Declining but not (yet) threatened: a challenge for avian conservation in Australia

ABSTRACT

Threatened species receive much attention in conservation science and practice. Species currently declining, but not yet listed as threatened, also deserve consideration to reduce their risk of sliding towards extinction and to maintain their functional roles in ecosystems. Information on declining bird species in Australia is available from four main sources: national databases, syntheses of historical change, regional monitoring programmes and summaries for guilds of species. Many species show evidence of decline; declines of species are occurring nation-wide, and they are ongoing. Trends for individual species vary geographically; they may be declining in part of their range but stable elsewhere. Common trajectories of population decline include: (a) a downward linear trend; (b) a marked downturn, sustained at a lower level; and (c) fluctuations through time associated with episodic events (e.g. drought) and incomplete recovery. Ongoing declines affect ecosystems through reduced species richness, homogenisation of bird communities, changes to interspecific interactions and ecosystem services, and contributing to extinction debt. Improving the conservation outlook for declining species requires systematic monitoring to know where, when and how much decline is occurring, together with protection of critical habitats and source populations, ambitious programmes of restoration, and identification and effective control of threats. Responding to declining species offers opportunities for community engagement, monitoring and action at a local and regional level. New ways are needed to incorporate such species in conservation planning and environmental regulation at a regional scale, to give them greater visibility and avoid local declines accumulating until taxa become nationally threatened.

POLICY HIGHLIGHTS

• Many bird species in Australia, not (yet) listed as threatened, are experiencing population decline and warrant greater attention in conservation planning and management.

• Patterns of change for individual species vary regionally, depending on past land-use and current pressures.

• Ongoing decline and loss of species has profound consequences for Australian ecosystems, including depleted and homogenised bird communities and reductions in the diverse ecosystem services that birds provide.

• New approaches are needed to provide formal recognition and appropriate management for declining species, particularly at a regional level, so they become more visible in environmental planning and can persist in regional landscapes.

KEYWORDS:

• Birds
• conservation status
• declining species
• extinction debt
• population trend
• threat listing

Introduction

Assessments of conservation status, preparation of threatened species lists, implementation of action plans to counter threats and global rankings that incorporate threatened species are prominent themes in conservation biology (Grenyer et al. 2006; Mace et al. 2008; Garnett and Baker 2021). Typically, species officially listed as threatened are accorded legal recognition and protection, thereby giving them high priority in the review of development proposals, in land-use planning, management plans for national parks and conservation reserves, preparation of state of the environment reports, and in government and philanthropic support for conservation. While conservation of threatened species is critical to the global effort to preserve biodiversity, there are complementary challenges that involve species not officially listed as threatened. For example, recent attention has highlighted the important role of common species in maintaining ecosystem function and in provision of ecosystem services (Gaston and Fuller 2008; Baker et al. 2019). Population declines in such common species may have profound consequences for conservation yet go largely unnoticed, as these species are all but invisible in environmental legislation (Garnett 2020).

Here, our focus is on Australian bird species that can be regarded as ‘declining’, but which are not officially listed as threatened. By ‘declining species’ we refer to those for which there is evidence (quantitative or qualitative) of current or recent (e.g. last 25 years) decline in geographic distribution and/or population size, but not at a level sufficient to meet formal criteria (see International Union for Conservation of Nature [IUCN] Standards and Petitions Committee 2022). Declining species may be widespread and common yet be experiencing a reduction in distribution and abundance; others may be rare and/or limited in distribution. Species may be regarded as declining in one or more regions, while being stable in distribution or population size in another region.

Greater attention to the conservation of declining species is warranted for several reasons. First, identifying species that are declining before they reach threatened status offers greater opportunity for an effective response. Recovery programmes for Endangered or Critically Endangered species, for example, are costly, challenging and require a long-term commitment (McCarthy et al. 2008; Garnett et al. 2018). Acting before a crisis point is reached has multiple benefits, economically and ecologically (Martin et al. 2012).

Second, attention to declining species is important to ensure their continued functional roles in both natural and modified ecosystems. Birds contribute to ecological processes that deliver many ecosystem services and economic benefits to human society: for example, pollination, seed dispersal, consumption of carrion, nutrient cycling, predation of invertebrates and vertebrates, and the transfer of nutrients between aquatic and terrestrial systems (Şekercioğlu et al. 2004; Whelan et al. 2008). Further, the mobility of many bird species, including migratory movements, means that they may link these processes across widely separated geographic areas. The greatest reduction in ecological processes is likely to occur when relatively common species decline in abundance (Baker et al. 2019): even a small percentage decline may result in a large absolute reduction in the number of individuals carrying out an ecological function.

Third, bird species experiencing decline may be important for cultural reasons. For example, some species (e.g. Wedge-tailed Eagle Aquila audax, Emu Dromaius novaehollandiae) have important cultural and totemic values for First Nations people in Australia (e.g. Clarke 2016; Cahir et al. 2018). Birds, including declining species, are among the wildlife that people encounter most frequently in daily life: they can contribute to a connection to nature, a sense of wellbeing and recognition of local heritage (Luck et al. 2011). At local and regional levels, birds of conservation concern can be valuable as flagship species to foster community support for conservation actions, such as the Grey-crowned Babbler Pomatostomus temporalis (Robinson 2006) and Little Penguin Eudyptula minor (Giling et al. 2008).

Finally, insights from research and management involving declining species can offer wider benefits, including for threatened species. Widespread but potentially declining species allow better opportunities to verify the prevalence, form and magnitude of declines, and to identify threats and causes of decline, compared with threatened species which often are rare and localised. Declining, but non-threatened, birds can serve as umbrella species (e.g. top predators such as forest owls), such that measures to enhance their conservation will benefit many other species (e.g. Loyn et al. 2001). Trials of conservation management options, such as experimental habitat manipulation and re-introductions (e.g. Bennett et al. 2013), can provide insights for ecosystem restoration that will also benefit species reaching critical stages of persistence.

Here, we review ‘declining species’ in the Australian avifauna by addressing four main questions:

1. What data sources are available to identify declining species in Australia?

2. What is the process of decline, and are there common trajectories of change?

3. What are the implications of decline of bird species for ecosystems?

4. What can be done to prevent, limit or reverse declines of bird species?

What data sources are available to identify declining species in Australia?

Knowledge on declines of bird species in Australia is documented in several forms (Table 1), each of which provides different insights into the nature and magnitude of decline, the species experiencing decline, apparent causal mechanisms for decline and the consequences of decline. Below, we summarise insights from these different types and sources of knowledge. Taxonomy and nomenclature throughout this paper follow BirdLife Australia (2022).

Table 1. Sources of information on changes in the status of bird species in Australia, relevant to identifying species considered to be declining (but not listed as of threatened conservation status).

Type of information	Description	Examples
1. National data sets and assessment of change at national, state or major bioregional levels	Use of national data sets, collected primarily by citizen scientists (e.g. Atlas projects) to analyse changes through time at national, state or major bioregion levels	Barrett et al. (2002) Comparison between Atlas 1 (1977–1981) and Atlas 2 (1998–2001).
		Ehmke et al. (2015) State of Australia’s Birds, supplements for selected bioregions
		Clemens et al. (2019) Developing waterbird indices for national reporting
2. Historical comparisons of change for a region or locality	Comparison of the historical status of bird species with the recent/present status for a locality, region or conservation reserve	Saunders (1989) – Wheatbelt, WA
		Date et al. (2002) - Pilliga Forest, NSW
		Conole (2002) – Inverleigh F&F reserve, Vic
		Woinarski and Catterall (2004) – Central Qld.
		Fulton (2013) - Dryandra Forest, WA
3. Monitoring at regional or landscape levels, based on repeated surveys of sites or locations	Analyses of change in the status of bird species (e.g. reporting rate, abundance) based on repeated surveys at locations through time	Szabo et al. (2011) - Mt Lofty Ranges, SA
		Kingsford et al. (2017) - Lake Eyre & Murray Darling Basins
		Lindenmayer et al. (2018) - SW Slopes, NSW
		Bounds et al. (2021) - Canberra area, ACT
		Reid and Nicholls (2020) – Cowra area, NSW
4. Syntheses for particular guilds, or ecosystem-based assemblages	Overviews of change in status for a selected guild or ecosystem-based assemblage of birds. Includes both descriptive and quantitative studies	Robinson and Traill (1996) - woodland birds
		Franklin (1999) - granivorous birds, northern Australia
		Clemens et al. (2016) - migratory shorebirds
		Williams et al. (2021) – wet tropics specialists

National data sets

Several national data sets offer insight into changes over time in the status of bird species in Australia: a comparison of two national Atlas projects (1977–1981 and 1998–2001) (Barrett et al. 2002); the State of Australian Birds (SOAB) analyses for the period 1999–2013 (Ehmke et al. 2015); and the Australian Bird Index analyses of waterbirds (Clemens et al. 2019).

