3.5.2 Ecological assets

Table 39 Summary of the number of ecological assets exceeding defined categories of modelled hydrological change for those assets that are in landscape groups where the hydrological response variable used for receptor impact modelling is groundwater drawdown within the zone of potential hydrological change

The assets potentially exposed to ≥0.2, ≥2 and ≥5 m additional drawdown is shown for the 5th, 50th and 95th percentile estimates of the maximum difference in drawdown (dmax) between the coal resource development pathway (CRDP) and baseline, due to additional coal resource development.

Data: Bioregional Assessment Programme (Dataset 1, Dataset 3)

Table 40 Summary of the number of ecological assets exceeding defined categories of modelled hydrological change for those assets that are in landscape groups where the hydrological response variable used for receptor impact modelling is increase in low-flow days per year within the zone of potential hydrological change

‘Low-flow days’ is the number of low-flow days per year. This is typically reported as the maximum change due to additional coal resource development over the 90-year period (from 2013 to 2102). The threshold for low-flow days is the 10th percentile from the simulated 90-year period.

Data: Bioregional Assessment Programme (Dataset 1, Dataset 3)

3.5.2.6 Assets and the ‘Springs’ landscape group

The ‘Springs’ landscape group is potentially impacted due to modelled additional coal resource development. The potential impacts and risks for the ‘Springs’ landscape group due to groundwater drawdown in the zone of potential hydrological change are detailed in Section 3.4.3. Section 3.4.3 also outlines the two interpretations of the hydrogeology for the source aquifer of the Doongmabulla Springs complex. Section 3.3.2 presents the two modelling conceptualisations that were used in the groundwater analytic element model (AEM), which are also relevant to understanding the potential impacts of coal resource development on these important ecological assets. In particular, the modelling results from the alternative AEM conceptualisation (Section 3.3.2) are considered to be more appropriate for assessing impacts to some specific points in the landscape (such as springs) that occur in areas where the uppermost Quaternary alluvium and Cenozoic sediment aquifer (the uppermost aquifer layer modelled in the AEM) is absent or not in direct (unimpeded) connection with the mining areas. Further information on the hydrological regimes and connectivity of the Doongmabulla Springs complex is also provided in Section 2.7.3 of companion product 2.7 for the Galilee subregion (Ickowicz et al., 2018).

The focus of this current section is on ecological assets that intersect with the ‘Springs’ landscape group and, in lieu of receptor impact modelling, on using the available lines of information to assess potential impacts on and risks to these assets. A large number of high-profile assets in the water-dependent asset register intersect with the ‘Springs’ landscape group. Although the 200 springs in this landscape group occupy less than 1% of the area of the zone of potential hydrological change, a total of 48 individual ecological assets (20% of all ecological assets in the zone) intersect with the ‘Springs’ landscape group (Table 39). Of these assets, 16 are confined entirely to the zone (i.e. do not occur anywhere outside the zone).

Most ecological assets associated with the ‘Springs’ landscape group are predominantly located in and around the Doongmabulla Springs complex. However, the Mellaluka Springs GDE is also an ecological asset listed in the asset register. Individual ecological assets located within the Doongmabulla Springs complex include the springs themselves (listed as a wetland complex in the ‘Surface water feature’ subgroup); the Doongmabulla Mound Springs Nature Refuge; habitat (potential species distribution) of an EPBC Act-listed TEC, ‘The community of native species dependent on natural discharge of groundwater from the Great Artesian Basin’; and habitat (potential species distribution) of two EPBC Act-listed threatened plant species: blue devil and salt pipewort (Figure 67 and Figure 72). Blue devil is an erect perennial herb, whereas salt pipewort is a small aquatic herb that grows in shallow water (in depths as low as 10 cm). In addition, a sociocultural asset, Doongmabulla Spring Natural Indicative Place, is located within the Doongmabulla Springs complex (see Section 3.5.4 for information about potential impacts to sociocultural assets).

Table 41 Ecological assets that intersect with the 'Springs' landscape group

^aTypology and punctuation are given as they are used in the asset database.

