Developing conceptual models is a critical step in hydrogeological studies that should utilise multiple lines of evidence and data types to minimise conceptual uncertainty, particularly in data-sparse systems. This study used new and existing major ion and isotope (O, H, Sr, C) data sets to refine a previous hydraulic-head-based conceptual model of the Galilee Basin (Australia). The analyses provide evidence for the locations of recharge and discharge areas and determine hydrochemical processes along flow paths to improve understanding of potential source waters to the Doongmabulla Springs Complex (DSC) and to infer mixing within, or exchange between aquifer units. There was good agreement between previously inferred recharge and discharge areas defined using hydraulic head data and interpretations from hydrochemical evolution along groundwater flow pathways, at least where data were available. Major ion and isotope data suggest that the DSC likely receives water from both a relatively shallow, local flow path and a deeper regional flow path. This observation is relevant to previous concerns about threats to the DSC, as mine-induced drawdown may impact the relative contributions to spring discharge from different recharge sources and aquifers. Silicate weathering in the deeper Clematis Formation and Dunda Beds, and evapotranspiration in the overlying Moolayember Formation have strong control on the total dissolved solids content. These findings suggest that the Clematis Formation and Dunda Beds are hydrochemically distinct from the Moolayember Formation, with limited exchange between these aquifers, which has important implications for model conceptualisation and ongoing monitoring of mining activities in the Galilee Basin.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Introduction
Developing conceptual models is a critical step in hydrogeological studies that should utilise multiple lines of evidence and data types, particularly in data-sparse systems. Most conceptual models have relied solely on groundwater elevations or geophysics (Enemark et al. 2019) and there are few examples of conceptual models that are informed by geochemical data. However, groundwater geochemistry provides valuable auxiliary information that helps define groundwater flow systems, determine the location of recharge and discharge areas, and help constrain interaquifer mixing (e.g., Carucci et al. 2012; Edmunds 2009; Han et al. 2014). Comprehensive conceptual models of the Galilee Basin (Queensland, Australia) are urgently needed given that groundwater monitoring is extremely sparse and the commencement of coal mining in this region with additional coal and gas projects undergoing regulatory approval processes (e.g., Evans et al. 2018b). These developments include the controversial Carmichael Coal Mine, an open-cut and underground coal mine covering a proposed area of ~28,000 ha (Currell et al. 2020). Concerns have been raised about potential adverse impacts from the Carmichael Coal Mine on the regional groundwater flow regime, rivers and streams, and the ecologically and culturally significant Doongmabulla Springs Complex (DSC; Currell et al. 2017, 2020; Keegan-Treloar et al. 2021). There are several alternate conceptual models of the regional flow system that have been presented in high-profile court cases, with ongoing disagreement as to which is the most likely (Currell et al. 2017). The Galilee Basin is remote, and until recently its subsurface geology and groundwater system remained largely untouched by anthropogenic activities. As such, existing knowledge of the hydrochemistry of the basin is based on limited data, mostly in mine lease areas along its eastern boundary. Given the limited understanding of the regional hydrogeology and the potential impact of coal mining and gas developments on the water resources, there is a clear need to improve the knowledge of the regional hydrogeology of the basin, including conceptual models of the primary hydrogeochemical processes (Fensham et al. 2016; Lewis et al. 2018).
The understanding of the hydrochemistry of the Galilee Basin is based mainly on the studies of Moya et al. (2015, 2016). Moya et al. (2015) examined the hydrochemistry of the Galilee and Eromanga Basins using multivariate statistical techniques. They observed that water in the main Triassic aquifer, the Clematis Formation, evolved from Na–Cl type in recharge areas to Na–HCO3 type with increasing distance from the eastern margin of the Galilee Basin. This was hypothesised to be due to carbonate dissolution occurring along flow paths, which tend to originate from sediment outcrop areas in the east. Moya et al. (2016) built upon this conceptualisation using environmental tracers and isotopes (i.e., δ2H, δ18O, δ13C, 36Cl, 87Sr/86Sr) to assess interaquifer connectivity and groundwater age along flow paths in the Galilee and Eromanga Basins. In the western area of the Galilee Basin, analyses of methane gas concentrations in groundwater suggested that there is hydraulic interconnectivity (likely due to faulting) between the deeper, Permian Betts Creek Beds and the shallower, Jurassic Hutton sandstone units. Moya et al. (2016) concluded that faults may play an important role in interaquifer connectivity in the central Galilee Basin.
Advertisement
The development of conceptual models is an important step in understanding regional groundwater flow processes and assessing the potential impacts of human activities on groundwater-dependent ecosystems. Keegan-Treloar et al. (2021, 2023) developed a conceptual model of the regional groundwater flow and recharge and discharge areas for the Triassic aquifers of the Galilee Basin using hydraulic head data and geostatistical methods. Data sparsity in the western areas of the Galilee Basin limited the interpretability of flow conditions over much of the basin. As the conceptual model presented by Keegan-Treloar et al. (2023) was based solely on head data, there is a need to extend and compare the interpretations by incorporating other data types, such as hydrochemical information. This study aims to improve and refine the hydrogeological conceptual model of the Triassic aquifers by examining hydrochemical evidence of recharge areas, the hydrochemical evolution of waters along flow paths, and the consistency of aquifer chemistry with discharge from springs. Furthermore, system understanding obtained from hydrochemical data will be compared with the previous hydraulic head-based study of Keegan-Treloar et al. (2023) to assess the validity of the hydraulic head-based recharge/discharge area analysis. The updated conceptual model aims to improve the understanding of the Triassic groundwater flow system by reducing the uncertainty in locations of regional recharge and discharge, which is critical for the protection of groundwater-dependent ecosystems, including the DSC, given that the Triassic aquifers are likely an important water source (Keegan-Treloar et al. 2021). Improved characterisation of the recharge/discharge areas and hydrochemical evolution along flow paths is required to inform management decision-making in terms of the mining and coal seam gas activities that are planned or underway in the basin. While this study is based in the Galilee Basin, the methodology used here (especially, the use of geochemistry) is applicable to constructing and assessing conceptual models more broadly.