Atlas projects

Comparison of the national Atlas projects allowed an assessment of change in the status of birds in Australia between 2 distinct periods, 20 years apart (Barrett et al. 2002). This comparison was challenging, not least because of different observer behaviour and different survey methods in each project, improved technology in the latter period (e.g. availability of GPS) and environmental variation in the intervening years. Analyses were based on reporting rates (i.e. the proportion of surveys in which a species was detected) in 1° grid blocks (n = 561), grouped by 80 biogeographic (IBRA) regions (Thackway and Cresswell 1995), for 411 species (excluding introduced species) with > 200 records in either Atlas. The first step in analysis tested for a regional difference between Atlas periods, and a second step then tested for change between Atlas periods (Table 2).

Table 2. Summary of changes in reporting rate for 411 bird species (not including introduced species) between Atlas 1 (1977–1981) and Atlas 2 (1998–2001) periods. Values are shown separately for threatened and non-threatened species. No change = no significant difference in reporting rate. Data are from Barrett et al. (2002), classification of threatened species as per Garnett and Baker (2021).

Variation between regions	Change in reporting rate 1977–1981 to 1998–2001	No. of non-threatened. species	No. of threatened species	Examples of (non-threatened) species
No regional variation	No change	64	6	Banded Stilt Cladorhynchus leucocephalus
				Bar-breasted Honeyeater Ramsayornis fasciatus
				Olive Whistler Pachycephala olivacea
	Decliners	15	5	Australian Bustard Ardeotis australis
				Flame Robin Petroica phoenicea
				Spotted Quail-thrush Cinclosoma punctatum
	Increasers	80	0	Brown Thornbill Acanthiza pusilla
				Spinifex Pigeon Geophaps plumifera
				Crescent Honeyeater Phylidonyris pyrrhopterus
Regional variation	No change	118	6	Black Swan Cygnus atratus
				Swamp Harrier Circus approximans
				Cockatiel Nymphicus hollandicus
	Decliners	41	2	Little Penguin Eudyptula minor
				Wedge-tailed Eagle Aquila audax
				White-faced Heron Egretta novaehollandiae
				Blue Bonnet Northiella haematogaster
	Increasers	74	0	Magpie Goose Anseranas semipalmata
				Superb Fairy-wren Malurus cyaneus
				Australian Shelduck Tadorna tadornoides

Most species (241/411, 58.6%) showed regional variation between Atlas periods: that is, there were differences between IBRA regions in changes in reporting rate between periods. Of those species modelled, the largest group showed no evidence of change between Atlas periods (194 species, 47.2%), while those showing an increase (154, 37.5%) outnumbered those showing a decline (63, 15.3%) (Table 2). Only 19 threatened species were included in this analysis (4.6% of the total) and these showed either declines (n = 7) or no change (n = 12) (Table 2). A major influence on changes between periods is likely to have been climatic variation. Rainfall was at least 25% greater in the Atlas 2 period than Atlas 1, with substantially higher rainfall across Western Australia, northern Australia and Cape York, and central eastern Australia (Barrett et al. 2002).

State of Australia’s birds

The SOAB analysed annual data from 1999 to 2013 to identify trends through time for individual species and for composite groups of species at the bioregional level (n = 9 bioregions for Australia) (Ehmke et al. 2015). For land birds, models were reported from four bioregions: Arid Zone, Eastern Mallee, South-east Mainland and East Coast (Ehmke et al. 2015). The following summary relates to non-threatened species. There were too few threatened species with sufficient data to model (0–2 species per region) and so a comparison of trends between threatened and non-threatened is not meaningful.

In all four regions, the largest group included species for which no change was detected, comprising 44–54% (mean 51%) of species modelled (Figure 1). The percentage of species found to have declined during this period (23–40%, mean 34%) was over twice the percentage identified as increasing (8–23%, mean 15%). There was marked variation between regions in trends for individual species. Only two species (Australian Pipit Anthus australis, Common Bronzewing Phaps chalcoptera) of those modelled in all four regions had the same trend in each (in this case, stable). Four species declined in three of the four regions (Southern Boobook Ninox boobook, Black-faced Cuckoo-shrike Coracina novaehollandiae, Magpie-lark Grallina cyanoleuca, Welcome Swallow Hirundo neoxena) and only one species increased in three of the four regions (Rufous Songlark Megalurus mathewsi).

Figure 1. Trends in the status of non-threatened species of land birds for four major bioregions from 1999 to 2013, as determined by the State of Australia’s Birds (Ehmke et al. 2015). Numbers above columns indicate the number of species modelled; numbers in parentheses indicate additional species with insufficient data to model.

Commonly, different trends were observed for the same species in different regions. For example, the Striated Pardalote Pardalotus striatus was reported to increase in the Arid Zone and South-eastern Mainland, decline in the East Coast and remain stable in the Eastern Mallee region. When comparing trends for species across pairs of regions, on average 60% (range 54–65%) of species differed in trend between regions. For example, only 26 of 62 species (42%) modelled for both the Arid Zone and Eastern Mallee had the same trend in each, whereas 58% differed. Some species showed markedly different trends between regions: for example, the Australian Magpie Cracticus tibicen showed an increase in the Arid Zone and Eastern Mallee, but declined in the South-east Mainland and East Coast regions. The Willie Wagtail Rhipidura leucophrys was reported to increase in the Arid Zone and South-east Mainland, but decline in the Eastern Mallee and East Coast regions.

Modelling trends for composite groups of species (e.g. aerial insectivores, carnivores) (Figure 2) helps to identify generality in population trends among similar species, includes species for which there are insufficient data to model individually and allows comparison of patterns of change within communities. Outcomes for three biogeographic regions (Figure 2) highlight marked geographic differences in long-term responses in bird communities: there was limited variation over time for bird groups in the Tropical Savanna but strong fluctuations were evident among some groups for the Eastern Mallee and Arid Zone. Notably, in each region, ecological groups respond in different ways to the same broad environmental conditions. Marked declines in some groups (e.g. woodland associated species in the Eastern Mallee, carnivores in the Arid Zone) signal important areas of conservation concern.

Figure 2. Examples of modelled trends through time of indices of occurrence for composite groups of bird species sharing ecological characteristics: a) Eastern Mallee region; b) Tropical Savanna and c) Arid Zone. Thicker lines indicate models which have greater precision. Plots are from Ehmke et al. (2015).

Australian Bird Index for waterbirds

Collation and analysis of data sets for the Australian Bird Index for waterbirds (Clemens et al. 2019) tested for trends in abundance of species over the long term (1983–2017), medium term (1997–2017) and short term (2005–2017). Challenges here included diverse data sets with spatially patchy coverage; the need for different monitoring approaches for species that favour different habitats (e.g. open freshwater, thick wetland vegetation, shorelines); the marked effects of ‘boom and bust’ cycles of drought and rain on many waterbirds, especially in inland Australia; and the movement patterns of many species, both local and long-distance. Key sources for long-term data were aerial surveys from the Eastern Australian Waterbird Survey (Kingsford and Porter 2009; Kingsford et al. 2017) and Shorebirds 2020 monitoring (e.g. Clemens et al. 2016).

