I will be speaking on “Science for Conservation Decisions”. The lecture will introduce environmental decision theory, and describe an experimental test of a particular example of this theory (optimal searching for a species; it seemed to work quite well in that instance). I will finish by briefly noting how decision theory helps clarify thinking, and that is perhaps its greatest strength.

If you are interested in seeing the slides I am planning to use, I have posted them here.

Approximately two-thirds of species in the world are thought to be insects, while vertebrates make up a tiny fraction (probably less than 1%). If insects and vertebrates had the same extinction risk there should be about 100 times as many threatened insects as vertebrates, but insects are relatively rare in threatened species lists.

Given the paucity of insects in threatened species lists, one might think that they are somehow extremely robust to extinction. Of course, that discrepancy is simply a function of the attention they have received. Species are only listed when there is sufficient information to assess them. Insects and other invertebrates, while loved by some, get much less attention than their abundance, biomass and functional roles deserve.

One of our QAEG post-docs, Melinda Moir, is doing her bit to redress the imbalance. Mel is studying the co-extinction risk of invertebrates that depend on threatened plants (e.g., Moir et al. 2010, 2011, in press a, in press b). As part of her previous research on insects, Mel and Gary Taylor identified some new psyllid species, one of which, Vesk’s plant louse Acizzia veski (Taylor and Moir 2009), has just been listed as vulnerable in Western Australia. Yes, it is a sapsucker, but not the yellow-bellied version, I believe.

One of the interesting things is that it is the first insect to be listed in Australia (perhaps more widely?) as threatened on the basis of co-extinction risk, being an affiliate of the threatened plant Acacia veronica, which is restricted to the Stirling Range near Albany in Western Australia. Co-extinction, the extinction of a species resulting from a decline of its host, is seen as a major threat to biodiversity (Koh et al. 2004), so it is interesting that co-extinction figured prominently in the listing of A. veski.

Another interesting aspect is that the species in question, Acizzia veski, has the honour of being named after one of QAEG’s principal researchers Peter Vesk. Now I’m sure the naming has nothing to do with any similarity of appearance or behaviour.

Photo credit for A. veski: Gary Taylor

But the common naming is rather appropriate because Peter is the primary author of a paper that describes a method to estimate host breadth, which is an important aspect when assessing co-extinction risk (Vesk et al. 2010).  One of the key features of this paper is that it accounts for imperfect detection of affiliate species on their hosts, providing a probability distribution for both the number of affiliates per host species, and the number host species per affiliate. QAEG has a range of research in the area of detectability, some of which is described in previous posts.

References (please email me if you need copies of our papers)

Burgman, M.A., and Lindenmayer, D.B. (1998). Conservation biology for the Australian environment. Surrey Beatty, Australia.

Koh, L.P. et al. (2004). Species coextinctions and the biodiversity crisis. Science 305, 1632-1634.

May, R.M. (1986). Biological diversity: How many species are there? Nature 324, 514 – 515.

Moir, M.L., Vesk, P.A., Brennan, K.E.C., Hughes, L., Keith, D.A., McCarthy, M.A., Coates, D.J., and Barrett, S. (in press a). A preliminary assessment of changes in plant-dwelling insects when threatened plants are translocated. Journal of Insect Conservation.

Moir, M.L., Vesk, P.A., Brennan, K.E.C., Keith, D.A., Hughes, L., and McCarthy, M.A. (2010). Current constraints and future directions in estimating coextinction. Conservation Biology 24: 682-690.

Moir, M.L., Vesk, P.A., Brennan, K.E.C., Keith, D.A., McCarthy, M.A., and Hughes, L. (2011). Identifying and managing threatened invertebrates through assessment of coextinction risk. Conservation Biology 25: 787-796.

Moir, M.L., Vesk, P.A., Brennan, K.E.C., Poulin, R., Hughes, L., Keith, D.A., McCarthy, M.A., and Coates, D.J. (in press b). Considering coextinction of dependent species during ex situ conservation and assisted migrations of threatened hosts. Conservation Biology.

Vesk, P.A., McCarthy, M.A., and Moir, M.L. (2010). How many hosts? Modelling host breadth from field samples. Methods in Ecology and Evolution 1: 292-299.

Taylor, G.S., and Moir, M.L. (2009). In threat of co-extinction: two new species of Acizzia Heslop-Harrison (Hemiptera: Psyllidae) from vulnerable species of Acacia and Pultenaea. Zootaxa 2249: 20-32.

