Mark Dobrowolski

The kwongan ecosystem is a biodiversity hotspot on the coastal sandplain of southwest Western Australia. It harbours many rare and endemic plant species adapted to nutrient impoverished soils, particularly to low phosphorus (P). In some kwongan regions surface strip-mining has drastically disturbed these ecosystems, requiring restoration by soil profile reconstruction and re-establishment of the native vegetation, often assisted by mineral fertiliser applications. We examined soil monitoring data from several post-mining sites of restored kwongan vegetation (n = 127), categorised by age groups (10, 20, 30 and 40-year-old sites) and fertilisation history (type of fertiliser applied). Thirty two soil variables were compared to the surrounding native soils (n = 135) by multivariate and univariate statistical approaches to determine whether different restoration ages and fertilisers had long-term impacts on soil physicochemical characteristics. Our analyses showed that restored soils differed from native, with substantially lower carbon (−41 %) and nitrogen (−15 %). Sites restored within the last 10 years had higher sodicity and much lower porosity, while 40-year-old sites presented nearly ten times more Total P. While recent non-fertilised soils showed similar P concentrations to the reference sites, the high legacy P (persistence of P through time) found in older restored soils is a clear consequence of past superphosphate application, greatly decreasing soil N:P stoichiometric ratios (low N and high P) compared to native soils. We show that, after decades, restored soils are still impacted by the past mining activities and fertiliser application. These changes will likely alter microbial and plant communities, impairing restoration trajectories towards the native kwongan vegetation, particularly for P-sensitive species. Despite benefits of P-fertilisation (e.g. improved plant cover), the low carbon stocks and N:P ratios indicate a divergent ecosystem from the native state. Such changes in ecosystem stoichiometry may affect plant species competitiveness, and alter the composition of other trophic levels. Adjusting future restoration practices to reduce P fertilisation is promising, but addressing long-term stoichiometric shifts and soil compaction remains crucial.
[Display omitted]

1. The restoration of diverse self-sustaining ecosystems requires re-establishment of functional interactions among species. For plant communities, pollinators are usually essential for pollination, seed set and seed quality. A common assumption in ecological restoration for plants pollinated by animals is one of ‘build it and they will come’, which is rarely tested. Beyond seed set, there may be negative genetic consequences for seed quality if pollinators and their behaviour do not reflect those in reference populations.

2. Here, we conduct an ecological genetic assessment of seed quality via mating system parameters in Lambertia multiflora (Proteaceae), a species dependent on nectar-feeding birds for pollination. Four populations of L. multiflora in disturbed sites that were rehabilitated following mineral sands mining were compared with four native reference populations, near Eneabba, Western Australia. In each population, approximately 10 offspring from each of 10 maternal plants were genotyped with 11 highly polymorphic microsatellite markers.

3. From these data, genetic diversity and mating system parameters were assessed, and found to be equivalent across all populations. Mean allelic diversity and heterozygosity across loci were very high. All populations were completely outcrossing with no bi-parental inbreeding. Mean correlated paternity, sibship and effective population size estimates for restored and natural populations were not significantly different and reflected uniformly high paternal diversity and wide outcrossing. Equivalent genetic results for restored and natural reference populations indicate successful restitution of bird-pollinator services for L. multiflora in these post-mining rehabilitation sites.

4. Synthesis and applications. Reviewing our results with other published studies to date suggests a resilience of bird-pollinator services in restored plant communities. These findings provide some reassurance to restoration practitioners working in these global south systems where bird pollination is a feature, at least for similar landscape scenarios. Our study also highlights the global contribution of ecological genetics to the objective assessment of functional species interactions in ecological restoration, an increasingly important goal of land managers and regulators seeking to improve restoration standards.

Understanding what makes a community vulnerable to invasion is integral to the successful management of invasive species. Our understanding of how characteristics of resident plant interactions, such as the network architecture of interactions, can affect the invasibility of plant communities is limited. Using a simulation model, we tested how successfully a new plant invader established in communities with different network architectures of species interactions. We also investigated whether species interaction networks lead to relationships between invasibility and other community properties also affected by species interaction networks, such as diversity, species dominance, compositional stability and the productivity of the resident community. We found that higher invasibility strongly related with a lower productivity of the resident community. Plant interaction networks influenced diversity and invasibility in ways that led to complex but clear relationships between the two. Heterospecific interactions that increased diversity tended to decrease invasibility. Negative conspecific interactions always increased diversity and invasibility, but increased invasibility more when they increased diversity less. This study provides new theoretical insights into the effects of plant interaction networks on community invasibility and relationships between diversity and invasibility. Combined with increasing empirical evidence, these insights could have useful implications for the management of invasive plant species.

