ecosystems within the Coorong, Lower Lakes and Murray Mouth

(1)

An assessment of

ecosystems within the Coorong, Lower Lakes and Murray Mouth

(CLLMM) region

DEWNR Technical report 2016/32

(2)

An assessment of ecosystems within the Coorong, Lower Lakes and Murray Mouth (CLLMM) region

Ronald S. Bonifacio, Trevor J. Hobbs, Daniel Rogers, Sacha Jellinek, Nigel Willoughby and David Thompson

Department of Environment, Water and Natural Resources August 2016

DEWNR Technical report 2016/32

(3)

i Department of Environment, Water and Natural Resources

GPO Box 1047, Adelaide SA 5001 Telephone National (08) 8463 6946

International +61 8 8463 6946 Fax National (08) 8463 6999

International +61 8 8463 6999 Website www.environment.sa.gov.au

Disclaimer

The Department of Environment, Water and Natural Resources and its employees do not warrant or make any representation regarding the use, or results of the use, of the information contained herein as regards to its correctness, accuracy, reliability, currency or otherwise. The Department of Environment, Water and Natural Resources and its employees expressly disclaims all liability or responsibility to any person using the information or advice. Information contained in this document is correct at the time of writing.

This work is licensed under the Creative Commons Attribution 4.0 International License.

To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

ISBN 978-1-925510-45-4

Preferred way to cite this publication

Bonifacio, R.S., Hobbs, T.J., Rogers, D., Jellinek, S., Willoughby, N., Thompson, D., 2016. An assessment of ecosystems within the Coorong, Lower Lakes and Murray Mouth (CLLMM) region. DEWNR Technical Note 2016/32, Government of South Australia, Department of Environment, Water and Natural Resources, Adelaide.

Download this document at https://data.environment.sa.gov.au

(4)

Foreword

The Department of Environment, Water and Natural Resources (DEWNR) is responsible for the management of the State’s natural resources, ranging from policy leadership to on-ground delivery in consultation with government, industry and communities.

High-quality science and effective monitoring provides the foundation for the successful management of our environment and natural resources. This is achieved through undertaking appropriate research, investigations, assessments, monitoring and evaluation.

DEWNR’s strong partnerships with educational and research institutions, industries, government agencies, Natural Resources Management Boards and the community ensures that there is continual capacity building across the sector, and that the best skills and expertise are used to inform decision making.

Sandy Pitcher CHIEF EXECUTIVE

DEPARTMENT OF ENVIRONMENT, WATER AND NATURAL RESOURCES

(5)

iii

Acknowledgements

The Coorong Lower Lakes and Murray Mouth (CLLMM) Recovery Project is funded by the South Australian Government’s Murray Futures program and the Australian Government. We greatly appreciate the support shown by Hafiz Stewart and Kym Rumblelow (DEWNR) from inception to completion of the project. Jody Gates (DEWNR) made significant inputs to the ecology of birds studied here. Tim Croft shared his expertise in pre-European vegetation and taxonomic issues. Terry Sim and Ken Strother also provided much information on the history of CLLMM natural systems. Thai Te (DEWNR) contributed to resolving plant species taxonomic issues. We thank Andrew West, Andy Harrison, Phil Pisanu, Kirsty Bevan, Ross Meffin and Simon Sherriff (All DEWNR) for providing insightful comments on earlier versions of this report.

(6)

6.3.1 Ecosystem 1: Pink Gum (Eucalyptus fasciculosa) Low Open Grassy Woodland (MLR sands) 63 6.3.2 Ecosystem 2: Stringybark (Eucalyptus baxteri) / Cup Gum (E. cosmophylla) Woodland (MLR hills) 66 6.3.3 Ecosystem 3: Mixed Shrubland (Coast Daisy-bush Olearia axillaris / Coast Beard-heath Leucopogon

parviflorus / Coastal Wattle Acacia longifolia ssp. sophorae) (coastal dunes) 69 6.3.4 Ecosystem 4: Coastal White Mallee (Eucalyptus diversifolia) (SE/LL sandy loams) 72 6.3.5 Ecosystem 5: Sheoak (Allocasuarina verticillata) Low Shrubby Woodland (SE/LL sandy loams) 79 6.3.6 Ecosystem 6: Mixed Eucalypt (Mallee Box / Peppermint Box / SA Blue Gum) Woodland / Mallee

(Ridge-fruited / Narrow-leaf Red Mallee) Ecosystem 83

6.3.7 Ecosystem 7: Reeds and Rushes (Common Reed Phragmites australis / Bullrush Typha domingensis /

Sea Rush Juncus kraussii) (freshwater fringes) 88

6.3.8 Ecosystem 8: Lignum (Muehlenbeckia florulenta) Shrubland (non-saline clays) 90 6.3.9 Ecosystem 9: Samphire (Tecticornia pergranulata / Suaeda australis / Sarcocornia quinqueflora) /

Paperbark (Melaleuca halmaturorum) Shrubland (saline clays) 92

6.3.10 Other ecosystems (10.1–10.5) 95

6.4 Associations between bird species and terrestrial Ecosystems within the CLLMM region of South

Australia 96

6.5 Cluster analysis of bird habitat requirements to identify Ecosystem Response Groups, and their

associations with Ecosystems, in the CLLMM region of South Australia 99

6.5.1 Cluster analysis dendrogram 99

6.5.2 Cluster analysis ordination plot 100

6.6 Bird species lists for each Ecosystem Response Group (ERG) within the CLLMM region of South

Australia 100

(8)

List of figures

Figure 1.1 The information components and linkages that comprise the landscape assessment approach (Rogers

et al. 2012) 4

Figure 1.2 Landscape assessment provides the situation assessment component of a planning process – represented by the green oval in this generic planning-process example (based on Margoluis and

Salafsky, 1998) 5

Figure 2.1 Topography and landforms of the CLLMM region of South Australia 8

Figure 2.2 Remnant native vegetation extent and conservation reserves in the CLLMM region of South Australia 9

Figure 2.3 Mean annual rainfall of the CLLMM region of South Australia 10

Figure 2.4 Mean annual temperature of the CLLMM region of South Australia 11

Figure 2.5 Soils groups of the CLLMM region of South Australia 12

Figure 2.6 Landscape subgroups of the CLLMM region of South Australia 13

Figure 2.7 Location of 100 ha areas used to analyse historic changes in bird occurrence within IBRA subregions

of the CLLMM region of South Australia 16

Figure 2.8 The graphical structure of the Bayesian Belief Network used in determining bird trends within the

CLLMM region of South Australia 17

Figure 3.1 Potential distribution of Ecosystem 1: Pink Gum Low Open Grassy Woodland (MLR sands) of the