It begins:

“A SUPPOSEDLY scientific study into grazing in Victoria’s high country has been exposed as a shambolic exercise in finding cow poo.”

Read the article at: http://www.heraldsun.com.au/news/more-news/high-country-research-a-load-of-bull/story-fn7x8me2-1226236876741

But first, I’ll provide some context. The coalition took a policy to the state election in late 2010 that cattle grazing would be permitted in the Alpine National Park after the previous government stopped it for economic and environmental reasons. Rather than honouring this election promise directly when elected, the state government decided to return cattle to the park under the pretext of a scientific trial to assess the effects of cattle grazing on fire risk. Based on activities to date, the scientific content of this so-called trial is questionable.

Given the presence of federally-listed threatened species and communities at the grazing sites, the federal government determined that the return of cattle should be referred under the Environmental Protection and Biodiversity Conservation Act. That referral has now been submitted by the Victorian Department of Sustainability and Environment (DSE), with a plan to return cattle in late 2012. Under the EPBC Act, the federal minister is required to consider impacts on Matters of National Environmental Significance, which include federally-listed species and communities.

We make five main points about the referral:

1. The assessment of flora and fauna at the proposed sites is inadequate;
2. Despite inadequate information in the referral document, significant negative impacts of the cattle grazing seem likely;
3. The absence of a research plan means that details of actions at the proposed sites have not been provided or justified;
4. The currently incomplete environmental management plan appears to contain most of the information about how environmental impacts will be managed, but the plan is not provided with the referral documents so its adequacy cannot be assessed; and
5. There are scientific alternatives for assessing impacts of cattle grazing on fire risk that do not require introduction of cattle to the Alpine National Park; these should be investigated thoroughly prior to allowing possible impacts on Matters of National Environmental Significance.

These points mean that potential impacts of the cattle grazing on Matters of National Environmental Significance are likely to have been overlooked. Additionally, based on current scientific knowledge, documented threats to Matters of National Environmental Significance and the availability of potential alternatives to address the research questions, it is hard to see how the proposed activities can be approved. We expand on our points below.

1. The assessment of flora and fauna at the proposed sites is inadequate

The document “Fauna and Flora Desktop Assessment” on Matters of National Environmental Significance, which was an attachment to the referral, clearly states the inadequacy of the fauna and flora assessment (Section 2.7). It notes the data that were used are limited in extent, other data are likely to exist that could not be compiled in the limited time allowed by DSE, and that “where data is available, it is limited and patchy” (p. 14). According to significant impact guidelines 1.1 for the Environment Protection and Biodiversity Conservation Act 1999, when determining whether an action is likely to have a significant impact on a Matter of National Environmental Significance, the “self-assessment should be as objective as possible and based on sufficient information to make an informed judgement” (p. 3). The failure of a state government department to compile available information rigorously and collect further relevant and necessary information in a timely manner before submitting a referral under the EPBC Act should be questioned.

The level of rigour used to assess potential impacts on listed species and communities does not seem to meet the standards required of other referrals under the EPBC Act. At a minimum, the following should be required: field surveys for compiling species’ lists, targeted surveys for listed species, and rapid assessments to assess vegetation and to ground-truth mapped vegetation types.

The referral document implicitly acknowledges that impacts cannot be assessed properly until further flora and fauna surveys are undertaken. The document states that “the EMP and associated risk assessment will be updated following analysis of the survey results. Analysis of this information will assist in determining site-specific mitigation measures that are appropriate to effectively manage impacts from cattle grazing to the identified species and communities” (p. 12). If further relevant information can be compiled to help determine site-specific management, it should be provided as part of the referral so that the adequacy of any management can be viewed publicly and assessed by the federal minister.

The referral document titled “Matters of National Environmental Significance Risk Assessment” (Attachment 7 Part b) also lacks important information. The document does not describe the information that was used to conduct the risk assessment. Subjective risk assessments of this type are prone to bias (Burgman 2005), which is best reduced by clear and comprehensive descriptions of the factors that underpin the categorizations of risk: the data, other information and the basic logic that is used. Without such descriptions, it is impossible to evaluate the veracity and plausibility of the assessments.

The incomplete nature of the flora and fauna assessment means that impacts on Matters of National Environmental Significance are likely to be overlooked in the referral document.