Methods
Study area
The Galilee Basin contains a sequence of Permian and Triassic sediments that directly underlie the Jurassic and Cretaceous strata of the Eromanga Basin (Fig. 1). The Triassic-Cretaceous units of both basins are part of the Great Artesian Basin (GAB) (Habermehl and Lau 1997; Kellett et al. 2003; Moya et al. 2016). Groundwater recharge to the Galilee Basin is generally believed to occur to the east where the basin sediments outcrop (Fig. 1). This is supported by groundwater 36Cl age estimates that increase along the inferred groundwater flow path from east to west (Moya et al. 2016). However, hydraulic head data (Fig. 1c) indicate that some parts of the basin margins may act as discharge zones, and localised recharge may be occurring to the west.
Fig. 1
a Location of the Galilee Basin and the Great Artesian Basin (GAB), Australia; b The Galilee Basin and surrounding geological basins within the study area (red rectangle); c The study area marked with wells from the Clematis/Dunda formations (black triangles) and Moolayember Formation (black rectangles), hydraulic head contours (blue lines, from Keegan-Treloar et al. 2023), and the location of the DSC (blue circles). The proposed extent of the Carmichael Coal Mine is shown in dark yellow. The major surface features and the extent of the Galilee and Eromanga Basins are labelled
×
The study area is semiarid with a mean annual rainfall of ~543 mm/year (Bureau of Meteorology station No. 35264). The topography is mostly subdued with a topographical high along the eastern margin of the Eromanga Basin, which forms a boundary for the catchments draining westwards to the large salt lakes, Lake Galilee and Lake Buchanan (Evans et al. 2018b). Creeks and rivers are mostly ephemeral, except for reaches of the Carmichael River that are believed to receive baseflow from the Triassic aquifers and the nearby DSC (Currell et al. 2017; Evans et al. 2018b). The DSC consists of permanent freshwater springs, supporting ~160 wetlands (Currell et al. 2017). These wetlands are critical habitat for endangered species, including the Black-Throated Finch and Waxy Cabbage palm, and collectively, the springs are considered of high cultural significance to the Wangan and Jagalingou People (Wangan and Jagalingou Family Council 2015; Currell et al. 2017).
Hydrogeology
The stratigraphy of the study area includes the Triassic Moolayember, Clematis, Dunda Beds and Rewan Formations, which overlie the Permian Betts Creek Beds and Joe Joe Group (Fig. 2). Overlying these are Tertiary and Quaternary sediments comprising alluvium, basalt flows and other undifferentiated sediments. The Moolayember Formation is composed of mostly siltstone and mudstone and is considered a low permeability unit (Jiang 2014). The Clematis Formation and Dunda Beds are considered the main Triassic aquifers and are comprised of sandstones, siltstones and mudstones (Van Heeswijck 2006). There is likely a high degree of connectivity between the Clematis Formation and Dunda Beds, although several nested groundwater wells completed in each unit show that there is some hydraulic separation. These aquifer units are underlain by the Rewan Formation, consisting of mudstone, siltstone and sandstone (Jiang 2014). The Rewan Formation is considered to be a low permeability unit with an average thickness of 160 m (Evans et al. 2018a). Beneath the Rewan Formation are the Permian Bandanna and Colinlea Formations, which are grouped as the Betts Creek Beds and are considered to be aquifers. The lithology of the Betts Creek Beds includes sandstone, siltstone and coal with several interbedded tuffs (Allen and Fielding 2007; Phillips et al. 2018). Notably, the coal seams of the Betts Creek Beds are the target of several active and proposed mining and coal seam gas operations (Currell et al. 2017, 2020). Beneath the Betts Creek Beds is the Joe Joe Group, which is composed of mudstones, siltstones and sandstones.
Fig. 2
The stratigraphy, lithology and thickness (mean, minimum and maximum) of the Triassic and Permian Formations of the Galilee Basin. Adapted from Allen and Fielding (2007) and Evans et al. (2018a)
×
Advertisement
Keegan-Treloar et al. (2023) suggested that groundwater flow in the Clematis Formation is largely sourced from recharge areas in the north-east and south-east of the Galilee Basin, where the Clematis Formation outcrops. Thus, groundwater flows in the direction of either the DSC/Carmichael River or to the west towards the Galilee Basin boundary where the basin sediments are in contact with faults and/or the overlying Eromanga Basin (Evans et al. 2018b; AECOM 2021; Keegan-Treloar et al. 2021; see Fig. 1). There is evidence of localised recharge along much of the eastern margin of the Galilee Basin, although the primary areas of recharge appear to be in the north-east and south-east (Evans et al. 2018b; Keegan-Treloar et al. 2021). In the DSC/Carmichael River area within the central-eastern part of the Galilee Basin, the potentiometric surface in the Triassic Formation aquifers is ~10 m higher than that in the underlying Permian Formation aquifers, with hydraulic heads in the order of ~240 and ~230 m AHD (metres above Australian Height Datum), respectively (Keegan-Treloar et al. 2021). As such, the Triassic aquifers have a higher likelihood than the Permian formations of having adequate hydraulic head to meet a threshold required to support spring flow at the DSC (Keegan-Treloar et al. 2021). However, the Permian formations cannot be ruled out of having hydraulic connection with the springs as there is very limited hydraulic head data or groundwater monitoring infrastructure with reliable measurements to the west of the Carmichael Coal Mine lease area. Additionally, the extent of groundwater discharge from different depths/units in the Triassic (e.g., shallow Clematis vs. deeper Dunda Beds) remains unclear.
