Considerations in the protection of marsupial gliders and other mature-forest dependent fauna in areas of intensive logging in the tall forests of Victoria, Australia

Keywords: ecologically sustainable forest management, hostile environments, intense wildfire, intensive logging, mature-forest dependent species, protection areas, Southern Greater Glider, tall eucalypt forest, Yellow-bellied Glider.

The tall eucalypt forests of Australia are important in the regulation of planetary processes, for biodiversity conservation,carbon sequestration, recreation, amenity, water and for the yield of timber (Lindenmayer and Franklin 2002; FAO 2016; Wardell-Johnson et al. 2017; Watson et al. 2018). In the State forests of Australia, these forests have been exploited both intensively (i.e. clearfelling, with or without seed tree or habitat tree retention), and extensively (i.e. over a wide area) in the decades following the advent of woodchipping around 1970 (see Routley and Routley 1974; Resource Assessment Commission 1991; Dargavel 1995; Lindenmayer et al. 2022a).

Since about 1970, most of the tall or wet sclerophyll (see Wardell-Johnson et al. 2017) forests in the State forests of Australia available for logging have been altered in structure, composition and function by the large scale (i.e. industrial scale) of this logging activity (Resource Assessment Commission 1991; Dargavel 1995; Lindenmayer et al. 2011a, 2011b, 2022a; Wardell-Johnson et al. 2019). Much of the remainder of the available mature forest are foreshadowed to be logged by 2030 (VEAC (Victorian Environmental Assessment Council) 2017; Forestry Corporation 2005–2019; Swann and Browne 2016). A trend towards conversion of mixed aged old growth (single or mixed species) to single aged and species regrowth has been notable in the fashioning of these forests. As a result, there has been a tendency towards shorter rotations for what has been claimed to be the sustained yield of forest products (see Resource Assessment Commission 1991; Dargavel 1995; Calver and Wardell-Johnson 2004; Swann and Browne 2016). The ongoing logging operation return times of 40–60 years are within the immature phase of the regenerating stand (see Jacobs 1955; Bradshaw and Rayner 1997).

The implications of this conversion of mature forests to regrowth has led to recognition of the imperative of a complementarity of State forests to the conservation reserve network (e.g. Lindenmayer and Franklin 2002; Lindenmayer et al. 2011b; Slade and Law 2017). Thus, within State forests, there is now an array of approaches to conservation, ranging from strategic set asides as ‘Special Protection Zones’ or Protection Areas (PAs) and ‘Road, River and Stream Zones’ to operational activities, such as the protection of ‘habitat’, ‘feed’ or ‘seed’ trees, retention of ‘minimum basal area’ commitments and ‘variable retention’ logging in the light of intensive logging carried out extensively (Lindenmayer and Franklin 2002; Lindenmayer et al. 2012b; Department of Environment and Primary Industries 2021). These are continuing to evolve to account for climate change events such as the Black Summer fires of 2019–2020 (e.g. Department of Environment and Primary Industries 2014, 2021; Office of the Conservation Regulator 2020).

Regardless of these operational considerations, logging return times that are more rapid than the return time of the mature structure of the forest have consequences for those components of the forest that take longer to regenerate or return (Lindenmayer et al. 2020). Thus, a particular subset of forest-associated species of wildlife, including mature-forest dependent and large hollow dependent species are disadvantaged at coupe (from a few 10s to a few 100s of ha); compartment or forest block (a few 1000s of ha and encompassing numerous coupes; see Kavanagh et al. 2004; Loyn 2004; Lindenmayer et al. 2020); and therefore, landscape and regional scale (Lindenmayer et al. 2020; Smith 2020) by these changes to the forest.

Several mature-forest and hollow-dependent species are disadvantaged by changes from mature to regrowth forest (e.g. Leadbeater’s Possum, Gymnobelideus leadbeateri; Powerful Owl, Ninox strenua; Greater Sooty Owl, Tyto tenebricosa; Yellow-bellied Glider (YBG), Petaurus australis and Southern Greater Glider (SGG), Petauroides volans). For example, glider surveys in Victoria show significant range and density declines of the YBG (Box 1) and SGG (Box 2) in the logged wet forests of the Central Highlands, Alpine Areas and dry forests of East Gippsland of 50–85% since European invasion. Most of the decline has occurred in the last 25 years, pointing to intensive logging as the primary cause of glider decline in Victoria (Lindenmayer et al. 2020; Smith 2020). Thus, if mature-forest dependent species are to persist, conservation attention is required at the coupe level in the context of timber production in State forests.

Attention on YBGs and SGGs has resulted in developments in survey methodology to define priority areas for PAs, and the setting aside of these areas from logging (Department of Environment and Primary Industries 2014, 2021), wherever threshold population levels are met. Thus, a PA of approximately 100 ha of suitable habitat is required to be applied where a relative abundance of more than 10 individual (i.e. about two individuals per ha) SGGs or Common Brushtail Possums (Trichosurus vulpecula), or five individual YBGs (i.e. more than 0.2 individuals per ha) have been recorded per spotlight kilometre (table 13 of the Standards, Department of Environment and Primary Industries 2021). These PAs are therefore established where populations are at their most dense in State forests.

There have also been developments in the introduction of ‘variable retention’, ‘basal area retention’ (see Parliament of Victoria Legislative Council Environment and Planning Committee 2021) and implementation of ‘set asides’ within logging coupes to limit the impact of logging on these species and other High Conservation Values (HCVs). Nevertheless, even with operational plans and administrative arrangements for the implementation of PAs and set asides, there is no certainty that the most appropriate areas, shapes, configuration and management of such areas will take place without clear guidance, principles or consideration of their characteristics.

Ecological principles have long guided forest management (e.g. Holt and Talbot 1978; Wardell-Johnson et al. 1989; Lindenmayer and Nix 1993), but to be effective, require assessable criteria for application (Calver et al. 1998). Further, shallow use of principles is counterproductive (see Abbott and Christensen 1994, 1996 and responses by Calver et al. 1996, 1998). As a rule, such principles should be guided both by theory, and the biology and ecology of the organism/s for which the management is targeted. Further, such principles or considerations require guidelines around definitions to be effectively implemented into management practice. Finally, a process is required to enable managers to be responsible for the effective management of these forests for the purposes for which they have been designated.