Long-term trends for most species were either a decline in abundance or no apparent trend; very few species increased in abundance (Table 3). Contrasts between long-term and medium-term trends for the same species highlight the value of a longer perspective: for example, most of the waterbirds identified as declining in the longer term (~34 years) showed no trend in the medium term (~20 years). The dynamic nature of droughts and floods and their impact on wetland habitats throughout Australia result in marked fluctuations in waterbird populations, requiring a long-term perspective for meaningful insight (e.g. Kingsford and Porter 2009; Hansen et al. 2015; Kingsford et al. 2017). Migratory shorebird species predominantly declined (Table 3), consistent with other analyses of long-term change (Clemens et al. 2016; Studds et al. 2017) and signalling a dire situation for this group, many of which are now listed as threatened. Numerous waterbird species were not included in analyses of long-term trends, a notable group being those associated with thick wetland vegetation (e.g. crakes, rails, bitterns), which are poorly captured in any large-scale monitoring programme and whose populations represent a ‘complete knowledge void in the context of population trends for Australia’s avifauna’ (Clemens et al. 2019).

Table 3. Summary of trends for species of waterbirds (both threatened and non-threatened species) for which long-term trends in abundance in Australia were modelled. Medium-term trends reported for the same species are also shown. Data from Clemens et al. (2019; Tables 5–7).

Taxa	Long-term trend 1983–2017		Medium-term trend 1997–2017	Examples
Waterbirds (excluding migratory shorebirds)	Decline	27*	5 Decline	Australasian Shoveler, Musk Duck
			22 No trend	Chestnut Teal, Australasian Swamphen
	No trend	26^+	2 Decline	Little Tern, Red-capped Plover
			23 No trend	Magpie Goose, Eurasian Coot
			1 Increase	Gull-billed Tern
	Increase	3	2 No trend	Cotton Pygmy-goose
			1 Increase	Cape Barren Goose
Migratory shorebirds	Decline	13^	10 Decline	Far Eastern Curlew, Ruddy Turnstone
			2 No trend	Red-necked Stint, Sharp-tailed Sandpiper
			1 Increase	Double-banded Plover
	No trend	6^#	2 Decline	Greater Sand Plover
			4 No trend	Whimbrel, Sanderling

Clemens et al. (2019) also note additional evidence of a different trend for: *3 species, ⁺5 species, ^1 species and ^#2 species.

Summary – national data sets

Five key insights emerge from these summaries of national data sets. First, there is compelling evidence of population decline for many bird species across Australia, despite limitations in data sets, the use of reporting rates rather than data on abundance, the complexities of analyses, and cautions by the analysts that modelling results be interpreted carefully (Barrett et al. 2002; Ehmke et al. 2015; Clemens et al. 2019). Second, for many additional species there were insufficient data to model trends effectively: for example, comparison between Atlas periods was not possible for > 300 species. For SOAB analyses, even after screening species to include in the modelling process, there were insufficient data for about a quarter of species in each region (mean 24.0% for four regions). Third, the inability to include all species in analyses, together with limited sample sizes and lack of statistical power to detect change among some of those modelled, suggest the number of declining species may be underestimated. Fourth, many species show geographic variation in trends through time, increasing or declining in one region while remaining unchanged in others. Uniform geographic trends are rare. Finally, a fundamental issue in analysing trends through time is the limitations of the underlying data sets, particularly the relatively small number of fixed sites or locations that have been surveyed repeatedly and consistently through time.

Historical changes

A range of studies have synthesised historical changes in avifaunal assemblages, from extensive regions to local conservation reserves. A selection of examples is summarised in Table 4. Discerning overall trends from these studies is complicated by differences in the historical period covered, size of the study area, type and degree of anthropogenic transformation, the assemblage considered (e.g. all species, landbird species, resident species) and subjective interpretations of change in status (Table 4). Nevertheless, they reveal clear evidence of long-term, dynamic changes in the Australian avifauna in response to human land-use.

Table 4. Examples of studies of historical change in avifaunal assemblages from a range of ecosystems in Australia. Studies are listed in order of decreasing size of the study area. In some cases, ‘no change’ can include species of unknown trend.

Author	Study area	Size (km^2)	Land-use changes	Period	Number of species	Locally extinct	Declined	No change (& unknown)	Increased
Smith and Smith (1994)	Western NSW, arid, semi-arid	325,000	Extensive grazing on native vegetation, land degradation	1840s-1990s (~150 yrs)	291 non-vagrants	6	97	105	79 (+6 introd)
Saunders (1989)	WA Wheatbelt, temp. to semi-arid	140,000	Clearing, fragmentation, cropping, pastoral grazing	1900s−1980s (~80 yrs)	148 landbirds	2	39	85	22
Saunders and Curry (1990)	Murchison River catchment, WA, arid.	80,000	Pastoral grazing on native vegetation	1890s−1980s (~90 yrs)	118 landbirds, non-vagrants	3	5	89	21
Noske and Briggs (2022)	Rockhampton, Qld, sub-tropical	10,500	>65% native veg cleared, farming, coastal development.	1888–2019 (131 yrs)	307 resident spp. (1975)	11	11	n/a	n/a
Saunders (1989)	Kellerberrin District, WA Wheatbelt.	~5,500	Loss of > 93% native vegetation, cropping, pastoralism	1900s−1980s (~80 yrs)	131 landbirds	15	23	75	18
Saunders (1989)	Kondanin District, WA Wheatbelt	~5,500	Extensive clearing (>50% vegetation), cropping	1900s−1980s (~80 yrs)	127 landbirds	4	32	74	17
Date et al. (2002)	Pilliga Forest, NSW. Forests, woodlands	>5,000	Extensive logging, grazing, altered fire regimes	1880s−1990s (~110 yrs)	165 woodland species	8	23	87 (+40 unknown)	7
Walter and Walter (2007)	Darling Downs, Qld. Forests, woodlands.	1087	Clearing, farming.	1972–2007 (~35 yrs)	188 non-vagrants	~3	26	~146	13
Woinarski and Catterall (2004)	Central Qld, sub-tropical woodland	454	Clearing (~43%), grazing stock, altered fire regimes	1870s−1999 (~120 yrs)	189 original, (classifiable)	18	68	82	13 (6 introd)
Fulton (2013)	Dryandra Forest, WA. temperate. woodland.	~270 (10 blocks)	Cons. reserves among farmland, plantings ~ 30%	1953–2008 (~55 yrs)	141	3	20	55 (+22 unknown)	41
Conole (2002)	Inverleigh, SW Vic, temperate woodland	10.5	Cons. reserve, plantings, isolated 50–100 yrs	1973–2000 (~27 yrs)	95 non-vagrants	13	9	n/a	n/a
Recher (2023)	Kings Park, Perth, WA, temperate forest.	4	Urban reserve, history of logging, altered fire regime	1928–2008 (~80 yrs)	61	6	15	24	16
Egan et al. (1997)	Longneck Lagoon NR, NSW, woodland	<2	Cons. reserve in extensively cleared ecosystem	1937–1995 (~58 yrs)	98 landbirds (residents)	7	>6	n/a	>3 (+6 introduced
Saunders (1989)	East Yorkrakine NR, WA. woodland	0.8	Cons. reserve isolated by farmland since ~ 1920s.	1975–1988 (13 yrs)	73 landbirds	3	n/a	n/a	n/a

Changes – including local extinctions, declines and increases – are evident across all studies and in diverse ecosystems (Table 4), though the paucity of historical records means the tropical north is less well represented. A key insight is that changes have occurred over a long period and are not just a recent phenomenon. For example, at the property Coomooboolaroo in central Queensland, Woinarski and Catterall (2004) concluded from historical records that 11 species had become locally extinct between 1873 and 1933, and a further 7 species from 1934 to 1999. Substantial numbers of species also declined in each period, as well as there being increases and establishment of new species. Likewise, historical losses, likely in the late 1800s, were reported for western NSW (Smith and Smith 1994). These examples caution against the risk of the ‘shifting baseline syndrome’ (Pauly 1995); studies that span later periods (see Table 4) are unlikely to capture earlier losses associated with post-European settlement (pre-1900) and therefore may underestimate the totality of change experienced in a region or locality.

