Towards a continent-wide ecological site condition database using calibrated expert evaluations.

A cost-effective way of undertaking comprehensive, continental-scale, assessments of ecological condition is needed to support large-scale conservation planning, monitoring, reporting, and decision-making. Currently, cross-jurisdictional inconsistency in assessment methods limits the capacity to scale-up monitoring. Here we present a novel way to build a coherent continent-wide site-level ecological condition dataset, using cross-calibration methods to integrate assessments from many observers. We focus on the use of condition assessments from individual expert observers, a currently untapped resource. Our approach has two components: (1) a simple online tool that captures expert assessments at specific locations; (2) a process of calibrating and rescaling disparate expert evaluations that can be applied to the data to provide a consistent dataset for use in conservation assessments. We describe a pilot study, involving 28 experts, who contributed 314 individual site condition assessments across a wide range of ecosystems and regions throughout continental Australia. A correction factor for each expert was used to rescale the contributed site condition assessment scores, based on a set of 77 photographic images, each scored for their condition by multiple experts, using a linear mixed model. Our approach shows strong promise for delivering the volumes of data required to develop continental-scale reference libraries of site condition assessments. Although developed from expert elicitation, the approach could also be used to harmonize the collation of existing condition datasets. The process we demonstrate can also facilitate online citizen scientists to make site condition assessments that can be cross-calibrated using contributed images.

A condition metric for Eucalyptus woodland derived from expert evaluations.

The evaluation of ecosystem quality is important for land-management and land-use planning. Evaluation is unavoidably subjective, and robust metrics must be based on consensus and the structured use of observations. We devised a transparent and repeatable process for building and testing ecosystem metrics based on expert data. We gathered quantitative evaluation data on the quality of hypothetical grassy woodland sites from experts. We used these data to train a model (an ensemble of 30 bagged regression trees) capable of predicting the perceived quality of similar hypothetical woodlands based on a set of 13 site variables as inputs (e.g., cover of shrubs, richness of native forbs). These variables can be measured at any site and the model implemented in a spreadsheet as a metric of woodland quality. We also investigated the number of experts required to produce an opinion data set sufficient for the construction of a metric. The model produced evaluations similar to those provided by experts, as shown by assessing the model's quality scores of expert-evaluated test sites not used to train the model. We applied the metric to 13 woodland conservation reserves and asked managers of these sites to independently evaluate their quality. To assess metric performance, we compared the model's evaluation of site quality with the managers' evaluations through multidimensional scaling. The metric performed relatively well, plotting close to the center of the space defined by the evaluators. Given the method provides data-driven consensus and repeatability, which no single human evaluator can provide, we suggest it is a valuable tool for evaluating ecosystem quality in real-world contexts. We believe our approach is applicable to any ecosystem.

Quantifying ecosystem quality by modeling multi-attribute expert opinion.

The evaluation of ecosystem quality is inherently subjective, requiring decisions about which variables to notice or measure, and how these variables are integrated into a coherent evaluation. Despite the central role of human judgment, few evaluation methods address the subjectivity that is inherent in their design. There are, however, advantages to directly using opinion to create an expert system where the metric is constructed around opinion data. These advantages include stakeholder inclusion and the encouragement of a dialogue of data-driven criticism rather than subjective counter-opinion. We create an expert system to express the quality of a grassland ecosystem in Australia. We use an ensemble of bagged regression trees trained on calibrated expert preference data, to model the perceived quality of this grassland using a set of eight site variables as inputs. The model provides useful predictions of grassland quality, producing predictions similar to real expert evaluations of independent synthetic test sites not used to train the model. We apply the model to real grassland sites ranging from pristine to highly degraded, and confirm that our model orders the sites according to their degree of modification. We demonstrate that the use of too few experts produces relatively poor results, and show that for our problem the use of data from over twenty experts is appropriate. The scaling approach we used to calibrate between-expert data is shown to be an appropriate mechanism for aggregating the opinions of multiple experts. The resultant model will be useful in many contexts, and can be used by managers as a tool to evaluate real sites. It can also be integrated into ecological models of change as a means of evaluating predicted changes, for example, as a measure of utility when combined with cost estimates. The basic approach demonstrated here is applicable to any ecosystem, and we discuss the opportunities and limitations of its wider use.

An endangered flightless grasshopper with strong genetic structure maintains population genetic variation despite extensive habitat loss.

