Getting the mainstream newspapers excited is easy. Just give them a quotable quote on an emotional subject like shale gas drilling in the Karoo by means of hydraulic fracturing (“fracking”), and enjoy the spectacle of watching them blindly and unquestioningly run with it.
“Pro-fracking Prof does U-turn,” yelled the Cape Times.“Karoo fracking could cause water disaster,” one-upped TimesLIVE. “Fracking ‘will cause water pollution’”, proclaimed Fin24 over the same wire copy.
Gerrit van Tonder, a professor at the Institute for Groundwater Studies at the University of the Free State, had been broadly dismissive of the exaggerated fears of anti-fracking lobbyists in the past. It was therefore surprising, and newsworthy, that he suddenly declared himself to be “‘100% certain’ that underground water in the Karoo basin flowed upwards, and that there were a vast number of natural ‘pathways’ along which water could flow upwards.”
Several readers alerted me to these stories, and asked my view. So here’s a radical idea. Let’s have a look at that paper, shall we? No, really. I actually have a copy, and not just of the abstract. Van Tonder tells me I’m the only journalist who even asked for one. As a consequence, my column may be a little longer than the breathless news reports of last week. Sorry.
Van Tonder’s paper, entitled “Vertical movement of hydraulic fracturing fluids in the Karoo formations of South Africa”, is in very rough draft form and has not been peer-reviewed or published in any academic journal. Presumably, this means he has yet to remove the quotations copied and pasted from the Wall Street Journal, and perhaps find a better diagram of the hydraulic fracturing process than a thorougly misleading propaganda piece produced by an anti-fracking group.
Van Tonder’s draft describes limited evidence that there might be confined brine aquifers between layers of impermeable rock (of which shale is the classic example), both above and below the Ecca Group shale formation from which gas can be obtained several kilometres below ground. It is argued that these aquifers give rise to upwardly-migrating “artesian” water. (See p.346 in Chapter 13 of Earth’s Dynamic Systems for a diagram. In fact, read the entire chapter for a good primer in groundwater systems.)
The paper continues by describing how known natural formations, such as intrusions of dolerite that interrupt horizontal rock layers, as well as gas well boreholes themselves, could form “preferential pathways” for this deep water to migrate not just upwards, but all the way to the shallow fresh-water aquifers, which in the Karoo lie no deeper than 300 metres. (The majority of Karoo wells are water table wells less than 100m deep.)
Van Tonder proceeds to create a computer model that maps the possible migration of deep saline water mixed with fracking fluid, via the preferential pathways, to the surface. It’s a colourful picture, but it comes with a disclaimer: “The model must be viewed as a very rough first attempt based on sparse information which cannot be validated or verified.”
Nobody in the media read this line, or paid it any attention, but doesn’t this seem a little odd in light of the “100%” certainty that they did quote?
Sure, the deep water might exist. If it does, it is probably artesian, like most deep aquifers confined by impermeable rock formations. Sure, pathways to the surface could cause water to migrate upwards to equalise the gravitational pressure from the recharge zone (which can be hundreds of kilometres away where the deep layer crops out, often in a mountain ridge surrounding the basin). Generally, however, water at these depths circulates on time-scales measured in millennia.
But a “model” is essentially a computer programme that takes a set of initial conditions and a set of assumptions about how these change over time, in order to produce a possible set of future ending conditions. The value of a model is entirely dependent on the accuracy of the initial conditions and the validity of the assumptions about the factors that influence how they change over time. Even one error can dramatically change the outcome of a numerical model of this kind. In fact, textbooks on numerical methods often warn not to extrapolate far beyond the end of the data, because the uncertainty of the predictions rises rapidly, and any inaccuracies can have a compounding effect on errors.
There are many “unvalidated and unverified” assumptions in Van Tonder’s model which bear questioning. The Karoo Groundwater Atlas of February 2012 was prepared as part of Shell’s environmental impact assessment by a group of scientists led by SRK Consulting. Van Tonder participated as a technical reviewer. This report contains most of the Karoo-specific information that Van Tonder uses in his paper. It says: “[As] was agreed by (the contributing scientists), there is so little data available below 300m that it is difficult to ‘populate’ the deep conceptual model, beyond making intelligent hypotheses. This uncertainty is indicated by the liberal use of question marks…The deeper geological/hydrogeological model is less well understood and further work, such as land and airborne geophysics and exploration drilling, would be required to obtain a better understanding of this environment.”
