This section describes the potential impacts on ecosystems that result from hydrological changes due to additional coal resource development. Ecosystems are represented by landscape classes, reflecting broad-scale patterns in geomorphology, soils, hydrology, and habitat and land use information for a diverse range of landscape features. The BA defines a set of landscape classes as ecosystems with characteristics that are expected to respond similarly to changes in groundwater and/or surface water due to coal resource development. The basis for the landscape groups and landscape classes is described in companion product 2.3 for the Namoi subregion (Herr et al., 2018). A schematic overview of the classification criteria and corresponding typology of landscape classes and landscape groups is shown in Figure 24. The landscape classes are organised into six landscape groups, which provide a basis to assess dependency on surface water and groundwater (Figure 24).
Figure 24 Schematic overview of the landscape classification for the Namoi assessment extent
The classification criteria are shown across the top row and the corresponding typology of landscape classes and groups is shown in the right-hand columns.
GAB = Great Artesian Basin, GDE = groundwater-dependent ecosystem
Source: companion product 2.3 for the Namoi subregion (Herr et al., 2018)
When developing ecological models for the Namoi subregion it was deemed necessary to develop a discrete model for defining potential ecological impacts for the Pilliga and Pilliga Outwash IBRA subregions (SEWPaC, 2012). The Pilliga and Pilliga Outwash IBRA subregions, simply termed the ‘Pilliga region’ here, represent a unique set of landscapes within the Namoi subregion. By comparison to the other landscapes across the Namoi subregion, the Pilliga region has many unique attributes in terms of its ecology, geomorphology, hydrogeology, soils and ecohydrology. Consistent with the structuring of the ecological models outlined in companion product 2.7 for the Namoi subregion (Ickowicz et al., 2018), the potential hydrological changes and ecosystem impacts are presented for both a separate Pilliga region of the zone of potential hydrological change and the remaining ‘non-Pilliga’ areas or reporting regions for the relevant landscape classes (see Section 3.3.2.2 for details on reporting regions).
A key consideration when identifying potential hydrological changes in landscape classes is that for some landscape elements there is limited or no surface water modelling. This is due to both the distribution of potential model nodes across the assessment extent and the limitations on interpolating model outputs across the stream network. Thus, changes in surface water hydrological response variables cannot be quantified for some parts of the zone of potential hydrological change. The extent of surface water modelling across the relevant stream-based landscape classes is presented when discussing hydrological and ecological impacts and was most limited for the upland riverine classes.
Results from regional-scale hydrological modelling indicate potential risks to about 1400 km2 of water-dependent ecosystems that intersect the zone of potential hydrological change (Table 16). The hydrological modelling assumed and included environmental releases in line with NSW legislation requirements. Therefore, there were no impacts on streamflow in regulated rivers (see Table 12 in companion product 2.6.1 for the Namoi subregion (Aryal et al., 2018)) since any changes were accounted for by changes in dam releases. Details can be found in Section 2.6.1.4 of companion product 2.6.1. The zone of potential hydrological change contains all four riverine landscape classes within the ‘Floodplain or lowland riverine’ landscape group and includes 61% of the entire extent of this group within the assessment extent (Table 16 and Figure 25). Among these four riverine classes, the largest contribution is from the ‘Temporary lowland stream’ (2062.2 km) and ‘Permanent lowland stream’ (979.6 km) landscape classes (Table 16). All six of the non-riverine classes occur within the zone of potential hydrological change (Table 16). The largest non-riverine landscape classes by area are the ‘Floodplain grassy woodland GDE’ (421.7 km2) and ‘Floodplain grassy woodland’ (121.3 km2) (Table 16). Almost half of all the ‘Floodplain riparian forest GDE’ in the assessment extent (148.7 km2) is included in the zone of potential hydrological change (Table 16). Most of the areas classified as ‘Floodplain wetland’ and ‘Floodplain wetland GDE’ in the assessment extent (30.1 and 151.8 km2, respectively) form part of the zone of potential hydrological change (Table 16). A qualitative mathematical model was developed for the ‘Floodplain or lowland riverine’ landscape group, capturing components of the riverine and riparian systems as well as the adjacent floodplain and its wetlands and vegetation (see Section 2.7.3 in companion product 2.7 for the Namoi subregion (Ickowicz et al., 2018)) (Table 16). Given the expertise and resources available at the receptor impact modelling workshop for the Namoi subregion, it was decided that receptor impact models be developed for a subset of landscape classes in this group. The details of these models are outlined in Section 2.7.3 of companion product 2.7 (Ickowicz et al., 2018) and in Section 3.4.3.3 (Table 16).