Figure 3.2 Potential distribution of Ecosystem 2: Stringybark / Cup Gum Woodland (MLR hills) of the CLLMM

region of South Australia 22

Figure 3.3 Potential distribution of Ecosystem 3: Mixed Shrubland (coastal dunes) of the CLLMM region of South

Australia 23

Figure 3.4 Potential distribution of Ecosystem 4: Coastal White Mallee (SE/LL sandy loams) of the CLLMM region

of South Australia 24

Figure 3.5 Potential distribution of Ecosystem 5: Sheoak Low Shrubby Woodland (SE/LL sandy loams) of the

Figure 3.6 Potential distribution of Ecosystem 6.1: Mallee Box Grassy Woodland (LL loams) of the CLLMM region

of South Australia 26

Figure 3.7 Potential distribution of Ecosystem 6.2: Peppermint Box Grassy Woodland (MLR loams) of the CLLMM

Figure 3.8 Potential distribution of Ecosystem 6.3: Ridge-fruited / Narrow-leaf Red Mallee (MLR sands) of the

Figure 3.9 Potential distribution of Ecosystem 6.4: SA Blue Gum Grassy Woodland (SE/LL loams) of the CLLMM

Figure 3.10 Potential distribution of Ecosystem 7: Reeds and Rushes (freshwater fringes) of the CLLMM region of

South Australia 30

Figure 3.11 Potential distribution of Ecosystem 8: Lignum Shrubland (non-saline clays) of the CLLMM region of

South Australia 31

Figure 3.12 Potential distribution of Ecosystem 9: Samphire / Paperbark Shrubland (saline clays) of the CLLMM

Figure 3.13 Potential distribution of Ecosystem 10.1: Chaffy Saw-sedge Swampland of the CLLMM region of South

Australia 33

Figure 3.14 Potential distribution of Ecosystem 10.2: Red Gum Grassy Woodland (MLR river flats) of the CLLMM

(9)

vii Figure 3.16 Potential distribution of Ecosystem 10.4: Sheoak / Native Pine Grassy Woodland (LL loams) of the

Figure 3.17 Potential distribution of Ecosystem 10.5: Agroecosystems (agricultural lands) of the CLLMM region of

South Australia 37

Figure 3.18 The average contribution of the method type on synthesis assessments of bird occurrence trends within the South Australian CLLMM region (greater entropy reduction values suggest greater

influence of the method type on determining occurrence trends) 42

Figure 6.1 Dendogram of the hierarchical cluster analysis showing vegetation survey sites (e.g. “4_G3”), similarity in vegetation species composition and cover values (e.g. lines, linkages), and their natural groupings to classify terrestrial ecosystems (e.g. “Eco1”) of the CLLMM region of South Australia. Very small groups (shaded) were discarded from the classification, but used to help inform the identification of

’Other’ terrestrial ecosystems 60

Figure 6.2 Dendogram of the hierarchical cluster analysis showing bird species (e.g. “SH”), similarity in habitat requirements (e.g. lines, linkages), and their natural groupings to classify ecosystem response groups

of the CLLMM region of South Australia 99

Figure 6.3 Plot of cluster analysis associations between bird species (e.g. “BrF”), similarity in habitat requirements (i.e. non-metric multidimensional scaling, NMDS) and their natural groupings to Ecosystems (e.g.

“10.5”) within the CLLMM region of South Australia 100

List of tables

Table 2.1 Description of landscape subgroups within the CLLMM region of South Australia 7 Table 2.2 Vegetation survey species cover classes and their corresponding numeric values (i.e. proportion cover)

used in cluster analyses of vegetation associations 14

Table 2.3 Conditional probability table behind the output node ‘Species Trend’ in the Bayesian Belief Network used to determine bird trends within the CLLMM region of South Australia 17

Table 3.1 Terrestrial ecosystems of the CLLMM region of South Australia 20

Table 3.2 Bird species and trends in occurrence (i.e. ‘synthesis bird assessment’) within the South Australian CLLMM region, including trend probabilities (numbers) and the most likely trend classification

(shaded cells) from BBN modelling 38

Table 3.3 Response groups and associated bird species and ecosystems within the CLLMM region of South

Australia 43

Table 4.1 Summary of recommendations for priority ecosystem groups for restoration within the landscapes of

the SA CLLMM region 52

(10)

Summary

One of South Australia’s Ramsar wetlands of international importance, the Coorong, Lower Lakes and Murray Mouth (CLLMM), has been in decline due to human-driven changes in its hydrology and the use of surrounding landscapes. In 2010, the Department of Environment, Water and Natural Resources (DEWNR) began a

revegetation program (CLLMM Vegetation Program) in an attempt to restore some of the region’s terrestrial ecosystems, create resilience in the system, and arrest declines in biodiversity. In order to maximise the effectiveness of the revegetation activities, the CLLMM Vegetation Program required guidance to prioritise investment and activities. The objective of this report was to inform that prioritisation with an ecological analysis of where such revegetation activities would be most effectively delivered to support these broader program objectives.

A landscape assessment (LA) was applied to the CLLMM region in order to identify ecosystems, provide indicators of biodiversity decline, and prioritise restoration efforts through the analysis, synthesis and interpretation of the following information:

 the nature of the ecosystems in the landscape

 status and trends of terrestrial bird species and their associations with ecosystems

 land use and native vegetation clearance history.

Quantitative analyses were augmented with expert knowledge to improve interpretation of the results.

Seventeen terrestrial ecosystems were identified within the region, and 40% of terrestrial bird species were found to have decreasing frequencies of occurrence in these landscapes. Bird decline was strongly correlated with ecosystem types with a long and extensive history of native vegetation clearance. Eight terrestrial bird Ecosystem Response Groups were identified and associated with ecosystems. This information was used to formulate management recommendations (focussing on revegetation) for each ecosystem, in the context of investment available from the CLLMM VP for on-ground activities.

This study suggests management activities should focus on ecosystem groups identified as those at greatest risk of biodiviersity loss via declining resilience or changing to undesirable states. Terrestrial ecosystems identified by this project as most at risk of biodiversity loss include:

1. Mallee communities of the eastern Mount Lofty Ranges, specifically in the proximity of larger remnants such as Ferries–McDonald Conservation Park (i.e. Ecosystem 6.3: Ridge-fruited / Narrow-leaf Red (Eucalyptus incrassata / leptophylla) Mallee (MLR sands))

2. Grassy woodland communities of the eastern Mount Lofty Ranges (i.e. Ecosystem 6.1: Mallee Box (Eucalyptus porosa) Grassy Woodland (LL loams); Ecosystem 6.2: Peppermint Box (E. odorata) Grassy Woodland (MLR loams); Ecosystem 6.4: SA Blue Gum (E. leucoxylon) Grassy Woodland (SE/LL loams);

Ecosystem 10.2: Red Gum Grassy Woodland (MLR river flats); and Ecosystem 10.4: Sheoak (Allocasuarina verticillata) / Native Pine (Callitris gracilis) Grassy Woodland (LL loams))

3. Samphire / Paperbark shrubland communities associated with saline wetlands (i.e. Ecosystem 9: Samphire (Tecticornia spp.) / Paperbark (Melaleuca halmaturorum) Shrubland (saline clays)).