In this paper, we outline strategic and operational considerations for application in the protection of HCVs (specifically YBGs and SGGs) in Victorian forests in the context of a State native forest-based timber industry. We ask:

1. Given that the requirement to set aside approximately 100 ha of suitable habitat as a PA for SGGs and/or YBGs, wherever threshold population levels are met, what principles or considerations should be applied to their location, design and management?

2. Will appropriately implemented PAs be sufficient to safeguard YBGs and SGGs in State forest? If not, how should these species by conserved elsewhere in State forests when the threshold population for setting aside PAs for these species is not met?

Prior to outlining 10 considerations for the design of these PAs, we provide definitions and background towards their development. We conclude by suggesting that a genuine commitment to ecologically sustainable forest management (ESFM) is required to support the conservation of mature-forest dependent species, and also the on-going viability of the timber industry. We conclude by outlining the commitment that entails.

At broad scales (landscape and region), logging can theoretically be carried out within the principles of a CAR (comprehensive, adequate and representative) reserve system (JANIS 1997), ESFM and for the long-term stability of forests and forest industries (Lindenmayer and Franklin 2002). ESFM, implies the sustainable management of all the products and values provided by forests (see Lindenmayer and Franklin 2002; Lindenmayer et al. 2000, 2012b), although the conceptual ambiguities and practical challenges of ecological forestry and hence, ESFM have been recognised (Batavia and Nelson 2016).

Nevertheless, high levels of resource commitment have potential to threaten both the conservation of SGGs and YBGs, and the native forest industry. That is because it compromises ESFM by reducing return times to forest coupes (i.e. in this case to between 40 and 60 years), leads to eras of timber supply bottlenecks, limits capacity for set asides in the context of HCVs, or limits options for basal area retention in forest coupes. For example, VEAC (Victorian Environmental Assessment Council) (2017) concluded (prior to the 2019–2020 Black Summer fires) that a decline in ash sawlog supply in Victoria indicates that alternative sources will be required to meet industry commitments (of 265 000 m³ per year of ash pulp logs) until 2030 based on legislated supply agreements. Further, that the situation applying in Victoria may have broader applicability is demonstrated by biomaterial reports from the forests of south-eastern New South Wales. Here, logging levels for the 25-year period 2005–2029 (excluding 2017–2020) are recognised by the NSW Forestry Corporation as being above the long-term sustainable yield (Forestry Corporation 2005–2019). This suggests similar or even more rapid return times to supply resource-based forest industries in south-eastern New South Wales.

Logging that provides a sustained yield of all the components provided by forests need not have significance, even at local scales. However, reviews have suggested that the sustainable exploitation of renewable resources has never been successfully achieved in industrial societies (e.g. Ludwig et al. 1993; Hilborn et al. 1995). Thus, the scale in space (impact increasing with intensity and extensiveness) and time (impact increasing with more rapid return times to the coupe), determines impact at all scales. Where the return time of intensive logging is faster than the recovery of the structure on what YBGs and SGGs depend (upwards of 150–200 years), then logging causes serious and irreversible impact at all spatial scales (Lindenmayer et al. 2020; Smith 2020).

Any extensive activity or process that fundamentally alters the long-term (i.e. >100 years) structure of the forest has impact on local wildlife at all spatial scales. However, intensive logging has much greater significance at the coupe scale than intense wildfire, and does not equate, mimic or resemble the disturbance of intense wildfire (Lindenmayer and McCarthy 2002; McLean et al. 2015; Karna et al. 2019; Wilson et al. 2021). This is because a much greater proportion of the biomass and structure of the forest required by mature-forest dependent species is removed by intensive logging than by intense wildfire. Thus, areas post the widespread 1939 wildfires in Victoria are not now generally regarded as hostile environments for mature-forest dependent species, or as unsuitable habitat for YBGs or SGGs. This is because they have retained a substantial proportion of the structural components of the forest that was present prior to these fires (see van der Ree and Loyn 2002).

By contrast, clearfelling removes the structure required by mature-forest dependent species within the coupe; a structure not regained for centuries. The retention of habitat, seed or feed trees, or other components can potentially reduce the return time of the structure required by mature-forest dependent species, but will not in itself ensure their conservation (see below). It should be noted that logging of forest that has been disturbed by wildfire (i.e. salvage logging) is particularly damaging to wildlife (Lindenmayer et al. 2012a).

Intensive logging of eucalypt forests changes their structure to a stand dominated by regrowth, making them more vulnerable to wildfire for upwards of 50 years, thereby impacting at the regional scale (Attiwill et al. 2014; Lindenmayer and Taylor 2020; Lindenmayer et al. 2022a, 2022b; Bowman et al. 2021; Zylstra et al. 2021). Thus, the flammability of tall eucalypt forests changes as a stand develops from regrowth to old growth, with flammability increasing for at least several decades after intensive logging (Furlaud et al. 2021). Further, intensive logging makes fires more severe and the resultant regrowth more fire sensitive (Taylor et al. 2014; Wilson et al. 2018; Lindenmayer et al. 2011a, 2021, 2022b; Furlaud et al. 2021). This is because logging initially redistributes flammable leaf and branch material from the canopy to ground level, increasing fuel load (see Zylstra et al. 2022), makes the physical environment of these areas hotter and drier (Lindenmayer et al. 2022b), and within about 10 years changes the forest to be more flammable (see Zylstra et al. 2016, 2021; Furlaud et al. 2021; Lindenmayer et al. 2021, 2022a; Zylstra et al. 2021). Thus, fire severity tends to peak in juvenile regrowth from about age 10–40 years in a range of forest types. Therefore, areas of regrowth following intensive logging less than 50 years old are hostile to mature-forest dependent species such as the YBG and SGG. Habitat may also be unsuitable (e.g. regrowth following intensive logging 50–100 years previously) and presumably transitional (e.g. regrowth following intensive logging 100–200 years previously).