Most riverine landscape classes within the ‘Non-floodplain or upland riverine’ landscape group in the zone of potential hydrological change are classified as ‘Temporary upland stream’ (745.1 km), reflecting the intermittent or ephemeral nature of many of the stream segments (Table 16). The remainder of the stream network is classified as ‘Permanent upland stream’ (92.6 km), ‘Temporary upland stream GDE’ (34.7 km) and ‘Permanent upland stream GDE’ (14.2 km) (Table 16). The ‘Upland riparian forest GDE’ landscape class makes up a very small area of the zone of potential hydrological change (2.9 km2) along with a small area of ‘Non-floodplain wetland’ (13.1 km2) and ‘Non-floodplain wetland GDE’ landscape classes (8.1 km2) (Table 16). Three different components of this landscape group were considered for the qualitative modelling given the general lack of hydrological and spatial connectivity between the landscape classes across the zone of potential hydrological change. The upland riverine model included all upland riverine landscape classes and the adjacent riparian vegetation (‘Upland riparian forest GDE’). Three receptor impact models were formulated based on this qualitative mathematical model and include the potential response of the riverine system through changes in macroinvertebrate assemblages, presence of tadpoles and changes in projected foliage cover in the riparian trees along the stream channel (Table 16). In addition to the upland riverine system, a qualitative mathematical model was developed for the non-floodplain wetlands in this landscape group (Table 16, further details are provided in Section 2.7.4 of companion product 2.7 for the Namoi subregion (Ickowicz et al., 2018)).
The Pilliga region within the zone of potential hydrological change contains both upland and lowland riverine reaches. ‘Temporary upland stream’ (530.4 km) and ‘Temporary lowland stream’ (624.6 km) landscape classes make up the majority of the riverine networks, reflecting the highly ephemeral and/or intermittent nature of the drainage network (Table 16 and Figure 25). Only a small fraction of the ‘Permanent lowland stream’ landscape class (14.3 km) intersects with the Pilliga region in the zone of potential hydrological change (Table 16). Given the unique characteristics of the Pilliga’s stream network (i.e. low relief, intermittent flow patterns), a qualitative model, in consultation with the local experts, was developed for both upland and lowland riverine classes – Pilliga riverine (Table 16; further details are provided in Section 2.7.3 and Section 2.7.4 of companion product 2.7 (Ickowicz et al., 2018)). This meant that both lowland and upland riverine landscape classes share a similar model that encompasses both the riverine and riparian systems. Two separate receptor impact models were formulated that use changes in projected foliage cover and assemblages of macroinvertebrates as receptor impact variables (Table 16).
The ‘Grassy woodland GDE’ landscape class makes up most of the non-riverine landscapes in the Pilliga region (561.7 km2) and only a small portion (72.8 km2) of the total 634.5 km2 of this landscape class across the entire zone of potential hydrological change of this landscape class is located outside of the Pilliga region (Table 16). This landscape class includes a collection of different vegetation communities and habitats, however, given the concentration of this landscape class in the Pilliga region of the zone of potential hydrological change, a qualitative model was developed by the workshop participants with a focus on the ecology of this region in mind (Table 16; further details are provided in Section 2.7.4 of companion product 2.7 (Ickowicz et al., 2018)). Given the limitations on resources at the receptor impact modelling workshop and the uncertainty surrounding the nature of groundwater dependency of vegetation in the Pilliga region, a receptor impact model was not formulated for this landscape class.
