”;

·      “Induced seismicity threatens both geologic containment and well integrity”; and

·      “According to the IPCC’s AR6, CCS is extremely expensive and will barely move the needle on climate change.”

 The report says, “The detailed EPA Class VI well permitting process cannot overcome the fundamental flaws and inherent uncertainty present in today’s best science and engineering.”  Report, p. 9.  The report summarizes some of the flaws and uncertainties as follows:

“(1) The discipline of reservoir characterization fails to correctly characterize traps and seals in 20% of the time. Relying on this discipline to protect a sole-source aquifer is too risky.

“(2) Simulation models used to define the Area of Review are notoriously bad at correctly simulating multiphase fluid flow. Original model-based predictions of CO2 plume migration in Decatur – made “with at least 90% confidence” – have already been proven inaccurate.

“(3) It is impossible to guarantee that subsurface faults – such as the fault along which injected CO2 has already migrated up to the primary sealing Eau Claire Shale – do not exist.

“(4) Even the best seismic data surveys provide us with images that are too coarse to let us identify fine-scale faults and fractures. Buoyant CO2 needs only very fine-scale cracks along which to migrate upward.

“(5) Geochemical analysis has shown that supercritical CO2 dissolved in brine weakens shale cap rocks.

“(6) Scientists and engineers do not fully understand the kinetics of the corrosion of steel by supercritical CO2 dissolved in brine. Recent leaks at ADM Decatur were caused by the unexpected corrosion of 13 Cr Steel, which is the oil-and-gas industry standard.

“(7) The Illinois basin is already pierced by many oil and gas and water wells: these man-made pathways through the subsurface provide myriad potential pathways for CO2-laced brine and its associated leached heavy metals to contaminate drinking water.

“(8) Illinois Basin geology is susceptible to earthquakes caused by subsurface pressure increases resulting from injection of supercritical CO2: induced seismicity already occurred as a result of the limited injection volumes at ADM’s CCS1. Total volume of CO2 under EPA Class VI permitting review for injection in the Illinois Basin is more than 100 times greater than the volume already injected.”

The report concludes, “Our descendants will drink water and use water for agriculture. Geologic carbon dioxide storage is an experimental project. Allowing private interests to risk the safety of public resources is not a risk that can be justified by our current state-of-the art or science or engineering.” Report, p. 9.

Dr. Hovorka presents an opposing view

The author of this article asked a representative of WVR if WVR or WCS would like to comment on Dr. Harvey’s report. The representative suggested contacting Dr. Hovorka. The author did so, and she provided a written narrative and also provided many comments posted on a pdf of Dr. Harvey’s report. She said,

“First a bit on my credentials;  I am  a geologist who has been leading a more than 16+ member team doing research on CO2 storage in deep saline formations (https://gccc.beg.utexas.edu/).  We are within the state geologic survey and The University of Texas therefore are qualified as a third party; we do work with and receive funding from diverse industries as well as state, federal, and other industries.  We have competed 8 studies monitoring large CO2 injection projects with several classified as experimental but SACROC, West Ranch, Hastings, and PetraNova were fully commercial.  We also have collaborated with an equal number of field injections internationally, mostly past experiments. We are now collaborating on elements of about 12 full scale commercial CO2 storage projects, many in Texas but with elements across the US and the world.  I started this work in 1998 and have published and presented results extensively to diverse groups including peers, regulators, NGO’s environmental activities, the Groundwater Protection Council, and the public. 

“Here is what I know from my personal data collection:

“CO2 injection is a fully mature technology operated under a mature CO2 permitting program that Congress set up under the Safe Drinking Water Act of 1974. This permitting program, known as underground injection control (UIC) is designed specifically to protect fresh water.  Many thousands of wells injecting hazardous and poor quality water and gases have been operated successfully for 50 years.  CO2 has been injected for the same time frame under UIC rules, initially as a solvent to enhance oil production in oil fields.  The first (still running) injection of CO2 outside into deep saline formations for the sole purpose of not releasing to atmosphere started in Norway in 1996 and the number of such projects has slowly grown until about 50 are now in operation. They have been heavily monitored and to date we know that none has harmed fresh or marine waters.  In the US, EPA decided to promulgate a new and more stringent rule in 2010, known as Class VI, that builds very sturdily on the past experience but for this new benefit to remediate the carbon cycle however in truth deep well injection is not ‘new’.  