Data sources
The current study collated hydrochemical and isotope data for the Galilee Basin from the published literature (Moya et al. 2016), industry reports (AECOM 2021; EMM 2021), government reports (Department of State Development 2018), environmental data reporting by Bravus Mining (formerly Adani Mining Group; Bravus Mining 2022) and the Queensland Groundwater Database (Queensland Government 2022). Where data were duplicated between reports/papers and the Queensland Groundwater Database, the Queensland Groundwater Database was selected as the preferred data source given that the data are in a consistent format and the database has been subjected to established quality control measures. It is worth noting that the attributed stratigraphy to some of the groundwater well production zones in the database warrants further investigation and review, particularly those wells completed at the transition between the Clematis and Moolayember formations.
Hydrochemistry and isotope data are available from groundwater monitoring wells completed in the Clematis Formation, Dunda Beds and Moolayember Formation, mainly along the eastern margin of the Galilee Basin. The available data become increasingly sparse towards the west as the depth of the Galilee Basin sediments increases beneath the Eromanga Basin. In this study, the Clematis Formation and Dunda Beds were considered as a single unit owing to data sparsity in the region. This point is further supported by similarities in hydraulic heads in these units (Keegan-Treloar et al. 2023). In total, 329 hydrochemistry samples were available from 34 wells in the Clematis Formation/Dunda Beds and 23 samples were available in the Moolayember Formation from 12 wells. The locations of these wells are shown in Fig. 1c. A single value was chosen for each sampling site by selecting the most recent sample with a charge balance error of less than 5%. Where no samples with a charge balance error less than 5% were available, the sample with the lowest charge balance was selected subject to selection criteria outlined in the following sections. The relevant dataset is provided in the electronic supplementary material (ESM). The choice to represent each well with a single measurement was considered reasonable because the total dissolved solids (TDS) remained approximately constant over time where time-series data were available. Some temporal variability in groundwater hydrochemistry from the basin is evident upon close inspection of the data; however, this was considered of limited relevance to the overall conceptualisation and beyond the scope of this study.
In addition to the data collated from other sources, samples were collected in 2019 and 2021 from Camp Spring, one of the main springs in the DSC. Camp Spring has a discrete discharge vent (in contrast to most other springs that are submerged beneath pooled water), and therefore spring water can be easily collected with minimal impacts from rainfall or evapotranspiration. This provided representative samples of the groundwater contributing to one of the major DSC springs. Further studies of hydrochemical-isotope indicators in additional springs in the DSC are ongoing as part of a larger research project.
Data quality and geochemical analyses
Bicarbonate (HCO3) and carbonate (CO3) concentrations were calculated from alkalinity (as CaCO3) and pH using PHREEQC (Parkhurst and Appelo 2013). TDS values were calculated based on the sum of major cations and anions. The electrical balance was calculated for all samples. Without compromising the data integrity of the sparse data set, several samples that had an electrical charge balance error greater than 5% were retained but were subjected to additional scrutiny. Where the error was due to missing analytes (e.g., alkalinity), the sample was censored in analyses dependent on the missing analyte (e.g., total dissolved solids, HCO3), but included where the analysis was not influenced by the missing analyte (e.g., Na/Cl ratios). Mineral saturation indices were calculated with PHREEQC using the available hydrochemistry data (Parkhurst and Appelo 2013).
Isotopic data
Stable isotopes of water (δ2H and δ18O), strontium isotope ratios (87Sr/86Sr) and δ13C were collated from datasets provided by Moya et al. (2016) and the Queensland Herbarium (2020). There was a total of 15 samples from the Clematis Formation/Dunda Beds, 8 samples from the Moolayember Formation and 1 sample from Camp Spring. Tables of isotope data are provided in the ESM.
As none of the rainfall stations are collecting stable isotope data in the Galilee Basin, the local meteoric water line (LMWL) from the nearest site, Charleville (IAEA/WMO 2021) was used. Although Charleville is ~400 km south of the Carmichael Coal Mine, it has a similar climate and is located a similar distance from the coastline, making it a logical choice to compare stable isotope data from the Galilee Basin (Moya et al. 2016; Hollins et al. 2018). The LMWL for Charleville is δ2H = 8.1 δ18O + 12.6, which has a similar slope to the global meteoric water line (GMWL) (Craig 1961).
Results
Major ion chemistry
Figure 3 shows the hydrochemical data together with a constructed map of likely groundwater recharge and discharge areas based on an analysis of the potentiometric surface in the study area by Keegan-Treloar et al. (2023). The hydraulic head data represent current hydraulic conditions (i.e., at the time of measurement), and can be used to inform the locations of possible recharge/discharge areas. In contrast, the hydrochemistry represents an integration of various physical and geochemical processes occurring over longer times and groundwater flow paths.