2. Despite the inadequate information in the referral document, significant negative impacts of cattle grazing on Matters of National Environmental Significance seem likely

Despite the limited survey work, nationally-listed flora, fauna and ecological communities have been recorded in or adjacent to the proposed sites. Detrimental impacts of cattle grazing on some of these are known to occur through consumption and trampling (e.g., Attachment 7 Part B). One example of this is the Leafy Greenhood Pterostylis cucullata, which is likely to be present at all the sites (p. 39 of Attachment 5 Part A). The national recovery plan for this species, which was prepared by DSE (Duncan 2010), recognizes grazing and trampling by cattle as a major threat in Tasmania; cattle grazing is likely to be a significant impact wherever it co-occurs with the species.

The possibility of impacts of cattle grazing on the listed species does not seem to have been properly addressed in the referral documents; impacts of cattle grazing are mentioned for only two species in the Fauna and Flora Desktop Assessment (Austral Toadflax Thesium austral and Maroon Leek-orchid Prasophyllum frenchii). Even in these cases, the comments are very brief, and information on the need and ability to manage possible impacts is absent. Despite this cursory treatment of impacts of cattle grazing, the residual risks (Attachment 7 Part B) are assessed as high and very high for nationally-listed species and communities even in the presence of the unclear “management controls”. In some cases, these management controls were not predicted to reduce high and very high risks posed by the cattle grazing to numerous nationally-listed species and communities (see Attachment 7 Part B, Long-footed Potoroo on p. 1, Smoky Mouse on p. 2, Alpine She-oak Skink on p. 4, Alpine Tree Frog and Spotted Tree Frog on p. 5, Bog Willow-herb on p. 7, Bogong Eyebright on p. 8, Alpine Sphagnum Bogs and Associated Fens on p. 10).

Given the presence of nationally-listed species and communities coupled with poor knowledge of their precise locations, it seems likely that impacts will occur, though they might remain undetected. This is especially true for some of the fauna species that have very cryptic habits that make them difficult to detect when present, but that are likely to occur at the sites (e.g., some of the listed frog species and the Spot-tailed Quoll). The same issue is also likely to apply to some of the difficult to detect plant species, such as orchids.

The study sites comprise a mostly very remote area of almost 280 km². A comprehensive flora and fauna assessment of this entire area, or even a fraction of it where cattle are proposed to be placed, has not been undertaken. Therefore, the claim in the Fauna and Flora Desktop Assessment that impacts are unlikely to occur seems overly-confident given the limited ability to determine locations of many species over such a large area, and the likely difficulties of monitoring and controlling cattle movements. At a minimum, comprehensive evidence needs to be provided to support the claim that the proposed environmental management plan can avoid significant impacts on Matters of National Environmental Significance when:

• the precise location of nationally-listed species cannot be determined reliably;
• only “a sample of cattle in each research site will be fitted with tracking collars” (p 10, Matters of national environmental significance risk assessment);
• the planned location of cattle grazing has not been determined; and
• possible impacts of cattle grazing on most species has not been considered.

3. The absence of a research plan

The referral document claims that the cattle are part of a scientific trial. However, the plan for this research has not been provided with the referral document, and it appears to be still in preparation. The continued absence of a research plan one year after cattle were introduced to the Alpine National Park is difficult to reconcile with claims that these activities are part of a scientific trial.

Such a research plan would detail the activities that would occur at the research sites, including the exact locations of grazed and ungrazed areas, the location, type and condition of any fencing, which environmental attributes will be measured, and how measurements will be made, analysed and used. The plan should justify the need and validity of stratifying site selection by elevation and aspect. In the absence of these details, it is not possible to precisely assess the impact of the activities on Matters of National Environmental Significance, or the rationale for choosing the particular sites and activities.

The absence of a detailed research plan also means that the ability of the proposed activities to estimate the fire management benefits of cattle grazing cannot be assessed. Indeed, even the attributes of the environment that are supposedly at risk from fire have not been identified.

