Impact statement

A diversity of conservation activities is needed to avert the loss of species threatened with extinction. Prioritisation of investments can enhance the transparency and defensibility of resource allocation and can also inform the funding required to reverse decline of species.

Core components of a prioritisation

At its most basic, prioritisation means to arrange in order of priority (Mace et al., Reference Mace, Possingham, Leader-Williams, Macdonald and Service2007). Prioritisation is underpinned by decision science or operations research (Hemming et al., Reference Hemming, Camaclang, Adams, Burgman, Carbeck, Carwardine, Chades, Chalifour, Converse, Davidson, Garrard, Finn, Fleri, Huard, Mayfield, Madden, Naujokaitis-Lewis, Possingham, Rumpff, Runge, Stewart, Tulloch, Walshe and Martin2022), and seeks to enhance the extent to which decisions (i.e., choices between alternatives given a stated objective) are informed, transparent and defensible. In the field of biodiversity conservation, prioritisation typically informs the allocation of conservation resources (i.e., funding, effort, time and attention).

There are several approaches to species prioritisation. They differ according to whether they prioritise:

• • species themselves (Chen, Reference Chen2007; Liu et al., Reference Liu, Kenney, Breman and Cossu2019),

• • their habitats or populations (Nielsen and Kenchington, Reference Nielsen and Kenchington2001; Clarkson et al., Reference Clarkson, Marsh and Dowling2012; Strimas-Mackey and Brodie, Reference Strimas-Mackey and Brodie2018),

• • conservation activities targeting species (Joseph et al., Reference Joseph, Maloney and Possingham2009; Wilson et al., Reference Wilson, Carwardine and Possingham2009; Rose et al., Reference Rose, Heard, Chee and Wintle2016; Brazill-Boast et al., Reference Brazill-Boast, Williams, Rickwood, Partridge, Bywater, Cumbo, Shannon, Probert, Ravallion, Possingham and Maloney2018; Gillespie et al., Reference Gillespie, Roberts, Hunter, Hoskin, Alford, Heard, Hines, Lemckert, Newell and Scheele2020),

• • abatement or mitigation of particular threats more generally (Carwardine et al., Reference Carwardine, Martin, Firn, Reyes, Nicol, Reeson, Grantham, Stratford, Kehoe and Chades2019), or

• • protection of areas of land critical for species protection (Sinclair et al., Reference Sinclair, Milner-Gulland, Smith, McIntosh, Possingham, Vercammen and Knight2018; Leclerc et al., Reference Leclerc, Magneville and Bellard2022).

While the focus here is on species, it is also important to consider the actions required to abate the threats to species persistence, the locations where these actions must be implemented, and the most effective timing for implementation (Game et al., Reference Game, Kareiva and Possingham2013). The actions of interest will have a cost and likelihood of success, and consideration of these more pragmatic components of prioritising species has triggered debate around the use of the term “triage” in conservation (see below).

The basic conservation prioritisation problem has the following core components (Figure 1).

Figure 1. Core components of a prioritisation problem.

Constraints

These may include a budget envelope, or alternatively, the minimum amount of conservation or benefit that is sought. This minimum amount might be specified as a target. Targets can be tailored to account for life history characteristics and the habitat needs of a species, such as the minimum viable population size for a species needed for it to persist in the wild. Alternatively, targets can be generalised across species (e.g., protect 30% of the range of each species of interest).

A generalised version of the conservation prioritisation problem subject to a fixed budget, b_t, is described below. Let x_jkt be the amount of money to be spent on action k in location j in time period t. Each year the cost of all the actions across all the locations must be less than our overall budget, so we have the constraint

$$ \sum \limits_{j=1}^N\sum \limits_{k=1}^P{x}_{jkt}\le {b}_t,\hskip1em \mathrm{for}\ \mathrm{every}\ \mathrm{year}\;t, $$

where N is the number of locations and P is the number of possible actions.

The overall aim is to find a solution (or multiple strong-performing solutions) through manipulation of the control variables that have the highest possible value of the objective function subject to our constraints. Each location j has a cost c_j and each asset (which might include species) i has a target r_i. The variable x_j equals 1 if location j is selected for investment, otherwise it equals 0. The objective can be to either minimise the cost.

$$ \sum \limits_{j\in P}{c}_j{x}_j $$

subject to the constraint that the targets are met for each asset of interest.