Fig. 3
a Spatial distribution of total dissolved solids (TDS) in the Clematis Formation and Dunda Beds (triangles) and Moolayember Formation (squares). Note the same colour gradation is used for both formation groupings both. a Black triangles were missing HCO3, which influences TDS. The colour bar in the lower left shows the proportion of realisations where the potentiometric surface indicated likely recharge (blue) or discharge areas (red) from Keegan-Treloar et al. (2023). Yellow shading identifies regions where the Clematis and Dunda Beds outcrop. Green-shaded areas are the mine footprint. b Inset map of the DSC region outlined in black (a). c Cross section of the transect A to A′ shown in (b) displaying TDS trends with depth and the hydraulic head contours (blue lines and values) that indicate groundwater discharge towards the DSC and the Carmichael River
×
Figure 3a shows proposed recharge (blue) and discharge areas (red) from Keegan-Treloar et al. (2023), who examined the curvature, in terms of concavity, of different equiprobable realisations of the potentiometric surface for the Triassic aquifers (Clematis Formation/Dunda Beds). Spatially, groundwater has low TDS in the southern outcrop areas, with TDS increasing to the west (see Fig. 3a). In the northern area where the Clematis and Dundas Beds outcrop, there is low TDS groundwater (223 mg/L), with TDS generally increasing towards the DSC. In the region of the DSC and Carmichael River (outlined in black Fig. 3a), TDS values range from 121 to 866 mg/L (Fig. 3b). An examination of the screen depths of these bores indicates that TDS is generally higher (474–866 mg/L) in the deeper groundwater (180–234 m below ground) compared to the TDS (121–186 mg/L) in the shallower groundwater (79–123 n below ground). Camp Spring had a TDS of 417 mg/L, which is similar to the TDS observed in three wells (413, 411 and 486 mg/L) near the DSC (Fig. 3c). This suggests the spring waters are consistent with either a mixture of groundwater from different depths or from the mid-depth of the Clematis Formation/Dunda Beds.
The groundwater in the Clematis Formation/Dunda Beds aquifer varies between Na–Cl and Na–HCO3 types (Fig. 4). In comparison, samples from the Moolayember Formation are entirely Na–Cl type and the sample from Camp Spring is Na–Cl type but has a higher proportion of HCO3 than the majority of other Na–Cl dominated groundwater samples.
Fig. 4
Piper plot of the hydrochemistry in the Clematis Formation/Dunda Beds (orange triangles), the overlying Moolayember Formation (red rectangles) and Camp Spring (green circle)
×
The TDS of groundwater in the Clematis Formation and Dunda Beds ranges from ~100 mg/L to ~900 mg/L and the molar Na/Cl ratios increase from ~0.8 to 3.5 with higher TDS, albeit without a clear linear trend (Fig. 5a). The Moolayember Formation has higher TDS and Na/Cl ratios that are approximately constant, with a median value of 0.82. Evapotranspiration is the likely cause of the trend of constant Na/Cl ratios with increasing TDS (e.g., Herczeg and Edmunds 2000) given that the average Na/Cl ratio in this trend (0.8–0.9) matches that of rainfall (0.9 at Charleville). Halite dissolution can also cause this trend, but is unlikely, as halite has not been recorded in bore logs of the Moolayember Formation (e.g., Moya et al. 2016; Evans et al. 2018b). Halite dissolution would also decrease the HCO3/(HCO3 + Cl + SO4) ratios in Fig. 5b as it adds Cl but none of the other anions. The Na/Cl ratios for the Clematis Formation and Dunda Beds mostly exceed 1 at low TDS values, and the ratio of HCO3/(HCO3 + CO3 + Cl + SO4) increases with higher Na/Cl ratios (Fig. 5b). Both these trends could be due to weathering of Na-silicates like the feldspar albite, which releases Na but not Cl or SO4 (Eq. 1). Camp Spring has a Na/Cl ratio of 1.3 and HCO3/(HCO3 + CO3 + Cl + SO4) ratio of 0.35, also suggesting that weathering of Na-silicate minerals is influencing the composition of the groundwater that is discharging at the DSC (Fig. 5a, b; after Herczeg and Edmunds 2000).
Fig. 5
Molar major ion ratios for samples from the Clematis Formation and Dunda Beds (orange triangles), Moolayember Formation (red squares) and Camp Spring (green circle) showing a Na/Cl ratios versus TDS and b HCO3/(HCO3 + CO3 + Cl + SO4) ratios versus Na/Cl ratios. Note the log scale on the x-axis (a). The evapotranspiration (ET) and weathering trend lines follow a similar approach to that described by Herczeg and Edmunds (2000)
Additional evidence for Na-silicate weathering comes from the correlation between pH and both HCO3 and Na in Clematis Formation/Dunda Beds groundwater (Fig. 6). Initial recharge (rainfall) is relatively acidic due to its high H2CO3 content. Equation (1) will cause a progressive increase in pH due to the conversion of H2CO3 to HCO3. Moya et al. (2015) also noted that Clematis Formation groundwater evolves from Na–Cl to Na–HCO3 downflow. Thus, mineral weathering is largely responsible for the rise in TDS along the flow path, as indicated in Fig. 5.