“I agree that fresh water resources must be protected, but the greatest hazards to them is from climate change itself.  Deep well injection to prevent release of CO2 to the atmosphere is needed to protect land and environment and is known to be safe.” 

As to Dr. Harvey’s paper, Dr. Hovorka says, “His paper repeats many misunderstandings that are commonly stated but incorrect. One-by-one rebuttal may not be useful, however I attach a few comments to the PDF [of his report]. Note that none of the objections he provides are specific to the Mahomet aquifer. It is not especially endangered by deep well injection of CO2 and does not need special protection.  All aquifers in all regions need protection as provided by the Safe Drinking Water Act (https://www.epa.gov/sdwa).  

“In our publications you will find an even longer list of risks evaluated, tested and assessed.  It is important not to consider this de-risking operation as a reason to stop work. I give some analogies to cars, which are many times more dangerous than deep well injection of CO2; both are nevertheless needed. The societal approach needs to be thoughtful and require rigorous and effective risk conceptualization and effective de-risking, which I and many other CCS researchers have done over the past 25 years.”  

In addition to her general remarks, Dr. Hovorka provided comments posted onto a pdf of Dr. Harvey’s report. Her comments, together with the relevant text from Dr. Harvey’s report are set out in an appendix to this article. In her first comment, Dr. Hovorka takes issue with Dr. Harvey’s statement, “As of 2024, worldwide, there existed only 12 operational CCS [carbon capture storage] facilities injecting CO2 into deep saline aquifers.” She cites the 2024 Global Status of CCS report.

That 2024 CCS report, at pages 57-58, lists a total of 50 operational carbon storage facilities in the world. Of those, only 12 have a storage code of “deep saline formation;” 2 have a storage code of “mineral carbonization;” and 36 have a storage code of “enhanced oil recovery.” Of the 12 facilities that have a storage code of “deep saline formation,” only 5 are located in the U.S. and the earliest one of those is ADM’s project in Decatur, Illinois, which become operational in 2017. 

Many of Dr. Hovorka’s other  comments say that the risks and uncertainties stated in Dr. Harvey’s report are addressed in the EPA’s UIC program that requires a robust evaluation of risks during the permitting process, requires monitoring to ensure safety, and requires corrective action if necessary.

Dr. Harvey’s report presents an opposing view about the effectiveness of the EPA’s permitting process to ensure long-term safety of water sources. His report says, “The U.S. Environmental Protection Agency’s (EPA’s) Class VI permitting process is rigorous and detailed; however, because the process cannot overcome the key limitations of fundamental uncertainties in state-of-the-art science and engineering, the EPA Class VI permitting cannot reliably guarantee safe injection of CO2, under sole-source aquifers and associated recharge areas.”  He adds, “[T]he conclusions provided here [in his report] are applicable to any jurisdiction considering the regulation of injection of CO2 near important water sources, especially but not limited to single-source aquifers.”

The author of this article asked Dr. Harvey if he had a response to Dr. Hovorka’s comments about his report. He replied:

“Many of Dr. Hovorka’s comments restate how she believes that permitting requirements will prevent leaks. I don’t think they will. For one reason, they rely on dubious numerical models. I point again to the simple 2-D laboratory experiment where the structure was exactly known (it was constructed in a laboratory!), yet modeling still failed to predict CO2 movement. And, this is from one of the strongest multi-phase modeling groups in the world — not consultants who do permit applications. 

“Dr. Hovorka also argues in several places that many experiments (100’s!) have been successful. I don’t know what she is referring to. Maybe modeling exercises? There have been very few large-scale injections into saline aquifers and there is very little field verification of numerical models. The Sleipner site has probably been studied the most, and models performed poorly there.  I have experience working with these models. 