The objective of the maximal-coverage problem is to maximise some measure of “benefit”, given a fixed budget. That is, the objective is to maximise

$$ \sum \limits_{i\in I}f\left({y}_i(x)\right).\hskip1em $$

Subject to the constraint

$$ \sum \limits_{j\in J}{c}_j{x}_j\le b.\hskip1em $$

A variety of algorithms can be used to prioritise investments in species (Hanson et al., Reference Hanson, Schuster, Strimas-Mackey and Bennett2019). These algorithms might generate optimal or near optimal solutions to a specified problem such as maximising coverage or minimising loss (Wilson et al., Reference Wilson, Carwardine and Possingham2009). Alternatively, scoring approaches can be used to rank alternative options according to specific criteria, such as cost-effectiveness or cost-utility (Wilson et al., Reference Wilson, Underwood, Morrison, Klausmeyer, Murdoch, Reyers, Wardell-Johnson, Marquet, Rundel, McBride, Pressey, Bode, Hoekstra, Andelman, Looker, Rondinini, Kareiva, Shaw and Possingham2007). Scoring approaches will provide inefficient solutions if they do not account for the complementarity between actions (Possingham et al., Reference Possingham, Wilson, Andelman, Vynne, Groom, Meffe and Carroll2006).

The key is to distinguish conservation problem formulation (the objective, control variables, etc.) as outlined above, the algorithms or criteria that can be used to find solutions to the problem (e.g., simulated annealing) and software packages (e.g., Marxan) that can package up all these components into an easy-to-use interface (Ball et al., Reference Ball, Possingham, Watts, Moilanen, Wilson and Possingham2009).

Box 1. Alliance for Zero Extinction.

Launched globally in 2005, the Alliance for Zero Extinction (AZE) focusses on sites for preventing global extinctions of species. AZE sites are often the last remaining refuges of one or more endangered or critically endangered species. The protection of such sites was aligned with the Convention on Biological Diversity’s Aichi Targets, namely Aichi Targets 11 and 12, with the overarching objective of improving the status of biodiversity by safeguarding ecosystems, species and genetic diversity. The AZE sites are also aligned with the draft Post-2020 Global Biodiversity Framework, specifically (CBD, 2021), milestone A.2 (the increase in the extinction rate is halted or reversed, and the extinction risk is reduced by at least 10%, with a decrease in the proportion of species that are threatened, and the abundance and distribution of populations of species are enhanced or at least maintained). There are over 850 AZE sites worldwide, and over half of these are at least partially protected. The alliance focusses on engaging with governments, institutions, community groups, and non-governmental organisations to improve and implement policy, deliver site conservation programs and progress research efforts to prevent extinctions. From the viewpoint of the alliance, the objective of species prioritisation is zero extinction (Wiedenfeld et al., Reference Wiedenfeld, Alberts, Angulo, Bennett, Byers, Contreras-MacBeath, Drummond, da Fonseca, Gascon, Harrison, Heard, Hochkirch, Konstant, Langhammer, Langrand, Launay, Lebbin, Lieberman, Long, Lu, Maunder, Mittermeier, Molur, al Mubarak, Parr, Ratsimbazafy, Rhodin, Rylands, Sanderson, Sechrest, Soorae, Supriatna, Upgren, Vie and Zhang2021).

Box 2. Threat interactions, co-extinction and other dependencies.

Species at risk of extinction are typically impacted by more than one threatening process. For example, in Australia at-risk fauna are impacted by multiple threatening processes (Allek et al., Reference Allek, Assis, Eiras, Amaral, Williams, Butt, Renwick, Bennett and Beyer2018). There are likely to be commonalities between required management actions to abate threats to species, and furthermore, the actions taken can interact; that is, the costs, benefits and feasibility of one action can change when another action is undertaken. Consider for example the control of invasive species. Actions taken to control an invasive species may release other pest species from competition or predation. In Australia, control of invasive rabbits alone may lead to intensified predation of native prey by foxes, whereas control of foxes alone may result in increased rabbit populations and competition with native herbivores. If these interactions are ignored, opportunities to enhance efficiency could be missed and targeted efforts could be compromised. Explicitly managing will likely alter decisions about what actions to invest in and where they should occur and has the potential to deliver increased investment efficiency (Auerbach et al., Reference Auerbach, Wilson, Tulloch, Rhodes, Hanson and Possingham2015; Figure 2). An extension of this approach is to assign the threats to species to particular industries or sectors in order to prioritise overarching sectoral improvements in management and ultimately accountability (Prugh et al., Reference Prugh, Sinclair, Hodges, Jacob and Wilcove2010).

Figure 2. Case study: Prioritisation accounting for threat management interactions. Adapted from Auerbach et al. (Reference Auerbach, Wilson, Tulloch, Rhodes, Hanson and Possingham2015).