(11)

2

1 Introduction

1.1 Background

The Coorong, Lower Lakes and Murray Mouth (CLLMM) region is the terminus for the drainage of the

Murray-Darling Basin, which covers about 14% (1 073 000 km²) of Australia (Cann and Barnett 2000). It features a complex mosaic of lakes, coastal lagoons, interconnecting channels and vegetation communities (Seaman 2003;

Fluin et al. 2007). The region contain wetlands of international importance, listed under the Ramsar Convention in 1985. Water extraction across three Australian states has reduced freshwater discharge from the River Murray to the sea by 75% (Cann et al. 2000). As a result, negative impacts on the unique ecology of the system have substantially increased since 2007 when it became evident that ecosystem processes were collapsing in

association with decreasing water levels within the Lakes (DEH 2009). In addition, terrestrial ecosystems associated with the Ramsar site have been perceived to be in a slow decline since European settlement, primarily due to preferential historic clearance of native vegetation in the region (Butcher and Rogers 2013). In response, the Australian and South Australian Governments funded DEWNR to deliver the “CLLMM Vegetation Program” as part of the CLLMM Recovery Project, whose broad objectives included increasing the ecological resilience of the region, primarily through revegetation activities. The landscape assessment presented here has the primary objective of providing the ecological information to support the prioritisation of this revegetation activity, such that it most effectively addresses this loss of ecological resilience.

1.2 Landscape assessment

Landscape assessment is an approach (or framework) for identifying priority ecosystems for restoration, while informing the development of evidence-based landscape-specific goals for nature conservation (Rogers et al.

2012). It facilitates the development of comprehensive, explicit and outcome-driven nature conservation strategies, and contributes to the maintenance of ecological resilience in South Australia’s landscapes. This approach is intended to guide managers beyond the simple concept of conserving the native extant biodiversity, to a more nuanced and prioritised suite of interventions that target those components of the landscape that have suffered loss of resilience, and are approaching thresholds that would cause transition to an undesirable state.

Landscape assessment is founded on the nested, hierarchical nature of biodiversity – operating on the principle that the conservation requirements at higher levels of organisation (e.g. landscapes, ecosystems) should meet the requirements of the majority of biodiversity at lower levels (e.g. species; Noss 1987; Hunter et al. 1988; Hunter 1991; 2005).

Broadly, the principle objective of landscape assessment is to identify landscape-scale systemic issues that are driving loss of ecological resilience, such that these underlying causes of decline can be addressed through management (Rogers et al. 2012). For the purposes of this assessment, the focus of the analyses are on identifying ecosystems within landscapes that are associated with decline. This ecosystem focus was adopted for two main reasons:

 An environmental history of the CLLMM region suggested that, overwhelmingly (although not universally), the systemic driver of biodiversity loss in the region’s terrestrial systems is the historic preferential

conversion of native vegetation for European agricultural systems (Paton et al. 1999). The pattern of clearance targeted some ecosystems over others, depending on the suitability of the underpinning environment (soil, climate, topography) to support these European agricultural activities, suggesting that ecosystem is a strong predictor of decline (as has been observed elsewhere, e.g. for the southern Mount Lofty Ranges, see Rogers 2011a)

 Given the a priori requirement for the CLLMM Vegetation Program to invest specifically in revegetation activities, the Program required particular information on where revegetation activity would provide the

(12)

most ecological benefit (rather than the broader suite of interventions that one might identify through a more comprehensive analysis of systemic drivers of decline).

 Broadly, landscape assessment (as defined by Rogers et al., 2012) relies on a synthesis of three elements (see Figure 1.1) to identify systematic patterns associated with biodiversity decline within landscapes. The key groups of components - Ecosystem Assessment, Species Assessment and Land Use Assessment are synthesised, primarily to identify the alternate states and trends of ecosystems within a landscape, and to identify the drivers (e.g. historic clearance of vegetation for intensive agriculture) for these alternative trends:

1. Ecosystem Assessment - An understanding of the ecosystems that comprise the landscape of interest.

Including information on the environmental settings of each ecosystem and the typical (or best remaining) ecological expression of that setting

2. Species Assessment - An understanding of the current state (i.e. species conservation status) and recent trajectory (i.e. declining, stable or increasing occurrence trends) of species (for which adequate information is available) within the landscape.

Information on the ecological requirements of each species, with particular reference to their association with the ecosystems - i.e. their preferred habitat types.

Using that information to determine groups of species with similar trajectory that can be associated with particular habitat types

3. Land Use Assessment - An understanding of the spatial and temporal variation in human modification of the landscape (e.g. the location and chronology of vegetation clearance/modification within

landscapes).

The landscape assessment approach has now been applied to several regions of South Australia (Rogers, 2010;

Willoughby, 2010; Rogers, 2011a, b; Willoughby et al., 2011; Rogers 2012a, b, c, d; Rogers et al., 2012; Gillespie et al., 2013), some of which overlap with the geographic boundary of concern in this study. Preliminary landscape assessment analyses of the region (Butcher and Rogers 2013) were based on existing generalised frameworks for conservation decision making (McIntyre and Hobbs, 1999, 2000). Butcher and Rogers (2013) provide an overview of the environmental history, patterns of vegetation clearance, and generic priorities for conservation investment or further research in terrestrial landscapes located with 5 km (entirely or partially) of the CLLMM.

Landscape assessment uses both spatial biological survey information and knowledge gathered via expert opinion or key informants (Northrip et al. 2008). The use of expert opinion or key informants as a technique in gathering information has been extensively used in the medical field (e.g. Muhit et al. 2007; Kalua et al. 2009) but has been adapted to investigate environmental problems (Wacker 2005; Bonifacio et al, 2010; http://www.unitedway- weld.org/compass/ environmental_issues.htm). Results using key informants (i.e. ‘local champions’) can be comparable with formal surveys but is financially more efficient (Pal et al. 1998).

In this study, bird species are used to represent ecosystem-scale processes and interactions (i.e. ‘systemic’ issues).

The decision to use avifauna for the region is based on both previous applications of the landscape assessment approach and the difficulties involved in obtaining useful information for most fauna that is informative about the state and trajectory of ecological communities. Birds are a visible, relatively diverse and relatively well-studied fauna occurring in most agricultural settings. This reflects the ease of collecting bird data compared to other taxa (Mac Nally et al. 2004). In addition, the spatial scale over which terrestrial bird populations operate is comparable to the scale over which human activities operate; thus the scale at which we define our landscapes may be comparable between terrestrial birds and human impacts (Major, 2010).