“Another way to look at this is to ask:  Why, if CCS developers are confident in their safety, do they insist on deals where government assumes responsibility for leaks after short time periods?

“One detail -- I cite a Schlumberger report for the ADM site that claimed “with at least 90% confidence” CO2 would not reach the primary seal within the project lifetime.  Despite Schlumberger’s calculated odds, CO2 did migrate upward, reaching the Eau Claire Shale.  Dr. Hovorka responded to this, “No migration above the Eau Claire confining zone is detected by  monitoring.” Schlumberger claimed 90% confidence that it would not reach the base of the Eau Claire, they did not give odds that it would migrate above the Eau Claire. It did reach the underside of the Eau Claire, and quickly, according the seismic reinterpretation.” 

EPA declines to comment on the report

The author of this article also asked the EPA if it would be willing to comment on Dr. Harvey’s report. An EPA spokesperson responded, “EPA cannot comment on reports prepared outside the agency. However, we would refer you to the agency’s scientific documentation on Class VI wells.”

As noted above, Dr. Harvey’s report was published two days after EPA Region 5 finalized the Class VI permits that allow WCS to store 20 million tons of carbon dioxide underground in Vermillion and Vigo Counties, Indiana, and two days after Region 5 prepared its May 15 memorandum containing its reasons and findings to shorten the post-injection monitoring period from the default 50 years to 10 years.  EPA Region 5 decided not to allow public comment on the many reasons and findings in its May 15 memorandum. Thus, the public had no opportunity to challenge Region 5’s decision to use a 10-year post-monitoring period using Dr. Harvey’s report.  And, Region 5 did not have the benefit of that report when it finalized the permits.

The map at the top of this article shows the aquifers that are above the pore space in Vermillion County.

………………………………

Link to Report of Charles F. Harvey, “Limitations of science and engineering relevant to protecting drinking water: The Mahomet Aquifer as a case study, 2025”:  https://ecojusticecollaborative.org/wp-content/uploads/2025/03/Limitations-in-science-and-engineering-related-to-Class-VI-permitting-process-4.pdf

Decision of the EPA Board of Appeals, dated March 3, 2025:   https://yosemite.epa.gov/oa/eab_web_docket.nsf/9a0c1c4939bdf18b852584be006bffec/e65f020c747473d085258c540050b693/$FILE/Wabash%20Order%20Remanding%20in%20Part%20and%20Denying%20in%20Part,%20EAD%20FINAL%202025.3.21.pdf

Links to other related articles in Vermillion Reports:

“Taking a look at the constitutionality of the law authorizing Wabash Valley Resources’ CO2 project”: https://vermillionreports.org/laws-concerning-c02-project

“Analysis and Viewpoint: The public was short-changed in the EPA public hearing and comment period on Wabash Carbon Services’ CO2 permits”: https://vermillionreports.org/public-shortchanged-epa-public-hearings

………………………………………..

Appendix  Containing Dr. Hovorka’s Comments to Dr. Harvey’s Report

This appendix contains the  comments that Dr. Susan D. Hovorka provided to Larry Gavin relating to specific portions of Dr. Charles D. Harvey’s report titled, “Limitations of science and engineering relevant to protecting drinking water: The Mahomet Aquifer as a case study.”  Dr. Hovorka posted the comments on a pdf of the report. The relevant portions of Dr. Harvey’s report are quoted and then Dr. Hovorka’s comments relating to that portion of the report are quoted immediately below.

This appendix, at the end, also contains an overall response that Dr. Harvey provided to Dr. Hovorka’s comments.

Dr. Harvey’s Report, p. 1: “As of 2024, worldwide, there existed only 12 operational CCS [carbon capture and storage] facilities injecting CO2 into deep saline aquifers.”

 Dr. Hovorka’s comment: “Incorrect, please google global ccs institute 2024 report.”