Prioritising species versus assessing extinction risk

The process of prioritising species to prevent extinction differs from assessing extinction risk per se. Extinction risk is typically assessed using criteria such as species life history characteristics often within a population viability analysis framework (Liu et al., Reference Liu, Yang, Zanatta, Lopes-Lima, Bogan, Zieritz, Ouyang and Wu2020), responsibility for protection (Martin et al., Reference Martin, Cardoso, Arechavaleta, Borges, Faria, Abreu, Aguiar, Carvalho, Costa, Cunha, Fernandes, Gabriel, Jardim, Lobo, Martins, Oliveira, Rodrigues, Silva, Teixeira, Amorim, Homem, Martins, Martins and Mendonca2010; Kricsfalusy and Trevisan, Reference Kricsfalusy and Trevisan2014), taxonomic uniqueness (Chen, Reference Chen2007), rarity (Toledo et al., Reference Toledo, Becker, Haddad and Zamudio2014), management feasibility (Martin et al., Reference Martin, Cardoso, Arechavaleta, Borges, Faria, Abreu, Aguiar, Carvalho, Costa, Cunha, Fernandes, Gabriel, Jardim, Lobo, Martins, Oliveira, Rodrigues, Silva, Teixeira, Amorim, Homem, Martins, Martins and Mendonca2010), recovery potential (Di Marco et al., Reference Di Marco, Cardillo, Possingham, Wilson, Blomberg, Boitani and Rondinini2012), species distribution (Liu et al., Reference Liu, Kenney, Breman and Cossu2019), threat status or some combination of these criteria.

The threat status of a species is often determined by the International Union for the Conservation of Nature (IUCN) Red Listing criteria, which use quantitative rules to assign the risk of extinction (Mace et al., Reference Mace, Collar, Gaston, Hilton-Taylor, Akcakaya, Leader-Williams, Milner-Gulland and Stuart2008) and ecosystem collapse (Keith et al., Reference Keith, Rodríguez, Rodríguez-Clark, Nicholson, Aapala, Alonso, Asmussen, Bachman, Basset, Barrow, Benson, Bishop, Bonifacio, Brooks, Burgman, Comer, Comín, Essl, Faber-Langendoen, Fairweather, Holdaway, Jennings, Kingsford, Lester, Nally, McCarthy, Moat, Oliveira-Miranda, Pisanu, Poulin, Regan, Riecken, Spalding and Zambrano-Martínez2013). Building on these static lists, the Red List Index evaluates trends in biodiversity (Butchart et al., Reference Butchart, Resit Akçakaya, Chanson, Baillie, Collen, Quader, Turner, Amin, Stuart and Hilton-Taylor2007). Red Lists are supported within countries through legislation that seeks to protect threatened and endangered species, such as the Environment Protection and Biodiversity Conservation Act 1999 in Australia. Red Lists have singly been used to prioritise species, but they are not a prioritisation in and of themselves (Miller et al., Reference Miller, Rodriguez, Aniskowicz-Fowler, Bambaradeniya, Boles, Eaton, Gardenfors, Keller, Molur, Walker and Pollock2006) – they do not identify specific actions, or quantify what would be involved to shift the conservation status of a species or ecosystem (Collen et al., Reference Collen, Dulvy, Gaston, Gärdenfors, Keith, Punt, Regan, Böhm, Hedges, Seddon, Butchart, Hilton-Taylor, Hoffmann, Bachman and Akçakaya2016; Kyrkjeeide et al., Reference Kyrkjeeide, Pedersen, Evju, Magnussen, Mair, Bolam, McGowan, Vestergaard, Braa and Rusch2021). Red List assessments are also not comprehensive across all species or even taxonomic groups (Walsh et al., Reference Walsh, Watson, Bottrill, Joseph and Possingham2013; Tingley et al., Reference Tingley, Meiri and Chapple2016; Tapley et al., Reference Tapley, Michaels, Gumbs, Bohm, Luedtke, Pearce-Kelly and Rowley2018).

The “Red to Green” framework was developed to translate the Red List Index into prioritisations based on extinction risk (Akçakaya et al., Reference Akçakaya, Bennett, Brooks, Grace, Heath, Hedges, Hilton-Taylor, Hoffmann, Keith, Long, Mallon, Meijaard, Milner-Gulland, Rodrigues, Rodriguez, Stephenson, Stuart and Young2018; Kyrkjeeide et al., Reference Kyrkjeeide, Pedersen, Evju, Magnussen, Mair, Bolam, McGowan, Vestergaard, Braa and Rusch2021). This approach uses quantitative criteria of risk assessment (i.e., extinction risk for species, risk of ecosystem collapse for habitat) to develop measurable objectives and targets for species and habitats (e.g., the improvement in the Red List category to be achieved by a target year). This is then followed by the identification of conservation actions needed to reach the goals and quantification of the costs of these actions and other constraints. This approach is much closer to a species prioritisation, compared to the use of threat status alone.