Despite the availability of bird data, presence-only data from a plethora of sources is difficult to use with respect to assigning trends to species (Elphick 2008). To help address this, a Bayesian Belief Network (BBN) was used to

(13)

4 Figure 1.1 The information components and linkages that comprise the landscape assessment approach (Rogers et al. 2012)

1.3 Study objectives

This landscape assessment study within the CLLMM region of South Australia has the following objectives:

1. to identify and describe the different terrestrial ecosystem types of the region 2. to assess changes in biodiversity within these ecosystems

3. identify drivers of change within these ecosystem

4. identify priority terrestrial ecosystems for conservation investment.

Thus, landscape assessment provides the situation assessment component of a generic planning process

(Figure 1.2). Landscape assessment alone does not provide detail on the specific interventions required to realise a conservation goal (e.g. Situation Model). Further, while landscape assessment is designed to help set

context-specific conservation goals, that process (setting, and acting on, conservation goals) requires a more inclusive approach for the diverse range of stakeholders involved (e.g. TNC 2007, CMP 2007).

Land Use Assessment Spatial, semi-

quantitative (e.g. Hundred, County Data)

Textual (e.g.

local histories)

Synthesis – land use history Species Assessment

Species ecology (ecosystem associations) Species state and trajectory

Species “response”

groups (functional groups)

Synthesis – ecological attributes associated with alternate trajectories Ecosystem Assessment

Vegetation classification Environmental settings

Ecosystem classification (and spatial

prediction of extent)

Landscape synthesis – identify alternate patterns and drivers of change

(14)

Situation assessment

Set work programs Define goals

Define targets, milestones and performance

measures

Implement

Evaluate and review Revise?

Monitor performance

measure(s) Situation

Model

Figure 1.2 Landscape assessment provides the situation assessment component of a planning process – represented by the green oval in this generic planning-process example (based on Margoluis and Salafsky, 1998)

(15)

6

2 Methods

2.1 Study area

This study considers terrestrial landscapes and associated ecosystems in close proximity (<5 km) to the estuarine Coorong and lacustrine Lower Lakes of the River Murray in South Australia (see Figure 2.1). The study area intersects three biogeographic regions (i.e. Kanmantoo, Murray–Darling Depression, Naracoorte Coastal Plain;

IBRA Version 7, DotE 2012) and includes five IBRA sub-regions (i.e. Fleurieu, Murray Mallee, Murray Lakes and Coorong, Tintinara, Bridgewater). These lands are dominated by annual cereal cropping and livestock grazing production systems, with smaller components of high intensity agriculture and conservation areas containing predominately native vegetation communities (Figure 2.2). The region experiences a Mediterranean climate with cool wet winters and warm dry summers. Mean annual rainfall (Figure 2.3) in the study area ranges between 352-734 mm/year, and mean annual temperature (Figure 2.4) between 14.3–16.3 °C (ANUCLIM Version 6.1, 1976 to 2005, Xu & Hutchison 2013). Topographic variation is low, with a maximum elevation of 180 m AHD on the south-eastern slopes of the Mount Lofty Ranges.

The natural vegetation is diverse – ranging from wetland-associated communities (e.g. reeds and sedges) to terrestrial communities (grassland, shrub/heath, mallee and grassy woodlands). Major terrestrial vegetation types of the region include open grassy woodlands (structurally dominated by Pink Gum Eucalyptus fasciculosa, Native Pine Callitris gracilis, SA Blue Gum E. leucoxylon, Mallee Box E. porosa and Peppermint Box E. odorata, woodlands with a shrubby understorey, Sheoak Allocasuarina verticillata, mallee communities (Coastal Mallee E. diversifolia, Ridge-fruited Mallee E. incrassata, Narrow-leaf Red Mallee E. leptophylla and Beaked Red Mallee E. socialis), and coastal or saline shrublands (Wattles Acacia spp., Samphire Tecticornia spp., Swamp Paperbark Melaleuca

halmaturorum). The region also contain smaller components of woodlands (Brown Stringybark E. baxteri, Cup Gum E. cosmophylla, Red Gum E. camaldulensis), native grasslands, sedgelands and fringing wetland communities.

2.2 Data

2.2.1 Landscapes and environment

Landsystems and soil types have been classified and mapped for the agricultural areas of South Australia (DWLBC Soil and Land Program 2007; Hall et al. 2009). Each mapped ‘Soil Landscape Unit’ (SLU) polygon represents a landscape with similar topographic and soil properties (DEWNR SDE ‘LANDSCAPE.SALAD_Soil_Subgroup’). SLU polygons can contain multiple landscape elements and soil types when the size of each component is lower than the spatial scale of original mapping. The estimated areal proportion of each component within the polygon is also documented. Soil attributes (e.g. depth, clay content) within components are described using semi- quantitative classes (DWLBC Soil and Land Program 2007; Hall et al. 2009). Soil landscape units (SLU) located wholly or partially within 5 km of Lake Alexandrina, Lake Albert and the Coorong provide a foundation (i.e. abiotic characteristics) for identifying ecosystems of the region (Figure 2.5). Landscape subgroups (Table 2.1) were identified by geographic regions with similar climate, topography and soil landscape units (Figure 2.6).

(16)

Table 2.1 Description of landscape subgroups within the CLLMM region of South Australia Landscape subgroups Description

Mount Lofty Ranges Terrestrial plains north and west of the Lake Alexandrina and Lake Albert, and extending into hills and slopes of the Mount Lofty Ranges

Lower Lakes a) Terrestrial plains and low hills surrounding Lake Alexandrina and Lake Albert b) Aquatic and periodically inundated areas fringing Lake Alexandrina and Lake Albert

Coastal Dunes Coastal dunefields and aquatic fringes of the Coorong lagoon South East Terrestrial landscapes southeast of the Coorong lagoon

2.2.2 Vegetation

Remnant native vegetation (Figure 2.2) has been surveyed, mapped and described over recent decades by DEWNR (DEWNR 2008, 2015, Heard and Channon 1997; e.g. DEWNR SDE ‘VEG.SAVegetation’). Vegetation surveys used a standard methodology (Heard and Channon 1997) to identify and describe the structure, crown cover class and species composition within each vegetation ‘Patch’ (i.e. DEWNR SDE ‘FLORA.SurveySites’). Vegetation survey (i.e.

unique ‘PatchID’) data from patches =< 900 m² in size or with fewer than four species were excluded to minimize errors in local ecosystem classifications.

Taxonomic issues resulting from data collected over many years by observers with differing skill levels were resolved, where possible, by natural historians with local knowledge.

2.2.3 Birds

Information on the presence and location of terrestrial birds between 1908 and 2013 within the region were compiled from Biological Databases of South Australia (BDBSA) and the database of BirdLife Australia (July 2013).

