2.7.3 'Riverine' landscape group

Summary

The two closely related landscape classes, ‘Perennial – gravel/cobble streams’ and ‘Intermittent – gravel/cobble streams’, overwhelmingly dominate potentially impacted stream reaches in the ‘Riverine’ landscape group (see Section 2.7.2). Hence, receptor impact modelling for the ‘Riverine’ landscape group focused on the ‘Perennial – gravel/cobble streams’ and ‘Intermittent – gravel/cobble streams’ landscape classes, and a conceptual model for the two classes is presented here.

Only 3% of the perennial streams and 2% of the intermittent streams within the zone of potential hydrological change were classed as being in good geomorphic condition. Of the intermittent streams within the zone, 51% were in moderate condition and 46% were in poor condition. Of the perennial streams within the zone, 83% were in moderate condition and 15% were in poor condition. The relatively poor geomorphic condition of the intermittent streams was reflected in riparian cover: 59% of the intermittent river reaches within the zone of potential hydrological change had some vegetation cover, while 88% of the perennial river reaches had some vegetation cover.

The qualitative model for the ‘Perennial – gravel/cobble streams’ landscape class identifies three surface water flow regimes, groundwater and precipitation as the main components of the hydrological regime that maintain and shape the ecosystem. The first four components are predicted to change due to coal resource development. Qualitative mathematical modelling of 14 plausible combinations of change in these components indicates a consistently negative effect across all the model’s variables.

The hydrological components were subsequently interpreted into a set of hydrological response variables, and some of the ecological components were chosen as receptor impact variables (with associated sample units), to assess the response of: (i) annual mean percent canopy cover of woody riparian vegetation to changes in dmaxRef (maximum difference in drawdown under the baseline future or under the coal resource development pathway future relative to the reference period (1983 to 2012)), tmaxRef (the year that the maximum difference in drawdown relative to the reference period (1983 to 2012) (dmaxRef) occurs), EventsR0.3 (mean annual number of events with a peak daily flow that is assumed to result in ‘overbench’ flow) and EventsR3.0 (mean annual number of events with a peak daily flow that is assumed to result in ‘overbank’ flow); (ii) mean density of larvae of the Hydropsychidae family (net-spinning caddisflies) to changes in mean annual number of zero-flow days (averaged over 30 years) (ZQD); and (iii) mean density of the eel-tailed catfish (Tandanus tandanus) to changes in ZQD and QBFI (baseflow index as described in Table 9).

The first receptor impact model suggests that the amount and rate of groundwater drawdown (dmaxRef and tmaxRef) have the strongest (relative to the other hydrological response variables) effect on annual mean percent canopy cover across the landscape class. If all other hydrological response variables are held at the mid-point of their elicitation range, then a 6-m reduction in groundwater levels (from their average conditions between 1983 and 2012), is predicted to lead to a roughly 20% decrease in mean percent canopy cover in both the short- (2042) and long- (2102) assessment years. The model also indicates that mean percent canopy cover in the future will be influenced by mean percent canopy cover in the reference year (2012).

This model also suggests that an increase in the frequency of overbench flows (EventsR0.3) will have a relatively small positive effect on mean percent canopy cover. However, the large uncertainty reflected in the 80% credible intervals does not preclude the small possibility of EventsR0.3 having a negligible effect on mean percent canopy cover. The summary statistics for the marginal distribution of the other model coefficients indicate that there is insufficient information in the expert-elicited data to determine the effect of overbank flows (EventsR3.0).

The second model strongly supports the hypothesis that an increase in ZQD will have a negative effect on the density of net-spinning caddisfly larvae. The model suggests that larval density can vary substantially across the landscape class (from <100 to 1000 per m²) under conditions of constant flow (ZQD = 0). As the number of zero-flow days increases, however, the model predicts that density will drop quite dramatically with values less than 1 per m² falling within the 80% credible interval under very intermittent flow conditions (ZQD >200 days).

The experts’ elicited values in the third model suggest that average density of the eel-tailed catfish will decline as ZQD increases, from about 5 individuals per 600 m² transect under continuous flow (ZQD = 0), holding all other covariates at their mid-values, to less than 1 individual in two transects as flow becomes more intermittent (increase in zero-flow days).

The qualitative mathematical model for the ‘Intermittent – gravel/cobble streams’ landscape class focused primarily on pool habitat and its role in providing refugia for fish and other aquatic organisms during periods of low flow. The model identifies surface water replenishment and groundwater input as the critical hydrological variables that maintain the pools’ ecology, and it examined the potential impacts of coal resource development on these hydrological variables, individually and in combination. The three resulting cumulative impact scenarios, reflecting combinations of decrease and no change in the hydrological variables, lead to predictions of a negative or zero (no change) response across most of the model’s variables.

The initial receptor impact modelling workshop for the Gloucester subregion was unable to address the ‘Intermittent – gravel/cobble streams’ landscape class. The Bioregional Assessment Programme addressed this omission by holding a second elicitation with a single expert. This expert, however, elected to address the subsurface fauna (hyporheic invertebrates) in riffle habitats (which are not represented in the qualitative model) and its response to the number of zero-flow days. This relationship was formalised into a receptor impact model that described the response of mean hyporheic invertebrate taxa richness to changes in ZQD.

The model reflects the expert’s view that increasing ZQD will have a negative effect on hyporheic taxa richness, despite lack of certainty about its average value. The model suggests that mean taxa richness can vary substantially across the landscape class from 10 to 20 per sampling unit (mean hyporheic invertebrate taxa richness in 6 L water pumped from a depth of 40 cm below the streambed) under conditions of constant flow (ZQD = 0). As the number of zero-flow days increases, however, the expert was of the opinion that density would drop to values from 1 to 8 under extremely intermittent flow conditions (ZQD >300 days).

It is important to recognise that many of the summary statements about the model described in this section simplify the (often more complicated) relationship between receptor impact variables and hydrological response variables captured by the receptor impact models. They are not risk or impact predictions for the Gloucester subregion. These predictions are provided in companion product 3-4 (impact and risk analysis) for the Gloucester subregion (Post et al., 2018).