The impact is exacerbated where the logging occurs over extensive areas within a relatively short period (i.e. less than about 50 years; see Fig. 1 for an example from Victoria). Large areas of relatively continuous fire-sensitive regrowth results in cascading fire impacts at a landscape scale (Bowman et al. 2014). Thus, a feedback loop is established so that the landscape becomes increasingly flammable with increasing fire incidence (the idea of landscape traps; Lindenmayer et al. 2011a, 2022a). Thus, large swathes of forest have been rendered more fire-prone, leading to a greater likelihood of high impact fires at a landscape scale. This profoundly impacts biodiversity, including mature-forest dependent species, but also has considerable implications for property, infrastructure and people (Lindenmayer et al. 2021, 2022a).

Fig. 1.  Fragmentation of tall forest and the establishment of hostile environments (50% of map area) for mature forest dependent species by intensive logging; an example based on the geographic context of the proposed ‘Wheel’ Coupe in the Cann River District, East Gippsland, Victoria. Proposed Special Protection Zone (as at 11 November 2021) and intensively logged coupes are shown (with year of logging). Note that this area was also burnt (at varying intensity) during the 2019–2020 Black Summer bushfires.

Logging immediately changes the canopy forest microclimate for several decades (Lindenmayer et al. 2022b). A mature forest has a more even and humid microclimate than a recently logged forest (see MacFarlane et al. 2010; Norris et al. 2012; Lindenmayer et al. 2022b). Although SGGs are particularly vulnerable to high summer temperatures during and immediately following logging (Box 2), the changed canopy microclimate persists for decades (i.e. until the overstorey canopy has re-established). This is one reason SGGs are considered especially vulnerable to the interactive effects of logging, fire and climate change (see Lindenmayer et al. 2020; Smith 2020), and why global warming is a particular threat to the SGG (Smith and Smith 2018; Wagner et al. 2020; Ward et al. 2020) in the context of intensive logging. Further, the occurrence of fire soon after logging exacerbates the threat (Lunney 1987; McLean et al. 2018; Lindenmayer et al. 2020).

SGGs (and YBGs) within logged areas tend to die during the logging operation, or soon thereafter (Tyndale-Biscoe and Smith 1969) from predation, exposure or starvation. Retained trees provide an important biodiversity contribution (Manning et al. 2006). However, edge effects and impacts of fragmentation are likely to override any potential beneficial effects of variable retention (e.g. VR1 spacing of retained trees around 35 m in the general coupe or VR2 – 25 m spacing) in the short to medium term (i.e. about 50 years). It should be noted that SGGs, like most sedentary mammals, tend not to disperse following destruction of their habitat, but instead die in situ (see Tyndale-Biscoe and Smith 1969).

Smith (2020) examined the adequacy of changed site operating conditions, such as retaining habitat trees, increasing widths of riparian zones and temporary protection of fire refuges as a way of mitigating logging impacts in the fire affected forests of New South Wales. Smith (2020) concluded that impacts would remain severe for SGGs and YBGs while the level of the cut (i.e. amount of timber removed) remained so far above sustainable levels. Further, these impacts should be regarded as irreversible, given that logging return times are projected to occur within the immature growth phase of the forest.

Given the high risk of local extinction of YBGs and SGGs within reserved old-growth remnants in typical forest compartments managed for timber production (McCarthy and Lindenmayer 1999; Lindenmayer et al. 2020; Smith 2020; Lefoe et al. 2022), the precautionary principle (see Read and O’Riordan 2017), need for future proofing (see Rich 2014) and for a biodiversity safety net (Dinerstein et al. 2020) demonstrates a need to reserve from logging, additional areas of suitable habitat in each forest compartment occupied by these species. McCarthy and Lindenmayer (1999) recommended an additional 650 ha per intensively managed forest block or compartment for SGGs. However, that recommendation needs to be reviewed in the light of the increased understanding of global warming effects in southern Australia (e.g. Garnaut 2008; Wardell-Johnson et al. 2011; IPCC 2022). For example, the severe broad-scale Black Summer fires of 2019–2020 (see Ward et al. 2020; Bilney et al. 2022; Lindenmayer et al. 2022a, 2022b) have in concert with intensive logging, severely impacted the YBG and SGG (see Lindenmayer et al. 2020; Smith 2020). It is widely understood that the impacts from global warming are accelerating and will not rapidly be stabilised (see Wardell-Johnson et al. 2011, 2017; IPCC 2022). Further, given the sensitivity of SGGs to heat stress (see Rübsamen et al. 1984; Smith and Smith 2018; Wagner et al. 2020; Box 2), and the large home range required by the YBG (Box 1), these species are particularly sensitive to current and projected interactive climate change impacts. At all scales, the synergistic impacts of climate change, fire and logging interact to make the overall impact of disturbance much greater than the sum of their individual effects (e.g. McComb and Cushman 2020).

The recognition of the impact of logging and wildfire has resulted in regulations (Department of Environment and Primary Industries 2014, 2021) for the setting aside of approximately 100 ha PAs whenever population numbers of YBGs and SGGs achieve triggering levels. Thus, the designation of PAs for YBGs and SGGs depends initially on spatially explicit and verifiable distribution records, as well as understanding of what defines locally suitable (also areas that are hostile, unsuitable and transitional) habitat. Suitable habitat includes a large area (i.e. 100 ha or more), with an appropriate configuration of food and den sites to enable seasonal requirements to be met, as well as providing year-round safe passage between these resources (see Boxes 1 and 2). This habitat tends to be physiognomically complex compared with regrowth forest. Protection of this habitat requires considerations to guide design of PAs for these species in the context of on-going disturbance and climate change.