Only indigenous species were included in the analysis. Each record contained the species taxonomy, common name, year, month, location (map coordinates), spatial accuracy (m) of each record and source of the data.

Records with a spatial accuracy of >1000 m were excluded from analyses to reduce errors in associations with ecosystems.

Taxonomic issues resulting from data collected over many years by observers with differing skill levels were resolved, where possible, by natural historians with local knowledge.

(17)

8 Figure 2.1 Topography and landforms of the CLLMM region of South Australia

(18)

Figure 2.2 Remnant native vegetation extent and conservation reserves in the CLLMM region of South Australia

(19)

10 Figure 2.3 Mean annual rainfall of the CLLMM region of South Australia

(20)

Figure 2.4 Mean annual temperature of the CLLMM region of South Australia

(21)

12 Figure 2.5 Soils groups of the CLLMM region of South Australia

(22)

Figure 2.6 Landscape subgroups of the CLLMM region of South Australia

(23)

14

2.3 Ecosystem assessment

2.3.1 Ecosystem classification

Ecosystems for the region were identified from vegetation survey site (i.e. ‘Patch’) floristic composition, plant species cover and abiotic characteristics (i.e. soil subgroups, mean annual rainfall, topographic slope, landscape subgroups) using hierarchical cluster analysis (Legendre and Legendre 1998). Cluster analyses used the ‘hclust’

function in the ‘vegan’ package, (Oksanen et al. 2011), with the number of groups informed by the ‘kgs’ function (White and Gramacy 2012). Clustering methods used Bray-Curtis dissimilarity measure and WPGMA

agglomeration. Non-metric Multidimensional Scale (NMDS) was also used to help with visualisation of dissimilarity between sites, using the metaMDS function in the vegan package (Oksanen et al. 2011). The ‘abundance’ measure used was based on species-level categorical cover class descriptions from vegetation surveys which converted to representative numeric values (Table 2.2). All analyses were done using R (R Core Team 2013).

Table 2.2 Vegetation survey species cover classes and their corresponding numeric values (i.e. proportion cover) used in cluster analyses of vegetation associations

Cover class description Proportion cover

Sparsely or very sparsely present - cover very small (less than 5%) 0.01

Not many, 1–10 individuals 0.02

Plentiful but of small cover (less than 5%) 0.03

Any number of individuals covering 5–25% of the area 0.05

Covering more than 75% of the area 0.75

2.3.2 Other ecosystems

Vegetation survey sites that did not strongly cluster to ecosystem classifications, and pre-1750 native vegetation types (DEWNR 2015; DEWNR SDE ‘VEG.PEVegetation’) that did not match ecosystem classifications from the cluster analysis, were used to identify additional ecosystems within the region. For each of these additional ecosystem types abiotic characteristics (i.e. soil subgroups, mean annual rainfall, topographic slope, landscape subgroups) for vegetation survey sites or DEWNR pre-1750 vegetation mapping units (i.e. DEWNR SDE

‘VEG.PEVegetation’) were used to define these additional ecosystems.

2.3.3 Ecosystem mapping

Ecosystem mapping for the region is constrained to the soil landscape units (SLU) located wholly or partially within 5 km of Lake Alexandrina, Lake Albert and the Coorong (Figure 2.5 & Figure 2.6) to minimize potential

misclassifications from insufficient calibration data. Soil and landscape characteristics (i.e. proportions of SLU soil subgroups, landscape subgroups) associated with each ecosystem were used to construct spatial domain models to represent their likely distribution (see Sect. 6.2). Additional ecosystems are mapped from their pre–1750 native vegetation extent (DEWNR 2015; DEWNR SDE ‘VEG.PEVegetation’), topographic data (DEWNR SDE

‘TOPO.WaterCourses’) or recent landuse mapping (DEWNR SDE ‘LANDSCAPE.LandUse2008’).

(24)

2.4 Species assessment

2.4.1 Trends in bird occurrence Biological databases

To quantify historic trends in bird species occurrence within the region all bird records (i.e. BDBSA + Bird Atlas data) were assigned to 100 ha hexagonal subdivisions of the study area (Figure 2.7) and changes in occurrence of each species within each 100 ha area between historic (1908–2013) and recent (2000–13) periods were analysed to identify species declines or increases (Franklin 1999). Hexagons without any bird species records were excluded as false negatives for any species. The degree of change (i.e. ‘Current vs All Time’) was calculated as the proportion of hexagons in which a species was recently recorded (≥2000), compared to the number of times it had been

recorded over the entire survey period (1908–2013). To reduce potential biases resulting from less reliable data this analysis excluded species with <10 records and/or those that had been recorded in <5 hexagons.

Linear regression was performed on the proportion of 100 ha areas that were occupied by each species per year (i.e. Number Observed’). Species were identified as declining in this analysis (i.e. ‘Trend Analysis’) if the results of the regression (i.e. ‘P Value’; ‘R-squared’) indicated p values of <0.1, and a negative slope value (i.e. ‘Slope’). Again, steps were taken to address issues of variable effort in species surveys. Lists with less than three species were removed (to reduce bias associated with surveys that target particular species), and years with less than five surveys were removed.

Expert bird assessment model

For each species, information on current status and trend of occurrence from prior DEWNR regional assessments (Gillam 2011; Gillam 2012), including results from the analysis of biological databases, were reviewed by panel of experts (i.e. eight ecologists). Species were given the following ‘status scores’ by these experts based on their assigned conservation status: Extinct = 6, Critically Endangered = 5, Endangered = 4, Vulnerable = 3, Rare = 2, Near Threatened = 1, Least Concern or Data Deficient = 0. Species trends were scored as: definite decline=-2, probable decline = -1, stable= 0, probable increase= +1, or definite increase= +2. Threat scores were generated from the sum of status and trend scores, where: 0–1 = Least Concern; 2 = Rare but Stable; 3 = Widespread but Declining; 4 = Rare and Declining; and 5 = Extinct (locally). The mean of each species threat score for each IBRA subregion (i.e. Fleurieu, Murray Mallee, Murray Lakes and Coorong, Tintinara, Bridgewater/Lucindale; Figure 2.7) and the number of IBRA subregions where the species once existed were calculated (i.e. ‘Status Trends Scores’,

‘Number of SubRegions’ ). Species that were assigned an overall threat score of 3 or 4 were identified as declining in this analysis.

