2.7.2.3 'GDE' landscape group

The total area of GDE landscape classes within the zone of potential hydrological change is 102 km², most of which (57%) is ‘Forested wetland’, with smaller areas of ‘Rainforest’ (23%), ‘Dry sclerophyll forest’ (14%), ‘Wet sclerophyll forest’ (4%), and ‘Freshwater wetland’ (1%) landscape classes. Receptor impact modelling of GDE landscape classes is described in Section 2.7.4. Qualitative and quantitative models were developed for the ‘Forested wetland’ (Section 2.7.4.2), ‘Dry sclerophyll forest’ and ‘Wet sclerophyll forest’ landscape classes. Based on feedback from the qualitative modelling workshop, ‘Wet sclerophyll forest’ and ‘Dry sclerophyll forest’ landscape classes were represented by a single qualitative model and quantitative model (wet and dry sclerophyll forests, Table 5; Section 2.7.4.3).

A qualitative model was developed for all forested wetlands at the qualitative modelling workshop, which represented vegetation classes typical of both the coastal and inland parts of the Hunter subregion. As noted in the previous section, riverine forested wetlands are the predominant riparian habitat along perennial streams in the Hunter subregion, and this interconnection is reflected via the riparian habitat node in the perennial stream qualitative model. Other points of connection, such as frogs, are also reflected in the two models. Ecosystems represented in the forested wetlands qualitative model can be sensitive to changes in the flow regime of the perennial streams they occur along.

During the expert elicitation of quantitative relationships for the forested wetlands receptor impact model, the experts were explicit that their assessment related to riverine forests in the western and central uplands of the Hunter subregion, typically comprising Eucalyptus camaldulensis (river red gum), E. tereticornis and Casuarina cunninghamii. These correspond to ‘Eastern riverine forest’ and ‘Coast and tableland riverine forest’ vegetation classes in Table 5 of companion product 2.3 for the Hunter subregion (Dawes et al., 2018). Thus, the quantitative model does not apply to ‘Coastal swamp forests’ and ‘Coastal floodplain wetlands’ vegetation classes. Riverine forests account for 50% of the ‘Forested wetland’ landscape class within the zone of potential hydrological change. The experts stipulated that the model should also not be applied to riverine forests along regulated rivers due to their altered flow regimes. In Section 3.4 of companion product 3-4 for the Hunter subregion (Herron et al., 2018), the potential impacts on ‘Coastal floodplain wetlands’ and ‘Coastal swamp forests’ vegetation classes are assessed using the qualitative model, while potential impacts on riverine forests along unregulated river sections are evaluated using results from the quantitative receptor impact modelling.

A qualitative model was developed for the ‘Rainforest’ landscape class (rainforests, Table 5; Section 2.7.4.4 and reflected the experts’ shared view that in this area rainforests depend on the concentration of rainwater in upland gullies for their water supply, rather than regional groundwater; hence, a quantitative model was not progressed as it was considered very unlikely (less than 5% chance) that coal resource developments of these rainforests would impact their water supply. The rainforests within the Hunter subregion’s zone of potential hydrological change are predominantly Keith’s ‘Northern warm temperate rainforests’ vegetation class (Keith, 2004). These occur in sheltered gullies and slopes in hilly-to-steep terrain of the coast and escarpment on moderately fertile soils in high rainfall areas, extending above 1000 m elevation, on granites, rhyolites, syenites or sedimentary substrates that yield acid soils with moderate levels of nutrients (OEH, 2016). However, 10 km² of the distribution of Keith’s (2004) ‘Northern warm temperate rainforests’ vegetation class overlaps with the alluvium in the Hunter subregion, which given their topographic position suggests they could have some dependence on alluvial groundwater in some circumstances. The rainforest model presented in Section 2.7.4.4 does not reflect this. The dependence on groundwater of riparian rainforests within this landscape class is considered a knowledge gap.