Drivers of decline in a particular region may be complex, diverse and affect different species in different ways. In the Pilliga region, for example, Date et al. (2002) noted that the region has experienced logging of varying intensity through time, altered fire regimes (including periods of fire exclusion), grazing by stock, droughts, changing densities of European Rabbits Oryctolagus cuniculus and episodic tree regeneration. Some types of historical events may have lasting impact in a region. For example, severe extended drought in western NSW (1895–1903) and central Queensland (1900–1902) were identified as pivotal events affecting the avifauna in these regions (Smith and Smith 1994; Woinarski and Catterall 2004).

The magnitude of reported changes to avifaunal assemblages, particularly local extinctions, appear to be related to the size of the study area and the extent of transformation. In extensive regions, >100,000 km², relatively fewer extinctions were reported than in some smaller areas (Table 4). In extensive areas, a species may persist somewhere in the region despite widespread geographic decline, whereas in smaller areas local decline leads to local loss. For example, Saunders (1989) reported the regional extinction of 2 landbird species for the entire Wheatbelt Region of WA; but, within that, the loss of 4 species for the Kondanin District (>50% cleared) and 15 species for the Kellerberrin district (>90% cleared). Areas subject to more extensive transformation, such as major urban development in cities, have also experienced substantial declines and losses (e.g. Keast 1995; Loyn and Menkhorst 2011).

In many cases, authors of historical studies infer that avifaunal changes are ongoing, and in some cases highlight species likely to decline further and become locally extinct (e.g. Conole 2002; Hewish et al. 2006; Szabo et al. 2011). For example, in addition to reporting the loss of 13 species of woodland birds from the Inverleigh Flora and Fauna reserve in south-western Victoria from 1973–2000, Conole (2002) identified a further 9 species that had declined and were considered vulnerable to loss in the near future.

As for analyses of national data sets, the number of species regarded as declining historically in a study area is likely to be an underestimate. Species may be listed as ‘unchanged’ based on status in remaining native vegetation, but clearing of extensive areas means that, overall, the historical population size has been greatly reduced. For example, in the Western Australian Wheatbelt, species such as the Rufous Whistler Pachycephala rufiventris and Weebill Smicrornis brevirostris, typically relatively common in native vegetation, were listed as not changed (Saunders 1989); yet the historical loss of > 80% of native vegetation in many districts implies that the absolute size of these populations has declined massively. Second, for many species, historical records were too sparse to determine a change in status reliably (e.g. for 40 of 165 woodland birds in the Pilliga region; Date et al. 2002). Such ‘unknown’ species will typically include those that are cryptic (e.g. Lewin’s Rail Lewinia pectoralis), irregular in occurrence (e.g. Little Button-quail Turnix velox) or occur at low density (e.g. Australian Masked Owl Tyto novaehollandiae). Declines of such species are easily overlooked. Third, the perception of the status of a species is influenced by where observations are made; for example, a tendency for observers to spend more time in conservation reserves and remnants of native vegetation than in extensive farmland will bias perceptions and reporting rates by reflecting occurrence in higher-quality habitats where decline is less likely.

A notable feature in reviewing studies of historical changes in Australian birds is the lack of studies that have sought to document knowledge held by long-term residents (but see Abbott 2008), and especially Indigenous people, in remote areas for which other knowledge sources are limited. Such studies have provided critical insights on marked changes in the mammal fauna of inland and northern Australia (Burbidge et al. 1988; Ziembicki et al. 2013).

Longer-term monitoring programmes at the landscape or regional level

Projects that systematically monitor bird species over time at fixed sites at a landscape or regional level have increased over the last two decades. Many are part of broader research projects investigating impacts of land use, landscape change or environmental variation (e.g. Lindenmayer et al. 2018; Reid et al. 2022), or involve citizen-science monitoring to track the conservation status of species in a region of interest (e.g. Szabo et al. 2011; Bounds et al. 2021). Examples of such projects are summarised in Table 5.

Table 5. Examples of monitoring studies at a regional or landscape level in Australia, summarising the time period, temporal span of the study, numbers of species analysed and the numbers for which there was evidence of decline, no change or increase over the period of study. Studies are ordered by year of publication.

				Year	No. years		Species	Number of species
Region	Author	No. sites	Time period	span	surveyed	Total species	analysed	Declined	No change	Increase
Brisbane, QLD	Catterall et al. (2010)	Forest 29 Suburb 38	1991–1993 vs 2006–2008	15	2 periods	Forest 78 Suburb 63	38 31	3 7	34 14	1 10
Mt Lofty Ranges, SA	Szabo et al. (2011)	170	1999–2007	9	9	133 all species	59 Stringybark	9	38	12
							59 Gum	11	36	10
SW Slopes, NSW	Lindenmayer and Cunningham (2011)	66	1998–2009	12	8	116 landbirds	76	4	40	32
Cape York Peninsula	Perry et al. (2011)	418	1998/2001–2008	9	2	267 all species	146	21	116	9
Kakadu NP, NT	Woinarski et al. (2012)	136	2001–2009	6	2 periods	138 all species	91	3	71	17
Nth. Wheatbelt, WA	Saunders and Doley (2013)	Farm	1987–2011	24	24	76 (excl. vagr.)	76	15	59	2
C Wheatbelt, NSW	Ellis and Taylor (2014)	142	2005–2012	8	2 periods	154 landbirds	61	16	23	22
North-Central Vic	Bennett et al. (2014)	240	2002–2012	10	3	n/a, landbirds	132	73	52	7
North-Central Vic	Bennett et al. (2014)	up to 139	1995–2011	15	3	n/a, landbirds	98	53	27	18
Murray Darling Basin	Kingsford et al. (2017)	transects	1983–2014	32	32	50 waterbirds	50	28	21	1
Lake Eyre Basin	Kingsford et al. (2017)	transects	1983–2014	32	32	50 waterbirds	50	1	45	5
Cumberland Pl., NSW	Saunders (2018)	n/a	1970s-2018	~60	6 periods	n/a, woodl. sp	59	20	15 (+11?)	12
SW slopes, NSW	Lindenmayer et al. (2018)	203	2002–2015	13	8	155 landbirds	108	30	64	14
Cowra region, NSW	Reid and Nicholls (2020)	121	2002–2018	17	17	167 landbirds	105	34	26	42
Canberra, ACT	Bounds et al. (2021)	142	1998–2019	21	21	168 landbirds	129	32	73	24
Wet Tropics, Qld	Williams et al. (2021)	114	2000–2016	17	17	54 landbirds	42	19	7	16
Murray floodplain, Vic	Reid et al. (2022)	14, 45	1998–2017	19	4	n/a, landbirds	45	21	23	1

A consistent outcome from these studies is that avifaunal communities are dynamic: in every case there are species that have declined and species that have increased, although the largest proportion remains unchanged (Table 5). Given the variation between studies, there was little statistical difference between the percentage of species analysed that declined over the course of the study (range 3–56%, mean 26.7%) compared with those increasing (2–42%, mean 19.6%) (Table 5, for landbird studies). In most cases, numerous additional species could not be modelled due to data limitations.

Interpretation of these studies and their implications must consider several factors. First, the duration (length) of the study influences the weight that can be given to the changes identified. Populations fluctuate through time in response to multiple factors (e.g. climatic variation, disturbance processes, vegetation secondary succession), and so the longer the period studied (e.g. ≥15 years) the greater the confidence that changes detected represent genuine change in conservation status. Second, the greater the frequency of data collection, the greater the confidence in the reliability of trends revealed and the capacity to interpret change processes. Annual monitoring over a lengthy period (e.g. annually for 17 years; Reid and Nicholls 2020; Williams and de la Fuentes 2021) allows a nuanced understanding of fluctuations across years as well as for the long-term trend. Comparisons based only on two or three survey periods separated by time also reveal change, but it is more difficult to interpret the driving processes because many kinds of environmental change could have occurred in the intervening period.