Synthesis bird assessment

The outputs of analyses of data from biological databases and expert assessment were added as a parent node into a Bayesian Belief Network (BBN; McCann et al. 2006) to determine whether birds were “increasing”, “stable”, or “decreasing” in the region. Uncertainties in biological data (e.g. BDBSA surveys are not standardised through time), and confidence levels of expert assessments of bird species status and trends, are incorporated into the BBN analysis. The structure of the model is shown in Figure 2.8 and the conditional probability table behind the output node is shown in Table 2.3. Sensitivity analyses of the BBN model using entropy reduction identified the

contribution of each assessment method on the synthesis bird assessment of trends in occurrence (Pearl 1988, Korb and Nicholson 2004, Marcot et al. 2006, Pollino et al. 2007, Smith et al. 2007). Entropy measures the degree of uncertainty in a variable. Entropy reduction describes the expected reduction, l ,in mutual information of a query variable Q due to a finding F and is calculated as:

where q is a state of the query variable Q, f a state of the findings variable F, and the summation refer to the sum



 _q _fP q f Pq f P f l ( , )log[ ( , )/ ( )]

(25)

16 Figure 2.7 Location of 100 ha areas used to analyse historic changes in bird occurrence within IBRA subregions of the CLLMM region of South Australia

(26)

Figure 2.8 The graphical structure of the Bayesian Belief Network used in determining bird trends within the CLLMM region of South Australia

Table 2.3 Conditional probability table behind the output node ‘Species Trend’ in the Bayesian Belief Network used to determine bird trends within the CLLMM region of South Australia

Parent node states Outcome states (bird trends)

Expert model Current vs All time Trend analysis Increase Stable Decrease

No Decline 0 to 0.7 Increasing 25 50 25

No Decline 0 to 0.7 Stable 0 60 40

No Decline 0 to 0.7 Decreasing 0 30 70

No Decline 0.7 to 1 Increasing 50 45 5

No Decline 0.7 to 1 Stable 30 60 10

No Decline 0.7 to 1 Decreasing 5 65 30

Decline 0 to 0.7 Increasing 10 30 60

Decline 0 to 0.7 Stable 0 40 60

Decline 0 to 0.7 Decreasing 0 10 90

Decline 0.7 to 1 Increasing 50 40 10

Decline 0.7 to 1 Stable 0 60 40

Decline 0.7 to 1 Decreasing 0 30 70

2.4.2 Response groups

As the targeted focus of this assessment related to the identification of ecosystems to be prioritised for revegetation, the focus of response groups related to the strength of association that bird species had with particular ecosystems or habitat types. Information on the habitat requirements for each species was gathered from literature (e.g. Handbook of Australian, New Zealand and Antarctic Birds series). This information was supplemented by expert assessments to identify associations between each bird species and ecosystem (Rogers et al. 2012). Each expert rated the strength of these associations: 0 = no likelihood of the species in the ecosystem;

StatusTrendsScores 0 to 2.5

2.5 to 4 50.0 50.0 2.25 ± 1.2

NumSubRegions 0 to 2

2 to 4 4 to 5

33.3 33.3 33.3 2.83 ± 1.5

r_squared 0 to 0.25 0.25 to 0.5 0.5 to 0.75 0.75 to 1

25.0 25.0 25.0 25.0 0.5 ± 0.29 Slope

Negative Positive

50.0 50.0 NumObserv

0 to 5 5 to 10 10 to 20 20 to 40

25.0 25.0 25.0 25.0 13.8 ± 11

P_Value 0 to 0.001 0.001 to 0.05 0.05 to 0.1 0.1 to 1

25.0 25.0 25.0 25.0 0.163 ± 0.26 ExptModel

No Decline Decline

55.0 45.0

CurrentVsAllTime 0 to 0.7

0.7 to 1 50.0 50.0 0.6 ± 0.29

SpeciesTrend Increase Stable Decrease

13.2 46.8 40.1

TrendAnalysis Increasing Stable Decreasing

25.8 48.6 25.6

(27)

18 ecosystem with a strong association with that ecosystem. Results across experts were averaged and consensus integer values were agreed for each species and ecosystem.

Based on these information sets, each bird species was classified to an ‘Ecosystem Response Groups’ (ERG), that grouped species based on common habitat type associations (as per ‘ecosystem groups’ of Chin et al. 2010).

Species were classified to ERG using hierarchical cluster analysis (Legendre and Legendre 1998). Cluster analyses used the ‘hclust’ function in the ‘vegan’ package, (Oksanen et al. 2011), with the number of groups informed by the ‘kgs’ function (White and Gramacy 2012). Clustering methods used Bray-Curtis dissimilarity measure and WPGMA agglomeration. Non-metric Multidimensional Scale (NMDS) was also used to help with visualisation of dissimilarity between species, using the metaMDS function in the vegan package (Oksanen et al. 2011). Natural breaks in similarity measures were checked against expert rankings to adjust some break points between clusters.

The outputs of the cluster analysis was further refined with qualitative analyses in cases where a large cluster was evident but thought to represent a number of smaller groups based on expert knowledge of birds in the region.

All analyses were done using R (R Core Team 2013).

2.5 Land use assessment

2.5.1 Land use history

A comprehensive synthesis of the post-European land use history of the CLLMM region was recently undertaken by Butcher and Rogers (2013) that provided a summary of the spatial and temporal history of environmental modification since European settlement. This summary was used in conjunction with spatial information (e.g.

Figure 2.2) to assess the timing, location and extent of native vegetation clearance, with particular reference to differences among landscapes and ecosystems within landscapes. This published information was augmented using two local historians (i.e. key informants) with extensive knowledge about pre-European vegetation in the study area. Estimates of the extent of change within each ecosystem and the confidence levels of this information were collated. All information gathered from key informants was classified by the informant on their confidence in reliability of the information (i.e. ‘Low’ 1–45%, ‘Medium’ 46–75%, ‘High’ 76–90%, ‘Very High’ >90%). Recent landuse mapping (DEWNR SDE ‘LANDSCAPE.LandUse2008’) was also used to identify current landuse activities with the region.

2.6 Landscape assessment

The analyses were synthesised to inform the state and trend of the systems within each CLLMM landscape, and the most likely drivers of these patterns of change. In summary, the analyses above provide the following information to this synthesis:

 Ecosystem Assessment – provides a description of the important ecosystems that comprise each landscape. This provides a framework for the inherent ecological variation that occurs in the system, and the ecological structure and function (e.g. habitat types) variability among systems. In addition, the drivers of environmental change that have resulted in loss of biodiversity can often vary with among these ecosystems (Paton et al. 1999). In agricultural landscapes in particular, understanding the nature and distribution of these different ecosystems is critical to understanding where intervention is required to reduce the risk of biodiversity loss. Ecosystem assessment thus provides the biophysical context on which the landscapes assessment is based.

 Species Assessment – by understanding the state and trend of individual species, and the strength of their association with different ecosystems, we are able to use groups of species as indicators of the state and trend of different ecosystems. This provides the foundation for assessing where (which ecosystems) intervention is most urgent to prevent biodiversity loss within a landscape.

