Border Ranges Biodiversity Management Plan: defining functional groups for use in resource-limited multi- recovery implementation scenarios

FLORA REPORT PREPARED FOR NSW DEC BY

Robert Kooyman and Maurizio Rossetto

National Herbarium of NSW, Botanic Gardens Trust, Mrs Macquaries Rd, Sydney NSW 2000, .

June 2007

Abstract This report provides an overview of the development of a bioregional approach to biodiversity assessment and management that uses trait-based plant functional groups as a basis for multi-species recovery planning. Multi-variate methods were used to extract and test emergent groups, and additional information fields related to species life history and distributional data were added to develop a biodiversity assessment matrix in spreadsheet format (Appendix 1). Tests of phylogenetic independence were undertaken and showed that phylogeny significantly affects the clustering of character states for nearly all the traits studied. Data rich samples were used to test the methods in one (rainforest) community type, and several species from one of our emergent groups were chosen from that sample to provide an example of the function of the biodiversity assessment tool. Relating emergent trait-based plant functional groups to habitat was found to be the most informative approach for the development of management recommendations and recovery planning related to landscape scale threat / risk categories. Appendix 2 provides a list of species representing the various groups that have been identified as the focus for additional research and information gathering. Appendix 3 provides species level information related to defining the realised niche (reflecting species distribution in relation to environmental variables) for twenty-nine species from the data rich sample that occur on the BRBMP list.

1 Table of Contents

Project Brief...... 3

Introduction...... 4

Methods...... 8 Study area and data compilation...... 8 Trait Information...... 9 Habitat Types and Initial Threat Assessment ...... 9 Data analysis methods for the target species ...... 10

Results...... 12 Identifying appropriate trait-based groups...... 12 BRBMP Project Listed Species ...... 13 Description of trait-based groups obtained from cluster analysis...... 14

Discussion...... 15 Multi-species planning based on trait-related groups: conceptual framework ...... 15

References...... 21

Appendix 1 - Biodiversity Assessment Matrix...... 34

Appendix 2 – Group priorities for research...... 37

Appendix 3 – Data Rich Community Example (SNMVF) ...... 40 Introduction...... 40 Methods...... 40 Data analysis methods for the data rich community example ...... 40 Trait relationships for rare and threatened taxa in the SNVF sample...... 41 Results...... 42 Discussion...... 43

2 Project Brief

• Develop a transferable conceptual approach to guide rare and threatened plant biodiversity management planning for use at bio-regional scales. • Using the available and accessible data develop a (parsimonious) trait-based approach defining plant functional groups that reflect biological / ecological / evolutionary and habitat factors. • Provide a spreadsheet based biodiversity assessment matrix that, once relevant data has been added, can integrate group and species level landscape-scale threats to help define RISK-based assessment categories that inform management priorities and the strategic allocation of resources. • Provide a broad and fine-scale data rich (species richness / abundance / environmental variables) and threatened species rich ecosystem-based example to allow exploratory analyses and testing of methods. • Provide guidance for future implementation needs and develop preliminary examples using taxa for which data is available.

3 Introduction

The study of the distribution of and their relative abundances in variable landscapes is central to the development of conservation planning. Determining how species make a living and how they interact with other species and environmental variables in their shared habitats is an important area of research linked to community ecology. Plant life history traits provide insights into both evolutionary and ecological processes and are regarded as a critically important area of research in plant science. From a conservation perspective, functional groups based on life-history characters have the potential to bring together taxa that are likely to respond to selective processes, environmental threats and potential management actions in similar ways. They also provide insights into the mechanisms behind species responses to land-use change (Verheyen et al., 2003; Kolb and Diekmann, 2005).

The focus of recovery planning in response to Australian state and federal threatened species legislation has, until recently, predominantly been on the development of single- species recovery plans that developed strategies for threat abatement and population preservation. However, the allocated resources available for the development of such plans has often been insufficient and exhausted before many of the identified recovery actions could be implemented. In response to this, there is increasing interest in developing management and recovery planning efforts that are focused on securing multiple species outcomes at bioregional scales. The aim of such plans is to identify threatening processes, and coordinate and prioritise recovery efforts across landscape, community, site, and species levels while simultaneously responding to the requirement for cost-effectiveness.

The need for effective approaches to guide the use of scarce resources for the conservation of rare and threatened species is well established and acknowledged as of broad interest to conservation scientists, managers and practitioners. Using plant functional traits to group species for research and management has attracted considerable research interest over many years (Cornelissen et al., 2003), and has resulted in a number

4 of plant strategy schemes being developed (eg., Grime 1977; Gitay and Noble 1997; Westoby 1998; Westoby et al., 2002).

The idea that plant species with different qualitative and quantitative traits can be more successful in different parts of the landscape is certainly not new (for example Schimper, 1903). It is also well known that plant species functional traits vary both along environmental gradients and within communities of species under similar conditions. The shift in trait frequencies along gradients, the mix of traits at a site, and trait correlations, can all provide important insights into community and species level vegetation processes, and species present day ecological competence. In that context, the physical environment can then be considered as filtering the kinds of species that can succeed at a given site (by providing the framework within which species interact), but not as the final determinant of the range of trait values present at a site (Westoby and Wright 2006).

Partial reviews of the history of development and efficacy of biodiversity recovery planning at various scales and in different contexts have been undertaken (for example, Harding et al., 2001; Hecht and Parkin, 2001; Abbitt and Scott, 2001; Clarke et al., 2002; Moore and Wooller, 2004; Roberge and Angelstam, 2004; Male and Bean, 2005). It is not our intention to compare or provide detailed descriptions of the various approaches to recovery planning, from the species to the systems level (refer to McNeely, 2006) however, the applicability of other recovery planning approaches to this case has been explored. It was concluded that:

Reverting to a compilation of single species population viability analyses (PVA) for all taxa in the list was considered inconsistent with the agency-defined multi-species planning objectives and unachievable within the project time frame or (more importantly) with the available resources.

Focusing on, or prioritising species on the basis of rarity (or population size) was found to be unsuitable (in this case), as it was previously shown that the local species with the smallest populations are not necessarily the most threatened (for example research in the

5 study area on hardeniana; Rossetto and Kooyman, 2005; and australis; Kooyman, 2005).

The surrogate biodiversity assessment approach uses environmental (abiotic) surrogates to model and predict species diversity across whole landscapes (Faith, 2003; Faith et al., 2004; but see Araújo et al., 2004). The methods presented here are predominantly focused on a subset of listed rare taxa in the study area only, and (except for the data rich example) do not reflect research into biodiversity patterns in assemblages at landscape scales. However, in this approach we do use species level preferred habitat as a secondary filtering process within the trait-based groupings, to assist the intended threat / risk assessment process (Appendix 1).

The ‘umbrella’ species approach assumes that survival and recovery of the chosen taxa reflects favourably on all co-occurring taxa (see Roberge and Angelstam, 2004; Lambeck, 1997). Ecologists generally agree that it is difficult to determine which species are the most significant (or informative) in different ecosystems. From a strictly functional perspective, species matter to the extent that their traits interact to maintain ecosystem function. Thus, the ‘umbrella’ species concept may not always be useful, or the most informative, as appropriate ‘umbrella’ species are not necessarily to be found on lists of recognised rare or threatened taxa. The exception may be in cases where multi- species approaches based on unrestricted focal species selection procedures from whole of biota are used (Roberge and Angelstam, 2004).

Comparison of rare and common taxa based on phylogenetic contrasts is another means to develop species groups for conservation assessment (for example, Farnsworth 2007), but such comparisons were not undertaken as part of this study. However, phylogenetic independent contrasts were undertaken to test for phylogenetic influence on trait groups within the species included in the project, and to test the usefulness of this approach for grouping species.

6 Here we aim to identify trait-based functional groups that provide a useful starting point for assembling and prioritising species as a means to explore the development of broader management options within a constrained (resource and time) scenario. The objective is to respond to legislative requirements and to the format required by local conservation agencies for multi-species biodiversity management and recovery plans, and hopefully to a widely acknowledged need for conservation actions in resource and time constrained management scenarios (Bawa et al., 2004; McNeely, 2006).

The methods and example described below emerged as a consequence of our process of exploration of a range of options, and analysis tools, to address what we consider an important question and challenge in biodiversity management. That is: what to do when confronted with the problem of managing numerous threatened species in a high biodiversity bioregion when few data and resources are available? Important distinctions from previous approaches are the constraint of the rare species list (that is, no contrasts or comparisons with more common taxa were undertaken) and the use of trait-based clustering methods to identify species groups within that list, thus avoiding a-priori sorting on the basis of potentially arbitrary or conflicting ‘expert opinion’. We establish a starting point for prioritising management actions based on plant groups that reflect biological and ecological aspects of species life histories and provide a preliminary example of how to use groupings within a management context. However, it is acknowledged, that for such approaches to succeed, conservation objectives need to be clearly defined (Nicholson and Possingham, 2006) and managing multiple species at multiple scales will always present unique and context dependent challenges (Meentemeyer and Box, 1987; Fischer et al., 2004).

7 Methods Study area and data compilation The study area for the plan presents a demanding array of challenges for conservation planning at the bioregional multi-species level. It straddles the Southeast Queensland and Northern New South Wales border, is geographically complex, bisected by a political boundary, and has (across that boundary) numerous shared but differentially rare species and habitats, including many restricted endemics (Fig. 1 and Appendix 1). The area also represents an important subtropical centre of endemism and diversity (Webb and Tracey, 1981; Webb et al., 1984) with representatives of both tropical and temperate floras (Burbidge, 1960), and over 50 endemic genera and more than 200 species occurring at their southern or northern limits (McDonald and Elsol, 1984).