Conservation research is dominated by vertebrate examples but the shorter generation times and high local population sizes of invertebrates may lead to very different management strategies, particularly for species with low movement rates. Here we investigate the genetic structure of an endangered flightless grasshopper, Keyacris scurra, which was used in classical evolutionary studies in the 1960s. It had a wide distribution across New South Wales (NSW) and Victoria in pre-European times but has now become threatened because of land clearing for agriculture and other activities. We revisited remnant sites of K. scurra, with populations now restricted to only one area in Victoria and a few small patches in NSW and the Australian Capital Territory (ACT). Using DArtseq to generate SNP markers as well as mtDNA sequence data, we show that the remaining Victorian populations in an isolated valley are genetically distinct from the NSW populations and that all populations tend to be genetically unique, with large F ST values up to 0.8 being detected for the SNP datasets. We also find that, with one notable exception, the NSW/ACT populations separate genetically into previously described chromosomal races (2n = 15 vs. 2n = 17). Isolation by distance was detected across both the SNP and mtDNA datasets, and there was substantial differentiation within chromosomal races. Genetic diversity as measured by heterozygosity was not correlated with the size of remaining habitat where the populations were found, with high variation present in some remnant cemetery sites. However, inbreeding correlated negatively with estimated habitat size at 25-500 m patch radius. These findings emphasize the importance of small habitat areas in conserving genetic variation in such species with low mobility, and they highlight populations suitable for future translocation efforts.

Genetic data and climate niche suitability models highlight the vulnerability of a functionally important plant species from south-eastern Australia.

Habitat fragmentation imperils the persistence of many functionally important species, with climate change a new threat to local persistence due to climate niche mismatching. Predicting the evolutionary trajectory of species essential to ecosystem function under future climates is challenging but necessary for prioritizing conservation investments. We use a combination of population genetics and niche suitability models to assess the trajectory of a functionally important, but highly fragmented, plant species from south-eastern Australia (Banksia marginata, Proteaceae). We demonstrate significant genetic structuring among, and high level of relatedness within, fragmented remnant populations, highlighting imminent risks of inbreeding. Population simulations, controlling for effective population size (N e), suggest that many remnant populations will suffer rapid declines in genetic diversity due to drift in the absence of intervention. Simulations were used to demonstrate how inbreeding and drift processes might be suppressed by assisted migration and population mixing approaches that enhance the size and connectivity of remnant populations. These analyses were complemented by niche suitability models that predicted substantial reductions of suitable habitat by 2080; ~30% of the current distribution of the species climate niche overlaps with the projected distribution of the species climate niche in the geographic region by the 2080s. Our study highlights the importance of conserving remnant populations and establishing new populations in areas likely to support B. marginata in the future, and adopting seed sourcing strategies that can help populations overcome the risks of inbreeding and maladaptation. We also argue that ecological replacement of B. marginata using climatically suited plant species might be needed in the future to maintain ecosystem processes where B. marginata cannot persist. We recommend the need for progressive revegetation policies and practices to prevent further deterioration of species such as B. marginata and the ecosystems they support.

Expert predictions of changes in vegetation condition reveal perceived risks in biodiversity offsetting.

Biodiversity offsetting typically involves the trade of certain losses of habitat with uncertain future conservation benefits. Predicting the latter requires estimates of two outcomes; the biodiversity losses without conservation management (averted loss), and the biodiversity gains with conservation management (management gain). However, because empirical data to inform these estimates are limited, they are normally guided by expert opinion, often derived via unstructured methods without consideration of uncertainty. Here we used a structured elicitation with 29 experts to gather estimates of averted loss and management gain at offset sites. We used two methods; (i) experts estimated change in an aggregate biodiversity value (vegetation condition) and; (ii) experts provided probabilistic estimates of change for individual vegetation condition attributes, such as the richness and cover of plant growth forms. On average, experts predicted there would be only modest improvements with conservation management, yet uncertainty and variation among experts was large; in some cases, conservation benefits were not predicted. Estimates of change in vegetation condition suggested that benefits were from both averted loss and management gains and were thought to most likely arise in cases where starting condition was low to moderate. Similar patterns were observed for individual vegetation condition attributes, with management gains, relative to a reference, tending to be negatively correlated with starting value. Our study finds that: (i) on average, gains at offset sites are expected to be small, (ii) at many sites, experts do not believe gains can be obtained, and (iii) experts' opinions can be divergent resulting in elevated levels of uncertainty. The potential for losses under conservation management highlights the need to: identify those components of biodiversity most likely to benefit from conservation management; better understand those situations when offset obligations are most likely to be met and conversely those situations with higher risk; and further develop offset mechanisms that encourage early or prior gains. These findings together with the global proliferation of biodiversity offsetting, provide a strong imperative to improve empirical data and investment in long-term, site-based monitoring of biodiversity outcomes at offset sites.