Post-mining scenarios present challenges for restoration in a wide range of environments, especially in the context of climate change. The source of seed for restoration has been an issue of intense focus, as seed provenance can impact plant fitness and restoration outcomes. However, post-mining landscapes require substrate reconstruction prior to vegetation re-establishment. Critically, the relative importance of provenance and substrate in ecosystem recovery has been rarely quantified in a statistically rigorous framework. We established a large provenance trial with Banksia attenuata and Eucalyptus todtiana at two Western Australian mine sites in post-mining reconstructed and adjoining un-mined substrates. We show that site and substrate were 4 and 26 times more important than provenance in explaining survival for B. attenuata and E. todtiana, respectively. At one site, there was 100% mortality in the post-mining substrate but high survival and no clear provenance effect in the un-mined substrate. At the second site, there was again no clear provenance effect, but E. todtiana survival was higher in the post-mining than un-mined substrate. Our results show that post-mining substrate changes can overwhelm provenance issues. Consequently, where substrates are highly impacted, alternative restoration targets and/or greater investment in substrate research are needed to improve restoration outcomes. Due to the thousands of mines across the world, this is an internationally relevant finding with important implications for investment into global ecosystem recovery.

Seed for restoring natural ecosystems is an expensive and finite resource. Yet seedling emergence is notoriously low in mining rehabilitation and restoration projects more generally. Unlocking this potential in broadcast seed has been critical to restoring a diverse kwongan ecosystem at Iluka’s Eneabba mineral sands mine, in the mid-west of Western Australia.

Using an innovative combination of rehabilitation practices in the sandy soils of Eneabba—seed burial, land imprinting, and an artificial soil crust—two-fold more seedlings emerged from broadcast seed in 2017 than from standard rehabilitation practice of ripping and seed broadcast. To combine and automate these practices in a single pass, we designed, built, and commissioned tractor-drawn equipment named ‘Flora Restorer’. Flora Restorer spreads fertiliser, scarifies the uneven soil, air-seeds a diverse size and shape seed mix, land imprints and thereby buries the seed, and sprays the soil surface with dilute bitumen emulsion forming an artificial soil crust to stop wind and water erosion.

Over the last four years of use, consistently two- to three-fold more seedlings have emerged from broadcast seed using Flora Restorer than from previous rehabilitation practice. Independently, long-term botanical monitoring has also demonstrated revegetation improvement using Flora Restorer, with species richness more than 50% higher and native plant cover more than double compared to rehabilitation at Eneabba over the previous 20 years. Flora Restorer demonstrates that large improvements in revegetation outcomes are possible in mining rehabilitation by automating seed sowing in combination with land imprinting and application of an artificial soil crust.

Background and Aims
Clonality is a key life-history strategy promoting on-spot persistence, space occupancy, resprouting after disturbance, and resource storage, sharing and foraging. These functions provided by clonality can be advantageous under different environmental conditions, including resource-paucity and fire-proneness, which define most mediterranean-type open ecosystems, such as southwest Australian shrublands. Studying clonality–environment links in underexplored mediterranean shrublands could therefore deepen our understanding of the role played by this essential strategy in open ecosystems globally.

Methods
We created a new dataset including 463 species, six traits related to clonal growth organs (CGOs; lignotubers, herbaceous and woody rhizomes, stolons, tubers, stem fragments), and edaphic predictors of soil water availability, nitrogen (N) and phosphorus (P) from 138 plots. Within two shrubland communities, we explored multivariate clonal patterns and how the diversity of CGOs, and abundance-weighted and unweighted proportions .of clonality in plots changed along with the edaphic gradients.

Key Results
We found clonality in 65 % of species; the most frequent were those with lignotubers (28 %) and herbaceous rhizomes (26 %). In multivariate space, plots clustered into two groups, one distinguished by sandy plots and plants with CGOs, the other by clayey plots and non-clonal species. CGO diversity did not vary along the edaphic gradients (only marginally with water availability). The abundance-weighted proportion of clonal species increased with N and decreased with P and water availability, yet these results were CGO-specific. We revealed almost no relationships for unweighted clonality.

Conclusions
Clonality is more widespread in shrublands than previously thought, and distinct plant communities are distinguished by specific suites (or lack) of CGOs. We show that weighting belowground traits by aboveground abundance affects the results, with implications for trait-based ecologists using abundance-weighting. We suggest unweighted approaches for belowground organs in open ecosystems until belowground abundance is quantifiable.

Plant traits—the morphological, anatomical, physiological, biochemical and phenological characteristics of plants—determine how plants respond to environmental factors, affect other trophic levels, and influence ecosystem properties and their benefits and detriments to people. Plant trait data thus represent the basis for a vast area of research spanning from evolutionary biology, community and functional ecology, to biodiversity conservation, ecosystem and landscape management, restoration, biogeography and earth system modelling. Since its foundation in 2007, the TRY database of plant traits has grown continuously. It now provides unprecedented data coverage under an open access data policy and is the main plant trait database used by the research community worldwide. Increasingly, the TRY database also supports new frontiers of trait‐based plant research, including the identification of data gaps and the subsequent mobilization or measurement of new data. To support this development, in this article we evaluate the extent of the trait data compiled in TRY and analyse emerging patterns of data coverage and representativeness. Best species coverage is achieved for categorical traits—almost complete coverage for ‘plant growth form’. However, most traits relevant for ecology and vegetation modelling are characterized by continuous intraspecific variation and trait–environmental relationships. These traits have to be measured on individual plants in their respective environment. Despite unprecedented data coverage, we observe a humbling lack of completeness and representativeness of these continuous traits in many aspects. We, therefore, conclude that reducing data gaps and biases in the TRY database remains a key challenge and requires a coordinated approach to data mobilization and trait measurements. This can only be achieved in collaboration with other initiatives.