Larry Gavin’s note: I reviewed the 2024 Global Status of CCS report, as suggested by  Dr. Hovorka. That report, at pages 57-58, lists a total of 50 operational carbon storage facilities in the world. Of those,  12  have a storage code of “deep saline formation;” 2  have a storage code of “mineral carbonization,” and 36 have a storage code of  “enhanced oil recovery.” The 12 facilities that have a storage code of deep saline formation are listed below, together with the country of location, and the operational year, all taken from the table in the 2024 Global Status of CCS report:

Sleipner CCS project, Norway, 1996

Equinor Snohvit, Norway, 2008

Shell Quest, Canada, 2015

ADM Illinois Industrial, U.S., 2017

Qatar Energy LNG, Qatar, 2019

Chevron Gorgon, Australia, 2019

Dark Horse Storage,  U.S., 2021

Red Trail Energy Richardton Ethanol, U.S. 2022

Entropy Glacier Gas Plant, Canada, 2022

Harvestone Blue Flint Ethanol, U.S., 2023

CNOOC Enping, China, 2023

Barnett Zerc CCS, U.S., 2023

Dr. Harvey’s Report, p. 1: Leaks from Archer Daniels Midland (ADM) wells in Decatur, Illinois, show that we cannot assume that the oil and gas industry’s technology and know-how will transfer to the carbon capture and storage (CCS) use case.

Dr. Hovorka’s comment: ADM had an out-of zone migration that allowed fluids to migrate up a well from 8000 to 6000 feet below ground. no leak occurred to surface and no impact to groundwater was detected by their monitoring program. https://www.adm.com/en-us/standalone-pages/adm-and-carbon-capture-and-storage/

Dr. Harvey’s Report, p. 2: Water for human consumption and agricultural use will be needed far beyond the project life of a CO2-injection facility: the public interest demands caution.

Dr. Hovorka’s comment: Yes, The ETA UIC program provides this caution explicitly.

Dr. Harvey’s Report, p.2: Because the models give uncertain results, they potentially underestimate the area at risk of contamination.

Dr. Hovorka’s comment: Models are the method to assess uncertainty. The required monitoring program is designed to reduce uncertainty and assure no endangerment of USDWM.

Dr. Harvey’s Report, p.2: In 2013, reservoir simulations used to support the Area of Review designation for ADM Decatur’s CCS1 project were used to assert “with at least 90% confidence” that CO2 would not reach the primary seal within the project lifetime; however, CO2 has already migrated along faults, reaching the Eau Claire Shale.

Dr. Hovorka’s comment: This has not occurred. A paper using geophysics to consider within Mt Simon migration has been published and limitations discussed. No migration above the Eau Claire confining zone is detected by monitoring.

Dr. Harvey’s Report, p.3:  The science used to characterize reservoirs and sealing rocks – reservoir geoscience – is unavoidably inexact. Reservoir geoscience cannot be relied upon to reliably, correctly identify cap rocks that will keep CO2 contained: at least 20% of such similar predictions made in the oil and gas industry to identify oil and gas deposits are wrong, a tolerable level failure for oil exploration, but a very high risk for sole source drinking water aquifers.

Dr. Hovorka’s comment: please see https://www.sciencedirect.com/science/article/pii/S1750583623000786 for rebuttal

Dr. Harvey’s Report, p.3:  Recent investigations have shown that supercritical CO2 dissolved in brine weakens shale cap rocks: further geochemical study is needed.

Dr. Hovorka’s comment: Study of rock-water CO2 reaction is explicitly required by Class VI rules and must prove to regulator that any reaction will not cause endangerment.

Dr. Harvey’s Report, p.3: Our engineering design practices cannot be guaranteed to prevent water contamination: Key uncertainties exist in core components of well design. For example, scientists lack a full understanding of the kinetics of pipe corrosion in the presence of supercritical CO2 dissolved in brine.

Dr. Hovorka’s comment: Similarly, UIC rules explicitly require a  robust demonstration of injectate - materials compatablity and require a monitoring program with coupons or alternatives.