GDE = groundwater-dependent ecosystem; GW = groundwater

Data: Bioregional Assessment Programme (Dataset 1, Dataset 2)

Fensham et al. (2016) described the morphology and distribution of the various spring groups that comprise the Doongmabulla Springs complex. In general, springs underlain by the Moolayember Formation are termed ‘discharge springs’ (see Figure 41 in Section 3.4), and some of these have a distinctive mounded morphological structure. At some locations though, the original morphology of the springs is uncertain due to various types of anthropogenic modification (e.g. Joshua Spring has been extensively modified through construction of a turkey’s nest dam for use on the local pastoral lease). Springs to the east of the discharge springs (Figure 41 in Section 3.4) are situated near the base of topographic slopes within or near areas of outcropping Triassic bedrock (i.e. Clematis Group or Dunda beds). These are termed ‘outcrop springs’ by Fensham et al. (2016). The Little Moses and Yukunna Kumoo spring groups are underlain by the Clematis Group, whereas Dusk and Surprise spring groups occur on the Dunda beds outcrop. Groundwater discharges from some springs and contributes baseflow to Dyllingo Creek and the Carmichael River, and into alluvium associated within these stream valleys.

The signed digraph models for the ‘Springs’ landscape group identify the rate of groundwater flow as a critical factor in maintaining the aquatic community of springs (see companion product 2.7 for the Galilee subregion (Ickowicz et al., 2018)). Groundwater flow from the source aquifer needs to be at a rate that maintains a damp or submerged state in the spring such that the surface does not dry. An increase in groundwater flow above this threshold supports a wetted-area regime around the perimeter and downstream of the spring. Qualitative mathematical modelling indicated a zero or ambiguous response for most of the biological variables in the system to depletion of groundwater and available subsurface water. The models predicted a positive response for macrophytes (submerged and floating) but only for the cumulative impact scenario where subsurface water decreased and groundwater drawdown did not occur (see companion product 2.7 for the Galilee subregion (Ickowicz et al., 2018)).

The use of the Landsat archive of Digital Earth Australia enables the qualitative modelling predictions for the ‘Springs’ landscape group to be examined through time (i.e. over the past 30 years) within the zone of potential hydrological change (see Section 3.2.3.3 for further information about Digital Earth Australia and its application to the BA). The distribution of surface water and wet vegetation (Figure 73) as identified by the tasselled cap wetness (TCW) index derived from the Landsat archive shows that wet areas in surface drainages persist along Dyllingo Creek and the Carmichael River (even in dry years). These occur in the vicinity of the Doongmabulla Springs complex as well as downstream along the Carmichael River towards its confluence with the Belyando River. Surface water and areas of wetted vegetation are not temporally persistent in most tributary streams, except for isolated ponds (e.g. the previously unmapped water features located along some streamlines in Figure 73). The different discharge spring groups also show wide variation in persistence of water and wetted vegetation at surface (Figure 73). For instance, spring pools at the Moses-Keelback and Wobbly springs have remained persistent over the last 30 years, whereas only a relatively minor response is evident in the wetness index for the Mouldy Crumpet springs. The difference in response may have some bearing on localised differences in the water budgets and local geomorphology for individual spring groups. For example, springs with more persistent occurrence of water at surface may have a discharge rate that generally exceeds the evaporation rate, with the local geomorphology allowing for water to be ponded at surface. In general, outcrop springs associated with the Clematis Group aquifer (e.g. Yukunna Kumoo spring group) appear to be temporally persistent and have a relatively high wetness index, whereas outcrop springs that occur on areas of outcropping Dunda beds (e.g. Surprise and Dusk spring groups) have a lower wetness index. Visual inspection of the TCW index mapping for this area indicates that there may be more spring vents in the vicinity of Surprise and Dusk springs than have previously been mapped (Figure 73).

Hovmöller time-series plots (see Section 3.2) provide a useful tool to evaluate how the Landsat response for different spring groups varies over a 30-year period (1987 to 2016). The time-series response for two transects (Transect 1 and Transect 2, as shown in Figure 73) is depicted for various springs in the Doongmabulla Springs complex, using the TCW index and normalised difference vegetation index (NVDI). Analysis of these plots provides clear indications that there are significant variations in the distribution and availability of water in the vicinity of different springs groups within the Doongmabulla Springs complex. Such internal complexity may complicate the downscaling of impacts and risks from the scale of a spring complex to that of a spring group, as well as for their ongoing monitoring and management.