The ‘Rainforest’ landscape group occupies a limited area within the zone of potential hydrological change, with the ‘Rainforest’ landscape class intersecting 4.0 km2 of the zone and the ‘Rainforest GDE’ landscape class intersecting only 0.3 km2 (Table 16). A qualitative mathematical model was developed for this landscape group given the conservation values surrounding the vegetation types common to this group (further details are provided in Section 2.7.5 of companion product 2.7 (Ickowicz et al., 2018)).
Two springs are known to occur within the zone of potential hydrological change, which are classified as ‘GAB springs’ based on their association with underlying sandstone formations. These two springs are located on the eastern edge of the Pilliga Basin and are thought to be primarily recharge springs, given their location on the eastern fringes of the Great Artesian Basin (GAB) (Fensham and Fairfax, 2003). A qualitative mathematical model was formulated for a typical recharge GAB spring (Table 16, further details are provided in Section 2.7.6 of companion product 2.7 (Ickowicz et al., 2018)).
Data: Bioregional Assessment Programme (Dataset 2)
Table 16 Extent of all landscape classes in the assessment extent and zone of potential hydrological change
The relevant reporting region, qualitative model and receptor impact model are given for each landscape class.
aValues for the extent in assessment extent are the same regardless of reporting region.
GDE = groundwater-dependent ecosystem; na = not applicable
Data: Bioregional Assessment Programme (Dataset 1)
Receptor impact modelling converts the potentially abstract information about hydrological changes to quantities that can be more readily understood and interpreted. In particular, outcomes of the modelling relate more closely to their ecological values and beliefs and therefore support community discussion and decision making about acceptable levels of coal resource development (see companion submethodology M08 (as listed in Table 1) for receptor impact modelling (Hosack et al., 2018)). Receptor impact models are not intended to make site-specific predictions, but rather to quantify the range of possible responses of selected receptor impact variables to a given change in hydrology. It is beyond the scope of a BA to make precise predictions at exact locations.
Receptor impact variables represent biological components of the ecosystem that experts have chosen as indicators of ecosystem condition, and which are considered likely to be sensitive to changes in the hydrology of that system (see companion submethodology M08 (as listed in Table 1) for receptor impact modelling (Hosack et al., 2018)). Changes in hydrology are represented in the model by hydrological response variables, chosen to reflect particular water requirements of the ecosystem. The magnitude of change in the chosen receptor impact variables to changes in one or more hydrological response variables, captured through an expert elicitation process, is an indicator of the magnitude of risk to the ecosystem as a result of hydrological perturbation. For example, a prediction that the number of riffle-breeding frog species is likely to decrease in a particular reach where zero-flow days are predicted to increase does not necessarily mean that there are riffle-breeding frogs present and that they will be impacted. Rather, it means that given the magnitude of hydrological change predicted in that reach, there is a specific risk to the habitat requirements of riffle-breeding frogs, and more generally a risk to the ecosystems represented by the landscape class the riffle-breeding frog inhabits. The receptor impact modelling results are provided at a landscape scale and should not be interpreted as exactly representing the local conditions of a particular site. Predictions of receptor impact variables are ultimately one line of evidence, and any assessment of risk, particularly at a local scale, needs to be considered in conjunction with the broader hydrological changes that may be experienced and the qualitative mathematical models, which assist in describing potential knock-on effects to the ecosystems.
In the following sections, the results from receptor impact models should be treated as indicating the experts’ pooled knowledge as to the likelihood and magnitude of ecological impacts in an ecosystem given a known hydrological change. Results also capture the uncertainties arising from lack of knowledge and the variability inherent in landscapes across short and long distances.