Both the YBG and SGG are mature-forest dependent species (see Boxes 1 and 2). Without human intervention, substantial areas within their geographic range are hostile or provide unsuitable or transient habitat, including native grassland, shrubland and woodland. Other areas, such as cleared agricultural and urban landscapes on formerly suitable habitat have effectively been rendered as permanently hostile. Still, other areas of altered habitat such as regenerating former habitat after intensive logging are in transition, possibly towards suitable habitat. While these coupes may theoretically transition from hostile environments through unsuitable, transient to suitable habitat over time (depending on age of the regeneration and the level of retention of habitat components), this transition can take centuries (see Jacobs 1955; Bradshaw and Rayner 1997). Therefore, given the accelerating impacts of global warming (IPCC 2022), there is uncertainty as to whether future vegetation will resemble the vegetation that it replaced (Wardell-Johnson et al. 2017).

In the context of forest regeneration, hostile environments may include limited food or den sites for some or all of the year and/or carry an enhanced risk of predation in the passage between den or food sites. Small areas of appropriately structured habitat or with appropriate feed trees that are not able to sustain a family group (YBG) or individuals (SGG), and narrow (<100 m wide) corridors are hostile as they act as traps or sinks (Lindenmayer et al. 2011a, 2022a). Notwithstanding, the biodiversity importance of isolated ‘paddock trees’, ‘habitat trees’ or ‘feed trees’ (Manning et al. 2006), their presence does little in the short to medium term (i.e. about 50 years) to change the circumstance of hostile environments for mature-forest dependent species.

The considerations introduced here have been summarised in Box 3. They can be grouped into Overall approach (Consideration 1), Survey records and habitat attributes (Considerations 2–4), Size and shape considerations (5–6), Management history (Considerations 7–8); and Boundary considerations (9–10).

To provide the best opportunity for the sustainability of populations of the YBG and SGG, a precautionary approach in forest management is required. (Consideration 1 – precaution, prevention and future proofing). This recognises the biodiversity risks and consequences associated with long-term compositional, structural and functional alteration and fragmentation impacts associated with intensive and extensive logging; and the synergistic effects associated with interactions of logging, climate change and stochastic factors. Therefore, Consideration 1 – precaution, prevention and future proofing is the overriding guide to the designation and design of PAs, recognising that long-term sustainability of populations of these species is the primary consideration of these PAs.

The precautionary principle (Harding and Fisher 1994; Deville and Harding 1997; Calver et al. 2011; Read and O’Riordan 2017) requires enacting when there is: (1) scientifically plausible risk; and (2) high uncertainty about the risk. The need for precaution rises with serious consequences of risk and level of uncertainty (Deville and Harding 1997). Because impacts of intensive logging, carried out extensively over a short time frame (i.e. within the immature phase of the relevant trees) on YBGs and SGGs have been established over decades (i.e. since Tyndale-Biscoe and Smith 1969), preventive measures require to be undertaken (i.e. negating a known risk). Any uncertainty is not about the impact of logging, but on the outcomes and extent to which such logging will exacerbate numerous interactive threats associated with climate change across the region. Thus, while the location of logging coupes can be designated, the occurrence of chance events, particularly associated with wildfire under changed climatic circumstances cannot. Therefore, to provide an increased likelihood for persistence of these species, retained vegetation in PAs should be provided with as much protection with respect to location, shape, size and local and regional management, against future chance events as possible. In other words, application of these considerations should always have as a first priority the conservation of these species.

The most important data for defining a PA are relatively high population densities of YBGs or SGGs. Precise, accurate, and verifiable procedures for sampling populations have been developed (i.e. Brown et al. 2011; MacHunter et al. 2011; Chick et al. 2020; Cripps et al. 2021). The approach adopted in Victoria (i.e. that of Chick et al. 2020) is well designed, repeatable and effective at recording the target species (i.e. SGG, YBG), without necessarily detecting all individuals. Thus, the approach is valid and reliable in obtaining a minimum estimate from a sample population. While areas of survey (as well as other activities such as resource extraction) may be informed by modelling, management activity tends to be based on verifiable records. Thus, designation of a PA depends upon actual records obtained by standard survey approaches. It should be noted that individual, isolated or ad hoc detections are likely to trigger different actions (see below). Given this trigger, adequate survey coverage of regions where logging is likely to be carried out is of high priority.

SGG density varies proportionally to the availability of hollow-bearing trees and they do not persist in areas of forest where such trees are absent (Box 2; Lindenmayer et al. 2017). Within a forest of suitable habitat, SGGs prefer higher overstorey basal areas of old-growth tree stands (Incoll et al. 2001; Lindenmayer et al. 2017) and the number of suitable hollows is an important factor affecting SGG and YBG distribution (Lindenmayer et al. 2004, 2017). These species use multiple den trees, adding to the complexity of habitat use, and explains why sites with many hollow trees are so important. Such sites may also be desirable habitat for a range of other mature forest dependent species (Loyn 2004; Lindenmayer et al. 2017). The ability of eucalypts to form hollows varies with species and management history (e.g. Wormington and Lamb 1999), with hollows starting to form by the mature growth stage (i.e. around 120 years in ash forests). Large-sized hollows required by SGGs and YBGs do not usually occur (at least in ash forests) until they are around 200 years old or when trees are beginning to reach their senescent stage (Gibbons and Lindenmayer 2002; Lindenmayer et al. 2017). However, there are forest types (e.g. mixed forests) or situations (i.e. following wildfire) where they occur in younger stands. Dominant trees in a eucalypt forest stand may be in the senescent stage for hundreds of years. Thus, a PA should include minimal area hostile (i.e. even-aged patches of juvenile regrowth) or unsuitable (i.e. regrowth following intensive logging 50–100 years old) to mature-forest dependent species, recognising that individuals and stands of most eucalypts naturally regenerate, particularly after fire or other disturbance.