Stepping Stone and Mouldy Crumpet spring groups are not readily apparent on the TCW Hovmöller plot (Figure 74). It could be that discharge at these springs is too low to support wetted vegetation or spring pools that are extensive enough to be detected by the Landsat sensors (at least for the past 30-year period of the Landsat record). The highest TCW response in Figure 74 occurs around the channel of Dyllingo Creek. On the northern bank of the creek, it appears the Joshua Springs turkey’s nest dam was empty prior to 1992 but has been relatively full since. The NDVI (Figure 74) plot shows that narrow bands of permanent and stable vegetation occur around the Cattle Creek and Dyllingo Creek channels. Vegetation around Stepping Stone and Mouldy Crumpet spring groups is patchy and varies on an annual basis with the greatest extent occurring during the wet season, which suggests that long-term groundwater supply may be less reliable. In the vicinity of Mouldy Crumpet springs, vegetation was most extensive during relatively wetter seasons that occurred in 1992, 1998 to 2001, and 2010 to 2011. Some changes in land cover or land use may have occurred up slope of the Joshua Spring turkey’s nest dam around 1990.

The Hovmöller time-series plot for Transect 2 (Figure 75, location shown in Figure 73) intersects the Moses, Keelback and Wobbly spring groups. From the TCW response, landscape wetness around the Moses spring group is generally only detectable during the wet season. However, a pool fed by outflows from the Keelback and Moses springs is apparent. The TCW response suggests that since 1994 this pool has been a significant feature, remaining fairly constant in extent (approximately 200 m long) year round (although this pool appears to have diminished between 1988 and 1994). Another spring-fed pool, downstream of Moses and Keelback springs at the confluence of Cattle and Dyllingo creeks, shows more seasonal variation with a lower wetness index than the pool upstream. However, it is still largely present year round. While minimal variation in greenness is evident in the NDVI response here (Figure 75), it appears that vegetation cover has been relatively stable since 1988 (when data acquisition commenced).

Combining these remotely sensed observations with the signed-digraph models and qualitative mathematical modelling for the ‘Springs’ landscape group indicates that there is likely to be considerable variability in response to groundwater depletion among the many individual springs and spring groups that comprise the Doongmabulla Springs complex. Further, this inherent variability will likely mean that different levels of groundwater drawdown may potentially impact the ecological functioning of different springs and spring groups. This is a noteworthy set of conclusions given that all 187 springs in the Doongmabulla Springs complex are considered for this BA as a single ecological asset.

As has been emphasised previously the potential impact of groundwater drawdown due to additional coal resource development on these assets depends on the interpretation of the source aquifer(s) for Doongmabulla Springs (Section 3.4 and companion product 2.6.2 (Peeters et al., 2018)). The identity of the source aquifer for the Doongmabulla Springs complex has been contentious and the subject of previous scientific and legal dispute. Further information on the competing arguments for the source aquifer of these springs is provided in Section 3.4. Other references include: JBT Consulting (2015), Webb (2015), Currell (2016), Currell et al. (2017), Evans et al. (2018b) and Fensham et al. (2016).

As detailed in Section 3.4, the Clematis Group aquifer is considered the most plausible primary source aquifer for the Doongmabulla Springs complex, with the Dunda beds contributing minor volumes of groundwater to the easternmost spring groups in the complex (Dusk and Surprise, Figure 73). Leakage through the Moolayember Formation aquitard from the regional groundwater system in confined portions of the Clematis Group aquifer is the likely source for the discharge springs in the western part of the complex (Figure 73). In contrast, more localised groundwater flow in unconfined parts of the Clematis Group aquifer is considered the main source for springs near areas of Clematis Group outcrop (Little Moses and Yukunna Kumoo, Figure 73). Overall, it is inferred from the available evidence that the Doongmabulla Springs complex may represent a regional groundwater discharge feature sourced from the Clematis Group aquifer.