(28)

 Land Use Assessment – understanding the history of environmental change, including the nature and extent of impact, provides information regarding the drivers of ecological change inferred from the species assessment, as well as correlative support for this ecological change. Understanding the nature of land-use change allows us to better understand the key systemic drivers of decline, in order to design objectives that address these drivers. A widespread example is the historic preferential clearance of different ecosystems in agricultural landscapes that are correlated with agricultural potential.

(29)

20

3 Results

3.1 Ecosystem assessment

Nine ecosystems were initially identified from hierarchical cluster analysis (see Table 3.1, Appendix 6.1, Figure 6.1) of species composition and vegetation cover data from biological surveys in the region (i.e. DEWNR SDE

‘FLORA.SurveySites’), and abiotic data. The cluster analysis identified anticipated (e.g. Gillespie et al 2013) associations between vegetation floristics, soil types and/or landscape subgroups (i.e. geomorphology, topography, climate). Comparisons between ecosystems resulting from the classification of vegetation survey sites, unclassified vegetation survey sites and pre-1750 native vegetation mapping (DEWNR 2015; DEWNR SDE

‘VEG.PEVegetation’), detected four subdivisions of Ecosystem 6 (i.e. ‘Mixed Eucalypt woodland / Mallee ecosystem’) and four additional native ecosystems. The modern ‘Agroecosystems (agricultural lands)’ was

recognised and included in the list of ecosystems for the region. The combination of results from cluster analyses using vegetation survey data, pre-1750 vegetation mapping, modern landuse mapping and expert opinion identified 17 Ecosystems within the CLLMM region of South Australia (Table 3.1).

The relationships between the 17 Ecosystems and soil groups (i.e. DEWNR soil mapping), Landscape subgroups (i.e. geographic regions with similar climate, topography and soil landscape units), pre-1750 vegetation mapping, topographic data or recent landuse mapping are given in Appendix 6.2. More detailed descriptive information, including vegetation composition are shown in Appendix 6.3.

Potential distribution maps for each ecosystem (Figure 3.1 to Figure 3.17) are constrained to the outer boundary of soil landscape units (SLU) polygons located wholly or partially within 5 km of Lake Alexandrina, Lake Albert and the Coorong (e.g. Landscape Subgroups; Figure 2.5 & Figure 2.6). Maps for each ecosystem include estimates of the proportion of each mapped soil-landscape unit, pre-1750 vegetation mapping unit or land use polygons that matches each ecosystem model’s criteria (see Appendix 6.2). Ecosystems 10.1, 10.2, 10.3 and 10.5 with models based on mapped polygons without proportional data were assigned a nominal value of 75%. Proportions were classified into four levels for mapping outputs: >60% (high); 30–60% (medium); 5–30% (low); and 0–5% (very low).

Table 3.1 Terrestrial ecosystems of the CLLMM region of South Australia

Ecosystems Methods used

1. Pink Gum Low Open Grassy Woodland (MLR sands) Cluster analysis

2. Stringybark / Cup Gum Woodland (MLR hills) Cluster analysis

3. Mixed Shrubland (coastal dunes) Cluster analysis

4. Coastal White Mallee (SE/LL sandy loams) Cluster analysis

5. Sheoak Low Shrubby Woodland (SE/LL sandy loams) Cluster analysis

6. Mixed Eucalypt woodland / Mallee ecosystem Cluster analysis,

pre-1750 vegetation mapping, expert knowledge 6.1 Mallee Box Grassy Woodland (LL loams)

6.2 Peppermint Box Grassy Woodland (MLR loams) 6.3 Ridge-fruited / Narrow-leaf Red Mallee (MLR sands) 6.4 SA Blue Gum Grassy Woodland (SE/LL loams)

7. Reeds and Rushes (freshwater fringes) Cluster analysis

8. Lignum Shrubland (non-saline clays) Cluster analysis

9. Samphire / Paperbark Shrubland (saline clays) Cluster analysis

10. Other ecosystems Pre-1750

vegetation mapping, expert knowledge, modern landuse mapping 10.1 Chaffy Saw-sedge Swampland

10.2 Red Gum Grassy Woodland (MLR river flats) 10.3 Tussock Grassland (dryland)

10.4 Sheoak / Native Pine Grassy Woodland (LL loams) 10.5 Agroecosystems (agricultural lands)

(30)

Figure 3.1 Potential distribution of Ecosystem 1: Pink Gum Low Open Grassy Woodland (MLR sands) of the CLLMM region of South Australia

(31)

22 Figure 3.2 Potential distribution of Ecosystem 2: Stringybark / Cup Gum Woodland (MLR hills) of the CLLMM region of South Australia

(32)

Figure 3.3 Potential distribution of Ecosystem 3: Mixed Shrubland (coastal dunes) of the CLLMM region of South Australia

(33)

24 Figure 3.4 Potential distribution of Ecosystem 4: Coastal White Mallee (SE/LL sandy loams) of the CLLMM region of South Australia

(34)

Figure 3.5 Potential distribution of Ecosystem 5: Sheoak Low Shrubby Woodland (SE/LL sandy loams) of the CLLMM region of South Australia

(35)

26 Figure 3.6 Potential distribution of Ecosystem 6.1: Mallee Box Grassy Woodland (LL loams) of the CLLMM region of South Australia

(36)

Figure 3.7 Potential distribution of Ecosystem 6.2: Peppermint Box Grassy Woodland (MLR loams) of the CLLMM region of South Australia

(37)

28 Figure 3.8 Potential distribution of Ecosystem 6.3: Ridge-fruited / Narrow-leaf Red Mallee (MLR sands) of the CLLMM region of South Australia

(38)

Figure 3.9 Potential distribution of Ecosystem 6.4: SA Blue Gum Grassy Woodland (SE/LL loams) of the CLLMM region of South Australia

(39)

30 Figure 3.10 Potential distribution of Ecosystem 7: Reeds and Rushes (freshwater fringes) of the CLLMM region of South Australia

(40)

Figure 3.11 Potential distribution of Ecosystem 8: Lignum Shrubland (non-saline clays) of the CLLMM region of South Australia

(41)

32 Figure 3.12 Potential distribution of Ecosystem 9: Samphire / Paperbark Shrubland (saline clays) of the CLLMM region of South Australia

(42)

Figure 3.13 Potential distribution of Ecosystem 10.1: Chaffy Saw-sedge Swampland of the CLLMM region of South Australia

(43)

34 Figure 3.14 Potential distribution of Ecosystem 10.2: Red Gum Grassy Woodland (MLR river flats) of the CLLMM region of South Australia

(44)

Figure 3.15 Potential distribution of Ecosystem 10.3: Tussock Grassland (dryland) of the CLLMM region of South Australia

(45)

36 Figure 3.16 Potential distribution of Ecosystem 10.4: Sheoak / Native Pine Grassy Woodland (LL loams) of the CLLMM region of South Australia

(46)

Figure 3.17 Potential distribution of Ecosystem 10.5: Agroecosystems (agricultural lands) of the CLLMM region of South Australia

(47)

38

3.2 Species assessment

3.2.1 Trends in bird occurrence

A total of 80 588 bird occurrence records from DEWNR and Bird Atlas databases were used to identify species and their occurrence between 1908 and 2013 within the region (138 taxa, see Table 3.2). Bayesian Belief Network models using trends in bird occurrence from biological databases (i.e. change over time ‘Current vs All Time’, regression slopes ‘Trend Analysis’) and expert scores (i.e. ‘Expert bird assessment model’) provide a synthesis assessment of trends in bird occurrence in the region (Table 3.2), including measures of confidence in

classifications. Uncertainties in the synthesis assessment were calculated from statistical variations in trends and regression slopes from analyses of biological databases, and confidence levels of expert scores, within a BBN model. Entropy analyses within BBN model provide measures of the importance of each assessment method on the synthesis assessment of trends in bird occurrence within the region (Figure 3.18).