A third aspect in interpreting change is the period in time when monitoring occurred in relation to wider environmental change. Notably, all studies reported here (Table 5) overlapped (at least in part) with the severe Millennium Drought (~2000–2009) (van Dijk et al. 2013). Some studies specifically focussed on the impact of drought on population trends (e.g. Ellis and Taylor 2014; Reid et al. 2022). Trends through time for many species, and for composite groups (e.g. Figure 2(a,c) above), showed a dip in reporting rate during the drought followed by a partial or complete recovery. Fluctuations through time are common, but of great concern is that for many species the recorded post-drought recovery was only partial, such that there was a net decline overall (Bennett et al. 2014; Reid et al. 2022). For example, of 45 landbird species of the Murray River floodplain (Reid et al. 2022), 78% (35) showed strong evidence of decline during the drought (i.e. between 1998 and 2009). Some subsequently rebounded to varying extents post-drought, but by 2017 (5 years post-drought) 51% (23 species) still showed an overall decline across the 19-year period (Reid et al. 2022). Decline followed by incomplete recovery in response to extreme events (drought, bushfire) is a pathway to endangerment. Alternatively, if full recovery subsequently occurs over a longer period (e.g. following the wet period of the 2021–2022 La Nina years), the alarming number of declines revealed in recent years (Table 5) may be an over-estimate.

Finally, a critical element in interpreting longer-term studies is which species have changed and whether patterns are consistent across locations. Table 6 lists species assessed as declining in at least three of nine monitoring studies from the temperate woodland region of south-eastern Australia (Lindenmayer and Cunningham 2011; Szabo et al. 2011; Bennett et al. 2014; Ellis and Taylor 2014; Lindenmayer et al. 2018; Saunders 2018; Reid and Nicholls 2020; Bounds et al. 2021; Reid et al. 2022). For broader comparison, 26 species regarded as ‘declining woodland birds’ in temperate woodlands (after Reid 1999; Watson 2011) are also shown (Table 6). There was limited consistency among studies: many species were recorded as both declining and increasing in different studies. Of the 26 ‘declining woodland birds’, a trend was reported in one or more studies for 22 of these species with 16/22 (73%) declining in at least one study. Overall, 12/22 (55%) species showed more declines than increases across the 9 studies. The strongest alignment was for the Brown Treecreeper Climacteris picumnus, Dusky Woodswallow Artamus cyanopterus, Eastern Shrike-tit Falcunculus frontatus, Restless Flycatcher Myiagra inquieta, Hooded Robin Melanodryas cucullata and Jacky Winter Microeca fascinans (Table 6), with the south-eastern subspecies of the treecreeper and robin now considered threatened (Garnett and Baker 2021).

Table 6. Change in status of species over time as determined from nine studies in temperate woodland ecosystems in south-eastern Australia (see text and Table 5 for details). Species listed include those regarded as ‘declining woodland birds’ (X, after Watson 2011; n = 26 species) plus species recorded as declining in three or more studies. The Australian International Union for Conservation of Nature (IUCN) category (after Garnett and Baker 2021) is also listed for each species (CR: Critically Endangered, E: Endangered, V: Vulnerable).

Species	Declining woodland birds	IUCN Category	Number of studies (n = 9) in which
			Declining	Unchanged	Increasing
Emu	X			1
Square-tailed Kite	X
Wedge-tailed Eagle			3	1
Painted Button-quail	X				1
Bush Stone-curlew	X			1
Crested Pigeon			3		3
Peaceful Dove			4	3	1
Horsfield’s Bronze-Cuckoo			3	1
Spotted Nightjar	X
Galah			3	4
Sulphur-crested Cockatoo			3	3	2
Red-rumped Parrot			6		1
Swift Parrot	X	CR	1
White-browed Treecreeper	X
White-throated Treecreeper			3		3
Brown Treecreeper	X		7	1	1
Superb Fairy-wren			4	2	2
Little Friarbird			3	1	3
Noisy Friarbird			4		2
Black-chinned Honeyeater			4	1
Regent Honeyeater	X	CR
White-plumed Honeyeater			5	3
Fuscous Honeyeater			4	2
Striated Pardalote			4		4
Speckled Warbler	X		1	3
Weebill			3	1	4
Chestnut-rumped Thornbill	X			1	1
Yellow-rumped Thornbill			3	2	3
Striated Thornbill			3	3	1
Buff-rumped Thornbill			3	5	1
Southern Whiteface	X	V	3	2
Grey-crowned Babbler	X			4
White-browed Babbler	X		2	1	1
White-browed Woodswallow	X		2	1
Dusky Woodswallow	X		5	1
Australian Magpie			3	2	3
Black-faced Cuckoo-shrike			5	4	0
White-winged Triller			4		2
Varied Sittella	X		2	3	2
Crested Bellbird	X		1
Eastern Shrike-tit	X		5	2
Rufous Whistler	X		1	4	3
Grey Shrike-thrush			4	1	3
Willie Wagtail			5	3
Restless Flycatcher	X		4	3
Australian Raven			4	2	2
Hooded Robin	X		4	1
Jacky Winter	X		4	2	1
Eastern Yellow Robin	X		3	2
Red-capped Robin	X			7
Fairy Martin			3	2
Tree Martin			3	3
Mistletoebird			4	3	2
Diamond Firetail	X	V	2	1	2

Syntheses for particular guilds or ecosystem-based assemblages

Evidence of decline, and also likely risk of future decline, in the Australian avifauna comes from syntheses of assemblages that are associated with an ecosystem type, or which share ecological characteristics. We summarise three illustrative examples below: other such groups with reported declines or risk of decline include granivores of the northern savannas (Franklin 1999), malurids (Skroblin and Murphy 2013), ‘mallee’ specialists and birds of prey.

International migratory shorebirds

Long-term monitoring (since 1974) of international migratory shorebirds associated with the East Asian–Australasian Flyway has revealed alarming declines in populations in Australia, where species occur during the non-breeding period (Clemens et al. 2016). For 1996–2014, analyses showed continental decreases in abundance for 12 of 19 species analysed (Clemens et al. 2016). In the southern half of the continent, 17 of 19 species showed population declines. Evidence strongly links population decline of numerous species with coastal development and destruction of tidal mudflats in the Yellow Sea region in Asia. These mudflats serve as critical stop-over habitat for birds to refuel along the migration pathway to breeding sites in the northern hemisphere. For 10 taxa that refuel on Yellow Sea mudflats, their degree of reliance on these mudflats was a much stronger predictor of flyway-level population trends than body size, migration distance, breeding range size or generation time (Studds et al. 2017). Taxa showing the greatest rates of decline are those with greatest dependence on the Yellow Sea. While effective management of shorebird habitats in Australia is important, halting and countering declines will depend largely on measures outside Australia, to protect and replace stop-over sites along the migratory pathway (Studds et al. 2017).

Montane rainforest species

Worldwide, species associated with montane environments are vulnerable to a changing climate due to the narrow environmental range of many montane specialists. Birds associated with rainforests of the Wet Tropics of Australia, a region of outstanding World Heritage significance, show clear evidence of such vulnerability (Williams and de la Fuente 2021; de la Fuente et al. 2023). Patterns of change revealed by monitoring (2000 to 2016) are complex, however, and not consistent across species. Overall, rainforest specialists (29 species) declined in local abundance by ~ 20% on average, whereas habitat generalists (13 species) showed an increase over this period of 9%. Comparisons of change in relation to elevational zones were revealing. Lowland (0–450 m) specialists (6 species) experienced a strong increase of 72% over the study period; mid-elevation (450–850 m) specialists (16 species) declined by 21%; while upland (>850 m) specialists (13 species) experienced a decline of 44% (Williams and de la Fuente 2021). Changes in abundance were strongly related to change in mean temperature, indicating climate as a primary driver. Change for species whose core elevation was lower tended to be positively related to temperature, whereas species with a higher core elevation tended to decline with increasing temperature (de la Fuente et al. 2023). Conservation of declining, high elevation, rainforest specialists is challenging because these declines are occurring within large intact tracts of protected World Heritage rainforest.