Fig. 6
Bicarbonate (circles) and sodium (crosses) versus pH for groundwater samples from the Clematis Formation and Dunda Beds
×
Na/Cl ratios and TDS in Clematis Formation/Dunda Beds groundwater are low in the outcrop areas to the north and also to the south, beyond the DSC and Carmichael River (i.e., 0.9–1.0 and 210–250 mg/L, Figs. 7a and 3a, respectively), indicative of limited evapotranspiration during relatively rapid recharge through the porous sandy soils that have developed on outcrops of these formations (predominantly sandstone). Na/Cl ratios and TDS in Clematis Formation/Dunda Beds groundwater generally increase from the north to the south-east and down the hydraulic gradient (Figs. 3 and 7), most likely because silicate weathering is adding Na and other dissolved species along this flow path. Na/Cl ratios in Moolayember Formation groundwater are 0.5–1.0 throughout the study area, suggesting that evapotranspiration during recharge is the key process influencing TDS (Fig. 7a). Moolayember Formation groundwater has a higher salinity than that in the Clematis Formation/Dunda Beds, reflecting a greater impact of evapotranspiration due to slower recharge through the clayey, less permeable soils that overlie this dominantly shaley formation.
Fig. 7
a Na/Cl ratios in the Clematis Formation/Dunda Beds (triangles) and the Moolayember Formation (squares). Colour of the markers denotes the range-category in Na/Cl ratios, while labelled values are the Na/Cl ratios. Wells with Na/Cl but no TDS measurements are unlabelled. Green-shaded areas are the mine footprint. b Inset map showing the black rectangle (a). c The transect A–A′ shown in (b) displays the Na/Cl ratio with elevation. Blue lines and values are hydraulic heads (m AHD) from Keegan-Treloar et al. (2023)
×
All the water samples are undersaturated with respect to anhydrite, halite and gypsum (Fig. 8a, e, f, respectively), although the saturation indices increase along the inferred groundwater flow paths, indicating that precipitation of these minerals is unlikely. Some samples, mostly in the northern and southern outcrop areas of the Clematis Formation/Dunda Beds, are saturated regarding calcite, aragonite, and dolomite (Fig. 8b–d, respectively). Although this could suggest that carbonate precipitation/dissolution is influencing groundwater composition, Sr isotope data (discussed below) indicates that this is unlikely. The SI values for Camp Spring are below saturation with respect to all the modelled minerals (mentioned previously) and within the range for groundwater in the Triassic aquifers.
Fig. 8
Mineral saturation indices in the Clematis Formation/Dunda Beds (orange triangles) and the Moolayember Formation (red squares) with the saturation index (SI) on the y-axis and the TDS on the x-axis (log-scale). a Anhydrite, b Calcite, c Aragonite, d Dolomite, e Halite and f Gypsum. Dashed lines indicate the limit of saturation (i.e., SI = 0)
×
Stable isotopes of water
The stable isotopes of groundwater in the Clematis Formation/Dunda Beds and the Moolayember Formation have the following ranges: δ2H = –47.8‰ to –23.1‰ and δ18O = –7.1‰ to –4.3‰ (Fig. 9a). All the samples (aside from well HD02) have lower δ2H and δ18O values than the weighted mean rainfall (δ2H = –28.5‰ and δ18O = –5.05‰), suggesting that the aquifers are likely recharged preferentially by seasonal rainfall with depleted isotopic signatures. This includes rainfall associated with tropical cyclone systems, in which evaporation is limited and the condensation efficiency is high (Hollins et al. 2018). Such rainfall systems may also be associated with an increased rainout effect before reaching the study area. A paleoclimatic signal (e.g., more depleted isotopic values associated with recharge during previous cooler geological epochs: Clark and Fritz 2013) also cannot be ruled out without further groundwater age information. There are also likely to be some differences between the groundwater and spring isotope data and the WMR due to the use of the LMWL from Charleville, which despite being the closest rainfall station with stable isotope data, is located ~600 km to the south of the study site.
Fig. 9
Stable isotopes of water. a δ2H versus δ18O for the Triassic units shown with linear lines of best fit (yellow indicates Clematis/Dunda Beds and red indicates Moolayember), GMWL, and LMWL and weighted mean rainfall (WMR; δ2H = –28.5‰ and δ18O = –5.05‰) for Charleville (Hollins et al. 2018). b δ2H versus Cl (log-scale) with annotated likely physical processes. Bore HD02 is labelled as this sample differs from other groundwater samples, with a higher δ2H and δ18O than the WMR
×
The samples from the Clematis Formation/Dunda Beds and Moolayember Formation both define trends with a lower slope (e.g., Moolayember Formation samples; δ2H = 4.8 δ18O + 11.2) than the LMWL and GMWL (Fig. 9a), suggesting that they have been influenced by evaporation (Coplen et al. 2000). The well HD02 was not included in the trend analysis (Fig. 9a) because it is situated alongside the Carmichael River and the enriched δ2H and δ18O values are likely indicative of infiltration of summer surface water from the Carmichael River. The sample from Camp Spring lies within the range of groundwater from the Clematis, Dunda Beds and Moolayember formations, suggesting a similar source of recharge.
The relative influence of evaporation and transpiration on the stable isotope data set was assessed using the relationship between δ2H and Cl (Fig. 9b). In the Clematis Formation/Dunda Beds, the groundwater samples become more enriched in 2H with increasing Cl until chloride concentrations approach 90 mg/L (Fig. 9b). At higher Cl concentrations, δ2H values are relatively constant. These trends likely represent the initial influence of evaporation during rainfall infiltration near the ground surface and subsequent transpiration prior to groundwater recharge. A similar, although somewhat noisier trend is observed in the Moolayember Formation.