The large home ranges of YBGs encompass dispersed and seasonally variable food resources (Box 1). They rely on forest dominated by the mature and senescent growth stages of particular forest types, but usually including a wide range of forest eucalypts. In a large stand of mature forest, they tend to favour forest at the transition between wet sclerophyll and lowland forest. This is due to the availability of a wide variety of feed trees in mixed vegetation types (see Eyre and Goldingay 2003, 2005). Disturbance may cause a shift in distribution towards unburnt gully refugia (Lunney 1987). As mature forest low in the landscape has greater capacity to act as refugia under scenarios of drought, climate change and fire events (Lindenmayer et al. 2013; Berry et al. 2015), these areas should be targeted as PAs. The SGG tends to prefer mid-elevation forests (optimal 845 m above sea level in Kavanagh’s (2000) study) and sites of highest fertility (Braithwaite et al. 1984; Wagner et al. 2021). Drought reduces the availability of young palatable leaves; it affects growth in eucalypts and thereby the availability of leaves for feeding. Remote sensing, survey and modelling are important tools in recognising the development of habitat over time (e.g. Karna et al. 2019; Armstrong et al. 2020). Nevertheless, approaches to survey and modelling should be both reliably and validly designed, so as to be able to detect an effect where one is likely (e.g. cf. Environmental Protection Authority 2021; Bilney et al. 2022).

The size of a PA has been mandated as approximately 100 ha (Department of Environment and Primary Industries 2021, tables 13). However, continuity of mature forest beyond the PA provides increased future proofing (Consideration 6 – fragmented landscapes, and Consideration 9 – context and boundaries). Edge effects are disadvantageous to SGGs and YBGs (Lindenmayer et al. 1993; Youngentob et al. 2012; Nelson et al. 1996). Thus, the formation of hostile edges increases predation risk (e.g. by Powerful and Sooty Owls; and by Carpet Pythons; see Kavanagh 1988). Hostile environments provide barriers and also isolates populations or family groups. This can in turn, reduce gene flow, decrease population viability and increase the risk of localised extinction due to random environmental and demographic events. Thus, PAs should be effectively as large as possible, and any area retaining mature forest and significant populations of mature-forest dependent species in fragmented habitat should be set aside as a PA (see Consideration 6 – fragmented landscapes). Because of the sensitivity of these species to fragmentation (Pope et al. 2004; Youngentob et al. 2013; Taylor and Lindenmayer 2020) and edge effects (Youngentob et al. 2012), PAs should tend towards round or square shape rather than be narrow or linear, particularly where they abut hostile environments. However (at least in mixed species or drier forest types), it is desirable to include gully areas in PAs, as they tend to be edaphically richer and moister habitat than upland sites. YBGs and SGGs persist in short (i.e. less than 1 km long) corridors of at least 100 m in width where there is suitable mature forest at either end (see Goldingay and Kavanagh 1991). Therefore (at least in mixed species or drier forest types), where possible, stream zone habitat should be included within PAs. This is due to the future proofing and multiple other benefits that wide stream zone reserves provide (e.g. Wardell-Johnson and Roberts 1991; Wardell-Johnson et al. 1991; Graziano et al. 2022). SGG and other solitary species can survive in linear strips and small patches retained after logging, at least in the short-term, better than colonial and social species such as the YBG because social species tend to consume widely dispersed food (Lindenmayer et al. 1993; Taylor et al. 2007).

For mature forest-dependent species of large home range such as YBGs or SGGs, intensive logging creates adverse effects through edges (Consideration 7 – logging history; see Youngentob et al. 2012) and hostile environments (Consideration 5 – size and shape considerations; see Lindenmayer et al. 2004; Taylor and Lindenmayer 2020). Following intensive logging (i.e. clearfelling), there is an approximate 150–200-year long time-frame before the re-establishment of the structure of tall forest, at least to be suitable for species such as YBGs and SGGs. Therefore, fragmentation is a long-term legacy of such activity carried out extensively (Pope et al. 2004; Youngentob et al. 2013; Taylor and Lindenmayer 2020). Landscape change and habitat fragmentation are major drivers of species loss globally (Saunders et al. 1991; Fahrig 2003; Fischer and Lindenmayer 2007; Crooks et al. 2017); increasingly affecting forests worldwide by reducing fragment size, increasing isolation of patches and creating more edge environment (Watson et al. 2018). In many areas of State forest, intensive logging has been carried out extensively (or is proposed to be carried out extensively). Thus, for areas where SGGs or YBGs have been detected, level of fragmentation from proposed forest coupes that have been rendered (or proposed to be rendered) as hostile (i.e. logged within 50 years) should be carried out prior to any proposals for logging activity. SGGs are particularly sensitive to variation in microclimate associated with fragmentation (Lindenmayer et al. 2022b), while the home range of a social group of YBGs can be upwards of 80 ha, and individuals are highly mobile, being able to move upwards of a kilometre from their home range. Therefore, under circumstances of extensive fragmentation (e.g. Fig. 1, an example of an operations proposal for ‘Wheel’ Coupe, which would render >50% of the forest as hostile within 1 km of the centre of the proposed coupe), Consideration 6 – fragmented landscapes should be enacted. Thus, where the conservation of mature-forest dependent species such as SGGs and YBGs is a goal in forest management, the impact of habitat fragmentation should be considered. For example, should a proposed logging operation increase fragmentation (i.e. render the environment hostile) beyond levels likely to allow these species to persist in the immediate area, then remaining nearby mature habitat should be set aside from logging.

Establishment of barriers to movement may be the greatest long-term effect of habitat alteration and reduction (Goldingay and Kavanagh 1991) and any hostile habitat within a PA wider than the distance over which an individual can glide (about 100 m) has potential to act as a dispersal barrier (Lindenmayer et al. 2020, Lindenmayer et al. 2022b). Thus, hostile environments should be avoided within a PA. The minimum area of immature regrowth (i.e. regrowth 50–100 years old) resulting from intensive logging should be included within a PA for YBGs or SGGs, as this regrowth would possesses few or no hollow or feed trees, and is therefore considered unsuitable. However, regrowth greater than 50 years old does pose a reduced fire severity risk compared with juvenile regrowth (Lindenmayer et al. 2022a, 2022b) and may allow dispersal. Ideally, however, a PA should largely be comprised of mature forest (of the appropriate vegetation types).