Dr. Harvey’s Report, p.3: T he geology of the Illinois Basin is conducive to induced seismicity: an increase in earthquakes has already resulted from injecting CO2 in central Illinois. This induced seismicity can damage both a cap rock’s ability to contain CO2 and well integrity.

Dr. Hovorka’s comment: Seismically is indeed a concern during deep well injection and has to be managed. Deep well injection is used in very seismogenic areas  such as japan and requires management and monitoring to assure safety.

Dr. Harvey’s Report p.2: Geologic carbon dioxide storage is an experimental project. Allowing private interests to risk the safety of public resources is not justified by science or engineering.

Dr. Hovorka’s comment:  experiment phase for this technology was completed by 2010, after more than a decade of intense experimentation (partly by me).

Dr. Harvey’s Report, p. 3: Engineers in the oil and gas industry have copious experience using computer models to simulate multiphase flow in the subsurface: that experience has repeatedly shown that reservoir simulation models provide unreliable and highly variable predictions of fluid flow unless they are extensively tuned to plentiful real-world bservations (e.g., Oliver and Chen, 2011; Lee et al., 2011; Valdez et al., 2020; Bahrami et al., 2022; Zhang et al., 2023).

Dr. Hovorka’s comment:  True. Class VI program therefor requires history matching and corrective action if needed.

Dr. Harvey’s Report p.4: Injecting large volumes of CO2 results in an extensive area of regional pressure increase that extends far beyond the CO2 plume itself (Zoback and Hennings, 2025). Currently, the EPA Class VI permitting process defines the AoR by the regional extent of the CO2 plume; however, the pressure changes from injection of CO2 are what cause earthquakes (‘induced seismicity’), which can damage both wellbores and cap rock integrity (see more detail in Section 6). After simulating a region-wide geologic carbon storage program in central Illinois, Birholzer and Zhou (2009) stated, “Our results suggest that the area that needs to be characterized in a permitting process may comprise a very large region within the basin if reservoir pressurization is considered.”

Dr. Hovorka’s comment: True. UIC program requires robust calculation of AoR followed by "tracking the pressure front" and corrective action if needed.

Dr. Harvey’s Report, p.4:  In 2013, researchers at ADM partner Schlumberger reported results of their pre-injection reservoir modeling for the Decatur sites. They stated, “Over the lifespan of simulations plume stayed below the middle Mount Simon which is 250 m below the primary seal Eau Claire with at least 90% confidence” (Senel and Chugunov, 2013). Reporting in early 2025 on seismic imaging of the CO2 plume from ADM injection well CCS1, Bukar et al. (2025) identified evidence that CO2 has already migrated upward along faults to reach the primary seal (i.e., the Eau Claire Shale)

Dr. Hovorka’s comments:  Reaching the confining zone is expected, and not a problem. See Bump paper cited above.

Dr. Harvey’s Report p.4:  As of 2024, worldwide, there existed only 12 operational carbon capture and storage (CCS) facilities injecting CO2 into deep saline aquifers, half of which began operating in 2021 or later (Global CCS Institute, 2024). Any injection of CO2 into brine must therefore be considered experimental.

Dr. Hovorka’s comment: My personal experience is otherwise.

Dr. Harvey’s Report, p.4:  n the US, ADM’s Decatur project stands alone as the only project to inject significant amounts of CO2 not used to enhance oil recovery from a petroleum reservoir. After violating its Class VI permit, ADM is not currently injecting CO2 at its Decatur facility.

Globally, many marquee geologic carbon storage projects have either been shut down or substantially reduced because of unexpected problems with injection of CO2 (e.g., Eiken et al., 2011). Algeria’s In Salah was shuttered in 2011 because of significant uplift observed at the surface, which indicated that CO2 injection was affecting subsurface faults and fractures, thereby increasing earthquake risk (White et al., 2014). In 2011, StatOil (now Equinor) stopped injecting CO2 at Snøhvit after the reservoir pressure increased much faster than engineers had predicted, approaching the pressure that would likely fracture the cap rock (Hansen et al., 2013). CO2 injection at Sleipner in Norway has been substantially slowed, presumably in part because of indications that the CO2 has not been contained as originally predicted (Cavanaugh and Haszeldine, 2014).

