Roadshow 2.1 Water Quality monitoring in the Beetaloo GBA region

Hi, everyone. My name is Lisa Golding from CSIRO Land and Water. On behalf on my co-researchers listed here, I'll be presenting research on the risks to aquatic life from hydraulic fracturing chemicals and flowback wastewaters, as part of the water quality program.

We had four objectives in this work. Objective one focused on a risk assessment of the hydraulic fracturing chemicals for the protection of freshwater biota. In objective two, we identified and quantified the organic, inorganic and radionuclide chemicals and elements that returned in the flowback water following hydraulic fracturing. In objective three, we conducted a direct toxicity assessment of the stored flowback wastewaters to determine the safe dilutions required to protect 95% of freshwater biota. Finally, in objective four, we modelled the unlikely scenario of a spill or seepage from a flowback wastewater tank and whether it would pose a risk to freshwater biota via transport through soil and groundwater.

So, today, I'll highlight some of the key messages from each objective and you'll see the pending publication listed at the bottom of each slide that you can refer to for more detail. For all objectives, we focused on two wells in the Beetaloo Sub-basin, located southeast of Katherine, the Kyalla and Tanumbirini wells. Now, these wells differed in their well type, with Kyalla being a horizontal well and Tanumbirini being a vertical well. They also differed in the depth and importantly, the target shale formations they accessed.

So, in objective one, we focused on the risk of hydraulic fracturing chemicals to freshwater biota. The chemicals make up a small percentage of the total hydraulic fracturing fluid volume, but because the volumes of fluid are large, the chemical concentrations can be high. We conducted a qualitative screening risk assessment based on literature values for the persistence by accumulation and toxicity of each chemical and categorised them as being low, potential and potentially high concern to freshwater biota. We also calculated the risk quotient for each chemical in the fluid based on the estimated downhole concentrations as the predicted environmental concentrations, or PEC and literature derived values for the predicted no-effect concentrations, or PNEC.

So, based on the qualitative screening risk assessment, we found that for the 42 chemicals used in the two wells combined, 30 were common to both wells, 10 were unique to Tanumbirini and two chemicals unique to Kyalla. The 42 chemicals were fairly evenly split between the categories of low, potential and potentially high concern. However, when the risk quotient approach was applied to these same chemicals, 63 to 75% of the chemicals were of potentially high concern and a total of 14 chemicals from both wells had risk quotients greater than 1,000. Some examples of potentially high concern chemicals were the biocide TTPC, hydrochloric acid and the ethoxylated alcohols.

Now, these risk quotients are conservative because they assume no attenuation processes or mitigation controls that will reduce their bioavailability to aquatic organisms. So, while there was a high inherent risk of effects to aquatic organisms, once the control mitigation measures were incorporated, there was a low residual risk.

In objective two, we chemically analysed the organics, inorganics and radionuclides in the flowback waters collected at different time points and there were two main trends in the flowback of chemicals over time, shown in these representative figures. For those chemicals where an initial peak occurred, followed by a decline over time, they were thought to be associated with the hydraulic fracturing fluid initially returning to the surface. For those chemicals that had a slow increase over time, they were thought to be gradually released from the target formation and of geogenic origin.

So, we know that the chemical concentrations in the flowback were dynamic and changing with time. We also know that the concentrations differed between the wells. Kyalla flowback was more saline and had higher metal and radionuclide concentrations than Tanumbirini and this demonstrates the site specific nature of risk assessment required. Contaminants of concern were identified when hazard quotients were greater than one, for example, with boron and zinc. One of the key points identified from the work was that many organic compounds could not be identified and remain undescribed. These require specialised methods to be developed and research labs to identify and quantify them.

In objective three, we addressed the potential risk from unidentified chemicals and the integrated effects from chemical mixtures in flowback wastewaters by using direct toxicity assessment. You can see here how the wastewaters differed in appearance. Just to add a reminder that freshwater organisms are highly unlikely to be exposed to flowback wastewaters in the Northern Territory because discharge to surface or groundwaters is prohibited and mitigation measures are in place for accidental release.

So, we used long-term chronic exposures to represent long-term seepages as the worst case scenario. Salinity controls were used to determine the contribution of hypersalinity to observe toxicity and eight test species were used to represent tropical and temperate climates, three trophic levels and seven taxonomic groups.

Both of the untreated flowback wastewaters were chronically toxic to all eight freshwater species, with adverse effects at the 10% level, known as the chronic EC10, occurring at dilutions as low as 0.084 to 36% of the flowback. The Kyalla flowback wastewater was significantly more toxic than Tanumbirini by fourfold and also had the highest salinity, metal and radionuclide concentrations.

We were able to identify or quantify the contribution of hypersalinity to the overall toxicity as being 16% and 55% for the Tanumbirini and Kyalla wells, respectively. Therefore, toxicity is not due to salinity alone, but likely due to a mixture of chemicals, including the unidentified organic compounds. Using sensitivity distributions, shown here, we derived the safe dilution, or PC95, for each flowback wastewater that would protect 95% of freshwater species as 1 in 1140 and 1 in 300. These can then be used as targets for managing these wastewaters in the unlikely event of spills and seepages.

Objective four enabled us to use the safe dilution targets from objective three and the measured water chemistry data for the flowback in objective two, to model the fate and risk of a spill or seepage scenario of wastewater from both wells. The scenario was a 0.1 megalitre spill of flowback wastewater from a large storage tank into the soil and groundwater to a receptor 500 metres from the source.

Now, a reminder that this scenario is highly unlikely given the control mitigation measures that are in place. Attenuation processes such as dilution, dispersion, biodegradation, adsorption and radioactive decay were applied in the model in a stepwise manner. The predicted dilution factors of the flowback wastewaters at the receptor were compared with the PC95 dilutions factors derived from the direct toxicity assessment in objective three and used in separate risk quotients to determine the risk to freshwater biota.

So, under the modelled scenario, all analytes in Tanumbirini had a risk quotient less than one, so were not a concern to aquatic life. Only conductivity had a risk quotient greater than one at Kyalla once all attenuation processes were included. This was due to the unretarded flow of chloride through the soil that could not be attenuated. The risk quotients based on individual chemicals alone, shown by the blue bars, underestimated the risk by 1.8 to 2.5 times, compared to the risk quotients based on the direct toxicity assessment safe dilutions, shown in the red bars, meaning that DTAs deliver a more robust quantification of risk. The incorporation of natural attenuation processes into the model, shown by the green bars, did markedly reduce the risk ratio for most analytes. These attenuation processes were dependent on time, with more time required for attenuation of inorganics and radionuclides than organics.

So, the information from these studies will inform the management of hydraulic fracturing chemicals and flowback wastewaters and I recommend you refer to these publications for more details. I'm happy to take your questions. Thank you.