Australasian Science: Australia's authority on science since 1938

The Immediate Risks of Gas Production to Water Resources

Credit: Mantis Design/Adobe

Credit: Mantis Design/Adobe

By Margaret Shanafield & Craig Simmons

Public concerns about unconventional gas production focus on contamination of aquifers deep below the surface, yet the most immediate risk to water resources is right before our eyes.

The term “unconventional” gas refers to natural gas trapped in geology with low permeability that cannot be extracted using the formation’s natural pressure to push the gas upwards in the production well. Hydraulic fracturing or other advanced technologies are therefore required to recover an economically viable quantity of gas from these formations.

Because a large part of the world’s remaining unconventional gas reserves are located in countries that have traditionally been gas importers, the increasing economic viability of these resources could allow these countries to gain increased gas independence. However, the very techniques that have allowed this expansion to take place have also met with widespread contention and public scepticism.

In response, controversial bans on hydraulic fracturing in some US and Canadian states – and even countries such as France, Germany, Tunisia, Ireland and Scotland – have pitted environmental and human protection groups against petroleum giants. The potential impacts to water resources, most notably groundwater contamination, and possible consequences to human health and livestock are at the heart of this controversy.

There have been countless reports proposing potential impacts to water resources, but what does the science actually tell us about the probability of these impacts occurring?

Quantifying the Impacts of Unconventional Gas Production on Water Supplies

It’s difficult to put a probability on the impacts of an unconventional well on surface waters and shallow aquifers. This is due to several factors, including lack of data to fully characterise the subsurface geology in the areas of interest, ongoing changes in the technology used for drilling, casing and completing wells, difficulties in assigning pathways to sub­surface impacts, and difficulties in compiling and making sense of regulatory information that has been collected.

Still, we need to be scientific; we need to be quantitative and we need to be evidenced-based and clinical. A recent meta-analysis of data from international scientific reports and papers showed that, despite the public focus on subsurface effects from hydraulic fracturing, the likelihood of contamination from spills at the surface is orders of magnitude higher (https://goo.gl/JeyfFV). For example, reviews of environmental violations in the Marcellus Shale in North America point to spill rates of around one in ten for any given well over its lifetime, and an order of magnitude lower for spills greater than 1500 litres. Our most educated guess at the likelihood of contamination of groundwater via subsurface pathways suggests that the probability of contamination leaking into the aquifer from the surface is between three and five orders of magnitude greater than any other potential water resource impact. This scientific finding is in line with other emerging reports by the EPA and others, but incongruent with current public and political outrage about the issue of impacts from unconventional gas.

Surprisingly, little attention is given to the high probability of problems such as surface spills. Is the fear of what lies beneath eclipsing the obvious issues that are right before our eyes, even though they should be preventable with improving technology and engineering, and tighter regulation?

The Black Swan in the Room?

Our aim here is not to discount potential subsurface problems. There is no such thing as zero risk. Although several studies linking aquifer contaminants with nearby gas production activities have been criticised for various assumptions or limitations, increasing evidence shows that, under the right conditions, fluids can escape from the targeted production formation and travel long distances.

The ground beneath us is composed of many layers of different rocks, soils and clays of varying quality, as well as liquids such as water and gas. Liquids travel at different speeds through all of these layers, and can get held up by less permeable layers called “aquitards”. Understanding the permeability, spatial continuity and location of cracks and holes in these aquitards is critical for quantifying the potential for contaminants to reach potable aquifers.

Using the range of values for an aquitard from previous studies, and applying this understanding of layers and aquitards, the probability of gas or fracking fluids travelling 2000 metres upwards from a production formation during several hours of hydraulic fracturing in a single well is one in 108. (You are almost ten times more likely to get eaten by a shark in any given year.) If the fluid only needs to travel 100 metres upwards to reach an aquifer, this probability goes up two orders of magnitude to one in a million. (Your odds of winning the Australian lottery are about one in 45 million.)

However, it is rare that a production formation this close to an aquifer would undergo hydraulic fracturing, especially because production formations this close to the surface are typically more permeable.

Another possibility for fluids to migrate in the ground stems from the fact that there might already be holes in the ground from previous oil and gas production activities, and pressurising the subsurface in the vicinity of these wells can cause fractures that permit the flow of fluids through unwanted pathways. This is where Australia is quite lucky to be a relative newcomer in natural gas production. We have very few existing wells, drastically reducing the likelihood of problems caused when natural or induced fractures allow fluids to migrate between new and old wells. In contrast, recent studies of Pennsylvania, where there is a long legacy of gas production, suggest there could be up to half a million abandoned wells in that state alone.

The final possibility for groundwater contamination to occur due to underground movement of fluids is something we need to consider for our future – the risk of leakage after production ceases. The lifespan of a gas production well is, at best, around 30 years, and typically shorter. After production ceases, these wells are capped or cemented in to prevent leakage. All cement breaks down over time, although it may take 1000 years or more.

Due to this long timescale, combined with estimates of the reduction in pressure that would take place after removing gas from underground over 20 years, we estimate that the probability of leakage over 2000 metres is less than one in a million for any given well (if proper end-of-life procedures are undertaken). However, for typical coal seam gas formation depths (~200 metres), the likelihood is closer to one in 400. Monitoring the integrity of all wells will be important for years after production has ceased.

Given these probabilities, subsurface pathways between production formations and aquifers could provide the theoretical “black swan” event, an occurrence of very low probability but of dire consequences if it does take place without anyone knowing it has happened. Once contamination reaches an aquifer it is difficult to remediate and impossible to fully remove.

The risks and benefits of gas production in individual regions are therefore beyond the scope of scientific determination, other than to quantify likelihood with as much information as possible by accurately mapping faults, fractures, existing wells and geologic conditions. It is clear that to minimise this risk we must use the best available well construction technology, understand the geology (faults and fractures and permeability), and monitor. Given the best possible data, it will still be extremely difficult to understand exactly what is going to happen at depths of several kilometres, and there will probably be some areas where risks are too high to proceed.

Directing Regulation and the Public Dialogue

Unfortunately, science doesn’t appear to be informing an evidence-based debate. There appears to be a significant disparity between the scientifically evaluated risk and the community’s perception of risk, which has not focused on surface spills – the single greatest risk exposed by the scientific evidence. We cannot ignore the potential for groundwater contamination, and while there are few scientifically documented cases where unconventional gas production has caused aquifer contamination through subsurface travel, absence of evidence is not the same as evidence of absence, especially given the long time scales involved in transport of contaminants in aquifers. We contend that increased attention to what is going on at the surface offers immediate benefits for humans and the ecosystem that supports us.

Finally, we should not underestimate the socioeconomic impacts on communities, even if no contamination occurs. These are outside the scope of our work, but can be as strong a consideration in whether or not to develop a region. Science is necessary but highly insufficient to gain a social license to operate.


Margaret Shanafield is a senior researcher at Flinders University, and Craig Simmons is the director of the National Centre for Groundwater Research and Training.