Dr. Hovorka’s comment: CO2 storage project are closely monitored and every non-compliance is corrected. This should give confidence, not concern. Each of these cases cited is only a negative headline not a data-driven analysis. This is equivalent to saying that a car that failed its safety inspection must be junked and all cars abandoned, when what is needed is new front tires.

He missed quite a few other corrected problems, however also missed many entirely trouble free operations.

Dr. Harvey’s Report, p.5:  Advocates for use of the EPA Class VI permitting process to protect single-source aquifers assert that reservoir characterization has been perfected over its many decades of application in the oil and gas industry. In part because the subsurface is vast, variable, and difficult to observe, reservoir geoscience is unavoidably imperfect. Reporting on a 20-year dataset of Exxon Mobil’s conventional wildcat exploration wells in proven plays, senior company geologists reported that more than 20% of the failures to find oil and gas were a direct result of a mischaracterization of the trap and seal (Rudolph and Goulding, 2017). That is, even with best-in-class skill and data, 20% of the time, reservoir geologists incorrectly surmised that a cap rock would contain oil and gas. We can reasonably assume a similar (or worse) accuracy rate when using reservoir characterization techniques to assess the potential of a cap rock to contain CO2.

Even were reservoir characterization for the oil and gas industry to have a perfect track record when identifying seals and traps, industry experience characterizing reservoirs for the purpose of oil and gas extraction does not fully inform the characterization of reservoirs for carbon sequestration. Oil and gas projects involve net removal of fluids and gases, thereby decreasing subsurface pressures, which tends to stabilize faults and fractures; however, carbon sequestration increases subsurface pressures, risks opening fractures, and potentially induces seismicity. Carbon sequestration is a brand new use case for reservoir characterization.

Dr. Hovorka’s comment: We modify from oil and gas experience, not use it 1:1.

Most of the experience on containment is from the UIC program. See Class 1 permits example https://repositories.lib.utexas.edu/items/d316626b-c798-4634-93d4-8c53ff8be1ef

In response to the Harvey report statement, “Carbon sequestration is a brand new use case for reservoir characterization,” she says, “only 29 years old.”

Dr. Harvey’s Report, p.5: The existing very limited experience with geologic carbon storage has shown that a priori reservoir-property characterization is insufficient to properly predict and un CO2 flow. When reporting on lessons learned from company experience with three different CCS injection facilities, StatOil (now Equinor) authors said, “The actual CO2 plume development has been strongly controlled by geological factors that we learned about during injection” and that “detailed CO2 site characterisation and monitoring is needed to prove significant practical CO2 storage capacity–on a case by case basis” (Eiken et al, 2011; emphasis added).

Dr. Hovorka’s comment:  monitoring and corrective action is required to use data gained during injection to de-risk the injection further.

Dr. Harvey’s Report, p. 5: Even were geoscientists able to perfectly identify sealing cap rocks, recent geochemical literature has shown that supercritical CO2 dissolved in brine can weaken shale cap rocks over time (e.g., Ilgen et al. 2018; Yang et al., 2023). Researchers evaluating how CO2 sequestration conditions affect cap rock found that “mechanical properties are significantly weakened when shale specimens are exposed to both [supercritical] CO2 and brine” (Choi et al. 2021). There is insufficient real-world experience to understand whether CO2 dissolved in brine will weaken shale cap rocks in real-world settings.

Dr. Hovorka’s comment: 100's of experiments have been done. Effects are small, well known to project designers and regulators.