The identity of the source aquifer determines the nature of the impact from additional coal resource development, specifically aquifer depressurisation, on the assets related to the Doongmabulla Springs complex. If the source aquifer is primarily the Clematis Group aquifer then the potential exists for hydrological and ecological impact. Results for cumulative drawdown for both conceptual models used for the Galilee subregion’s AEM are presented in Section 3.4. Probabilistic modelling results based on both groundwater model conceptualisations show that 181 of the 187 springs in the Doongmabulla Springs complex have a 5% chance of experiencing additional groundwater drawdown in excess of 0.2 m. The original conceptualisation also predicts that there is also a 50% chance of these 181 springs experiencing this level of drawdown. In comparison, the alternative conceptualisation predicts that none of the springs in the Doongmabulla Springs complex would experience greater than 0.2 m of drawdown at the 50th percentile (median) of all model runs. Further, the original conceptualisation predicts that 120 springs in the Doongmabulla Springs complex have a less than 5% chance (very unlikely) of experiencing additional groundwater drawdown in excess of 2 m (there are no model results that exceed 2 m of drawdown at either the 5th, 50th or 95th percentile using the alternative conceptualisation). Section 3.4 suggests that results for the alternative groundwater modelling conceptualisation are more applicable for assessing drawdown at the Doongmabulla Springs complex (the rationale for this assessment is detailed in Section 3.3.2). In addition, and although not directly comparable due to major differences in modelling approach and development scenario, the peak drawdown at the springs from the proposed Carmichael Coal mine development (the mine nearest to these springs) will be about 0.1 to 0.3 m (Currell et al., 2017; GHD, 2013a).

In summary, the following statements can be made about the group of ecological assets that intersects the Doongmabulla Springs complex. First, the weight of available scientific evidence presented in this product, and in companion products 2.1-2.2 (Evans et al., 2018a) and 2.7 (Ickowicz et al., 2018), indicates that the source aquifer for these assets is primarily the Clematis Group aquifer. Second, both groundwater model conceptualisations used in this BA predict that 181 of the 187 springs in the Doongmabulla Springs complex have a 5% chance of experiencing additional groundwater drawdown in excess of 0.2 m. Last, consideration of multiple lines of evidence – including signed digraph models, qualitative mathematical modelling, archived Landsat imagery and ecological knowledge of the threatened vegetation species – indicates that this predicted level of drawdown will impact the ecological functioning of some ecological assets; however, there will be considerable variation in response across different springs and spring complexes. This final point underscores the need for additional local-scale analysis using more detailed geological, hydrological and ecological data to better understand the potential impact responses for individual springs and spring complexes.

Figure 73 Tasselled cap wetness index map - Doongmabulla Springs complex

The ‘wetness index’ represents the percentage of time observed as ‘wet’, based on the tasselled cap wetness (TCW) index. In the period 1987 to 2016, it is the proportion of times a pixel in the Landsat image archive exceeds a TCW index threshold. Each pixel represents a 25 x 25 m square area. A value of ‘1’ represents permanent water, whereas ‘0’ represents dry land and vegetation. Cenozoic sediment cover occurs in areas of this map where outcropping Triassic rocks of the Moolayember Formation, Clematis Sandstone and Dunda beds are not shown.

Data: Queensland Department of Natural Resources and Mines (Dataset 5); Bioregional Assessment Programme (Dataset 6); Queensland Department of State Development Infrastructure and Planning (Dataset 7)

Figure 74 Time - Landsat transect (Transect 1) through the Stepping Stone, Mouldy Crumpet and Joshua spring groups - Doongmabulla Springs complex

Middle panel shows wetness index response above a threshold of –400. Right-hand panel shows normalised vegetation difference index (NVDI) response. Location of Transect 1 on Figure 73.

Data: Bioregional Assessment Programme (Dataset 6)

Figure 75 Time - Landsat transect (Transect 2) through the Moses, Keelback and Wobbly spring groups - Doongmabulla Springs complex

Middle panel shows wetness index response above a threshold of –400. Right-hand panel shows normalised vegetation difference index (NVDI) response. Location of Transect 1 on Figure 73.