Strontium and carbon isotopes
The two main sources of Sr in the groundwater within the Triassic aquifers are recharge (rainfall) and mineral weathering. Rainfall collected near the DSC in April 2021 had an 87Sr/86Sr ratio of 0.7125, which is greater than modern seawater (0.7090; Aberg et al. 1989). The 87Sr/86Sr ratio in rainfall typically increases inland from seawater values near the coast due to leaching of windblown dust (Raiber et al. 2009). 87Sr/86Sr in rainfall is also influenced by changes in climatic conditions and the predominant wind directions, which causes temporal variability in the 87Sr/86Sr ratios over geological time (McNutt 2000).
87Sr/86Sr ratios of groundwater in the Triassic aquifers in the study area are displayed spatially in Fig. 10a. These are mostly similar to the ratio in recent rainfall at the site (87Sr/86Sr = 0.7125). Many of these samples come from wells within the likely recharge areas to the south-west and south-east of the DSC. The trend of constant 87Sr/86Sr and Sr/Na ratios with increasing Sr concentrations in these groundwater samples is due to evapotranspiration from the unsaturated zone prior to groundwater recharge (Fig. 10b). Rock Sr isotope analyses in the area show that the Rewan, Clematis and Dunda Beds have ratios of 0.714–0.73 (AECOM 2021), and therefore, the more radiogenic samples reflect the input of Sr from silicate weathering (De Caritat et al. 2005). 87Sr/86Sr ratios generally increase with distance along the inferred groundwater flow paths, indicating the increasing influence of silicate weathering. The low Sr/Na ratios (Fig. 10c) reflect the release of Na from the weathering of Na-silicate minerals.
Fig. 10
87Sr/86Sr ratios in the study area. a Spatial maps of 87Sr/86Sr ratios in the Clematis Formation/Dunda Beds (triangles) and the Moolayember Formation (squares), and scatterplots showing b87Sr/86Sr vs Sr, c87Sr/86Sr vs Sr/Na, and d87Sr/86Sr vs δ13C. b–d Dotted lines denote the modern seawater 87Sr/86Sr ratio
×
The higher Sr/Na ratios and lower 87Sr/86Sr ratios (typically < 0.710) of some samples (Fig. 10c) could be an indicator of carbonate dissolution (De Caritat et al. 2005); however, the Sr/Na ratios are lower than expected for this source, and carbonate dissolution would be expected to lead to decreased 87Sr/86Sr ratios with increased δ13C. This trend is not apparent in the available data (Fig. 10d); furthermore, at the time of deposition (Triassic), the 87Sr/86Sr ratio of seawater was 0.7075–0.708 (McArthur et al. 2001). Thus, rainfall and therefore freshwater carbonates would have had a similar Sr isotopic signature. Dissolution of these carbonates could provide the 87Sr/86Sr ratio of the least radiogenic samples (0.706); thus, carbonate dissolution is not likely a dominant process along the groundwater flow path. Instead, the samples with the least radiogenic Sr isotope composition have probably interacted with the Betts Creek Beds, and have a lower 87Sr/86Sr ratio than rainfall and seawater (0.707; AECOM 2021) because of weathering of the interbedded tuffs that are present within this formation (Phillips et al. 2018).
The δ13C values in the groundwater probably reflect the relative proportions of C3 and C4 plants in the local vegetation. C4 plants are warm-season grasses, sedges and a few halophytic shrubs, and C3 plants are trees, most shrubs and herbs, and cool-season grasses (Küҫükuysal et al. 2012). The vegetation in the study area is probably dominated by C3 plants, particularly trees, but with a component of C4 warm season grasses, which are common in parts of Australia with strong seasonal rainfall (Hattersley 1983) like the study area. Groundwater beneath C3-dominated ecosystems typically has δ13C values from –10 to –18‰, but these will rise as the proportion of C4 plants increases (Mook 2001), and could easily reach –9‰, spanning the range of δ13C values encountered in the present study (Fig. 10d).
Discussion
Hydrogeochemical conceptual model
The interpretation of recharge and discharge areas by Keegan-Treloar et al. (2023) is annotated in Fig. 11 to indicate the primary hydrogeochemical processes occurring along the expected groundwater flow paths. In the recharge areas r1, r3 and r7 (Fig. 11), the groundwater samples from the Clematis Formation/Dunda Beds have a low TDS (< 250 mg/L), low Na/Cl ratios (< 1.5) and Na/Cl water type. Additionally, recharge areas r3 and r7, where 87Sr/86Sr samples were available, have low 87Sr/86Sr ratios (0.708–0.712) that reflect a signature similar to rainfall. Throughout the study area, the stable isotopes of water in the Clematis Formation/Dunda Beds are mostly consistent with preferential recharge during heavy rains with a relatively depleted isotopic signature; however, a contribution from palaeowaters is also possible (see Fig. 9a). Despite scatter in the stable isotope groundwater data set, the δ2H vs Cl trends indicate that evaporation has an initial effect on infiltrating water. As water infiltrates beneath the evaporation extinction depth, the dominant salinization process becomes water use by vegetation, as signified by a transpiration trend (Fig. 9b), as well as the addition of Na due to silicate weathering. Overall, these findings support the hypothesised recharge areas r1, r3 and r7. The significance of r5 is uncertain as it is the only bore in this area. It has the least radiogenic Sr isotope measurement, similar to rainfall, but a high Na/Cl ratio much greater than rainfall. Hypotheses for this result require further sampling and analysis. Similarly, there is insufficient data to assess the recharge areas r2, r4 or r6 in the western region of the study area.