A study of impacts of cattle grazing in New South Wales (Professor Mark Adams’ HighFire project) has so far not been able to reliably estimate the effect of cattle grazing on fire, and concludes that “the combined effects of fire and grazing may take decades to be fully apparent” (http://www.bushfirecrc.com/projects/b6/highfire-project). Assessing the likelihood that research will yield useful results is a standard aspect of experimental design and could be undertaken with the available information. Such an assessment has not been conducted. Without detailed information about potential benefits, it is difficult for the federal minister to weigh them against the likely environmental impacts. However, given results to date from Professor Adams’ HighFire project, it is hard to see how the proposed activities will provide comprehensive data or rigorous conclusions about the effects of cattle grazing on fire risk, especially in the absence of a research plan that shows otherwise.

4. The environmental management plan is incomplete and is not provided

The referral document refers to a draft environmental management plan (EMP) that will detail how impacts on Matters of National Environmental Significance will be managed. This draft plan is, by definition, incomplete, but it is also not provided in the referral. Therefore, it is impossible to assess whether impacts on Matters of National Environmental Significance will be reliably identified by the EMP if such impacts occur, or whether any proposed management actions are likely to be sufficiently comprehensive to manage possible impacts. This is another example where the referral document seems incomplete.

5. Scientific alternatives to determine the effect of cattle grazing on fire risk have not been considered

A large body of national and international scientific literature already exists about fire behavior and its response to fuel load, fuel type and environmental conditions. On the basis of this information, detailed models of fire risk have been developed. One of these models (Phoenix RapidFire) is already used by DSE for strategic fire planning and tactical fire management. This model could be used to assess likely impacts of cattle grazing of different intensities on fire risk by examining how changes in vegetation would influence fire behavior in alpine environments. Such a model could be used to identify the possible impacts of cattle grazing on fire risk, without the need for field trials. It could also estimate the magnitude of changes in vegetation and fuel loads required to achieve useful changes in fire behavior. Such a study would help to determine the need for field trials of the effects of cattle grazing. This simple example illustrates that there are alternative scientific approaches for assessing impacts of cattle grazing on fire risk that do not have any potential impacts on Matters of National Environmental Significance. Alternatives such as this or other options that avoid at least some impacts on Matters of National Environmental Significance (e.g., locating the study outside the Alpine National Park) should be rigorously assessed and documented prior to considering the proposed activities. According to the referral document, no alternatives were considered.

In summary, the referral document seems incomplete, but the information provided suggests that significant impacts on Matters of National Environmental Significance are likely. Until alternatives are considered, Matters of National Environmental Significance are assessed properly, and the planned activities and management methods are described in detail, a thorough assessment of the impacts of the proposed activities is not possible. Further, the benefit of the proposed activities, as measured by the quality of information provided about the effect of cattle grazing on fire risk, cannot be assessed until the research plan is finalized and its ability to estimate the effect of cattle grazing is determined.

Both the negative environmental impacts and the potential positive benefits of the proposed activities have not been assessed properly in the referral document. Based on the availability of potential alternatives to address the research questions, and threats to nationally-listed species and communities, it is hard to see how the proposed cattle-grazing activities can approved under the EPBC Act.

Literature cited

Burgman, M. (2005). Risks and decisions for conservation and environmental management. Cambridge University Press, Cambridge.

Duncan, M. (2010). National Recovery Plan for the Leafy Greenhood Pterostylis cucullata. Department of Sustainability and Environment, Victoria.

Some environment groups seem concerned about these changes. Locking in timber supply on longer-term contracts is viewed by some as irresponsible, especially in the face of climate change; regeneration of the main timber species, mountain ash, might be compromised in hotter and drier climates. However, longer-term contracts need to be viewed from a perspective that considers current timber management in Victoria.

The current approach to timber management in Victoria has been to estimate timber growth and yield in the absence of fires, and then re-assess timber supply and contracts over time, and especially after any major fires. It is an approach that has been criticised for some time. Ignoring unplanned fires would seemingly lead to ever diminishing estimates of the sustainable timber yield.

It is possible that longer-term contracts will force a more strategic assessment of timber yields that explicitly accounts for losses from fire. It is possible to do some simple calculations to determine how wrong estimates of timber yield might be if fires are ignored. Here I show how.

The aim here is to determine what proportion of forest stands reach their rotation age prior to a fire – that is the age of the forest stand at which it is planned to be harvested for timber. For the sake of simple illustration, I will make some simplifying assumptions, but these could be relaxed to include more realism (at the cost of less generality).

The first assumption is that the annual probability of a stand-replacing fire is constant from year to year, regardless of the age of the stand. I’ll call the annual probability of fire p, and 1/p is the average time between fires. In Victorian mountain ash forest, p is approximately 0.01.