In Victoria, as elsewhere in south-eastern Australia, YBG and SGG habitat is periodically impacted by large-scale intense wildfires (e.g. 2019–2020 Black Summer fires). YBGs and SGGs have been adversely affected by these events, especially where their habitat abuts hostile environments (Lindenmayer et al. 2013, 2020; Lindenmayer and Taylor 2020). Any loss or fragmentation of habitat, including a change in spatial configuration, can result in a loss of critical food resources or alteration to the mosaic of forest types necessary to ensure a continual supply of food resources (Goldingay and Kavanagh 1991). Even where individuals survive, it may increase effort required by an individual to forage for dispersed food resources (Recher et al. 1987). Such an increase may have long-term effects including reproductive failure, decline in group size or decrease in density. These habitats should continue to be monitored, as YBGs or SGGs may have persisted in fire refugia or where mature trees have survived in these habitats (Lindenmayer et al. 2013). Abundance of arboreal mammals, in general, 2–3 years after severe burning is influenced by landscape context, such that abundance increases with increasing amount of understorey-only burnt forest within a 1 km radius (Chia et al. 2015). Spatial heterogeneity generated by fire patterns is an important aspect of colonisation and persistence following wildfire (Lindenmayer et al. 2013; Ward et al. 2020). While mature forest habitat dominated by resprouting eucalypts, burnt at high severity may soon again become suitable for YBGs and SGGs due to the retention of forest structure (cf. intensive logging), Lindenmayer et al. (2020) found that some species, including the SGG may continue to decline 10+ years after wildfire. Nevertheless, YBGs and SGGs have capacity to recolonise, once habitat again becomes suitable, depending on the availability of source populations and corridors for dispersal. Thus, while areas less impacted by fire should be preferentially targeted as PAs for SGGs and YBGs, severe fire history should not preclude the setting aside of an area as a PA.

YBGs and SGGs are adversely affected where their habitat abuts hostile environments (Lindenmayer et al. 2013, 2020b, 2022a, 2022b; Lindenmayer and Taylor 2020). Thus, a PA will have sustainability advantages where it abuts other secure areas of mature forest, particularly if that habitat is also suitable (evidenced by structure and verifiable records of SGGs and YBGs), or potentially suitable (i.e. mature forest, but without verifiable records) for SGGs and/or YBGs. It is desirable to maximise the distance to the edges of recent, or planned intensive logging activity (i.e. hostile environments) from records of SGGs or clumped records of YBGs in the PA. For the SGG and YBG, the disadvantage of proximity to edge increases with the recentness of the activity (e.g. van der Ree and Loyn 2002; Lindenmayer et al. 2022b). Thus, it is desirable to maximise the habitat integrity of any PA edge and to avoid current or planned intensive management edges. As SGGs and YBGs require many large hollows, this management edge lasts longer than the projected forest logging rotation in south-eastern Australia forests (see Goldingay and Kavanagh 1991).

Management boundaries, where trafficable, can lead to ecological damage or have unintended consequences as conduits for weed, pest and pathogen invasion. It is desirable to ensure management boundaries are, where possible, not trafficable, and at a distance commensurate with the stream order (i.e. greater distances from larger or higher order streams) from stream zones. They should also be on flat or mild slopes (<20°) to avoid erosion or other unintended ecological consequences at the edges of PAs (and streams) where these boundaries are accompanied by management edges.

Where State forests do not have high complementarity to the protected area network, the conservation of mature-forest dependent species such as SGGs and YBGs is compromised. Unfortunately, areas of HCV for gliders and other mature-forest dependent species are often targeted for logging (Taylor and Lindenmayer 2020), limiting this opportunity. Nevertheless, the application of these considerations has potential to increase the effectiveness of management concerning these HCV species.

Much of the tall forest of Victoria available for logging has been fragmented by recent intensive (i.e. within 50 years or the period over which the environment remains hostile) logging activity (e.g. Fig. 1). For example, given the continued presence of mature-forest dependent species, including YBGs in the proposed ‘Wheel’ coupe (i.e. 14 detections over six nights from 2 September 2019 to 16 October 2019 in VicForests Forest Protection Survey Program, FPSP; Fig. 2a), Consideration 6 – fragmented landscapes would require the setting aside of this area from logging to form an (albeit poorly shaped) PA over this area. The alternative, of logging, even at low intensities (i.e. with high basal area retention) would see a high probability of these species being rapidly and irretrievably lost from this compartment and landscape.

Fig. 2.  Operational planning and VicForests Forest Protection Survey Program (FPSP) results for planned logging coupes in Nowa Nowa District, Gippsland (‘Tiger’ and ‘Lior’ coupes), Cann River District, East Gippsland (‘Wheel’ coupe), and Marysville District, Central Highlands (‘Rookery’ and ‘Troop’ coupes), Victoria. Conservation Reserves, Protection Areas, proposed coupe boundaries, Variable Retention 1 logging (VR1 – minimum density nine trees per ha or about 35 m spacing), Variable Retention 2 logging (VR2 – minimum density of 14 trees per ha or about 25 m spacing), and planned set asides internal to coupes are shown. Locations of streams, very large or giant trees, habitat trees, and hazardous trees are also shown (summarised as ‘Large tree’ on the figures), along with detections of High Conservation Value (HCV) species (mostly mature-forest dependent), detected in the FPSP, as part of VicForests coupe planning process. FPSP species detections (as at 11 November 2021) in and near the proposed ‘Wheel’ coupe (a), proposed ‘Rookery’ and ‘Troop’ (as at 23 November 2021) coupes (b), and in and near the proposed ‘Tiger’ (as at 12 August 2021) and ‘Lior’ (as at 25 February 2021) coupes (c and d) are shown. VicForests FPSP surveys are acknowledged as providing incomplete geographic coverage of planned logging coupes.

