Lance Wilkie

Complementarity, biodiversity viability analysis, and policy-based algorithms for conservation

Abstract Biodiversity conservation “area-selection” strategies include not only trade-offs among society’s needs in land-use allocation, but also allocation of economic instruments such as incentives, levies, and biodiversity credits. For these applications, the key property of an area is its “complementarity”—the context-dependent, marginal gain in biodiversity provided by the area. Given that there has been little implementation of whole-sets of areas generated by the popular computer-based selection methods, we suggest that analogous “policy-based algorithms” would be a more effective real-world application of complementarity. Areas would be “selected” for conservation over time as a consequence of policies in which dynamic complementarity values influence application of economic instruments. These integrated biodiversity/economic strategies can use an extended form of complementarity reflecting marginal changes in regional probability of persistence of biodiversity. While probabilistic measures of biodiversity viability have been explored in area-selection for some time, it remains difficult to make viability statements about “all of biodiversity.” New approaches that use biodiversity surrogate information for “biodiversity viability analysis” (BVA) can take advantage of a general quantitative biodiversity framework in which pattern-based relationships among areas allow predictions at the species level. A standard assumption of “unimodal” species responses to environmental gradients yields an expected distribution of species in an ordination pattern, and allows sampling of inferred species. Based on environmental correlates, inferred species can be mapped in geographic space, forming distribution fragments. This information, when linked to species persistence models, may allow ongoing calculation of areas’ complementarity values. An example illustrates application of these ordination models to museum collection data for lizards from New South Wales (NSW), Australia.

Can Volunteers Collect Data that are Comparable to Professional Scientists? A Study of Variables Used in Monitoring the Outcomes of Ecosystem Rehabilitation

Having volunteers collect data can be a cost-effective strategy to complement or replace those collected by scientists. The quality of these data is essential where field-collected data are used to monitor progress against predetermined standards because they provide decision makers with confidence that choices they make will not cause more harm than good. The integrity of volunteer-collected data is often doubted. In this study, we made estimates of seven vegetation attributes and a composite measure of six of those seven, to simulate benchmark values. These attributes are routinely recorded as part of rehabilitation projects in Australia and elsewhere in the world. The degree of agreement in data collected by volunteers was compared with those recorded by professional scientists. Combined results showed that scientists collected data that was in closer agreement with benchmarks than those of volunteers, but when data collected by individuals were analyzed, some volunteers collected data that were in similar or closer agreement, than scientists. Both groups’ estimates were in closer agreement for particular attributes than others, suggesting that some attributes are more difficult to estimate than others, or that some are more subjective than others. There are a number of ways in which higher degrees of agreement could be achieved and introducing these will no doubt result in better, more effective programs, to monitor rehabilitation activities. Alternatively, less subjective measures should be sought when developing monitoring protocols. Quality assurance should be part of developing monitoring methods and explicitly budgeted for in project planning to prevent misleading declarations of rehabilitation success.

A quality control protocol for terrestrial invertebrate biodiversity assessment

A procedure for the implementation of quality control for laboratorysorting and identification of invertebrate specimens collected in biodiversityresearch is described. The procedure is based on process control sampling, aconcept of statistical process control (SPC) used widely in the manufacturingand information technology industries, and adapted to suit the tasks andproducts of biodiversity sorting procedures. The major advantages of processcontrol over other quality control mechanisms are that it is more stringent, andcontinuous. Hence, errors are detected and corrected as they occur, avoidingproliferation in the data set. The procedure is also highly interactive,offering the technicians the opportunity to learn as they work. Protocols havebeen developed while sorting material collected as part of a study into theimpacts associated with invasion of a habitat (coastal heath) by an exotic weed(bitou bush – Chrysanthemoides monilifera) on thecentral coast of New South Wales, Australia. Major findings from the analysis ofmaterial processed include: that errors may have a variety of causes andsubsequent implications for data quality, levels of identification errors can besignificant even at higher taxonomic levels (e.g. sorting insects to order),initial training periods on their own are insufficient to ensure errorminimisation, and even with stringent protocols the ratio of technician tospecialist effort can be maintained at a level of around 5:1. The need forincorporating effective quality control procedures into invertebratebiodiversity data compilations is emphasised.

The effects on terrestrial arthropod communities of invasion of a coastal heath ecosystem by the exotic weed bitou bush (Chrysanthemoides monilifera ssp. rotundata L.)

Bitou bush (Chrysanthemoides monilifera ssp. rotundata) is a major environmental weed of littoral habitats on the southeastern coast of Australia. This study investigates the impacts upon selected arthropod assemblages of habitat invasion by this weed. Sixteen sites were placed at four geographic localities within nature reserves between Forster and Budgewoi on the coast of New South Wales. The sampling design included two spatial scales (between and within localities) and eight repeat samples (taken at two scales of temporal separation). Arthropods were collected from both arboreal and epigaeic micro-habitats. Specimens of ants (Hymenoptera: Formicidae), beetles (Coleoptera), true bugs (Heteroptera) and spiders (Araneae) were identified to species level. Differences in α diversity and species abundance distributions between the taxonomic assemblages are described, along with comparisons of data contrasts between bitou bush-free (“control”) and bitou bush-invaded (“impact”) habitats and between geographic localities. A subsidiary impact associated with herbicide application for control of the weed is also examined. Analysis of environmental differences indicates that bitou bush acts as a dampening agent, reducing the degree of seasonal fluctuation in factors such as leaf litter cover. Arboreal Heteroptera were the only group to show consistent evidence of significant differences in taxon richness or abundance between control and impact treatments within a locality, seen as a disruption of normal seasonal variation in diversity in bitou bush-impacted sites. Significant differences between geographic localities were more common, suggesting an effect at broader spatial scales. Evidence for arthropod assemblages characteristic of specific vegetation types was detected for several groups, as were changes in arthropod assemblage composition following application of herbicide for bitou bush control.

Using spider web types as a substitute for assessing web-building spider biodiversity and the success of habitat restoration

Arthropods have been regarded as good indicators of habitat quality due to their sensitivity to changes in habitat state. However, there are many constraints to working with arthropods that make them inaccessible to land managers and most volunteer-driven initiatives. Our study examined a novel approach for detecting changes in web-building spider communities by focussing on the types of webs that spiders build rather than the spider itself. This method was cost-effective, easy-to-use, and importantly, we found a strong congruency between the diversity of web architecture and the diversity of web-building spider genera. The metrics derived from this method could distinguish differences in web-building communities among habitat types that represented a successional gradient, and thus we concluded that the method was useful for monitoring the progress of restoration. Many other applications for the method are possible such as environmental impact assessment and agricultural pest management, and we encourage development in these areas.

The ratio of exotic-to-native dung beetles can indicate habitat quality in riparian restoration

1. Replanting natives on cleared riparian land is a common form of restoration. While most assessments of success are focussed on flora, the impact on fauna is often unknown.