The high numbers of rare and threatened species in the study area belong to a broad range of lineages, with generally very little known about their biology and ecology. This reflects a pattern that has been identified more broadly in Australia, and in many places around the world (see for example, Root et al., 2003). These taxa occupy habitats that vary from larger protected areas, to fairly well protected but isolated remnants, to small remnants in poor condition within mostly cleared land. The varying levels of landscape disturbance and degradation, and potential / extant threats combine to create a complex mosaic of risk.

A final list including 159 rare, threatened (scheduled) and of concern (unscheduled) plant species to be included within the biodiversity management plan was determined by the relevant government conservation agencies and BRBMP vegetation committee. Many of these species are local endemic and by definition rare, and generally their life histories and distribution patterns are poorly known with little published information / data available. Because of the restriction in available ecological information and the lack of time and resources to significantly increase our knowledge, we selected a range of study traits for which information was (reasonably) easily accessible and that are recognised more broadly as important indicators of species and functional group level performance (Westoby and Wright, 2006).

8

Trait Information Trait information was extracted from published information sources provided as species taxonomic descriptions (for example, Floyd, 1989; Harden, 2001), herbarium specimens and databases, species recovery plans and other sources, and added in spreadsheet format to develop a matrix of 159 species by ten (ranked) traits. The traits chosen included those associated with life history strategies, dispersal distances, reproductive biology, energy balance, hydraulic architecture, and persistence potential. The specific traits for which information was compiled were seed size, fruit size, fruit type, dispersal mode, breeding system, pollination mechanism, size, wood density, resprouting / clonality, and maximum height. Five traits were subsequently removed. The rationale for doing so reflected the fact that fruit size was correlated with seed size, sex (breeding) system and pollination data were found to be unsuitable for analysis (binary form only or incomplete information), and wood density and maximum height were confounded (in this case) because of the inclusion of (non-woody) herbs, epiphytes and other life forms. Despite this reduction, the five remaining traits represent important features of plant ecological strategies (Westoby et al., 2002; Ackerly, 2004). Trait selection, in this case, also reflects the availability of species-level information, and the likelihood of their being informative (see Rossetto and Kooyman, 2005; Westoby and Wright, 2006) relative to threats. Trait rankings follow Rossetto and Kooyman (2005), and wherever possible were based on measured data from published floras.

Habitat Types and Initial Threat Assessment A second matrix was developed that included six habitat types, and determined species level altitudinal distribution categories (lowland only; upland only; and combination of both). In the study area the lowlands and some mid-altitude forest habitats have been most affected by human-impacts, while the uplands are largely intact. Subsequently, a threat assessment team associated with the project determined that species in the upland cool (less disturbed and better reserved forest areas) may be subject to different threats and (perhaps) a lower risk category (setting aside climate change; refer to Westoby and

9 Burgman, 2006) than lowland species in the agricultural and human-development prone landscape matrix.

Species were allocated to six habitat types that reflect the vegetation mapping criteria of the project, and represent the broad community types. For convenience, and to be consistent with the terminology used in Queensland, the structural and physiognomic rain forest typology of Webb (1978) that uses life form diversity and structural complexity with leaf size terminology to characterise types was used. 1 = Microphyll Vine Forest (MVFF), for example Nothofagus dominated forest often referred to as Cool ; 2 = Simple Notophyll – Microphyll Vine Forest (SNVF-SNMVF), for example Warm Temperate rainforest dominated by Ceratopetalum; 3 = Complex – Notophyll Vine Forest (CNVF-NVF), lowland to upland Sub-tropical rainforest); 4 = Araucarian Notophyll – Microphyll Vine Forest (ANVF-ANMVF) representing the drier vine forests ± Araucarian emergents; 5 = Eucalypt dominated Wet and Drier Sclerophyll forest types; 6 = rocky outcrop, cliff, mountain-top, habitat specialists.

Data analysis methods for the target species Multivariate analysis methods are based on the recommended methods detailed in Clarke and Worley (2006) for different data types; including linearly ranked (trait) data. The species by trait data were normalised and then classified by grouping similar taxa using a simple numerical hierarchical agglomerative clustering process, and the Gower association measure in the Primer v6 package. Similarity profile permutation tests (Simprof) were used to test the groupings. Similarity among quadrats/sites was further investigated through non-metric multidimensional scaling (nMDS) ordination using the underlying resemblance matrix as input. This provides a direct representation of the underlying classification in ordination space. Principal component analysis (PCA) was subsequently used to visualise and examine the position of group members in component space relative to the influence of trait variables (as axes). The Euclidean distance measure was used in the PCA analyses.

10 The relative merits of the various distance and association measures used in multivariate analysis are discussed in more detail in Clarke et al. (2006a,b); Clarke and Gorley (2006). In the case of the data rich (SNVF) example cited below that included site by species by abundance by environmental data, the Bray-Curtis distance measure was used for classification and ordination of sites.

In all cases additional analyses were undertaken to test the initial results. These included ANOSIM permutation tests (for the R statistic), and a modified MRT (multivariate regression ) analysis (De’ath, 2002) referred to as the Linkage Tree procedure in the PRIMER v6 package (Clarke and Gorley, 2006; Clarke et al., 2006c; and see Clarke and Warwick, 2001). The outputs from these analyses provided additional opportunities to interrogate and test the pattern of relationships of the species to life history traits and trait combinations, and to groupings and habitats. The assessment matrix developed uses the (Simprof tested) Primer groups generated by cluster analysis (classification) and shown in the nMDS ordination.

Phylogenetic independent contrast tests In order to identify potential management groups we also explored the co-distribution of traits among taxa. However, the study did not set out to test evolutionary hypotheses relating to phenotypic evolution or lineage diversification, or to use phylogenetic diversity at community or habitat levels as a potential measure of priority, or as a means to measure achievement (sensu Barker, 2002). Nevertheless, we did investigate the impact of phylogeny on the distribution of the chosen character states (traits) across the rare and threatened flora included (in the project) in order to provide preliminary background information for future studies on the evolutionary emergence of the selected traits and the evolutionary and biodiversity dynamics of the local flora.

To that end, a super-tree including all the BRBMP rare and threatened plant genera as terminal taxa was developed using Phylomatic (Webb and Donoghue, 2005). This is an automated process that builds a hypothetical tree based on the most recently published phylogenetic data. Genera rather than single species were used because species-level

11 phylogenies are too poorly understood to produce a useful tree. Intra-generic relationships were not always resolved and as a result clades for which insufficient information existed were presented as polytomies (soft), and unknown branch length were given unit length. The multi-state traits were traced over the tree using MacClade 4.03 (Maddison and Maddison, 2001). Phylogenetic conservatism, or clustering of traits, was examined for each trait by randomly reshuffling character states among the taxa 1000 times (Chazdon et al., 2003; Gross, 2005). Using character tracing, the actual number of steps for each character was compared to those obtained across the 1000 randomisations. If the actual number of steps ranked within the lowest 5% of those obtained in the reshuffled , the character was considered as significantly phylogenetically clustered.

Results Identifying appropriate trait-based groups Figure 2 shows the nMDS Primer emergent groups (G1-G5) based on the five selected traits for all species. Consistent with the Simprof tested dendrogram (not presented) this shows the life history trait groupings and identifies the relationship (similarity / dissimilarity) of trait groups. Some boundary overlap is evident in several groups, and there are several outliers (notably in G3). Figures 3 (nMDS) and 4 (PCA) show the distribution of species in ordination space relative to the influence of broad habitat types (1-6). Clustering relative to the interaction of traits (see also Figs. 5a-e) and habitats is evident, particularly in relation to seed size, fruit type and dispersal, and re-sprouting (refer to Rossetto and Kooyman, 2005). The additional analyses undertaken (including Simprof tests of the clustering analysis groups, and Regression Tree Analysis, in the Primer package) confirmed the relative strength of factors influencing the clustering, provided statistical tests of the results at a group level (Table 1), and (in the case of the Simprof tested clustering) provided the comparative group column for the multi-species assessment matrix (Appendix 1).

The distribution of species in ordination space relative to the influence of species altitudinal range shows considerable overlap with little or no clustering relative to the

12 selected traits in relation to species distributions along the altitudinal gradient. The altitudinal ordinations are (therefore) not presented.

Figures 5a-e (nMDS) show the distribution of trait influence across species and species groups in ordination space. When viewed in tandem with the habitat ordinations they provide insight into the patterns of distribution of traits in habitats (refer to Appendix 1 for more detail). They confirm the influence of seed size, dispersal mode, and the influence of re-sprouting / clonality (as persistence).

The phylogenetic conservatism test showed that phylogeny significantly affects the clustering of character states for nearly all the traits studied, with the exception of resprouting. However, we found that phylogenetic groups were not as useful as functional trait groups because species in the same family or genera often did not share the same combination of life history traits necessary to inform and develop threat / risk assessments at group and habitat level(s).