Birds of temperate woodlands

The status of birds of temperate woodlands has been of growing concern over recent decades (e.g. Robinson and Traill 1996; Reid 1999; Ford et al. 2001; Watson 2011; Rayner et al. 2014; Fraser et al. 2019). A set of 26 species have been identified as ‘declining’ (see Table 6) (Reid 1999; Watson 2011) while more recently Fraser et al. (2019) proposed that woodland birds, as a community, met criteria for listing under the national Environment Protection and Biodiversity Conservation Act 1999. Large-scale clearing of wooded vegetation in temperate southern Australia, particularly in the ‘sheep-wheat belt’, is undoubtedly the primary cause of population declines. At least four other causal processes are associated, at least in part, with landscape change from clearing: disproportionate loss of particular woodland types, especially those on more fertile soils; fragmentation and isolation of woodland habitats; degradation of wooded vegetation and loss of key resources for foraging, shelter or nesting; and the effects of the hyperaggressive Noisy Miner Manorina melanocephala on other bird species (Ford 2011; Watson 2011; Maron et al. 2013). Decline among woodland species has been reported to disproportionately affect those that are resident, small bodied, ground foraging and insectivorous (e.g. Reid 1999; Watson 2011). However, attempts to quantitatively analyse the characteristics of species identified as declining in regional studies have found it difficult to pinpoint consistent ecological traits (e.g. Mac Nally et al. 2009).

What is the process of decline, and are there common trajectories of change?

Decline is a complex process

Decline of a species is a complex process: it extends over time and typically involves a diverse suite of interacting factors. Such declines have two inter-related components, change in distribution and change in abundance (Figure 3). Assessments for the IUCN Red List use explicit criteria for both the distribution of a species (i.e. geographic range size, total extent of occupied habitat, degree of fragmentation of habitat) and abundance (i.e. total population size, type and magnitude of change in population size, population fluctuations) (IUCN Standards and Petitions Committee 2022; Garnett and Baker 2021). Conceptually, a similar approach could be used to identify declining species but, in practice, judgements generally are based on a combination of change in both distribution and abundance. For example, a decline in reporting rate of a species could arise from reduced distribution (fewer sites suitable), reduced population density (less frequent recording at sites) or a combination of both.

Figure 3. Diagrammatic illustration of the process of decline for a species, involving hypothetical changes in both distribution (e.g. reduced geographic range, fragmentation of populations) and abundance (e.g. overall reduction in population size due to range restrictions, reduced habitat quality in remaining parts of the range).

Change in distribution includes both a reduction in overall geographic range and change in areas occupied within the range (Figure 3). Commonly, decline is a consequence of local habitats becoming unsuitable such that a species no longer persists, with these effects accumulating from local to regional scales. Loss of suitable habitat can occur, for example, through land clearing, draining of wetlands, river impoundment, coastal development and climate-driven change in habitat. Temporal loss of habitat may arise from fires, logging of forests, excessive water extraction from rivers and contraction to refuges during extended drought. For example, loss of the Yellow-plumed Honeyeater Ptilotula ornata from the Kellerberrin district in the Wheatbelt of WA was associated with clearing > 90% of wooded vegetation (Saunders 1989); while wildfires from 1989 to 2005 are thought to have extirpated (at least temporarily) the Mallee Emu-wren Stipiturus mallee from most of its geographic range in South Australia (Brown et al. 2009).

Change in overall population size of a species is influenced by a reduction in distribution, but also by processes that operate within the occupied range (Figure 3). Degradation in the quality of habitats, impacts of exotic species, biotic interactions with other native species, extraction of water resources and environmental fluctuations can all have detrimental effects. For example, cattle grazing reduces the density of numerous bird species associated with ground and low shrub layers in forests and woodlands (Martin and Possingham 2005); biotic interactions with the hyper-aggressive Noisy Miner reduces the density of small insectivores across vast areas of fragmented woodland in eastern Australia (Mac Nally et al. 2012; Maron et al. 2013); feral and domestic cats are estimated to kill > 370 million birds per year in Australia (Woinarski et al. 2017); and water use and diversion in the Murray Darling Basin has resulted in population decline of many waterbird species (Kingsford et al. 2017).

Changes in both distribution and local population size have compounding impacts for bird species. Fragmentation and isolation of occupied habitats affects the capacity for effective dispersal and gene flow between local populations, as shown for the Blue-breasted Fairy-wren Malurus pulcherrimus in remnant vegetation among farmland in the WA Wheatbelt (e.g. Brooker and Brooker 2002). Where populations are small, they become increasingly vulnerable to disturbance events, loss of genetic diversity and stochastic environmental change. The net result is even further population decline.

Trajectories of change through time

Visual portrayal of trends through time is a powerful tool for understanding and communicating the changing conservation status of a species or assemblage (e.g. Gregory et al. 2004; Bayraktarov et al. 2021). Analyses of trajectories of change require long-term data sets. In addition to the SOAB (Ehmke et al. 2015) (see Figure 2), several regional monitoring projects based on citizen science surveys have analysed longer-term trends: 17 years for the Cowra region, NSW (2002–2018; Reid and Nicholls 2020) and 22 years for the ACT region (1998–2019; Bounds et al. 2021).

A discrete set of trajectories of change can be recognised, despite subtle variation among species. We identified 7 trends (Figure 4) common to multiple studies and illustrate their frequency of occurrence (and cite examples below) based on models of long-term change for 129 species in the ACT (Bounds et al. 2021). In that study, 56 species showed significant directional trends over time, 47 showed no evidence of change and for 26 the data were too limited to be confident of a trend. Trajectories of decline (Figure 4(b)) take three general forms: (i) a linear decline indicating ongoing unfavourable conditions or interacting stressors (e.g. Hooded Robin, Rufous Whistler Pachycephala rufiventris, Grey Currawong Strepera versicolor); (ii) a marked decline then stabilising at a lower level, indicative of a disturbance event from which the species shows little or no recovery (e.g. Willie Wagtail, Eastern Shrike-tit); and (iii) a fluctuating decline in which there are periods of both decline and increase but with net decline overall, a response to periodic disturbance (such as drought) but with incomplete recovery (e.g. Noisy Friarbird Philemon corniculatus, Superb Fairy-wren).

Figure 4. Generalised trajectories of change in reporting rate through time for bird species, including: (a) increasing trends; (b) declining trends; (c) stable trends; and d) the percentage of bird species (n = 129) showing each trend type from long-term monitoring of landbirds in the ACT (1998–2019) (Bounds et al. 2021).

Trajectories of increase (Figure 4(a)) follow two main forms: (iv) a linear increase, indicating ongoing favourable conditions, adaptation to transformed landscapes or use of new resources (e.g. Australian Raven Corvus coronoides, Australian King Parrot Alisterus scapularis); and (v) a fluctuating increase with periods of both increase and decline but net increase overall (e.g. Galah Eolophus roseicapilla, Pied Currawong Strepera graculina). Likewise, there are two main forms of ‘no change’ over time (Figure 4(c): (vi) a relatively stable (linear) population trend (e.g. Laughing Kookaburra Dacelo novaeguineae, Nankeen Kestrel Falco cenchroides); and (vii) a fluctuating pattern but with no apparent long-term change over time (e.g. Common Bronzewing Phaps chalcoptera, Red-capped Robin Petroica goodenovii). The relative frequency of these trajectories differs, but for the ACT monitoring programme (Figure 4(d)) both linear and fluctuating forms of decline and increase were well represented over the 22-year period (Bounds et al. 2021). Notably, fluctuating trends are characteristic of species for which resources follow a ‘boom and bust’ cycle (e.g. waterbirds, arid-zone nomads).

The trajectory of change for a species in a particular region and time period should not be directly extrapolated to other regions or time periods without verification. Interpreting temporal trends requires consideration of the nature of the data, the location and duration of monitoring, and the specific time period and pressures experienced. For example, trends in the ACT monitoring programme, particularly the fluctuating patterns of increase or decline, appear to have been influenced by the Millennium Drought (~2000–2009). The trend in a particular region may also be influenced by nearby or distant environmental conditions. During prolonged drought, for example, major population shifts can occur from drought-affected environments to other less-affected regions. Population trajectories of greatest concern from a conservation perspective are those showing a marked decline or linear decline, both requiring an urgency to determine the pressures causing the decline and to respond.