Fig. 11
Regional-scale hydrogeochemical conceptual model of the study area in the Galilee Basin, showing the relative percentage of realisations where the hydraulic head surface indicates discharge (red) or recharge (blue). The descriptions labelled in blue identify potential recharge zones r1–r7 and potential discharge zones d1–d5 (as per Keegan-Treloar et al. 2023). Black arrows indicate groundwater flow directions, and hydrogeochemical processes affecting the chemistry along inferred groundwater flow paths and with increased residence times are labelled in red
×
The location of the diffuse groundwater discharge area (d2) likely represents water discharging to the DSC and the Carmichael River from a northern flow path (r1) and a combination of a regional flow path from r7 and more localised recharge from r3 from the south (although the data in this region are relatively sparse, making it hard to assess the importance of this flow path to DSC discharge). This is supported by the vertical stratification and discernible differences in TDS and Na/Cl ratios in the south (see Figs. 3c and 7c). The hydrochemical signature from Camp Spring appears to be a mixture of water from weathered and less-weathered rocks (see Figs. 4 and 5). Figures 3c and 7c indicate that this spring is fed by a combination of shallow and deeper groundwater from both the south and north. There are few groundwater wells within the vicinity of the likely discharge area d3, but based on surface features, d3 may discharge to Lake Galilee, which is saline. Discharge area d4 is based on a single bore and may be an artefact, in which case r3 and r7 represent different parts of the same area where the Clematis and Dundas Beds outcrop and where there is likely to be recharge. This infers a continuous flow path from r7 to d2 following the topographic gradient. There are trends of increasing TDS and Na/Cl ratio along this flow path. Groundwater close to the likely discharge area d5 has a high TDS and low Na/Cl ratios, suggesting that evapotranspiration is occurring along the inferred flow path from the recharge area r7. There is insufficient hydrochemical data available to assess the likely discharge area d1, located at the western margin of the study area.
Multiple lines of evidence suggest that the dominant weathering process along groundwater flow paths within the Triassic-age aquifers is silicate weathering. The increasing Na/Cl ratios and 87Sr/86Sr ratios with distance from the recharge areas imply that silicate mineral weathering has occurred along the groundwater flow paths. These findings differ from those of Moya et al. (2016), who suggested carbonate weathering was the dominant process. This finding is likely due to this study examining a wider range of analytes on a smaller spatial scale, enabling differentiation between carbonate and silicate weathering.
In contrast to the deeper formations, evapotranspiration appears to be the main process responsible for the TDS variation in the Moolayember Formation (see Figs. 5a and 9b), either during recharge, or due to tree water use in areas where the water table is shallow. The influence of evapotranspiration is likely to be highly variable (in space) because the landscape in this region is very diverse, ranging from dense dryland forest near the drainage lines, and open woodland and low shrub vegetation elsewhere. The high evapotranspiration rates likely mask weathering trends in the Moolayember Formation.
Implications
This study found that where hydrochemistry and isotope data were available, there was good agreement between the conceptual model of recharge and discharge areas from hydraulic head modelling (Keegan-Treloar et al. 2023) and those identified from the hydrochemistry and isotope data sets. This agreement reduces the uncertainty of the hydrogeological conceptual model for the Triassic aquifers in the study area of the Galilee Basin. As recharge and discharge areas inferred by the hydrochemistry data are time-integrated, the agreement with the hydraulic and hydrochemical data suggests that the regional groundwater flow system has remained largely stable over long timescales. This observation is perhaps not surprising given that prior to recent mine dewatering activities (commencing in 2019), there has been limited groundwater abstraction in the Galilee Basin.
Hydrochemical evidence (including relationships between TDS and Na/Cl ratios with depth) suggests that there may be a layered flow system where the discharge area d2 (see Fig. 11) receives a mixture of older groundwater from a longer, deeper flow path (likely from r7) that shows silicate weathering and more recent recharge from a shorter, shallower flow path (likely r3). Major ion chemistry suggests that the water discharging from Camp Spring is consistent with a mixture of the low TDS waters from the shallow southern flow path (r3) and the higher TDS waters from the south-eastern deeper flow path (r7) and the northern flow path from recharge area r1. This observation indicates that the shallow, southern flow from r3 may be integral to maintaining the low salinity of the spring discharge at the DSC. This is particularly important because the conceptual model upon which mine impacts have been determined presumes that the DSC are sourced only from the Clematis and Dundas Beds (Currell et al. 2017, 2020). If mining operations that are located to the east of DSC impact this shallower flow path, the salinity of spring discharge may increase together with a decline in aquifer pressure, potentially impacting the sensitive ecosystems receiving water from the DSC. Given the current and proposed mine-dewatering in the Galilee Basin, future studies that investigate how changes in the contribution to spring discharge from local and deeper regional scale flow paths impact the quality of spring discharge will be of major benefit.