BRBMP Project Listed Species The species included in the BRBMP are predominantly rainforest species however the list includes taxa from outside the rainforest habitat and represents at least four rainforest habitat types. The range of life forms is also wide, with trees, shrubs, herbs, sedges, orchids, and epiphytes represented. Plant functional traits are known to vary along environmental gradients (reflecting both within and between habitat variation), and within communities of species occupying habitats with similar conditions. In this case the focus was on potential species level trait variation in relation to habitat variation and species altitudinal distributions, and these were the key factors used to determine and test emergent trait-based functional groups. Potential or identified threat and risk categories in the final plan reflect (and are influenced by) both landscape position (upland or lowland) and subsequent susceptibility to threatening processes, and are related to species life forms, life history traits, and life cycle aspects reflected in the groups.

13 Description of trait-based groups obtained from cluster analysis Figure 2 (nMDS) provides an overview of the relationship of trait-groups (G1-5), while Appendix 1 provides the list of species in each group (column heading - G) as part of the multi-species assessment matrix. Group descriptions (1-5) are provided below to assist interpretation.

Group 1: Species in this group have large fruits, are often dispersal limited (dispersal modes include gravity and/or rodents; refer to Rossetto and Kooyman, 2005), are mostly large canopy to medium to small persistent trees with the capacity to resprout, and are mature phase shade tolerant rainforest species. An outlier in the group (extreme right side of group) is the large herb Doryanthes palmeri (rocky outcrop specialist with big and big fruits), but habitat based sub-grouping resolves this. Most of these species are persistent and consequently potentially resistant to a range of threats (Rossetto and Kooyman, 2005). Some are found only within the existing reserve system (for example and sedentarius ms.), while others are more vulnerable to anthropogenic threats on private land (notably the two species).

Group 2: Species in this group are mostly small seeded herbs, sedges, shrubs and trees. The group is split almost in half (left and right) between rainforest habitats (2-4), and Wet Sclerophyll (5) and Cliff and Rock Outcrop (6) habitat specialists. Habitat provides a robust secondary explanation for the allocation of species in the sub-groups, relative to potential management actions. The cliff-top and rocky outcrop habitat group is dominated by shrubs and herbs; the wet sclerophyll habitat group includes sedges and shrubs; while the rainforest habitat group includes a variety of life forms (herbs, vines, a sedge, shrubs, and trees). A number of the latter group occur on the lowlands and are threatened by a variety of factors. The high altitude rocky outcrop group has several species that are potentially at risk from tourism activities and infrastructure (for example the cliff-top herb Euphrasia bella).

Group 3: Species in this group are mostly ferns, orchids, and epiphytes (representing distinct life form groupings) with mostly very small seeds. The group does include

14 several woody plants including one Gymnosperm ( baileyi) and one Asteraceae shrub (Cassinia collina). The outliers include several cliff top herbs and two sedges. The rocky outcrop (cliff) specialists from higher altitudes could (therefore) be grouped with those from the previous group for management convenience, though there is some variation on the basis of traits related to dispersal.

Group 4: The species in this group are mostly woody vines, trees, shrubs and parasitic mistletoes, with the two sub-groupings (interestingly sharing a similar mix of life forms) representing a split in species habitat preference for moist rainforest habitats (2-3) and the drier vine forest (4) and other habitats (5-6).

Group 5: The species in this group are mostly persistent trees and shrubs including clonal and re-sprouting species, with most species from the rainforest habitat types (1-4), and a few shrubs from the wet sclerophyll and rocky outcrop habitats.

Discussion Multi-species planning based on trait-related groups: conceptual framework Trait-based grouping has two main advantages. First, by cross-referencing to relevant taxa within the same group, it facilitates the delivery of preliminary recovery actions even for threatened species for which little ecological information is available. Second, it enables improved prioritisation of the taxa in need of further research (eg. representative taxa from poorly researched groups) and identifies the type of data needed (refer to Appendix 1). For many species, detailed data on distribution, population size, population dynamics and demographics, population biology and ecology (inclusive of genetic aspects), habitat aspects inclusive of environmental gradients, and community species richness and relative abundances are absent. In this case, the assessment matrix has identified the knowledge gaps, and a separate implementation process will fill those gaps through time relative to nominated Recovery Plan criteria, targets, and time-lines. As the assessment matrix, and group models, are populated with additional data through time, specific management based responses to identified (landscape scale) threats can be developed, tested, and modelled at the species and multi-species levels (both by group

15 and habitat/community; Appendix 1). Furthermore, once the process has started and the model is developed, adding new species to the system is relatively simple.

The accumulation of species level information for selected taxa will target representative and potentially informative species within the identified groups. It is intended that information accumulation at the species level should also inform the functional group and multi-species levels. Research results may then be more quickly available, informative, and broadly applicable than current species level approaches allow. This progressive implementation approach theoretically enables management actions to be taken on an increasing number of taxa as new information becomes available.

The approach suggested here offers a balance between the need for urgent action, both on-ground and in terms of the requirements and timelines associated with threatened species legislation, and the requirement for a sound scientific background on which to base biodiversity management and recovery planning. The trait-based framework for the multi-species planning approach proposed allows for a targeted and progressive accumulation of species level data (biological, ecological and evolutionary) through time without compromising the capacity of agencies and authorities to undertake initial threat/risk management analyses based on available information. That is, in contrast to a resource constrained species-level scenario, the group-based approach allows agencies to identify and fill significant knowledge gaps more quickly and evenly, and as a consequence theoretically provides the opportunity to take appropriate conservation action in a reduced time-frame across a larger number of taxa in a habitat.

Implementing threatened flora management strategies We suggest that data reflecting trait-based functional group composition in relation to community composition and environmental gradients provide the means to develop improved management approaches and responses to identified threats. Put simply, the relationship of biologically and ecologically derived groups to environmental gradients, combined with a measure of the relative susceptibility of species in those groups to threats, can enable more informed extrapolation to data-poor taxa. We believe that this

16 may be the ‘core’ information required to effectively implement biodiversity management plans using a multi-species approach. Unlike other surrogate type approaches, it provides information that links management to all species. Further to this, the approach works well even with artificial and heterogeneous species lists, because consulting the assessment matrix immediately provides information on relevant within- group distinctions (see example below).

We propose a three-stage approach for the development of multi-species recovery plans based on the use of trait-based groups developed from a limited database and relying on limited resources.

1. The first stage involves the accumulation of existing species level information based (as a minimum) on the traits used here, and reflecting the dispersal and persistence components of species life histories. In drier habitats this would include factors such as soil seed reserve and on-plant seed storage, as examples.

2. The second stage determines trait-based functional groupings on the basis of that information, compiled either as ranked or continuously measured trait data, by using the multivariate clustering methods recommended (and tested) in this study. This information is then added to the assessment matrix provided in Appendix 1 that identifies the life history trait-based groups and further sorts the groups by habitat types. The matrix also provides an overview of available information related to conservation status, level of threat, and other factors that can inform conservation management. The information in the assessment matrix, in combination with threat susceptibility ranking, will determine species and group level priority based on risk of decline and extinction.

3. The third stage identifies knowledge gaps and targets selected taxa representative of each group for additional data collection through time. The additional information that needs to be collected for representative species from each functional group during the implementation phase of the multi-species biodiversity management plan

17 includes data related to species distributions, population structure, population size, population demographics, genetic diversity and structure, reproductive biology and dispersal. This is the information that will help predict possible group responses to identified threats.

Developing multi-species recovery strategies: a preliminary example As a preliminary example, a number of the species in Group 1 (including introrsa, Eidothea hardeniana and Elaeocarpus sedentarius ms.) occur within SNVF- SNMVF, and have almost the whole of their distributional extent represented therein. Two of these species (E. hardeniana and E. sedentarius ms.) have been studied in considerable detail and adequate life history and habitat (community) information is available (refer to Appendix 1) to use them in a representative group example. The SNVF species occur in larger areas of reserved ‘natural’ habitat, and have very small population sizes, patchy and limited distribution, are known to be dispersal limited, and have limited capacity for both population expansion and/or response to disturbance events that further reduce population sizes. The group also includes two Macadamia species that have a similar combination of life history traits but occur predominantly in lowland sub-tropical rainforest (CNVF) habitats that are threatened by agricultural activities and development proposals. Dispersal of all these species is based predominantly on gravity, and rodent species that are known seed predators and have limited home ranges.

All the woody species in the group are relatively persistent (capable of resprouting) mature phase species with relatively slow turnover (and growth) rates. This suggests that despite small population size and subsequent heightened susceptibility to stochastic events, the combination of in-situ persistence and storage can mitigate the influence of unfavourable periods of time (or events) for reproduction (Higgins et al., 2000; Bond and Midgley, 2001; Rossetto and Kooyman, 2005). In this case, only persistent or extreme factors (such as total removal of available habitat by, for example, land clearing for development or identified risk factors such as pollen pollution by commercial orchard grown Macadamia) will remove such species completely.

18 An illustrative example of generalised management options for Group 1 species 1. Highland / protected populations (species): Translocation: unlikely to be necessary in populations that are within protected areas as these are protected and even small populations can contain sufficient diversity to be viable through mechanisms including long-term persistence of genetic individuals and preferentially outcrossed breeding systems (for example, Eidothea hardeniana, Elaeocarpus sedentarius, Endiandra introrsa, E. globosa, Niemeyera whitei, and likely pinnatifolia). While it is acknowledged that low genetic variability is likely in species with excessive reliance on vegetative reproduction (for example, the lowland species Elaeocarpus williamsianus, and Davidsonia johnsonii, both present in Group 5) the species in Group 1 show no tendency to clonality despite several species having small population sizes (for example, Eidothea hardeniana). In cases where translocation might be regarded as necessary care should to be taken with selection of material as molecular studies have shown that significant genetic spatial autocorrelation can result in considerable between-population genetic structure in large-fruited species, especially in the presence of local landscape barriers (for example, Eidothea hardeniana, Elaeocarpus sedentarius).