What are the implications of decline of bird species for ecosystems?

Birds are a visible and prominent part of Australia’s biodiversity, they are readily recorded by community observers and there are more data on bird distributions and trends in bird populations than for other taxa. They may serve as an indicator of wider patterns of biodiversity change, for taxa for which decline is known to be occurring but data are sparse and patchy (e.g. invertebrate groups; Braby et al. 2021). Indeed, declines in other taxa, may be even more severe than for birds. In Kakadu NP, for example, systematic repeated surveys (~2001–2009) at the same sites showed much greater declines for small mammal species than for birds (Woinarski et al. 2010, 2012).

Changes in the status of bird species (and other taxa) have diverse implications for the structure and function of ecosystems. First, trends are not uniform across regions and ecosystems (Figure 2): understanding where declines are most severe and which ecosystems are most vulnerable is critical for targeting conservation effort. Generally, greater declines occur in ecosystems that have experienced the greatest anthropogenic change. At a regional scale, for example, changes in western NSW (Smith and Smith 1994) have been greater in semi-arid than arid ecosystems, because semi-arid areas have experienced greater agricultural clearing. In turn, greater change appears to have occurred in arid pastoral rangelands in western NSW (Smith and Smith 1994) than in similar rangelands in northern Western Australia where there has been a shorter history of exploitation (Saunders and Curry 1990) (Table 4). Declines of waterbirds are more prevalent in the Murray Darling Basin than the Lake Eyre Basin (Kingsford et al. 2017), the former having experienced more severe impacts from water extraction and diversion. Even in the same general location, impacts vary between ecosystems. In the Cowra region, NSW, trends from 2002 to 2018 for species in open forests of the ranges tended to be mostly increases; whereas for species in remnant woodlands on more fertile, productive landscapes, declining trends were common (Reid and Nicholls 2020).

Second, decline and local loss of species alter the structure of bird communities, including a reduction in species richness and simplification and homogenising of the community. Historical syntheses (Table 4) and regional monitoring studies (Table 5) report both declines and increases in species for locations studied, but generally declines outweigh increases. Reduced species richness of bird communities is most evident where environmental transformation has been greatest, such as in cities and cleared farmland. In rural landscapes, those with greater loss of native vegetation have fewer bird species, especially forest and woodland species (Bennett and Ford 1997; Radford et al. 2005; Cunningham et al. 2014). Reductions also occur at the patch level: in Gippsland, Victoria, repeat surveys of forest birds in remnant forest patches 22 years after initial surveys (MacHunter et al. 2006) revealed a consistent reduction in the number of species per patch, such that the species-area relationship shifted significantly downwards.

In many locations, loss of species results in a shift towards a more homogenised bird community, dominated by a more uniform set of species. In agricultural landscapes, a reduction in forest and woodland species typically is accompanied by increased prevalence of widespread, ‘open country’ species able to thrive in farmland with scattered trees or small wooded patches and strips (e.g. Lynch and Saunders 1991; Barrett et al. 1994). Some evidence suggests a negative feedback loop, such that increased prevalence of these species may further affect woodland species: for example, through greater abundance of predators (butcherbirds, currawongs), competitors for tree hollows (e.g. Galah) and aggressive interspecific competition for food sources (e.g. Noisy Miner, Yellow-throated Miner Manorina flavigula). Sites with miner colonies support fewer species and more homogenous communities (Robertson et al. 2013; Kutt et al. 2016).

A third implication for ecosystems is changes to ecological processes when declining species are less able to fulfil their functional roles. While birds contribute to many ecological processes (e.g. insectivory, pollination, seed dispersal) (Ford and Paton 1986), there are relatively few studies that demonstrate the consequences of population decline on such processes. Experimental work in the Mt Lofty Ranges (Paton 2000; Paton et al. 2004), for example, showed that in the absence of bird pollinators, seed production of several bird-pollinated shrubs is depressed. Even where birds have access, there is pollinator limitation, likely due to reduced abundance of pollinators (honeyeaters) arising from disproportionate clearing of woodlands on more fertile soils (Paton et al. 2004). The abundance of insectivorous birds has been linked with tree health, particularly where insectivore populations are reduced by colonies of miners. Loyn et al. (1983) found that following experimental removal of Bell Miners Manorina melanophrys, there was an immediate influx of insectivorous birds which rapidly consumed psyllids (nymphs and lerps) that infest tree foliage and affect tree health, such that there was a 15% increase in epicormic growth of foliage on trees. Ecosystem engineering by the Superb Lyrebird Menura novaehollandiae, through its foraging activity on the forest floor, has profound effects on wet forest ecosystems. Lyrebirds annually turn over ~ 155 tonnes per ha of litter and soil, reduce soil compaction, increase litter depth and modify ground-layer vegetation (Maisey et al. 2021, 2022).

Fourth, there are time-lags in ecosystems experiencing the full consequences of declining bird populations. Extinction debt refers to future losses anticipated as a consequence of past actions. For example, decline of forest and woodland birds from historic land clearing plays out over time as local populations become isolated, are no longer self-sustaining and gradually disappear. In the Mt Lofty Ranges, South Australia, where ~ 90% of original forests and woodlands have been cleared, Ford and Howe (1980) used species-area theory to predict that up to 35–50 species might be expected to become locally extinct. By 2000, the region had lost 8 of 120 species of the pre-European avifauna, and by 2011 a further 10 species were judged to be ‘not viable’ over the long term, and another 10 were showing ‘troubling declines’ (Szabo et al. 2011). Paying the extinction debt from past land clearing and isolation is likely to be a major contributor to future declines and extinctions (e.g. Conole 2002; Ford et al. 2009). The challenge is to avert or delay payment of such debt by actively engaging in habitat restoration, both to increase the quality of existing habitats and greatly expand the amount of habitat (e.g. Paton et al. 2004).

What can be done to prevent, limit and reverse declines of bird species?

While recognising the breadth of the challenge and range of species with evidence of decline, conservation is best served by focussing on opportunities to make achievable gains, rather than dwelling on ‘gloom and doom’ messages of decline (Table 7). There is much potential to learn from experience in different parts of Australia and other countries because the conservation challenge of declining bird species is not unique. There is much experience in Europe, for example, in developing conservation responses to declines of species such as farmland birds, including effective monitoring, use of indices to communicate trends, diagnosis of causes of decline, experimental tests of interventions, and implementing broad-scale schemes to enhance habitats (e.g. Donald et al. 2001; Newton 2004; Baker et al. 2012).

Table 7. Measures to conserve declining species of birds in Australia.