Dr. Harvey’s Report, p. 6: Proponents of the existing process may point to seismic data surveys as evidence of detailed characterization of the subsurface. Seismic data gives a coarse estimate of the geology under the ground, but it cannot highlight the myriad potential fine-grained pathways through which CO2 might migrate (Williams-Stroud et al., 2020). The only way to know for sure that there are no sub-seismic-scale faults along which CO2 can migrate out is to go ahead and inject it, then wait and see if contamination occurs. Given the Mahomet’s single-source aquifer designation (EPA, 2015), this approach risks irreparable contamination of the region’s water source.

Dr. Hovorka’s comment: Pressure is the main tool used in combination with seismic. Sub seismic faults are by definition small and would not transmit fluids from depth.

Dr. Harvey’s Report, p.6:  CO2 migrates easily along natural faults: indeed, CO2 injected at ADM Decatur has already started migrating upward, beyond the containment zone, along pre-existing faults (Bukar et al. 2025). Over geologic time, both compressional and extensional stress has been applied to the Illinois Basin, in which the Mahomet Aquifer now sits. These tectonic episodes have resulted in the generation of geologic faults (Zemansky, 2025). Although the northern portion of the Illinois Basin has fewer faults than the southern portion, faults beneath the Mahomet do exist (e.g., Williams-Stroud et al., 2020). Panno et al. (2022) assert, “The presence of geologic structures in the vicinity of the injection sites could compromise protective caprocks and allow CO2 to migrate back to the surface.”

Dr. Hovorka’s comment:  Everywhere on Earth has faults. Most are known to be sealing. UIC rules require evaluation of faults for deep well injection to show retention.

Dr. Harvey’s Report, p. 6: CO2 can also migrate upward along man-made pathways. More than 400 existing wells are known to cross the Mahomet (Figure 2); the median completion year for wells crossing the Mahomet is 1966 (ISGS, 2024). See further detail on ways in which aging and abandoned wellbores can result in aquifer contamination in Section 5.2.

Dr. Hovorka’s comment: Wells also exist most places. Class VI has a robust and successful program for managing them better than Class II.

Dr. Harvey’s Report, p. 6:  Corrosion of 13 Cr steel caused the recent leaks in Decatur monitoring wells (Snider and Lefebvre, 2024), which were permitted as part of EPA’s existing Class VI process. Despite being the oil-and-gas-industry standard, 13 Cr Steel pipe is not recommended for conditions in which high-chloride brines combine with CO2 (Craig and Smith, 2011). Even were future CCS projects to routinely use more expensive, more-corrosion-resistant piping materials, it is important to note that scientific and engineering communities currently lack full understanding of how supercritical CO2 influences corrosion. Reviewing the state of current corrosion research, Sun et al. (2023) write, “The understanding of corrosion kinetics laws in s[upercritical]-CO2 environments is currently inadequate, and there is considerable variability in reported corrosion outcomes.”

Dr. Hovorka’s comment: UIC program requires a robust evaluation of corrosion and a suitable monitoring program of both old and existing well penetration. I will be giving at talk in August in The Hague on this topic.

Dr. Harvey’s Report, p. 6: Decades-old wellbores and completions cannot be guaranteed to maintain the structural integrity necessary to prevent CO2 migration (Gasda et al., 2004; Celia and Norbotten, 2009). Figure 3 show ways in which a single compromised wellbore can enable contamination of many layers above a CO2 injection zone; Figure 4 provides a schematic of the many potential ways in which an aging wellbore can provide pathways for CO2 leakage.

Dr. Hovorka’s comment:  Legacy wells in AoR are a major focus of UIC rules. Class I has been highly successful in managing wells. Less stringent rules of Class II have had local failures (we are writing a report on this topic now!). Class VI follows the successful Class I processes.

Dr. Harvey’s Report, p. 8: As repeatedly shown in central Oklahoma, fluid pressure increases from injection are enough to triggersubstantial induced seismicity (Langenbruch and Zoback, 2016; Langenbruch et al., 2018). When CO2 is injected, it displaces pore water into the already water saturated rock around the injection, increasing pressures far beyond the plume of CO2 itself (e.g., Birkholzer and Zhou, 2009; Bandila et al, 2012). However, the EPA defines the Area of Review by the modeled extent of the plume of CO2. The permitting process therefore does not require detailed structural analysis in regions that will be impacted by the increased pressure resulting from CO2 injection.