Actions: As a result of this improved understanding, conservation interventions based on maximising within-species diversity need to be done with awareness of landscape and habitat features that can act as ‘natural’ barriers to gene flow and protect localised within species diversity.

2. Lowland populations (species): Translocation: could be necessary in lowlands where fragmentation is an issue, and where single individuals, or very small populations are found within remnants (for example, Endiandra floydii, praealta, eerwah, Neisosperma poweri). The main difference with the previous group is that these small populations are likely to be a consequence of anthropogenic activity, rather than being at equilibrium. Species in the Macadamia have been identified as being vulnerable to pollen limitation (Pisanu, 2001), and the two lowland Macadamia

19 species in Group 1 must be regarded as vulnerable to pollen pollution from commercial orchard grown Macadamia cultivars, particularly when natural sources are not available (i.e. within small populations).

Buffer zone: In the context of these lowland species it might be essential to the survival of smaller natural and translocated populations to be protected from factors such as pollen pollution (for Macadamia), and disturbance and climatic events and variables, by appropriate scale buffers.

Habitat expansion: The small scale of some of the isolated remnants and remnant populations of some of these lowland species suggests that it may be important to both increase the population size and within species diversity of the populations and the size of the area of forest habitat, to secure a future for the species.

Management approaches for the species within the group are therefore similar in the same habitats but differ between habitats according to relative threat susceptibility. This shows that the trait-based groupings, in combination with habitat and distributional information, can provide a workable and effective approach to define, develop, and deliver appropriate conservation management actions. This remains contingent on adequate ecological and biological information being available for representative species within groups, and in relation to habitats and distribution relative to landscape threats.

Acknowledgements This project was partly funded by the New South Wales Department of Environment and Conservation, Andrew Hall, and Rainforest Rescue. We acknowledge the support of the BRBMP team from NSW DEC, Paul Houlder for providing the map and Chris Allen, Doug Benson, Caroline Gross and Bob Makinson for providing comments on earlier drafts of the manuscript.

20 References Abbitt, R.J.F. and Scott, J.M. (2001) Examining differences between recovered and declining endangered species. Conservation Biology 15(5), 1274-1284. Adam, P. (2001) A role for restoration ecologists in endangered community conservation? Ecological Management and Restoration 2(3), 165-166. Araújo, M.B., Densham, P.J. and Williams, P.H. (2004) Representing species in reserves from patterns of assemblage diversity. Journal of Biogeography 31, 1-14. Barker, G.M. (2002) Phylogenetic diversity: a quantitative framework for measurement of priority and achievement in biodiversity conservation. Biological Journal of the Linnean Society 76, 165-194. Bawa, K.S., Seidler, R. and Raven, P.H. (2004) Reconciling conservation paradigms. Conservation Biology 18(4), 859-860. Belbin, L. & Collins, A. (2004) PATN version 3.02 and 3.03. Blatant Fabrications Pty. Ltd. Hobart. Bond, W.J. and Midgley, J.J. (2001) Ecology of sprouting in woody plants: the persistence niche. Trends in Ecology and Evolution 16, 45-51. Braun-Blanquet, J. (1932) Plant Phytosociology. McGraw-Hill, New York. Burbidge, N.T. (1960) The phytogeography of the Australian region. Australian Journal of , 8, 75-211. Burgman, M.A., Possingham, H.P., Lynch, A.J.J., Keith, D.A., McCarthy, D.A., Hopper, S.D., Drury, W.L., Passioura, J.A. and Devries, R.J. (2001) A method for setting the size of plant conservation target areas. Conservation Biology 15(3), 603-616. Chazdon, R.L., Careaga, S., Webb, C. and Vargas, O. (2003) Community and phylogenetic structure of reproductive traits of woody species in wet tropical forests. Ecological Monographs 73: 331-348. Clark, J.A., Hoekstra, J.M., Dee Boersma, P. and Kareiva, P. (2002) Improving U.S. Endangered Species Act recovery plans: key findings and recommendations of the SCB Recovery Plan Project. Conservation Biology 16(6), 1510-1519. Clarke, K.R. and Warwick, R.M. (2001) Change in marine communities. An approach to statistical analysis and interpretation (2nd Edition). Primer-E Ltd. Plymouth Marine Laboratory. Plymouth. Clarke, K.R. and Gorley, R.N. (2006) Primer v6 user manual. Primer-E Ltd. Plymouth Marine Laboratory. Plymouth. Clarke, K.R., Chapman, M.G., Somerfield, P.J. and Needham, H.R (2006a) Dispersion- based weighting of species counts in assemblage analyses. Marine Ecology Progress Series 320, 11-27. Clarke, K.R., Somerfield, P.J., Airoldi, L. and Warwick, R.M. (2006b) Exploring interactions by second-stage community analyses. Journal of Experimental Marine Biology and Ecology 338, 179-192). Clarke, K.R., Somerfield, P.J. and Warwick, R.M. (2006c) On resemblance measures for ecological studies, including taxonomic dissimilarities and a zero-adjusted Bray-Curtis coefficient for denuded assemblages. Journal of Experimental Marine Biology and Ecology 330, 55-80. Cornelissen, J.H.C., Lavorel, S., Garnier, E., Diaz, S., Buchmann, S., Gurvich, D.E., Reich, P.B., ter Steege, H., Morgan, H.D., van der Heijden, M.G.A., Pausas, J.G. and

21 Poorter, H. (2003) A handbook of protocols for standardised and easy measurement of plant functional traits worldwide. Australian Journal of Botany 51, 335-380. De’ath, G. (2002) Multivariate regression trees: a new technique for modelling species- environment relationships. Ecology 83, 1105-1117. Faith, D.P. (2003) Environmental diversity (ED) as surrogate information for species- level biodiversity. Ecography 26(3), 374-379. Faith, D.P., Ferrier, S. and Walker, P.A. (2004) The ED strategy: how species-level surrogates indicate general biodiversity patterns through an ‘environmental diversity’ perspective. Journal of Biogeography 31, 1207-1217. Farnsworth, E. (2007) Plant life history traits of rare versus frequent plant taxa of sandplains: Implications for research and management trials. Biological Conservation 136, 44-52. Fischer, J., Lindenmayer, D.B. and Cowling, A. (2004) The challenge of managing multiple species at multiple scales: reptiles in an Australian grazing landscape. Journal of Applied Ecology 41, 32-44. Floyd, A.G. (1989) Rainforest Trees of Mainland South-eastern Australia. Inkata Press, Sydney. Forest, F., Grenyer, R., Rouget, M., Davies, T.J., Cowling, R.M., Faith, D.P., Balmford, A., Manning, J.C., Proches, S., van der Bank, M., Reeves, G., Hedderson, T.A.J., Savolainen, V. (2007) Preserving the evolutionary potential of floras in biodiversity hotspots. Nature 445, 757-760. Gitay, H., Noble, I.R. (1997) What are functional types and how should we seek them? In: Smith, T.M., Shugart, H.H., Woodward, F.I. (Eds.), Plant Functional Types, Cambridge University Press, Cambridge, pp. 3-19. Grime, J.P. (1977) Evidence for the existence of three primary strategies in plants and its relevance to ecological and evolutionary theory. American Naturalist 111, 1169-1194. Gross, C.L. (2005) A comparison of the sexual systems in the trees from the Australian tropics with other tropical biomes – more monoecy but why? American Journal of Botany 92, 907-919. Hagen, A.M. and Hodges, K.E. (2006) Resolving critical habitat designation problems: reconciling law, policy and biology. Conservation Biology 20(2), 399-407. Harden, G. (1990-2002) Flora of New South Wales, Vol. 1-4 (with revisions). University of New South Wales Press, Sydney, NSW, Australia. Harding, E.K., Crone, E., Elderd, B.D., Hoekstra, J.M., McKerrow, A.J., Perrine, J.D., Regetz, J., Rissler. L.J., Stanley, A.G., Walters, E.L. and NCEAS Habitat Conservation Plan Working Group (2001) The scientific foundations of habitat conservation plans: a quantitative assessment. Conservation Biology 15(2), 488-500. Harvey, P.H., Read, A.F. and Nee, S. (1995) Why ecologists need to be phylogenetically challenged. Journal of Ecology 83, 535-536. Hecht, A. and Parkin, M.J. (2001) Improving peer review of listings and recovery plans under the Endangered Species Act. Conservation Biology 15(5), 1269-1273. Higgins, S.I., Pickett, S.T.A. and Bond, W. (2000) Predicting extinction risks for plants: environmental stochasticity can save declining populations. Trends in Ecology and Evolution 15, 516-520. Kooyman, R.M. (2005) The Ecology and Population Biology of A.J. Scott. MSc Thesis, UNE, Armidale.