Issue	Actions	Explanation
Increase awareness in the Australian community that many species are declining, the process of decline is ongoing and there are consequences for ecosystems and cultural values.	Build recognition that declines are occurring for a wide range of species. Ensure management agencies consider declining species in conservation plans, strategies and programmes. Document the status of species to avoid shifting baseline syndrome. Identify and protect species of cultural significance (especially to Indigenous Australians).	Reversing declines requires community awareness that ongoing declines are real, and have implications for extinction risk, loss of ecological function and culture. Community support provides a foundation for more attention to the conservation of declining species.
Avoid and prevent ongoing loss and degradation of habitats.	Ensure ongoing national, state and local efforts to prevent loss and degradation of habitats in all ecosystems – terrestrial, freshwater, coastal and marine.	Loss and degradation of habitats is a fundamental cause of population decline and change in distribution.
Protect habitats critical for core populations and a capacity to respond to disturbance and extreme events, such as those associated with climate change.	Give priority to key habitats including: (a) specialised, restricted habitats; (b) productive environments (e.g. riparian zones, floodplains); (c) large intact areas that serve as demographic sources; and (d) refuges from extreme events (e.g. drought, wildfire).	Conservation of declining species requires retention of high-quality habitats, and a capacity for population recovery from disturbance and extreme events (drought, wildfire, extreme temperatures).
Restore habitats, particularly in modified environments.	Support, promote and implement restoration programmes in terrestrial and aquatic ecosystems (e.g. large-scale revegetation, natural regeneration, reduced grazing pressure, environmental watering).	Restoration will help reverse the loss of habitats, improve the quality of existing habitats and thereby increase population sizes and resilience.
Identify and ameliorate other threats that drive bird declines	Incorporate measures to reduce diverse threats (e.g. pesticides and pollutants, exotic species) through conservation action and management plans.	Diverse factors that directly (e.g. feral predators) or indirectly (pesticides that reduce invertebrate food resources) affect populations contribute to decline.
Engage internationally to protect habitats of migratory species.	Work through international treaties and inter-governmental negotiation to protect critical migratory habitats, especially stop-over points.	Environments in other countries are critical for intercontinental migrants, including terrestrial and shorebird species.
Expand applied research programmes to test outcomes of land management and conservation interventions; and forecast future change.	Use experimental approaches to test outcomes of management and conservation interventions. Forecast trajectories likely with climate change, extinction debt and other factors, to identify species sensitive to decline and needing prospective actions.	Effective conservation depends on knowledge of what works and for which species, what does not, and why species respond to change in particular ways. Knowledge is needed to inform effective pro-active programmes.
Enhance systematic monitoring at fixed sites, stratified across land use and environmental gradients. Include regions for which knowledge is limited.	Coordinate design of monitoring programs for different bird groups at the national scale; implement at regional and local scales in a consistent way, encouraging citizen science. Document species trends at local and regional scales to support local action.	Long-term monitoring data are critical to assess: (a) changes in distribution and abundance of species; (b) threat status; and (c) outcomes from land-use change and management actions. Important at both national and local scales.
Enhance the official recognition of ‘declining’ bird species at national, state and regional scales.	Identify and set criteria by which population decline at a regional level can trigger recognition of a species as ‘declining’.	Species not listed as ‘threatened’ can be invisible in environmental legislation and planning, and receive less consideration in planning schemes and land uses.

Building community awareness, recognition and involvement

Progress with halting and reversing declines of bird species depends on recognition in the wider community that declines are real, are happening now and will affect the quality of life and the natural and cultural heritage of Australians (and the world). Such community awareness is an essential foundation for wider conservation measures (Table 7). Valuable elements of community engagement include positive stories of success and achievement (e.g. case studies of habitat restoration, recovery of local populations), demonstration of specific actions that individuals and groups can undertake (e.g. via field days, group activities) and opportunities for active participation (e.g. monitoring, habitat restoration). Birds are visible and can be seen every day, and many Australian species are visually attractive and have fascinating behaviours and life-history. Community support for the conservation of birds can generate wider benefits for nature conservation and other taxa that occupy the same habitats. Importantly, community support adds weight to the call for governments to give greater attention to nature conservation through more effective environmental policy and legislation, funding, and management of land and water for conservation.

Habitat protection and restoration

The persistence of species experiencing decline depends fundamentally on the availability of suitable habitat of sufficient extent, location and quality (Table 7). This requires maintaining habitats (terrestrial, aquatic, coastal) in and among human land uses (e.g. urban, agricultural areas), and in extensive natural areas with less direct human impact. In cities and agricultural regions, the emphasis is on protection of natural remnants and aquatic systems, and restoration of habitats via ambitious programmes of revegetation and natural regeneration. Multiple mechanisms are needed, including voluntary contributions, stewardship schemes, land purchase, and approaches based on carbon and biodiversity markets. Attwood et al. (2009) proposed that Australia could incorporate aspects of the European agri-environment approach into stewardship schemes in rural environments; particularly the longer-term nature of agreements, payments to landholders more commensurate with loss of production, and actions specifically targeting agricultural lands (as well as remnant vegetation).

Existing natural areas, particularly extensive areas in parks and conservation reserves, multi-purpose forests, rangelands and wetland systems, are critical as reservoirs for larger populations and as demographic sources from which dispersal and recolonisation can occur. Key issues include the management of natural areas such that any use of resources (e.g. grazing by stock, timber harvesting, mining) does not cause declines; and to maintain resilience of ecosystems in the face of major disturbance events (e.g. floods, drought, wildfire). Areas of particular importance for sustaining resilient populations include large continuous tracts, productive environments in moister areas and on more-fertile soils, and areas that function as refuges from extreme events (drought, wildfire).

Monitoring and research

A range of monitoring programmes at national and regional scales (e.g. Tables 2, 3, 5) make valuable contributions to understanding change in the Australian avifauna. However, while monitoring of threatened species has improved over the last 30 years (Verdon et al. 2024), the absence of a coordinated national programme that systematically monitors all species across the continent in relation to environmental gradients and land uses severely hampers understanding change. A carefully designed national programme could include and complement regional monitoring programmes (such as that in the Cowra region; Reid and Nicholls 2020), capitalise on the interest and commitment of volunteer observers and provide a wealth of information at both national and regional levels on distribution, habitat use, effects of landscape change and temporal population trends. It would require dedicated funding for coordination, analysis and communication. Citizen-science observations (e.g. eBird, Atlas of Living Australia) could also contribute to regional and national monitoring programmes, especially when using a structured survey protocol (Callaghan et al. 2019). Citizen-science data sets have made important contributions to assessing the status of both threatened (e.g. Rogers et al. 2021) and non-threatened species (Lee et al. 2023).

While monitoring programmes document distributional patterns and temporal trends, further applied research is needed to determine why particular species are more (or less) vulnerable to decline (e.g. see Olah et al. 2024), the nature of population changes that occur, and the outcomes and effectiveness of management programmes and conservation actions (e.g. Walsh et al. 2023). In an era of climate change and a likely increased prevalence of large-scale disturbances, generating an understanding of how and why species respond to such events can help with pro-active planning for the future. Research is also needed to develop and refine new approaches to study species for which traditional techniques have limited value. Automated acoustic monitoring (e.g. Rowe et al. 2023) and identification of eDNA (e.g. Neice and McRae 2021) to detect cryptic species, for example, offer new opportunities, as does the use of molecular genetic techniques to evaluate changes in population demography and conservation status of species in modified environments (e.g. Harrisson et al. 2012).

Increased recognition for declining species

Species officially listed as threatened (i.e. Critically Endangered, Endangered, Vulnerable) receive special consideration in national, state and territory environmental legislation, and generally also in conservation strategies, planning schemes and other processes. Species regarded as declining do not receive such recognition and may effectively be ‘invisible’ (Garnett 2020). Greater recognition could make an important contribution towards better protection and restoration.

At a national level, ‘Near Threatened’ is recognised in assessments of the IUCN conservation status of Australian birds (Garnett and Baker 2021), whereby ‘A taxon is Near Threatened when it has been evaluated against the criteria but does not qualify for Critically Endangered, Endangered or Vulnerable now, but is close to qualifying for or is likely to qualify for a threatened category in the near future’ (IUCN 2022). However, ‘Near Threatened’ is not recognised in the Australian EPBC Act. Inclusion of such a category would give greater recognition for some declining species. In the 2020 review of Australian birds (Garnett and Baker 2021), 35 taxa (22 species) were assessed as Near Threatened, compared with 201 taxa (92 species) in threatened categories.

Given geographic variation in population trends, a regional perspective is also needed. A species may be experiencing severe decline in some regions but not in others, such that a single national assessment is not representative and may obscure active processes of decline towards threatened status. Official recognition at a sub-national level would allow greater scrutiny and consideration in regions where change is occurring. At a state level, conservation classifications can include recognition of species not formally regarded as threatened at a national level, such as species listed in Victoria under the Flora and Fauna Guarantee Act. Recognition at a regional level (e.g. IBRA natural regions; Thackway and Cresswell 1995), based on determined criteria and evidence, could support local awareness and community actions to limit further decline.

Acknowledgements

We warmly acknowledge all those who have contributed to monitoring programmes and analyses of monitoring data sets at national, regional and local scales. These provide an essential foundation for understanding the distribution and population trends of Australian bird species. We thank two reviewers for thoughtful comments on the ms.

Disclosure statement

No potential conflict of interest was reported by the authors.

Correction Statement

This article has been corrected with minor changes. These changes do not impact the academic content of the article.