In his first response to my request for a copy of his paper, Van Tonder wondered who was going to pay him for the additional field work that is required. This indicates that the research proposed in the Groundwater Atlas has not been done, and uncertainties therefore remain as high as ever.
So how did Van Tonder come, a mere three months later, to build a model that made him “100% certain”? Isn’t this a question a sceptical journalist should have asked?
In his defence, he only claimed to be certain of an upward flow gradient, and not, as the newspapers would have you believe, of a “water disaster”, “water pollution” or even that gas drilling should be disallowed. But it is troubling that he constructed a model based on “intelligent hypotheses” and caused these to be described with such overwhelming certainty in the media.
One assumption in the model that strikes me as questionable, for example, is that the potentiometric surface of the confined saline aquifers trapped below the Ecca shale – that is, the level to which the water would rise if unconfined by the impermeable rock layers above it – lies near the fresh-water aquifers at the surface. If it doesn’t, even a wide-open “preferential pathway” wouldn’t get it to the surface. You’d need a borehole pump. And even if it does, it is far from a given that the water produced from this depth will not be retarded or dissipated by the intermediate brackish aquifers which we know it will encounter along the way. In short, even if deep aquifers produce artesian water, how far above the top of the aquifer will that water rise? A little way? A long way? All the way to the surface?
In a letter published in the print edition of the Cape Times, Philip Lloyd, a research professor at the Energy Institute of the Cape Peninsula University of Technology, calls Van Tonder’s claim “just plain wrong”. In his reply, Van Tonder mentions the high-yielding fresh-water boreholes drilled in the Karoo along dolerite intrusions as illustrations of why he describes those intrusions as “preferential pathways”.
However, if those intrusions do conduct deep saline water to the surface, and this is the water Van Tonder fears might become polluted with fracturing fluid, then surely that problem would already have emerged in the existing research data? After all, by contrast with the deep formations, the Groundwater Atlas says “[t]he shallow (<300m depth) aquifer is well understood and documented,” and my bet is that Karoo farmers who routinely drill near dolerite formations don’t routinely expect to find salt water.
None of the newspapers asked such questions. I’m somewhat sympathetic, because it does require a modicum of hydrogeographic knowledge that journalists would need to acquire by, say, reading textbooks or asking other scientists on the subject. Doing so is far more work than slapping sensationalist headlines over wire copy, and probably gets you far worse readership data to flog to advertisers.
In my email interview with him, Van Tonder repeatedly highlights the problem of well construction. In some cases of well-sleeve failure, which are uncommon, gas migration into near-surface aquifers has indeed been detected. Oil company executives, environmental campaigners and government regulators all agree that well-construction standards can and should be improved, as this Wall Street Journal article makes clear. This is not a new issue, and proposals for better regulations to prevent well failures have been in the works for some years.
No doubt, the South African government will adopt international best practices in the conditions they impose on any exploration permits they might issue. Besides, as Lloyd points out, well failures are problematic for gas drillers too, since it prevents them from pressurising the well sufficiently to begin hydraulic fracturing in the first place. Also, cleanup operations and damages lawsuits are expensive. Well-sleeve failure is an expensive affair that gas production companies have every profit-driven reason (and some environmental ones) to want to avoid.
Van Tonder is specifically concerned about the fracturing fluid or contaminants leached from deep rock strata. Only inasmuch as these pollute the underground water does he think they might pose a threat to shallow groundwater.
Let’s assume that this deep water does reach fresh-water aquifers, as he speculates. He says he does not know to what extent so-called “naturally-occurring radioactive materials” are present in the Karoo, but depending on this, the use of non-toxic fracturing fluids, such as the “food-grade” concoction produced by Halliburton known as CleanStim, might reduce the risk to zero. And that’s before taking into account any physical filtration or chemical conversion which would serve to neutralise contaminants along the way.