22 Kolb, A. and Diekmann, M. (2005) Effects of life-history traits on responses of plant species to forest fragmentation. Conservation Biology 19(3), 929-938. Lambeck, R.J. (1997) Focal species: a multi-species umbrella for nature conservation. Conservation Biology 11(4), 849-856. Maddison, D.R., Maddison, W.P. (2001) MacClade 4.03. Sinauer Associates, inc., Sunderland, Massachusetts, USA. Male, T.D. and Bean, M.J. (2005) Measuring progress in US endangered species conservation. Ecology Letters 8, 986-992. Mc Donald, W.J.F. & Elsol, J.A. (1984) Moreton Region map series: summary report and species checklist for Caloundra, Brisbane, Beenleigh and Murwillumbah sheets. McNeely, J.A. (2006) Systems or Species? Approaches to conservation for the 21st century. Integrative Zoology 2, 86-95. Meentemeyer, V. and Box, E.O. (1987) Scale effects in landscape studies. Landscape, Heterogeneity and Disturbance (ed. M.G. Turner), pp. 15-34. Springer-Verlag, New York, NY. Mooers, A.Ø. (2007) The diversity of biodiversity. Nature 445, 717-718. Moore, S.A. and Wooller, S. (2004) Review of Landscape, Multi- and Single-species Recovery Planning for Threatened Species. Report for WWF. Nicholson, E, and Possingham, H.P. (2006) Objectives for multiple-species conservation planning. Conservation Biology 20(3), 871-881. Pisanu, P.C. (2001) Survivorship of the Threatened Subtropical Rainforest Tree L. Johnson () in Small Habitat Fragments. Doctoral Thesis, UNE, Armidale. Poorter, L., Bongers, L. and Bongers, F. (2006) Architecture of 54 moist forest tree species: traits, tradeoffs, and functional groups. Evolution 87(5), 1289-1301. Roberge, JM. and Angelstam, P. (2004) Usefulness of the umbrella species concept as a conservation tool. Conservation Biology 18(1), 76-85. Rondinini, C., Wilson, K.A., Boitani, L., Grantham, H. and Possingham, H.P. (2006) Tradeoffs of different types of species occurrence data for use in systematic conservation planning. Ecology Letters 9, 1136-1145. Root, K.V., Akcakaya, H.R. and Ginzburg, A.L. (2003) A multi-species approach to ecological valuation and conservation. Conservation Biology 17(1), 196-206. Rossetto, M., Gross, C.L., Jones, R. & Hunter, J. (2004a) The impact of clonality on an endangered tree (Elaeocarpus williamsianus) in a fragmented rainforest. Biological Conservation, 117, 33-39. Rossetto, M., Jones, R. & Hunter, J. (2004b) Genetics effects of rainforest fragmentation in an early successional tree (Elaeocarpus grandis). Heredity, 93, 610-618. Rossetto, M. and Kooyman, R.M. (2005) The tension between dispersal and persistence regulates the current distribution of rare palaeo-endemic rain forest flora: a case study. Journal of Ecology 93, 906-917. Schimper, A.F.W. (1903) Plant geography upon a Physiological Basis. Clarendon Press, Oxford. Soulé, M.E., Estes, J.A., Berger, J. and Martinez Del Rio, C. (2003) Ecological effectiveness: conservation goals for interactive species. Conservation Biology 17(5), 1238-1250.

23 Verheyen, K., Honnay, O., Motzkin, G., Hermy, M. and Foster, D.R. (2003) Response of forest plant species to land-use change: a life-history trait-based approach. Journal of Ecology 91, 563-577. Webb, L.J. and Tracey, J.G. (1981) Australian rain forests: patterns and change. Ecological Biogeography of Australia (ed. A. Keast), pp. 606-669. Dr W. Junk, The Hague. Webb, L.J., Tracey, J.G. and Williams, W.T. (1984) A floristic framework of Australian rain forests. Australian Journal of Ecology 9, 169-198. Westoby, M., Leishman, M.R. and Lord, J.M. (1995) On misinterpreting the “phylogenetic correction.” Journal of Ecology 83, 531-534. Westoby, M. (1998) A leaf-height-seed (LHS) plant ecology strategy scheme. Plant Soil 199, 213-227. Westoby, M., Falster, D.S., Moles, A.T., Vesk, P.A. and Wright, I.J. (2002) Plant ecological strategies: some leading dimensions of variation between species. Annual Review of Ecology and Systematics 33, 125-159. Westoby, M. and Wright, I.J. (2006) Land-plant ecology on the basis of functional traits. Trends in Ecology and Evolution 21(5), 261-268. Westoby, M. and Burgman, M. (2006) Climate change as a threatening process. Austral Ecology 31, 549-550. Westoby, M. (2006) Phylogenetic ecology at world scale, a new fusion between ecology and evolution. Ecology 87(7), S163-S165.

24 Figure 1: Map of the study area.

Figure 2: Non-metric Multidimensional Scaling (nMDS) ordination showing 159 species by 5 traits. Simprof tested trait-groups (1-5) are shown (G1-G5). Trait groups represent those influenced by seed size (larger)/ fruit type and the dispersal dimension; and the largely independent traits of resprouting (persistence) and leaf size. Stress in ordination: 0.13.

Figure 3: Non-metric Multidimensional Scaling (nMDS) ordination showing 159 species by 5 traits and habitat a-priori group distribution(s). A-priori groups 1-4 (rainforest) with 1=MVFF; 2=SNMVF; 3=CNVF; 4=ANMVF; and groups 5-6 (non-rainforest), with 5=Eucalypt forest (wet and drier sclerophyll); 6=rocky/cliffs, habitat specialists. The ordination shows the relationship of trait groups to habitat types. Refer to Figs. 2 and 7A-E. Stress in ordination: 0.13.

Figure 4: PCA (principal component analysis) ordination showing 159 species by 5 traits by 6 habitats. Some overlap between habitat groups is evident. Consistent with the nMDS analyses, these represent trait groups influenced (primarily) by seed size/fruit type and the dispersal dimension; resprouting (persistence); and leaf size. A-priori groups 1-4 (rainforest) with 1=MVFF; 2=SNMVF; 3=CNVF; 4=ANMVF; and groups 5-6 (non-rainforest), with 5=Eucalypt forest (wet and drier sclerophyll); 6=rocky/cliffs, habitat specialists.

Figure 5: Non-metric Multidimensional Scaling (nMDS) ordination showing 159 species by 5 traits by altitudinal group distribution(s); (1=lowland; 2=upland and lowland; 3=upland). Species are distributed in the ordination relative to seed size (larger)/ fruit type and the dispersal dimension; and resprouting (persistence). Refer to Figs 7A-E. The ordination shows that the selected traits and trait combinations are distributed across the range of species representing altitudinal distribution groups and confirms that, in this case, altitude has only a minor influence on trait variation. Stress in ordination: 0.13.

Figure 6: PCA (principal component analysis) ordination showing 159 species by 5 traits by 3 altitudinal groups. Considerable overlap across the altitudinal distribution groups is evident. This confirms that habitat types are the most influential variable and altitude has only a minor influence on trait variation (in this case). 1=lowland; 2=upland and lowland; 3=upland.

Figures 7 (a-e): Non-metric Multidimensional Scaling (nMDS) ordination (bubble-plots) showing 159 species by five traits (PRIMER v6); in relation to the prevalence and distribution of the correlated traits of dispersal mode / fruit type and seed size (a-c); and the largely independent variables of re-sprout (d) and leaf size (e). 7a. Dispersal Mode: 1 represents frugivore dispersed; 2 wind dispersed, 3 mammal; 4 ant (secondary); and 5 gravity. 7b Seed Size: 1 represents the smallest seeds and 5 the largest.7c Fruit Type: 1 represents fleshy fruits; 2 coloured arils; 3 non fleshy (dry); 4 spores; and 5 other.7d Re- sprout (persistence): 1 represents no resprouting; 2 intermediate disturbance response only; and 3 yes (as mature phase stem replacement; refer to Rossetto and Kooyman 2005).7e Leaf Size: 1 represents smallest (nanophyll); 2 microphyll; 3 notophyll; 4 mesophyll; and 5 macrophyll.

25 Fig 1.

26

Fig. 2

Fig. 3

27

Fig. 4

Fig. 5

28

Fig. 6

29 7a. Dispersal Mode 1 represents frugivore dispersed; 2 wind dispersed, 3 mammal; 4 ant (secondary); and 5 gravity.

7b. Seed Size 1 represents the smallest seeds and 5 the largest.

30

7c. Fruit Type 1 represents fleshy fruits; 2 coloured arils; 3 non fleshy (dry); 4 spores; and 5 other.

7d. Re-sprout (persistence) 1 represents no resprouting; 2 intermediate disturbance response only; and 3 yes (as mature phase stem replacement; refer to Rossetto and Kooyman 2005).

31

7e. Leaf Size 1 represents smallest (nanophyll); 2 microphyll; 3 notophyll; 4 mesophyll; and 5 macrophyll.

32 Table 1: Results of LINKTREE (classification and regression tree analysis) Simprof test in PRIMER v6 (1000 permutations) showing major splits (A-G) in link-tree dendrogram (not presented) by R-value (non-parametric measure of multivariate distance) and B% (absolute measure of group differences) by most significant trait(s).

Group R B% Trait A 0.63 86.2 dispersal mode B 0.73 67.3 dispersal mode C 0.70 48.6 resprout D 0.69 30.6 seed size E 0.88 24.8 leaf size F 0.84 81.1 seed size G 0.72 65.7 dispersal mode

Table 2: Results from the trait clustering tests, showing the actual number of steps, range of steps in randomized trees, and proportion of trees with a number of steps lower or equal to the actual number.