Vegetation maps are models of the real vegetation patterns and are considered important tools in conservation and management planning. Maps created through traditional methods can be expensive and time-consuming, thus, new more efficient approaches are needed. The prediction of vegetation patterns using machine learning shows promise, but many factors may impact on its performance. One important factor is the nature of the vegetation-environment relationship assessed and ecological redundancy. We used two datasets with known ecological redundancy levels (strength of the vegetation-environment relationship) to evaluate the performance of four machine learning (ML) classifiers (classification trees, random forests, support vector machines, and nearest neighbor). These models used climatic and soil variables as environmental predictors with pretreatment of the datasets (principal component analysis and feature selection) and involved three spatial scales. We show that the ML classifiers produced more reliable results in regions where the vegetation-environment relationship is stronger as opposed to regions characterized by redundant vegetation patterns. The pretreatment of datasets and reduction in prediction scale had a substantial influence on the predictive performance of the classifiers. The use of ML classifiers to create potential vegetation maps shows promise as a more efficient way of vegetation modeling. The difference in performance between areas with poorly versus well-structured vegetation-environment relationships shows that some level of understanding of the ecology of the target region is required prior to their application. Even in areas with poorly structured vegetation-environment relationships, it is possible to improve classifier performance by either pretreating the dataset or reducing the spatial scale of the predictions.

Questions

The community patterns in kwongan, a mediterranean-type scrub on nutrient-poor soils occurring in Western Australia, are poorly understood due to only few, focused studies using disparate sampling designs. We aimed to determine whether (a) classification of the kwongan vegetation of the Eneabba Sandplains leads to an ecologically informative classification scheme, and (b) we could identify environmental drivers underpinning the plant community patterns.

Location

Township of Eneabba (29°82′S, 115°27′E), approximately 250 km north of Perth, Western Australia, covering 1,210 km2.

Methods

We used a data set consisting of 512 relevés, collected following the standard field methodology of the Braun-Blanquet approach, and accompanied by an extensive set of environmental data consisting of 94 variables representing climate, fire, soil and terrain properties across 189 relevés. The data were classified and ordinated by a series of multivariate analyses. OptimClass assisted in the selection of the most robust classification procedure. Distance-based redundancy analysis (db-RDA) inferred the major ecological drivers of the vegetation patterns.

Results

Numerical classification, nonmetric multidimensional scaling, and syntaxonomic tabular analysis revealed two major community groups (MCG A and B), eight community groups, and 17 communities in the kwongan vegetation of the study area. All vegetation units are characterised in terms of floristic composition and position along major ecological gradients in the studied area. The MCGs separated along a composite gradient of soil-texture and exchangeable cations. The first two db-RDA axes explained 21% of the total variance which is very low, considering the high number of environmental variables used.

Conclusions

Our study provides a comprehensive insight into the high variability of vegetation types in hyper-diverse kwongan scrub at a landscape spatial scale; it is the first syntaxonomic account of the Western Australian kwongan vegetation, presenting a complete tabular comparative analysis. The studied major community groups segregate along a soil-texture and an exchangeable-cation content gradient. At a community scale, environmental filtering explained a small fraction of the vegetation–environment relationship. We suggest that the unexplained portion of the vegetation-environment relationship might be a product of slow-acting neutral processes in this hyper-diverse system; this assertion is amenable to rigorous future testing.

Premise of the study: Microsatellite markers were developed for population genetic analysis in the rare shrub Styphelia longissima (Ericaceae).

Methods and Results: We generated ca. 2.5 million sequence reads using a Personal Genome Machine semiconductor sequencer. Using the QDD pipeline, we designed primers for >12,000 sequences with PCR product lengths of 80–480 bp. From these, 30 primer pairs were selected and screened using PCR; of these, 16 loci were found to be polymorphic, four loci were monomorphic, and 10 loci did not amplify reliably for S. longissima. For a sample of 57 plants from the only known population, the number of alleles observed for these 16 loci ranged from two to 21 and expected heterozygosity ranged from 0.49 to 0.91. These markers were also amplified in Astroloma xerophyllum, a closely related species.

Conclusions: These markers will be used to characterize population genetic variation, spatial genetic structure, mating system parameters, and dispersal to aid in the management and conservation of the rare shrub S. longissima.