Data: Bioregional Assessment Programme (Dataset 6)

3.5.2.7 Assessing potential impacts on individual ecological water-dependent assets: a case study focused on the potential distribution of black ironbox

Within the operational constraints of the BAs it is not possible to assess potential impacts on each of the 241 ecological water-dependent assets within the zone of potential hydrological change. Instead, this section provides a case study to illustrate how multiple lines of available hydrological and ecological evidence may be used to assess potential impacts on individual assets. The detailed asset analysis presented here may assist future users of this BA to undertake a similar type of assessment, in order to develop a better understanding of potential impacts on, and risks to, a particular asset of interest.

The ecological asset selected for the detailed assessment presented here is the ‘potential distribution of Black Ironbox (Eucalyptus raveretiana)’. This is asset identification number 2126 in the water-dependent asset register for the Galilee subregion (Bioregional Assessment Programme, 2017). Black ironbox is listed as vulnerable nationally under the EPBC Act. Black ironbox grows to 25 m in height and mostly occurs along watercourses and on floodplains (Figure 76), although it may also occur (less commonly) in open woodland away from watercourses.

Black ironbox does not grow in pure stands, but rather is co-dominant with other tree species, including river red gum (E. camaldulensis), Moreton Bay ash (Corymbia tessellaris), river oak (Casuarina cunninghamiana) and weeping paperbark (Melaleuca fluviatilis). Black ironbox occurs mostly in coastal and sub-coastal parts of central Queensland, but is also recorded along the Suttor River (and its upper tributaries) in the Galilee assessment extent. It grows on soils that include sands, loams, light clays and cracking clays at up to 300 m above sea level.

The ‘potential distribution of Black Ironbox (Eucalyptus raveretiana)’ intersects with about 87 km² of vegetation ecosystems and 145 km of stream network in the northern part of the zone of potential hydrological change for the Galilee subregion (Figure 76). As noted previously, the potential geographic extent of this and other species within the zone is based on maximum entropy (MAXENT) modelling that relies largely on physical parameters and past observations of the presence and absence of the species. This means that the modelled extent of this asset within the zone has not been validated by field-based studies undertaken for this BA. Further work, which is beyond the scope of this assessment, would be needed to confirm the actual presence and distribution of black ironbox in the areas of the zone where it is modelled to occur.

The ‘potential distribution of Black Ironbox (Eucalyptus raveretiana)’ overlies parts of the northernmost area of the surface water zone of potential hydrological change, along sections of the Suttor River just upstream of Lake Dalrymple (Figure 21). However, this asset extent does not intersect with the groundwater zone of potential hydrological change (Figure 20). The asset extent within the surface water zone intersects areas that may experience hydrological changes due to additional coal resource development, in particular an increase in the number of zero-flow days per year (Figure 28). This includes the area within the zone of potential hydrological change where surface water modelling predicts the largest increases in the number of zero-flow days (Section 3.3.3). These modelled hydrological changes occur in the main channel of the Belyando River, and the Suttor River downstream of its junction with the Belyando, in particular, the approximately 250 km stretch of this river network from downstream of the junction with Native Companion Creek to Lake Dalrymple (Burdekin Falls Dam).

Figure 76 Intersection of 'potential distribution of Black Ironbox (Eucalyptus raveretiana)' and landscape groups in the northern part of the zone of potential hydrological change for the Galilee subregion

The ‘potential distribution of Black Ironbox (Eucalyptus raveretiana)’ is asset identification number 2126 in the water-dependent asset register for the Galilee subregion

Data: Bioregional Assessment Programme (Dataset 1, Dataset 2)

Along the northernmost part of the Suttor River (above the junction with Lake Dalrymple), there is a 5% chance that low-flow days will increase by 3 to 20 days per year in both the short-term period (2013 to 2042) and the long-term period (2073 to 2102). In addition, the modelling indicates that there is a 5% chance that low-flow spells will increase by 3 to 10 days per year in the long-term period (2073 to 2102) and that overbank flows will decrease by 0.02 to 0.05 events per year over the short-term period (2013 to 2042) (meaning one fewer overbank flow every 20 to 50 years on average) (Figure 48 and Figure 49). Hydrological changes (groundwater or surface water) are not predicted along the Cape River upstream of the junction with the Suttor River. As mentioned, changes to the groundwater system in areas where the black ironbox occurs in the zone are very unlikely as the asset extent does not intersect with the groundwater zone of potential hydrological change for the Galilee subregion (Figure 20).