Trait Observed steps Range of steps in 1000 Significance randomisations Seed size 24 24-42 0.001 Fruit type 25 25-42 0.001 Dispersal 26 26-46 0.001 Leaf size 32 30-45 0.006 Resprouting 23 20-33 0.6

33 Appendix 1 - Border Ranges Biodiversity Assessment Matrix List of species included in the BRBMP ordered according to the placement within the five trait-based groups (G). Additional available ecological and environmental information useful for decision-making relating to threat management and conservation is also listed (H: habitat type; LF: life form; AD: altitudinal distribution; PS: population size). The final group of columns lists the availability (Y: data available; -: data not available; ip: research in progress) of ecological data relating to major areas of scientific research useful to the conservation and management of species (PSt: population structure; DS: demographic structure; LP: landscape patchiness; DE: distributional extent; FO: frequency of occurrence in habitat; GD: genetic diversity; GS: genetic structure; BM: breeding mechanisms). Other symbols represent: H (habitats 1 to 6); LF (t:tree; st:small tree; s:shrub; v:vine; h:herb; se:sedge; f:fern; fa:fern ally; e:epiphyte; p:parasitic mistletoe; go:ground orchid); AD (1:lowland only; 2:lowland and upland; 3:upland); PS (s:small; m:medium; l:large).

Species Family G H LF AD PS PSt DS LP DE FO GD GS BM Eidothea hardeniana Proteaceae 1 2 t 3 s Y Y Y Y Y Y Y ip Elaeocarpus sedentarius 1 2 t 3 s Y Y Y Y Y Y Y ip Endiandra globosa 1 2 t 2 l ------Endiandra introrsa Lauraceae 1 2 t 3 l Y Y Y Y Y - - - Niemeyera whitei 1 2 t 2 l - - Y Y - - - - Corynocarpus rupestris subsp. rupestris Corynocarpaceae 1 3 t 2 s ------Endiandra compressa Lauraceae 1 3 t 1 s ------Endiandra floydii Lauraceae 1 3 t 1 m ------Floydia praealta Proteaceae 1 3 t 1 m ------Hicksbeachia pinnatifolia Proteaceae 1 3 t 2 l - - Y Y - - - - Macadamia tetraphylla Proteaceae 1 3 st 1 m Y Y Y Y Y Y Y Y Neisosperma poweri Apocynaceae 1 3 st 1 m ------Pouteria eerwah Sapotaceae 1 4 t 1 s ------ Proteaceae 1 4 st 1 m Y Y Y Y Y Y Y Y Doryanthes palmeri Doryanthaceae 1 6 h 3 m - - Y Y - ip ip - 2 2 t 3 l ------Carex hubbardii Cyperaceae 2 2 se 3 m ------Cassia brewsteri var. marksiana Fabaceae 2 2 t 1 s ------Olearia heterocarpa Asteraceae 2 2 s 3 m ------ pentacocca var. pentacocca 2 3 t 2 m ------Bosistoa selwynii Rutaceae 2 3 t 1 s ------Bosistoa transversa Rutaceae 2 3 t 1 s ------Desmodium acanthocladum Fabaceae 2 3 s 1 l ------Harnieria hygrophiloides Acanthaceae 2 3 s 1 m ------Helmholtzia glaberrima Phylidraceae 2 3 h 3 l ------Isoglossa eranthemoides Acanthaceae 2 3 h 1 s ------Senna acclinis Fabaceae 2 3 s 2 s ------ cunninghamii 2 4 s 2 s - - Y Y - Y Y - Rhynchosia acuminatissima Fabaceae 2 4 v 2 s ------Cyperus semifertilis Cyperaceae 2 5 se 3 m ------Lepidosperma clipeicola 2 5 se 3 l ------Sophora fraseri Fabaceae 2 5 s 2 s ------Westringia blakeana Lamiaceae 2 5 s 3 l ------Zieria collina Rutaceae 2 5 s 2 m ------Zieria southwellii Rutaceae 2 5 s 3 l ------Argophyllum nullumense Escalloniaceae 2 6 s 2 l ------Cyperus rupicola Cyperaceae 2 6 se 3 m ------Euphrasia bella Scrophulariaceae 2 6 h 3 s ------Gaultheria viridicarpa subsp. merinoensis Ericaceae 2 6 s 3 s ------Huperzia varia Lycopodiaceae 2 6 fa 3 m ------Leionema gracile Rutaceae 2 6 s 3 m ------Wahlenbergia glabra Campanulaceae 2 6 h 3 s ------Wahlenbergia scopulicola Campanulaceae 2 6 h 3 s ------Xerochrysum bracteatum subsp. Mt Merino Asteraceae 2 6 h 3 s ------Lastreopsis silvestris Dryopteridaceae 3 1 f 3 s ------Bulbophyllum caldericola 3 2 e 3 m ------34 stenophylla Grammitidaceae 3 2 f 3 m ------Lindsaea brachypoda Dennstaedtiaceae 3 2 f 3 s ------Angiopteris evecta Marrattiaceae 3 3 f 1 s ------Belvisia mucronata 3 3 e 2 s ------Bulbophyllum argyropus Orchidaceae 3 3 e 3 m ------Bulbophyllum globuliforme Orchidaceae 3 3 e 3 m ------Clematis fawcettii Ranunculaceae 3 3 v 2 m ------Crepidomanes vitiense Hymenophyllaceae 3 3 f 3 s ------Cyathea cunninghamii Cyatheaceae 3 3 f 2 m ------Dendrobium schneiderae var. schneiderae Orchidaceae 3 3 e 3 s ------Oberonia complanata Orchidaceae 3 3 go 1 s ------Peristeranthus hillii Orchidaceae 3 3 e 1 s ------Psilotum complanatum Psilotaceae 3 3 e 1 s ------ dilatatus Orchidaceae 3 3 e 2 m ------Sarcochilus fitzgeraldii Orchidaceae 3 3 e 2 s ------Sarcochilus hartmannii Orchidaceae 3 3 e 2 s ------Sarcochilus weinthalii Orchidaceae 3 3 e 3 s ------Bulbophyllum weinthalii Orchidaceae 3 4 e 3 m ------Callitris baileyi 3 5 t 1 s ------Cassinia collina Asteraceae 3 5 s 3 m ------Asplenium harmanii Aspleniaceae 3 6 e 3 m ------Drynaria rigidula Polypodiaceae 3 6 f 2 m ------Gen.(Aq247974) sp. Mt Merino Asteraceae 3 6 h 3 s ------Ozothamnus vagans Asteraceae 3 6 s 3 l ------Podolepis monticola Asteraceae 3 6 h 3 m ------Acronychia baeuerlenii Rutaceae 4 2 s,st 3 m Y ? Y Y Y - - - Austrobuxus swainii 4 2 t 3 l Y ? Y Y Y - - - ferruginea Proteaceae 4 2 st 3 l Y - Y Y Y - - - Pandorea baileyana Bignoniaceae 4 2 v 3 l ------Pararistolochia laheyana Aristolochiaceae 4 2 v 3 l - - Y Y Y - - - Streptothamnus moorei Salicaceae 4 2 v 3 l Y - Y Y Y - - - Symplocos baeuerlenii Symplocaceae 4 2 s,st 3 l Y - Y Y Y - - - pinnatum Proteaceae 4 3 st 3 m Y - Y Y - - - - Amyema flexialabstra Loranthaceae 4 3 p 1 s Y - Y Y Y - - - Amyema plicatula Loranthaceae 4 3 p 1 s ------Archidendron hendersonii Fabaceae 4 3 t 1 m Y - Y Y Y - - - Archidendron muellerianum Fabaceae 4 3 st 2 l Y - Y Y Y - - - marmorata Euphorbiaceae 4 3 st 1 s Y - Y Y Y - - - congesta Agavaceae 4 3 s 1 m Y - Y Y Y - - - foetida Lauraceae 4 3 st 1 m Y - Y Y Y - - - flagelliformis var. australis Sapindaceae 4 3 t 2 m Y - Y Y Y - - - Cupaniopsis newmanii Sapindaceae 4 3 s,st 2 l Y - Y Y Y - - - Davidsonia jerseyana Cunoniaceae 4 3 st 1 m Y Y Y Y Y ip ip ip Dendrocnide moroides Urticaceae 4 3 s,st 2 s ------ mabacea 4 3 s,st 1 s Y - Y Y Y - - - Diospyros major var. ebenus Ebenaceae 4 3 st 1 s Y - Y Y Y - - - campbellii Sapindaceae 4 3 t 1 s Y - Y Y Y - - - Endiandra hayesii Lauraceae 4 3 t 2 m Y - Y Y Y - - - Endiandra muelleri subsp. bracteata Lauraceae 4 3 t 2 s Y - Y Y Y - - - hilliana Proteaceae 4 3 t 1 m Y - Y Y - - - - Hypserpa decumbens Menispermaceae 4 3 v 1 m ------Jasminum jenniae Oleaceae 4 3 v 2 m ------ pulchella Sapindaceae 4 3 se 2 m Y - Y Y Y - - - Marsdenia coronata Apocynaceae 4 3 v 2 s ------Marsdenia hemiptera Apocynaceae 4 3 v 1 s ------Melicope vitiflora Rutaceae 4 3 st 1 s Y - Y Y Y - - - Owenia cepiodora Meliaceae 4 3 t 2 s Y - Y Y Y - - - Pararistolochia praevenosa Aristolochiaceae 4 3 v 1 m Y - Y Y Y - - - 35 Parsonsia tenuis Apocynaceae 4 3 v 3 m ------ moorei 4 3 st 1 m Y - Y Y Y - - - Syzygium hodgkinsoniae 4 3 t 2 m Y - Y Y Y - - - Syzygium moorei Myrtaceae 4 3 t 1 m Y - Y Y Y - - - Tinospora tinosporoides Menispermaceae 4 3 v 1 l Y - Y Y Y - - - Acacia bakeri Fabaceae 4 4 t 2 s Y - Y Y - - - - Brachychiton sp. Ormeau Malvaceae 4 4 t 1 s ------Cupaniopsis serrata Sapindaceae 4 4 st 1 s ------Cupaniopsis tomentella Sapindaceae 4 4 st 2 m ------Marsdenia longiloba Asclepiadaceae 4 4 v 1 s ------Muellerina myrtifolia Loranthaceae 4 4 p 2 s ------Tarenna cameronii Rubiaceae 4 4 st 1 s ------Tinospora smilacina Menispermaceae 4 4 v 2 s - - Y Y - - - - Turraea pubescens Meliaceae 4 4 s,st 2 s ------ dunnii Myrtaceae 4 5 t 1 m Y - Y Y - - - - Solanum limitare Solanaceae 4 5 s 3 s ------Tylophora woollsii Apocynaceae 4 5 v 2 s Y - Y Y - - - - Leucopogon sp. Lamington Ericaceae 4 6 s,st 3 m ------Pittosporum oreillyanum Pittosporaceae 5 1 s 3 m Y - Y Y - - - - Corokia whiteana Escalloniaceae 5 2 s 3 l Y - Y Y Y - - - Cryptocarya meisneriana Lauraceae 5 2 s 3 l Y - Y Y Y - - - Daphnandra tenuipes Monimiaceae 5 2 t 3 l Y - Y Y Y - - - Lenwebbia prominens Myrtaceae 5 2 st 3 l Y - Y Y Y - - - Uromyrtus australis Myrtaceae 5 2 st 3 l Y Y Y Y Y - - Y Acronychia littoralis Rutaceae 5 3 st 1 s Y - Y Y Y - - - Actephila grandifolia Euphorbiaceae 5 3 s 2 m - - Y Y - - - - Ardisia bakeri Myrsinaceae 5 3 s 2 m ------Davidsonia johnsonii Cunoniaceae 5 3 st 1 s Y Y Y Y Y ip ip ip Elaeocarpus williamsianus Elaeocarpaceae 5 3 t 1 s Y Y Y Y Y Y Y Y Eucryphia jinksii Cunoniaceae 5 3 t 3 s Y Y Y Y Y - - - australis Euphorbiaceae 5 3 st 2 m - - - - - Y Y - Fontainea oraria Euphorbiaceae 5 3 st 1 s Y Y Y Y Y Y Y ip Lenwebbia sp. Main Range Myrtaceae 5 3 st 3 m ------Mischocarpus lachnocarpus Sapindaceae 5 3 st 2 m ------Ochrosia moorei Apocynaceae 5 3 st 2 s - - Y Y Y - - - Phyllanthus microcladus Euphorbiaceae 5 3 s 1 s ------Quassia sp. Mt Nardi Simaroubaceae 5 3 s 2 m - - Y Y - - - - Rhodamnia maideniana Myrtaceae 5 3 s,st 1 m - - Y Y - - - - Uromyrtus lamingtonensis Myrtaceae 5 3 st 3 l ? Y Y Y Y - - - Wilkiea austroqueenslandica Monimiaceae 5 3 s,st 2 l - - Y Y Y - - - Xylosma terrae-reginae Salicaceae 5 3 st 1 m - - Y Y - - - - Acalypha eremorum Euphorbiaceae 5 4 s 1 s ------Choricarpia subargentea Myrtaceae 5 4 t 1 s Y Y Y Y Y - - - Citrus australasica Rutaceae 5 4 s,st 1 m - - Y Y - - - - Coatesia paniculata Rutaceae 5 4 st 1 s ------Cryptocarya floydii Lauraceae 5 4 st 3 m ------Fontainea venosa Euphorbiaceae 5 4 st 1 s ------ fragrantissima Myrtaceae 5 4 st 1 s Y Y Y Y Y - - - Pouteria cotinifolia var. cotinifolia Sapotaceae 5 4 st 1 s ------Hibbertia hexandra Dilleniaceae 5 5 s 3 l Y Y Y Y Y - - - Leionema elatius subsp. beckleri Rutaceae 5 5 s,st 3 l Y - Y Y Y - - - Myrsine richmondensis Myrsinaceae 5 5 s,st 2 s ------Pomaderris notata Rhamnaceae 5 5 s 3 m ------Pimelea umbratica Thymelaeaceae 5 6 s 3 m ------Plectranthus nitidus Lamiaceae 5 6 l 3 l - - Y Y - - - -