The ‘synthesis bird assessment’ of trends in occurrence identified 56 declining species (i.e. 41% of total species), 73 stable species (53%) and 2 increasing species (1%) within the region (Table 3.2). Two species (1%) are borderline decreasing and 5 species (4%) are borderline increasing in the region.

Table 3.2 Bird species and trends in occurrence (i.e. ‘synthesis bird assessment’) within the South Australian CLLMM region, including trend probabilities (numbers) and the most likely trend classification (shaded cells) from BBN

modelling

Common name Scientific name Probability of trend

Declining Stable Increasing

Australasian Pipit Anthus australis 0.190 0.623 0.188

Australian Bustard Ardeotis australis 0.595 0.388 0.018

Australian Hobby Falco longipennis 0.180 0.620 0.200

Australian Magpie Gymnorhina tibicen 0.053 0.458 0.490

Australian Owlet-nightjar Aegotheles cristatus 0.855 0.145 0.000

Australian Raven Corvus coronoides 0.150 0.613 0.237

Australian Reed-Warbler Acrocephalus australis 0.085 0.555 0.360

Australian Ringneck Barnardius zonarius 0.685 0.315 0.000

Barn Owl Tyto javanica 0.270 0.642 0.088

Beautiful Firetail Stagonopleura bella 0.845 0.155 0.000

Black Falcon Falco subniger 0.270 0.642 0.088

Black Kite Milvus migrans 0.655 0.345 0.000

Black-chinned Honeyeater Melithreptus gularis 0.800 0.200 0.000

Black-faced Cuckoo-shrike Coracina novaehollandiae 0.130 0.608 0.262

Black-shouldered Kite Elanus axillaris 0.085 0.555 0.360

Blue Bonnet Northiella haematogaster 0.800 0.200 0.000

Blue-winged Parrot Neophema chrysostoma 0.655 0.345 0.000

Brown Falcon Falco berigora 0.130 0.608 0.262

Brown Goshawk Accipiter fasciatus 0.655 0.345 0.000

Brown Quail Coturnix ypsilophora 0.063 0.487 0.450

Brown Songlark Cincloramphus cruralis 0.190 0.623 0.188

Brown Thornbill Acanthiza pusilla 0.190 0.623 0.188

Brown Treecreeper Climacteris picumnus 0.655 0.345 0.000

Brown-headed Honeyeater Melithreptus brevirostris 0.190 0.623 0.188

Brush Bronzewing Phaps elegans 0.655 0.345 0.000

(48)

Common name Scientific name Probability of trend

Declining Stable Increasing

Budgerigar Melopsittacus undulatus 0.580 0.420 0.000

Buff-rumped Thornbill Acanthiza reguloides 0.685 0.315 0.000

Chestnut-rumped Heathwren Calamanthus pyrrhopygius parkeri

0.488 0.472 0.040

Cockatiel Nymphicus hollandicus 0.640 0.360 0.000

Collared Sparrowhawk Accipiter cirrhocephalus 0.290 0.647 0.063

Common Bronzewing Phaps chalcoptera 0.085 0.555 0.360

Crescent Honeyeater Phylidonyris pyrrhoptera 0.369 0.581 0.050

Crested Pigeon Ocyphaps lophotes 0.058 0.472 0.470

Crested Shrike-tit Falcunculus frontatus 0.885 0.115 0.000

Crimson Rosella Platycercus elegans 0.130 0.608 0.262

Diamond Firetail Stagonopleura guttata 0.640 0.360 0.000

Dusky Woodswallow Artamus cyanopterus 0.655 0.345 0.000

Eastern Rosella Platycercus eximius 0.270 0.642 0.088

Eastern Spinebill Acanthorhynchus tenuirostris 0.290 0.647 0.063

Eastern Yellow Robin Eopsaltria australis 0.675 0.325 0.000

Elegant Parrot Neophema elegans 0.130 0.608 0.262

Emu Dromaius novaehollandiae 0.535 0.465 0.000

Fan-tailed Cuckoo Cacomantis flabelliformis 0.190 0.623 0.188

Galah Eolophus roseicapilla 0.060 0.480 0.460

Golden Whistler Pachycephala pectoralis 0.130 0.608 0.262

Golden-headed Cisticola Cisticola exilis 0.445 0.555 0.000

Grey Butcherbird Cracticus torquatus 0.655 0.345 0.000

Grey Currawong Strepera versicolor 0.190 0.623 0.188

Grey Fantail Rhipidura albiscapa 0.085 0.555 0.360

Grey Shrike-thrush Colluricincla harmonica 0.085 0.555 0.360

Hooded Robin Melanodryas cucullata 0.655 0.345 0.000

Horsfield's Bronze-Cuckoo Chalcites basalis 0.445 0.555 0.000

Jacky Winter Microeca fascinans 0.885 0.115 0.000

Laughing Kookaburra Dacelo novaeguineae 0.260 0.640 0.100

Little Button-quail Turnix velox 0.565 0.435 0.000

Little Corella Cacatua sanguinea 0.130 0.608 0.262

Little Eagle Hieraaetus morphnoides 0.885 0.115 0.000

Little Grassbird Megalurus gramineus 0.130 0.608 0.262

Little Raven Corvus mellori 0.080 0.540 0.380

Little Wattlebird Anthochaera chrysoptera 0.260 0.640 0.100

Magpie-lark Grallina cyanoleuca 0.080 0.540 0.380

Malleefowl Leipoa ocellata 0.825 0.175 0.000

Masked Woodswallow Artamus personatus 0.655 0.345 0.000

Mistletoebird Dicaeum hirundinaceum 0.270 0.642 0.088

Musk Lorikeet Glossopsitta concinna 0.200 0.625 0.175

Nankeen Kestrel Falco cenchroides 0.078 0.532 0.390

ecosystems within the Coorong, Lower Lakes and Murray Mouth

An assessment of