In this case study, the intersection of landscape groups with the extent of individual assets is used to assess potential impacts to the natural and human-modified ecosystems represented by each landscape group. Seven landscape groups are contained within the asset extent of the ‘potential distribution of Black Ironbox (Eucalyptus raveretiana)’ within the zone of potential hydrological change (Table 42). This includes a relatively small area of vegetation (14.3 km²) that is not considered to be water-dependent as it overlaps with the ‘Dryland’ and ‘Floodplain, non-wetland’ landscape groups. This small area of non-water dependent vegetation within the asset extent likely reflects the challenges of integrating datasets and assessment methods at different scales to estimate the geographic distribution of the asset and the overlapping landscape groups.

Most of the remaining asset extent within the zone is classified as groundwater-dependent vegetation (65.2 km²) (the ‘Floodplain, terrestrial GDE‘ and the ‘Floodplain, wetland GDE’ landscape groups) or groundwater-dependent streams (145 km) (‘Streams, GDE’). Two of the four receptor impact models developed for this assessment (‘Floodplain, terrestrial GDE’ and ‘Woody riparian vegetation’) are relevant when assessing potential impacts on the ‘potential distribution of Black Ironbox (Eucalyptus raveretiana)’ within the zone. Further details about the development of these receptor impact models is in companion product 2.7 for the Galilee subregion (Ickowicz et al., 2018), with the application of these models discussed in Section 3.4.

The ‘Floodplain, terrestrial GDE’ receptor impact model predicts changes to percent foliage cover of floodplain trees, such as Eucalyptus, Corymbia or Acacia species that dominate the alluvial river and creek flats in the ‘Floodplain, terrestrial GDE’ landscape group. These species co-occur with Eucalyptus raveretiana (as described above). The asset occurs in areas of the ‘Floodplain, terrestrial GDE’ landscape group that are considered to be ‘at minimal risk’ due to additional coal resource development (Figure 62).

Within the zone of potential hydrological change, most streams within the asset extent are classified in the ‘Streams, GDE’ (129 km or 89% of streams) landscape group. The ‘Woody riparian vegetation’ receptor impact model predicts changes to the percent foliage cover of Eucalyptus camaldulensis and Melaleuca spp. in the ‘Streams, GDE’ landscape group. The ‘High-flow macroinvertebrate’ receptor impact model is relevant to both the ‘Streams, GDE’ and ‘Streams, non-GDE’ landscape groups. However, this receptor impact model is not relevant to assessing potential impacts to the black ironbox asset as it predicts changes to the density of mayfly nymphs (order Ephemeroptera in the family Baetidae of the genus Offadens) in riffle habitat, 3 months after the end of the wet season.

Considering the multiple lines of evidence generated through this BA, as well as the existing knowledge base about the hydrological dependence and ecological characteristics of the black ironbox, allows for an assessment of potential impacts on this asset due to additional coal resource development. Integrating this information with the experts’ opinions developed through the receptor impact modelling process provides strong evidence that potential impacts to the ‘potential distribution of Black Ironbox (Eucalyptus raveretiana)’ are very unlikely within the zone of potential hydrological change as:

Table 42 Landscape groups within the zone of potential hydrological change for the Galilee subregion and the extent of their overlap with the ‘potential distribution of Black Ironbox (Eucalyptus raveretiana)’

The ‘potential distribution of Black Ironbox (Eucalyptus raveretiana)’ is an ecological water-dependent asset of the ‘Vegetation’ subgroup, and is listed as asset identification number 2126 in the water-dependent asset register for the Galilee subregion. The three non-floodplain landscape groups that occur in the Galilee assessment extent are not included in this table as they do not intersect with the black ironbox asset extent in the zone of potential hydrological change.

GDE = groundwater-dependent ecosystem

Data: Bioregional Assessment Programme (Dataset 1)