As the hydrochemistry of the Moolayember Formation is distinctly different from that of the Clematis Formation and Dunda Beds (i.e., notably, the TDS in the Moolayember Formation is up to an order of magnitude larger than that of the Clematis Formation and the Dunda Beds), there is likely limited connectivity between the Moolayember Formation and the Clematis Formation/Dunda Beds. This observation provides justification to treat the Moolayember Formation as a separate unit from the Clematis Formation/Dunda Beds, which has important implications for future modelling of interconnectivity between the major aquifer systems. The Clematis Formation and Dunda Beds are hydrochemically indistinguishable using the available data, suggesting similar mineralogy and/or hydraulic connectivity between these units. Consequently, the importance of relative groundwater flow contributions from these two units to the DSC remains unknown.
Uncertainties and future directions
The scarcity of hydrochemical and isotopic data in the Galilee Basin remains an issue for the development of the hydrogeological conceptual model of this multilayered aquifer system, particularly in the western part of the study area. Data scarcity in the study area prevented the assessment of hypothesised recharge areas r2, r4 and r6 and the hypothesised discharge areas d1, d3, d4 and d5, arising from the hydraulic head investigation of Keegan-Treloar et al. (2023). Further groundwater sampling near the mine site is likely to inform any potential impacts from mine dewatering, whereas sampling to the west will better inform the regional hydrogeology. Ideally, future groundwater sampling would include major/minor ions, stable isotope, and age-tracer data. Sampling will likely occur as more mining and coal seam gas activities begin in the Galilee Basin; however, it is emphasised that there is an urgent need for further sampling campaigns throughout the Galilee Basin to properly characterise the baseline hydrochemistry and understand timescales and pathways of flow to the DSC prior to substantial modification of the flow regime through mining activity. Methods like those adopted in the current investigation will assist that effort.
In this study, limited data indicating the age/residence time of groundwater were available, which prevented the determination of age variations along flow paths in the study area. An improved understanding of the distribution of groundwater age would be valuable to assess recharge and discharge areas, and to identify water sources of the DSC and Carmichael River. Assessing the age along flow paths would also be useful to approximate groundwater velocities, which would inform future numerical modelling of the region.
The interaquifer connectivity between the Triassic and Permian units is expected to be significant where there is faulting across the Rewan Formation (e.g., Moya et al. 2015; Evans et al. 2018b). This study focussed solely on the hydrochemistry of the Triassic aquifers, where the Permian units were not examined due to poor data availability. Thus, it was not possible to investigate important questions concerning interaquifer connectivity between Triassic and Permian formations in this study. Future studies could utilise major and minor ion chemistry and vertical hydraulic head gradients to characterise interaquifer mixing. However, data sparsity remains a challenge in the Galilee Basin, and additional bores, particularly in the Permian units, would need to be installed. In the meantime, novel techniques, including inverse hydrochemical modelling from spring discharge or geophysical studies (i.e., as described in Keegan-Treloar et al. 2023), may improve the understanding of the relationships between the Triassic and Permian aquifers.
Conclusions
The hydrochemistry and isotopic data utilised in this study supported several of the proposed locations of groundwater recharge and discharge areas previously identified using hydraulic head data. The main recharge areas identified in this study are in the southern and northern outcrop areas, with localised recharge to the south of the DSC and Carmichael River and south of Lake Galilee. Evapotranspiration was apparent in the hydrochemical signature in the recharge areas, and weathering of silicate minerals was apparent along groundwater flow paths towards the discharge areas. These observations reduce the uncertainty in the regional conceptual model of recharge and discharge areas, although uncertainty remains due to data sparsity, particularly in the western region of the study area of the Galilee Basin. These uncertainties could be reduced through additional field sampling in the western regions of the Galilee Basin.
The hydrochemistry of the Moolayember Formation is distinctly different from that of the Clematis Formation and Dunda Beds, particularly regarding Na/Cl ratios and TDS. These differences in the hydrochemical signature have implications for how the Triassic aquifers in the region are characterised and considered in hydrogeological conceptual models and inferred flow and hydraulic connectivity between aquifers. The findings suggest that the Moolayember Formation should be treated as a separate unit. Treating the Moolayember as a separate unit from the Clematis Formation and Dunda Beds differs from previous conceptualisations that have treated the Triassic units as a single aquifer.
Vertical stratification of the hydrochemical signatures in the Triassic aquifers within the study area of the DSC and Carmichael River suggests a combination of local and regional-scale flow paths with differing salinities, further supporting the need for further investigation. The impacts of adjacent mining activities and mine-induced drawdown in these aquifers may influence the hydraulic gradients and the contributions of the local and regional-scale flow paths as well as the salinity of waters discharging to receiving environments. This finding has important implications for the spring-dependent DSC and Carmichael River ecosystems, where changes in salinity may impact the functioning of these ecosystems.
More broadly, this study highlights the utility of using hydrochemistry to constrain conceptual models. The previous conceptual model of the Galilee Basin, which is largely unmodified, was based on sparse hydraulic head information. The combination of hydrochemistry and head information allowed for a more robust understanding of recharge and discharge areas, and mixing between overlying hydrostratigraphic units than was revealed from head data alone. In regions where substantive existing modification has occurred (e.g., via groundwater abstraction), the longer response times of water constituents (relative to heads) allow hydrochemistry to be used to inform the predevelopment state of the system as an invaluable tool in creating conceptual models.
Acknowledgements
We would like to acknowledge Angus Campbell for early reviews of the manuscript and field data. We would also like to acknowledge Boris Laffineur and Rod Fensham for providing hydrochemical data to assist the research and the support from Derec Davies who represents the project partner organisation Business Services of Coast and Country Incorporated.
Declarations
Conflict of interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.