Operations planning requires knowledge of the HCVs in and near coupes likely to be available for logging. Thus, surveys appropriate to the species under consideration, where necessary informed by effective modelling (or remote sensing, cf. Bilney et al. 2022) can guide management activity, including set asides or additional PAs. However, the FPSP of VicForests is not designed to detect all SGGs or other HCVs within coupes, or nearby to coupes. Thus, SGGs in particular, are not reliably detected beyond about 25 m from a transect, and are therefore often overlooked in such surveys, which tend to focus on perimeter tracks. This limits the effectiveness of the PA system and also the approaches to conserving these species in the context of timber production in Victoria’s State forests.

Operations planning requires survey at the local level, not only for the timber resource but also for HCV species. Thus, any ‘harvesting plan’ for proposed coupes requires addressing the considerations associated with HCVs, including YBGs and SGGs. For example, were the proposed ‘harvesting plans’ for ‘Wheel’, ‘Rookery’, ‘Troop’, Tiger’ and ‘Lior’ coupes (Fig. 2) to take account of these HCVs, then their locations in and around this proposed coupe would be known (Consideration 2 – presence of YBGs or SGGs). Indeed, FPSP surveys by VicForests identified numerous HCV species including YBGs and SGGs in and near the planned ‘Wheel’ (Fig. 2a), ‘Rookery’ and ‘Troop’ (Fig. 2b), and ‘Tiger’ and ‘Loir’ coupes (Fig. 2c), despite inadequate survey coverage, within and surrounding these coupes. In fact, high numbers of HCV species in these proposed coupes suggests proposed PAs in the vicinity of these coupes would benefit from application of the 10 considerations here presented.

Further, vegetation mapping or modelling showing the preferred and less preferred habitat of these species (i.e. Consideration 3 – habitat components) would also be included on plans. This has not been carried out for these coupes, leaving operational planning uncertain about the survival of these species in nearby unlogged forest following conversion of the proposed coupes to hostile environments. Given that for most targeted coupes, nearby coupes have been logged within the last 50 years (e.g. see Fig. 1), edge effects and fragmentation would be considered in the ‘harvest’ plan (i.e. Consideration 5 – size and shape considerations, and Consideration 6 – fragmented landscapes). While prescriptions generally consider safeguarding of individual trees during and after the logging operation, they rarely outline how such trees or HCV species are to be monitored following the operation. Further, there is usually little reference to prescriptions in nearby or adjacent coupes, such that lessons concerning the loss of HCV species from the logging of one coupe apparently do not inform prescriptions to better safeguard HCVs in subsequent adjacent or nearby coupes.

Because of the long time-scale required for the recovery of vegetation structure for SGGs following intensive logging, it is necessary to know the suitability or otherwise of surrounding areas, including proposed PAs for these species. Further, the shape (often long and narrow; e.g. Fig. 2b–d) of PAs does not always suggest sustainable habitat for these species. This limits opportunities for recolonisation from, as yet unknown source colonies. Thus, while detections of HCV species may be presented in operations plans (imperfect though they be), within or immediately adjacent coupes proposed to be logged, such detections are rarely presented for adjacent or nearby proposed PAs. Thus, Consideration 1 – precaution, prevention and future proofing requires a much more thorough consideration of the long-term survival of these species.

Rather than safeguard those proposed coupes with an indeterminate, but high level of HCV species from logging, VicForests often proposes retention of small patches within these coupes and use of Variable Retention logging for parts of the remainder of these coupes (e.g. Fig. 2b–d). Retained patches less than three hectares are unlikely to allow the survival of YBGs or SGGs in the context of the logging operation (Pope et al. 2004; Lindenmayer et al. 2004). Research dating back to Tyndale-Biscoe and Smith (1969) has demonstrated that this approach to logging will directly lead to the death of most or all SGGs and YBGs present in these coupes soon after the logging operation. Further, the time-scale required for the re-establishment of suitable habitat suggests that replacement of habitat with an environment hostile for these species effectively leads to a serious or irreversible damage to the environment.

Habitat of mature-forest dependent species is still logged in Victoria, despite these species recognised, both as HCVs, and as endangered by intensive logging activity. Thus, Consideration 1 – precaution, prevention and future proofing need not be invoked for action to be taken to prevent a known threat. Effective application of the considerations provided here would no doubt increase the survival chances of these species in the context of a native forest timber industry. However, because of the scale of logging, it is unlikely to be sufficient to safeguard these species in the context of intensive logging carried out extensively.

In the context of State forest, the approach of Chick et al. (2020) is a suitable survey methodology to determine threshold densities for YBGs and SGGs, provided that these surveys have sufficient geographic coverage and extent. The approach satisfies the requirements for validity and reliability to designate the establishment of PAs of approximately 100 ha (i.e. table 13 of Department of Environment and Primary Industries 2021). However, triggering densities of SGGs and YBGs for the setting of PAs are much higher than the average densities of these species in the forests where they occur in Victoria (Boxes 1 and 2). Therefore, because of the high level of threat from intensive logging at the landscape, compartment and coupe scale (Lindenmayer et al. 2020b, 2022a, 2022b; Smith 2020), forest managers need to be aware of the locations of SGGs and YBGs in areas proposed to be logged. Such knowledge would provide the information enabling operations planning for HCVs, as well as for the timber resource in State forests. Modelling the geographic coverage of this habitat (and timber values) is useful for determining appropriate survey locations (provided that this modelling is both valid and reliable cf. Bilney et al. 2022). Regardless, thorough survey is required to determine known locations, so that appropriate operations planning can be instituted (e.g. for internal set asides).

This knowledge can be achieved using the approach advocated by Chick et al. (2020), with modification to ensure thorough coverage of the coupe and surrounds, once approximate boundaries of proposed logging coupes are defined. Because YBGs are reliably detectable at distances of up to 150 m, transects (with playback) around boundaries or approximately through the centre of proposed coupes are sufficient for this species (unless the coupe is greater than 300 m in width). However, SGGs are not reliably detected beyond 25 m in mature forest (Chick et al. 2020). Therefore, approaches to ensure detection of SGGs within and immediately adjacent proposed coupes are required throughout the native forests of south-eastern Australia. This involves multiple transects at approximately 50 m width or equivalent, taking into account Occupational Health and Safety considerations, but otherwise using the methodology detailed by Chick et al. (2020).