Then, the probability that the stand will escape fire for R years, the rotation age of the forest, is (1-p)^R. So, the probability that the stand will be burnt prior to the rotation age is 1 – (1-p)^R. For R = 80, and p = 0.01, the chance of the stand burning is 0.55; there is worse than a 50% chance that the stand will reach rotation age. Not accounting for timber from salvage harvesting, this simple calculation suggests that approximately half the timber predicted to occur in the absence of fire will not be available in the presence of unplanned fires.

The same calculation can be repeated for different rotation ages and fire probabilities. Ignoring fires will lead to consistent over-estimates of timber availability, even for short rotation ages and low annual fire risks:

Figure: The probability of a forest stand reaching rotation age as a function of the rotation age, and for three different annual probabilities of fire (0.005, 0.01, 0.02), which correspond to average fire intervals of 200, 100 and 50 years.

Hopefully, longer-term contracts will account for these uncertain losses so that greater certainty can indeed be provided to Victoria’s timber industry. But this is also a nice example of how simple models can help focus thoughts on the issues.

We have six new post-doc/research fellow jobs available in the Quantitative and Applied Ecology Group (qaeco.com). Information about the positions, including how to apply, is available at The University of Melbourne’s jobs page.

The research will develop and test structured approaches to conservation investment, environmental decision making and performance monitoring.

Email me if you would like a copy of my chapter (and I posted a copy here as part of #pdftribute), which describes how adaptive management can assist reintroductions. It has an interesting meta-analysis of the reintroduction survival of translocated mammals.

Fig. 8.2. Post-release survival of reintroduced carnivorous mammals plotted versus body mass, based on the data compiled in Jule et al. (2008). Post-release survival was measured over periods of around one year (6–18 months) after release. Open circles are estimates of survival for individuals raised in captivity for each project and closed circles are estimates for wild-born individuals. Uncertainty in each estimate is not shown for the sake of clarity. These estimates are compared with the annual survival for non-translocated wild adult carnivores from McCarthy et al. (2008), represented by the median (solid lines) and 95% prediction interval (dashed lines).

List of Chapters

1. Animal Translocations: What Are They and Why Do We Do Them?

Philip J. Seddon, W. Maartin Strauss and John Innes

2. A Tale of Two Islands: The Rescue and Recovery of Endemic Birds in New Zealand and Mauritius

Carl G. Jones and Don V. Merton

3. Selecting Suitable Habitats for Reintroductions: Variation, Change and the Role of Species Distribution Modelling

Patrick E. Osborne and Philip J. Seddon

4. The Theory and Practice of Catching, Holding, Moving and Releasing Animals

Kevin A. Parker, Molly J. Dickens, Rohan H. Clarke and Tim G. Lovegrove

5. Dispersal and Habitat Selection: Behavioural and Spatial Constraints for Animal Translocations

Pascaline Le Gouar, Jean-Baptiste Mihoub and Francois Sarrazin

6. Modelling Reintroduced Populations: The State of the Art and Future Directions

Doug P. Armstrong and Michelle H. Reynolds

7. Monitoring for Reintroductions

James D. Nichols and Doug P. Armstrong

8. Adaptive Management of Reintroduction

Michael A. McCarthy, Doug P. Armstrong and Michael C. Runge

9. Empirical Consideration of Parasites and Health in Reintroduction

John G. Ewen, Karina Acevedo-Whitehouse, Maurice R. Alley, Claudia Carraro, Anthony W. Sainsbury, Kirsty Swinnerton and Rosie Woodroffe

10. Methods of Disease Risk Analysis for Reintroduction Programmes

Anthony W. Sainsbury, Doug P. Armstrong and John G. Ewen

11. The Genetics of Reintroductions: Inbreeding and Genetic Drift

Lukas F. Keller, Iris Biebach, Steven R. Ewing and Paquita E.A. Hoeck

12. Genetic Consequences of Reintroductions and Insights from Population History

Jim J. Groombridge, Claire Raisin, Rachel Bristol and David S. Richardson

13. Managing Genetic Issues in Reintroduction Biology

Ian G. Jamieson and Robert C. Lacy

14. Summary

Philip J. Seddon, Doug P. Armstrong, Kevin A. Parker and John G. Ewen