The ‘Freshwater wetland’ landscape class within the zone of potential hydrological change is represented entirely by Keith’s (2004) ‘Coastal freshwater lagoons’ vegetation class. Keith’s (2004) ‘Coastal heath swamps’, also represented by this landscape class, are outside the zone and very unlikely to be impacted by hydrological changes due to additional coal resource development. A qualitative model was developed for Keith’s (2004) ‘Coastal freshwater lagoons’ vegetation class (see Section 2.7.4.5). Experts at the workshop were unable to agree whether groundwater dependence of these lagoons is regional or local, and hence whether hydrological changes from underground coal mining higher up in the catchment would affect lagoon hydrology. Experts thought that tidal fluctuations might be influencing water levels in the lagoons. A drop in regional groundwater was considered likely to lead to seawater intrusion and hence have little impact on lagoon water levels, but potentially cause large increases in the salinity of the wetland water and catastrophic change in its ecology. Understanding whether these lagoons are sustained by connection to a rainfall-fed local groundwater aquifer or to a fresh regional aquifer, and the connectivity between local and regional aquifers is important for determining possible impacts due to mining. For the assembled experts, the water-dependencies of the coastal freshwater lagoons in the Hunter subregion were a knowledge gap. A review of the literature was undertaken to obtain further insight.

Claus et al. (2011) developed two pertinent conceptual models for wetland types under the NSW Wetland Monitoring, Evaluation and Reporting program: one for ‘Coastal freshwater lagoons and lakes’ and the other for ‘Coastal dune lakes or lagoons’. The former model identified overbank flow as the major source of water for the wetland, while the latter ‘are perched above the general water-table in dune hollows created by wind action and sealed by organically cemented sand.’ Both of these models indicate that the wetlands are unlikely to be dependent on regional groundwater aquifers. This is consistent with a study of Porters Creek Wetland in the Hunter subregion, which found that the wetland was associated with a shallow watertable that was replenished by rainfall (Payne et al., 2012). These studies suggest that coastal freshwater lagoons are likely to be dependent on local, perched groundwater systems that are disconnected from regional groundwater, and hence not particularly sensitive to drawdown caused by coal mining. Therefore, the assessment team decided that development of a quantitative model of Keith’s (2004) ‘Coastal freshwater lagoons’ vegetation class was not a major priority for this bioregional assessment. However, since the potential for adverse impacts from drawdown of regional groundwater might be lagoon specific, local assessments of the groundwater dependency of each lagoon within this landscape class are recommended to better resolve this issue.

Only small areas of the ‘Heathland’ landscape class were present in the zone of potential hydrological change (0.2 km²) and it was the view of the experts in the qualitative modelling workshop that these were likely to be dependent on local, perched groundwater systems. Wet heathland and swamp sclerophyll shrubland occur in open depressions where the watertable is shallow after periods of high rainfall, whereas sedgeland is found in closed depressions where standing water accumulates (Rutherford et al., 2013). Hence, no qualitative or quantitative models were developed for the ‘Heathland’ landscape class.

Only very small areas of ‘Grassy woodland’ (0.2 km²) and ‘Semi-arid woodland’ (<0.1 km²) landscape classes were present in the zone of potential hydrological change, and it was the view of the experts in the qualitative modelling workshop that these were likely to only have an opportunistic dependence on groundwater. Grassy woodlands are associated with fertile soils on the western slopes of the Great Dividing Range (Keith, 2004) and include the Commonwealth’s Environment Protection and Biodiversity Conservation Act 1999 (EPBC Act) critically endangered community ‘White Box – Yellow Box – Blakely’s Red Gum Grassy Woodland and Derived Native Grassland’. Salinity and rising groundwater levels threaten remnants of these woodlands at low elevations within the landscape (Yates and Hobbs, 1997).

The ‘Semi-arid woodland’ landscape class within the zone of potential hydrological change falls within Keith’s (2004) ‘Riverine plain woodlands’ vegetation class. These acacia-dominated woodlands occur on grey clay soils on flats and shallow depressions of the riverine plain far from active drainage channels; they are typically beyond the extent of modern-day floods. These include the EPBC Act-listed endangered ‘Weeping Myall Open Woodland’, which may have some groundwater dependence. This is based on the Draft conservation advice for Hunter Valley Weeping Myall (Acacia pendula) Woodland of the Sydney Basin Bioregion (Department of the Environment, 2014), which states ‘Alteration of groundwater and construction of roads through expansions to coal mining, or coal seam gas exploration and extraction … poses a potential clearing-related threat’. However, it appears likely to depend on accumulation of surface run-off in local lows and break-of-slope areas, indicated by this description from White et al. (2002):

In higher rainfall areas it typically forms an open woodland. As rainfall decreases the ecological community becomes increasingly restricted, tending to sparse or scattered stands of woodland occurring in discrete bands fringing better-watered country. It can also occur as relatively narrow strips on the margins of floodplain woodland or on minor depressions or run-on areas adjacent to sandhills.