The International Energy Agency recently laid out a set of Golden Rules which, it says, would earn the shale gas industry a “social licence to operate” by addressing environmental and social concerns in return for increasing the development cost of a typical shale gas well by 7%.
So the newspaper headlines massively overstated the case. Even if there is something to Van Tonder’s fears, which is far from certain, those risks can be mitigated relatively cheaply.
In his reply to Lloyd, Van Tonder explains his need to rush to the media to talk about the content of an unreviewed rough draft. His argument is that the moratorium on issuing exploration licences might be lifted in the next two months, while publication in a peer-reviewed journal can take much longer.
This is true, but as Van Tonder himself explained in a presentation he gave in March 2012 (when he was still pro-fracking himself), exploration is likely to be a nine-year process, the first two years of which will be used to produce a detailed environmental, social and health impact assessment, including magneto-telluric and seismic surveys designed to find answers to the very questions he raises. It seems curious that a scientist would act in a manner that seems designed to prevent the very research that is needed.
He tells me all he wants at this stage is an “open technical debate”, but by going to the media before his work is even available, he is undermining that debate and leading journalists, public opinion, and perhaps even government policymakers, to foregone conclusions.
I approached SRK Consulting to ask Millie Goes and Peter Rosewarne, the senior hydrologists on the project to map the Karoo’s groundwater, for detailed comments on Van Tonder’s claims. I had hoped to discover how many other scientists agreed and what they thought his motives might be for bypassing the ongoing research project by going directly to the media. Six days later, they had yet to respond.
But Nick Grealy of the NoHotAir blog did get a response, from one of the foremost researchers in the field. Terry Engelder of the department of geosciences at Penn State University had this to say about Van Tonder’s change of heart: “I have submitted a short paper to the Proceedings of the National Academy of Sciences that will, I hope, introduce (to the public) the possibility that gas shale permanently sequesters residual treatment water [that is, fracturing fluid that does not flow back up the well]…This note was written in response to a paper that claimed that natural brine was continuously leaking from a gas shale. If 60% to 90% of all treatment water is sequestered inside a rock formation that is mantled by a capillary seal (gas in the Marcellus has been held in place for 200+ million years), it is then highly unlikely that gas shale leaks brine at rates that matter to mankind. The same can be said for the Ordovician-Silurian Rocks gas shale of central Europe, the Mesozoic gas shale of the Paris Basin, and the Karoo of South Africa.”
It is true that Engelder might be wrong, but then, so might Van Tonder. This is how science works: by evaluating competing scientific claims in light of empirical evidence.
Van Tonder responded by challenging Engelder, who visited South Africa in March 2012 to host the Second Annual Shale Gas Forum, to a live debate on television. To fly to South Africa for this purpose, Van Tonder proposed that Engelder request funding from Shell. This is clever. If Engelder declines, Van Tonder has the moral high ground. If Engelder accepts, it would discredit him as being in the pocket of the oil and gas industry. (Full disclosure: I spoke, unpaid, about the media’s uncritical sensationalism at that conference.)
Grealy, for his part, makes a harsh assessment of Van Tonder: “The problem to the industry is that as Robert Howarth of Cornell or University of Manchester Tyndall Centre have proven in the past, it only takes one academic, no matter how minor, to achieve world fame by contradicting every other scientist in their field.”
Asking for research funding, demanding a live debate with a famous scientist on television, and garnering sensational headlines in most major South African newspapers, this Van Tonder has done. Even though nobody – other than a certain freelance columnist – has seen his paper.
Maybe Van Tonder is right and the authorities ought to be aware of the risk he claims to have identified. Maybe there ought to be a debate. On the other hand, maybe Grealy has a point.
Either way, the alarmist and uncritical headlines designed to influence public opinion and government policy are out there. The harm has been done. By the time other scientists get an opportunity to even read Van Tonder’s paper, what’s the bet their verdict will be moot? At best, it will merit a small correction among the cheap adverts in the back of the paper. DM