36 Appendix 2

Priority integrative research identified by the Assessment Matrix as necessary for the implementation of the Border Ranges Biodiversity Management Plan

The trait-based conceptual framework for the multi-species planning approach allows for a targeted and progressive accumulation of species level data (biological, ecological and evolutionary) through time without compromising the capacity of agencies and authorities to undertake initial threat/risk management analyses based on available information. In particular, the Assessment Matrix (Appendix 1) clearly identifies priority areas for integrative research implementation in relation to obtaining relevant demographic, genetic and breeding systems information. Furthermore, two of the trait-related groups are conspicuous for the near total absence of relevant biological information, which is likely to hinder any meaningful management process. These are Group 2 (mostly represented by small seeded herbs, sedges, shrubs, and trees from five out of six of the habitats included within this plan) and Group 3 (including mostly ferns, orchids, and epiphytes with generally very small seeds and related dispersal modes and capabilities).

Implementation actions should ensure that relevant information on distribution, population size, population dynamics and demographics, population biology and ecology (inclusive of genetic aspects), habitat aspects inclusive of environmental gradients, and community species richness and relative abundances is obtained (and compiled) for selected representative taxa (see priority list below). This way, the biodiversity assessment tool, and group models, will be populated with the essential additional data that will identify specific management based responses to threats at the species and ecological community levels. This is the basis of the progressive implementation approach presented that enables implementation actions to be taken on an increasing number of taxa as new information becomes available.

Appendix 3 shows, as a simplified example, part of the community-level information that can be efficiently obtained through well planned and integrated data gathering. It describes species distributions and abundance, and species richness in communities in relation to the influence of environmental gradients. Such data is both interpretative and predictive, particularly in regard to describing the realised niche, and, when combined with relevant genetic and population dynamics data, will be essential for the management and restoration of endangered ecological communities.

37

The additional information that needs to be collected for the representative species from each functional group listed below includes data related to species distributions, population structure, population size, population demographics, genetic diversity and structure, reproductive biology and dispersal. Because of recent technical and analytical developments and existing local experience, this information can be gathered relatively rapidly and efficiently, and will be essential to predict group responses to identified threats.

Table 1 List of taxa that need to be prioritised for the integrative research approach described above. Existing and available information are listed (PSt: population structure; DS: demographic structure; LP: landscape patchiness; DE: distributional extent; FO: frequency of occurrence in habitat; GD: genetic diversity; GS: genetic structure; BM: breeding mechanisms). * high priority taxa

Species Family PSt DS LP DE FO GD GS BM Group 1 *Endiandra introrsa Lauraceae y y y y y x x x *Endiandra floydii Lauraceae x x x x x x x x *Hicksbeachia pinnatifolia Proteaceae x x y y x x x x Niemeyera whitei Sapotaceae x x y y x x x x *Pouteria eerwah Sapotaceae x x y y x x x x

Group 2 Olearia heterocarpa Asteraceae x x x x x x x x Bosistoa pentacocca var. pentacocca Rutaceae x x x x x x x x *Bosistoa transversa Rutaceae x x x x x x x x Corchorus cunninghamii Malvaceae x x y y x y y X *Rhynchosia acuminatissima Fabaceae x x x x x x x x *Senna acclinis Fabaceae x x x x x x x x Sophora fraseri Fabaceae x x x x x x x x *Harnieria hygrophiloides Acanthaceae x x x x x x x x Zieria collina Rutaceae x x x x x x x x Zieria southwellii Rutaceae x x x x x x x x *Euphrasia bella Scrophulariaceae y x y x x x x x *Gaultheria viridicarpa subsp. merinoensis Ericaceae x x y y x x x x Wahlenbergia glabra Campanulaceae x x x x x x x x Wahlenbergia scopulicola Campanulaceae x x x x x x x x Cyperus rupicola Cyperaceae x x x x x x x x Lepidosperma clipeicola Cyperaceae y x y y y x x x

Group 3 *Bulbophyllum caldericola Orchidaceae x x x x x x x x *Bulbophyllum globuliforme Orchidaceae x x x x x x x x *Clematis fawcettii Ranunculaceae x x x x x x x x Dendrobium schneiderae var. schneiderae Orchidaceae x x x x x x x x *Sarcochilus fitzgeraldii Orchidaceae x x x x x x x x Sarcochilus hartmannii Orchidaceae x x x x x x x x *Sarcochilus weinthalii Orchidaceae x x x x x x x x Ozothamnus vagans Asteraceae x x x x x x x x Podolepis monticola Asteraceae x x x x x x x x

Group 4 Helicia ferruginea Proteaceae y x y y y x x x *Archidendron hendersonii Fabaceae y x y y y x x x *Baloghia marmorata Euphorbiaceae y x y y y x x x *Davidsonia jerseyana Cunoniaceae y y y y y ip ip ip Diospyros major var. ebenus Ebenaceae y x y y y x x x Diploglottis campbellii Sapindaceae y x y y y x x x *Endiandra hayesii Lauraceae y x y y y x x x *Endiandra muelleri subsp. bracteata Lauraceae y x y y y x x x Lauraceae y x y y x x x x