Dr. Hovorka’s comment: not correct. AOR is defined by pressure front.

Dr. Harvey’s Report, p.14: Pressures continue to rise as long as CO2 is injected and so the subsurface experiences more and more severe risks of induced fracturing and seismicity as the project continues. It is impossible to use observations at lower pressures to predict what will happen as pressures continuously rise.

Dr. Hovorka’s comment: Substantial past deep well injection provides good predictive capability. Remaining concern over upscaling is an area of current research, this is a normal pathway for all technologies, from land use to computation. Concern is useful, stopping progress because of excessive fear is not.

Dr. Harvey’s Report, p. 8: In the most recent iteration of the IPCC’s international report, carbon capture and storage is identified as by far the most expensive mitigation option among many options, with the average cost of mitigation exceeding $100 per ton of CO2 kept from the atmosphere (Figure SPM.7, IPCC 2023).Despite its expense, carbon capture and storage is unlikely to significantly move the needle on climate change, with a total contribution to net emission reduction far below other mitigation options (Figure SPM.7, IPCC 2023). More global heating would be averted if the money spent for CCS were instead used for one of the many less expensive options for reducing atmospheric greenhouse gasses (Jacobson et al., 2025). It is difficult to argue that the risks of CCS are justified by the need to reduce emissions because other means could reduce emissions far more without the associated risks.

Dr. Hovorka’s comment:  Jacobson is a very vocal CCS opposition guy. The IPPC reports always need CCS to attain climate goals at low cost and wide application, assigning 10-15% of emissions reduction to this tech. CCS is from the beginning seen as part of a portfolio, available what it is the best option. Other options for GHG reduction also have risks, for example concern about battery safety or use of lands for solar all need balance. Options are useful.

Dr. Harvey’s Report, p.9: Our descendants will still drink water and use water for agriculture. Geologic carbon dioxide storage is an experimental project. Allowing private interests to risk the safety of public resources is not a risk that can be justified by our current state-of-the-art science or engineering.

Dr. Hovorka’s comment: note that CO2 is present naturally in many aquifers and is not harmful.

Overall response by Dr. Harvey to Dr. Hovorka’s comments:

The author of this article asked Dr. Harvey if he had a response to Dr. Hovorka’s comments about his report. He replied by email:

“Many of Dr. Hovorka’s comments restate how she believes that permitting requirements will prevent leaks.  I don’t think they will.  For one reason, they rely on dubious numerical models. I point again to the simple 2-D laboratory experiment where the structure was exactly known (it was constructed in a laboratory!), yet modeling still failed to predict CO2 movement. And, this is from one of the strongest multi-phase modeling groups in the world — not consultants who do permit applications. 
“Dr. Hovorka also argues in several places that many experiments (100’s!) have been successful. I don’t know what she is referring to. Maybe modeling exercises? There have been very few large-scale injections into saline aquifers and there is very little field verification of numerical models. The Sleipner site has probably been studied the most, and models performed poorly there.  I have experience working with these models. 
“Another way to look at this is to ask:  Why, if CCS developers are confident in their safety, do they insist on deals where government assumes responsibility for leaks after short time periods?
“One detail -- I cite a Schlumberger report for the ADM site that claimed “with at least 90% confidence” CO2 would not reach the primary seal within the project lifetime. Despite Schlumberger’s calculated odds, CO2 did migrate upward, reaching the Eau Claire Shale.  Dr. Hovorka responded to this, “No migration above the Eau Claire confining zone is detected by monitoring.”  Schlumberger claimed 90% confidence that it would not reach the base of the Eau Claire, they did not give odds that it would migrate above the Eau Claire.  It did reach the underside of the Eau Claire, and quickly, according the seismic reinterpretation.”