Given the high density of YBGs or SGGs required before a PA is triggered and incomplete survey in areas proposed to be rendered as hostile for mature-forest dependent species, it is also necessary to consider conservation where a PA is not triggered. Where logging return times are more rapid than the return time of the required forest structure, and conservation of mature-forest dependent species is a goal, set asides are necessary wherever YBGs or SGGs have established home ranges. Survey quality will determine the area and location of set asides. However, set asides less than 3 ha in area internal to a coupe are unlikely to allow survival of individual SGGs or YBGs in the context of logging, with or without basal area retention (Pope et al. 2004; Lindenmayer et al. 2004).

The FPSP for HCVs carried out by VicForests provides a low level of coverage in areas proposed for logging (i.e. many SGGs will not be detected). It is therefore prudent to facilitate effective survey by others, as well as including a significant buffer (i.e. at least 100 m) around any SGG detected in the FPSP program as protection within these proposed coupes. This implies a set aside of 240 m radius (assuming a minimum home range of approximately 1.5 ha) for any SGG detection in mature forest in proposed logging coupes. The alternative of setting aside only the home range (with variable retention elsewhere in the coupe) for all detected SGGs requires a much greater commitment to effective survey in proposed coupes and their surroundings.

Individual SGGs and YBGs are unlikely to survive in areas of variable retention silviculture, unless a small proportion of their home range (i.e. most occurs in adjacent retained habitat) is impacted by the logging event (see Tyndale-Biscoe and Smith 1969). In the context of set asides, PAs and variable retention with retained patches, it is necessary to consider the population context for the coupe, compartment and landscape to ensure dispersal back into the coupe once the forest again becomes suitable for YBGs and/or SGGs.

Because of the more rapid return of large-hollow dependent species (e.g. SGG and YBG) and other biodiversity components, following intense wildfire than following intensive logging (Lindenmayer and McCarthy 2002; van der Ree and Loyn 2002; Nelson et al. 1996; McLean et al. 2015, 2018), Parliament of Victoria Legislative Council Environment and Planning Committee (2021) recommended establishment of variable retention silviculture as the default to maintain a portion of the structure of the logged forest, and thus improve biodiversity recovery in the post-logged forest (Recommendation 21). Nevertheless, because of the high impact of edge and fragmentation effects (Saunders et al. 1991; Fischer and Lindenmayer 2007) on SGGs and YBGs (Kavanagh 2000; Kavanagh and Webb 1998; Taylor and Lindenmayer 2020; Lindenmayer et al. 2020b; Smith 2020), this form of silviculture is inconsistent with the conservation of SGGs or YBGs (or indeed other mature-forest dependent species) in mature forest.

Outcomes associated with the scale of logging and coverage of forest compartments in tall and wet forest in Victoria suggests that a high proportion of available merchantable forest in forest compartments of these areas would be logged within the 60-year period 1970–2030, regardless of approaches concerning variable retention. Regardless, the application of variable retention approaches is insufficient to conserve SGGs or YBGs in the context of extensive areas of intensively logged forest over such a period. It should be noted, however, that while variable retention approaches are ineffective for reducing the risk to SGGs and YBGs, retention of habitat, seed trees and particular levels of basal area, does serve important other biodiversity functions in the context of this scale of logging.

At present, logging return times are much more rapid than the recovery of the habitat required by mature-forest dependent species such as the YBG and SGG in the State forests of Victoria (as elsewhere in south-eastern Australia). Therefore, logging should not occur in mature forest where it does serious or irreversible damage to YBGs, SGGs, their habitat and the environment. This implies concentrating logging activity within areas already transformed (and largely in the immature growth phase). The alternative of managing for the sustainability of mature-forest dependent species implies much longer logging return times. Regardless, concentration on regrowth forests or sustainably managing mature forest would, by necessity, reduce the level of resource supply to industry. However, it would also enable the sustainability of the industry as well as the conservation of the HCV components of the forest. The conversion of mature forest to regrowth managed within the immature growth phase is not a sustainable use of the forest.

A fundamental structural change to the forest is brought about by the removal from site of the largest components over an extensive area (Lindenmayer and McCarthy 2002). This will always impact on the species dependent on that structure (i.e. the mature forest-dependent species). This impact is process-driven and is not ameliorated by small changes to prescriptions in local forest coupes. This was recognised more than 50 years ago (see Tyndale-Biscoe and Smith 1969) for the SGG, and becomes much more problematic in the context of the extensive landscape scale changes carried out within the immature phase of the growth of these forest trees (i.e. within a 120-year period). The issue becomes of greater significance in the context of global warming and associated follow-on effects on climate change and the interactions of these changes with logging (see Wardell-Johnson et al. 2017; Ward et al. 2020; IPCC 2022).

Without re-reckoning of sustainable yield, the State forests of Victoria are likely to become dangerously fire-prone environments, bereft of mature-forest dependent and hollow-dependent species. Given the status of YBGs and SGGs as umbrella species (Kavanagh 1991; Kavanagh et al. 2004), and the known capacity of old-growth and older regrowth forest to use less water (Macfarlane et al. 2010), provide a greater dampener to temperature fluctuations (Norris et al. 2012; Lindenmayer et al. 2022b), store more carbon (e.g. Dean and Wardell-Johnson 2010; Dean et al. 2012; Watson et al. 2018) and be less severely affected by intense fires than regrowth (e.g. Lindenmayer et al. 2011a, 2011b, 2020, 2022a; Lindenmayer and Taylor 2020; Zylstra et al. 2021; P. Zylstra, G. Wardell-Johnson, D. Falster, M. Howe, N. McQuoid, S. Neville, unpubl. data), multiple additional benefits will accrue from such protection (see also Wardell-Johnson et al. 2017; Watson et al. 2018; Lindenmayer and Taylor 2020). These forests may then be able to continue to provide for an on-going timber industry, focused on the longer-term sustainability of the environment, the industry and the communities depending on it.

This paper is a review and synthesis, requiring no new data.

This research did not receive any specific funding.