38 Melicope vitiflora Rutaceae y x y y y x x x Owenia cepiodora Meliaceae y x y y y x x x Syzygium hodgkinsoniae Myrtaceae y x y y y x x x Syzygium moorei Myrtaceae y x y y y x x x Tinospora tinosporoides Menispermaceae y x y y y x x x Tinospora smilacina Menispermaceae y x y y x x x x *Tylophora woollsii Asclepiadaceae y x y y x x x x *Marsdenia longiloba Asclepiadaceae x x x x x x x x Muellerina myrtifolia Loranthaceae x x x x x x x x Amyema plicatula Loranthaceae x x x x x x x x

Group 5 *Davidsonia johnsonii Cunoniaceae y y y y y ip ip ip *Uromyrtus australis Myrtaceae y y y y y x x y Uromyrtus lamingtonensis Myrtaceae x y y y y x x x *Gossia fragrantissima Myrtaceae y y y y y x x x Fontainea oraria Euphorbiaceae y y y y y y y y *Fontainea australis Euphorbiaceae x x x x x y y x Pouteria cotinifolia var. cotinifolia Sapotaceae x x x x x x x x *Phyllanthus microcladus Euphorbiaceae x x x x x x x x Ardisia bakeri Primulaceae x x x x x x x x Myrsine richmondensis Primulaceae x x x x x x x x

39 Appendix 3 – Data Rich Community Example (SNMVF)

Introduction The single data rich community sample included here contains nearly a fifth (20%) of the species listed in the BRBMP and provides an example of how additional (detailed) information can inform the management planning process and modelling. Only ‘woody’ species are represented in the sample. Refer to Table 1 below.

Methods A data rich example for a single community type (SNVF-SNMVF, inclusive of the Wet Sclerophyll elements adjacent) from previous research undertaken by us is provided to test the efficacy and applicability of the methods in situations where data is available to explore species richness, abundance, and distribution in a community relative to selected traits and environmental variables. Refer to Rossetto & Kooyman (2005).

Data analysis methods for the data rich community example The data rich community sample covers 9.2 hectares and is comprised of 92 / 50 x 20m quadrat samples (with nested 20 x 20m subplots). The quadrat data were entered into a matrix consisting of 92 sites (objects) and 258 species (attributes). The data represents the specific rainforest types (SNVF- SNMVF) and areas of interaction and overlap with adjacent vegetation communities, and covered the whole of the available habitat area on the southern flanks of the Mt Warning study area with rhyolite- derived soils (154 km2).

All species that occurred within a plot were identified and recorded to species level. Species cover codes (modified from Braun-Blanquet 1932) were entered as a cover abundance scale. No transformation (or weighting) of the species abundance measures was undertaken as the intention was to preserve as much of the information in the full floristic samples as possible (refer to Clarke et al. 2006a).

The environmental variables used for ordination analyses were derived from the field-collected environmental data. The relative values and allocated rankings of the environmental variables and cover codes, along with detailed description of the methods are provided in Rossetto & Kooyman

40 (2005). In this case, transformation of variables included replacement of all measured environmental variables by ranks for PCA, and the use of the Spearman Rank coefficient for subsequent tests of results (Global Tests) and multivariate regression tree analyses (Clarke & Gorley 2006).

The floristic data were classified by grouping similar plots using a numerical hierarchical agglomerative classification process, the Bray-Curtis association measure, and similarity profile permutation tests (Simprof) of clusters (1000 permutations).

Additional analyses included PCA (principal component analysis), nMDS (non-metric multidimensional scaling), ANOSIM permutation tests (for the R statistic), the Global BEST match test, and a modified MRT (multivariate regression tree) analysis (De’ath 2002) referred to as the Linkage Tree procedure, all in the PRIMER v6 package (Clarke & Gorley 2006; Clarke et al. 2006c; and see Clarke & Warwick 2001). The outputs from these analyses provided additional opportunities to interrogate and test the pattern of relationships of the chosen environmental (abiotic) variables to assemblage patterns (floristic variation) in the data. The Global BEST match test procedure used the underlying resemblance matrix from the site by species data, Spearman rank correlation, and the Euclidean resemblance measure for the site by environmental variables data (Clarke & Gorley 2006).

Trait relationships for rare and threatened taxa in the SNVF sample Trait-based analyses for all the species (258) in this data followed the methods described above in the methods section for the listed species (159) by traits section. The data rich quadrat sample included a total of 30 of the 159 listed rare and threatened species in the project. That total included 4 restricted endemics that occur only in the study area, and a total of 25 species (including the 4 endemics) that have what we consider an adequate sample of the species distribution(s) and habitat preference(s) to interpret species level patterns and describe the ‘realised niche’ (Table 1). The results of the multivariate analyses (as ordinations) for the 30 target species are presented below. In this case the analsyses were undertaken in the PATN package (Belbin & Collins 2004) and the ordination diagram (Fig. 1) generated therein.

41 Table 1 (Appendix 2) List of (30) BRMSRP listed species in SNVF-SNMVF data rich example. *species with sufficient data to interpret patterns and define the realised niche.

*Species data: Species Habitat (SNVF-SNMVF-WS) sufficient? Acacia orites Yes Disturbance related Acronychia baeuerlenii Yes Mature, rainforest Archidendron muellerianum Yes Mature, rainforest Austrobuxus swainii Yes Mature, rainforest Corokia whiteana Yes Mature, rainforest Cryptocarya meisneriana Yes Mature, rainforest Daphnandra tenuipes Yes Mature, rainforest Eidothea hardeniana Yes Mature, rainforest Elaeocarpus sedentarius Yes WS edges, rainforest Endiandra globosa No WS, rainforest Endiandra hayesii Yes Mature, rainforest Endiandra introrsa Yes Mature, rainforest Helicia ferruginea Yes Mature, rainforest Helmholtzia glaberrima Yes Wet areas, riparian Hibbertia hexandra Yes WS, rainforest edges Hicksbeachia pinnatifolia Yes WS edges, rainforest Leionema elatius subsp. beckleri Yes WS, rainforest (lower canopy) Lenwebbia prominens Yes Mature, rainforest (wet) Lepidosperma clipeicola Yes WS, rocky, wet Melicope vitiflora Yes Lower alt., mature, rainforest Niemeyera whitei Yes Lower alt., mature, rainforest Pandorea baileyana Yes Mature, rainforest Pararistolochia laheyana Yes Mature, rainforest Quassia sp. Mt Nardi No Low alt., mature, rainforest Streptothamnus moorei Yes Mature, rainforest Symplocos baeuerlenii Yes Mature, rainforest Syzygium hodgkinsoniae No Lower alt., mature, rainforest Uromyrtus australis Yes Mature, lower rainforest Wilkiea austroqueenslandica No Mature, rainforest Zieria southwellii Yes WS, rainforest edge

Results Figure 1 shows the results of the species (258) by traits (5) ordination (refer to Rossetto & Kooyman 2005). Consistent with the findings of that study, the ordination presented here shows the influence of the dispersal dimension (inclusive of seed size, fruit type and dispersal mode) and persistence (as resprouting / clonality).

42

Fig. 1 Constrained ordination of 258 species in relation to five life history traits. Kruskal-Wallis values for the five traits are: fruit type (199.503), dispersal mode (156.1388), seed size (100.3892), resprout (99.4091), leaf size (15.3688). To provide a less congested graph, only the (30) rare and threatened species listed for inclusion in the BRBMP are shown. The influence of the fruit-type / seed-size / dispersal dimension is apparent on the ordination of the rare and threatened species, with most of these positioned towards the larger seed end of the size gradient. Resprouting is the other important trait dimension influencing the ordination of the rare taxa. Overall stress in the ordination (0.1115).

The results of the site by species abundance by environmental variables analyses (nMDS and PCA ordinations) for the (30) listed species present in the data for this community type are provided below. The figures show the distribution and abundance of the subset of rare and threatened species in the community based 92-plot sample relative to the influence of selected environmental variables. This equates with a presentation of the ‘realised niche’ for 24 of the species with sufficient data.

Discussion The results of the multivariate analyses of the SNMVF (Warm Temperate Rainforest) of the southern Mt. Warning caldera area indicates the potential for data rich plot based samples, particularly when complemented by detailed population genetic studies, to provide important information to support biodiversity management planning at the species, multi-species, community and habitat (ecosystem) levels. The real value of this type of data is the obvious additional power it provides for both interpretation and prediction, particularly in regard to ‘describing’ the realised niche. In the case of species with large or complete samples of population distributions within the sample, it represents species occurrence data without the frequently identified problems of commission and omission errors

43 (refer to Rondinini et al. 2006). Most importantly the data describes species distributions and abundance, and species richness in communities in relation to the influence of environmental gradients. This has the potential to link both species and trait-based functional groups to gradients, habitats, and threat responses to improve management responses. We suggest that such information will also be critical for the endangered ecological community restoration (see for example, Adam 2001) and management components of the project.

Figures 2-18 below represent species level data showing abundance and distribution relative to environmental variables. Figs. 2&3 provides an example of four palaeo-endemic species in the data set for which whole of population data is available (or almost so). Figs. 4-18 provide information for 30 species, with 29 of those listed as BRBMP species.

44

45

46

47

48

49

50

51

52