Some have approached species prioritisation by combining evolutionary data with measures of extinction risk. The focus here is to prioritise threatened species that represent large amounts of phylogenetic/functional trait diversity using metrics such as evolutionary distinctiveness (ED; Faith, Reference Faith2008; Cadotte and Davies, Reference Cadotte and Davies2010; Gumbs et al., Reference Gumbs, Gray, Wearn and Owen2018). Metrics such as ED ideally require species-level phylogenies to calculate the individual contribution of each species to the total phylogenetic diversity of a clade (Isaac et al., Reference Isaac, Turvey, Collen, Waterman and Baillie2007). Such genetic data is not routinely available and species-level phylogenies are often incomplete, although imputation methods exist (Curnick et al., Reference Curnick, Head, Huang, Crabbe, Gollock, Hoeksema, Johnson, Jones, Koldewey, Obura, Rosen, Smith, Taylor, Turner, Wren and Redding2015; Gumbs et al., Reference Gumbs, Gray, Wearn and Owen2018; Weedop et al., Reference Weedop, Mooers, Tucker and Pearse2019) and scenario-based approaches to account for uncertainties in extinction probability values have also been applied (Billionnet, Reference Billionnet2017). While providing a more comprehensive assessment of the benefit of taking action to mitigate the threats to a species, this approach alone falls short of a comprehensive prioritisation if it does not have a specified objective or account for what needs to be done where and when in order to secure the persistence of the species of interest.

Triage controversy

Prioritising species also carries with it an ethical dilemma (Wilson and Law, Reference Wilson and Law2016), reflected in the controversial debate about the use of the term “triage” in conservation. Triage in the field of emergency medicine is a process of prioritisation under severely constrained resources, where the needs of a few, resource-intensive, critical cases are “sacrificed” so that resources can be distributed to a greater number of less critical cases. Controversy associated with the use of the term in conservation has been well summarised previously (Bottrill et al., Reference Bottrill, Joseph, Carwardine, Bode, Cook, Game, Grantham, Kark, Linke, McDonald-Madden, Pressey, Walker, Wilson and Possingham2009) and it is clear from the discussion that has ensued over the past decade that the fundamental concern about the use of the word “triage” in conservation relates to how the term is interpreted.

“Triage” in conservation has typically been interpreted as allowing some critically endangered species to go extinct to save others. Its enaction is assumed to reveal the species that have been relegated to extinction, with some arguing that the concept of triage should be avoided altogether in conservation so as not to preclude opportunities to expand the resources available for conservation (Wiedenfeld et al., Reference Wiedenfeld, Alberts, Angulo, Bennett, Byers, Contreras-MacBeath, Drummond, da Fonseca, Gascon, Harrison, Heard, Hochkirch, Konstant, Langhammer, Langrand, Launay, Lebbin, Lieberman, Long, Lu, Maunder, Mittermeier, Molur, al Mubarak, Parr, Ratsimbazafy, Rhodin, Rylands, Sanderson, Sechrest, Soorae, Supriatna, Upgren, Vie and Zhang2021), whether it be for protection or monitoring (Wheeler et al., Reference Wheeler, Berteaux, Furgal, Pariee, Yoccoz and Gremillet2016). Alternatively, we should view triage as a process that analyses the expected outcomes of investments in different species, which is then used to prioritise investment based on what can be achieved with different levels of investment or effort. Such an analysis can then be used to determine the investment needed to prevent most species from going extinct, instigating efforts to fill any gap, such as through conservation marketing and fundraising campaigns (see Box 3). But while triage can inform, and motivate, investment in conservation, ultimately the level of resources available to prevent extinctions is a socio-political decision.

Asset maps shows what we value, but are not necessarily priorities

The identification of priority places for species conservation has received significant attention in both the peer-reviewed and grey literature (Moilanen et al., Reference Moilanen, Wilson and Possingham2009). The focus here has been on mapping the distribution of biodiversity and its patterns, such as centres of species endemism (Orme et al., Reference Orme, Davies, Burgess, Eigenbrod, Pickup, Olson, Webster, Ding, Rasmussen, Ridgely, Stattersfield, Bennett, Blackburn, Gaston and Owens2005), uniqueness (Eken et al., Reference Eken, Bennun, Brooks, Darwall, Fishpool, Foster, Knox, Langhammer, Matiku, Radford, Salaman, Sechrest, Smith, Spector and Tordoff2004), biodiversity hotspots (Myers et al., Reference Myers, Mittermeier, Mittermeier, da Fonseca and Kent2000) and various measures of diversity (Grenyer et al., Reference Grenyer, Orme, Jackson, Thomas, Davies, Davies, Jones, Olson, Ridgely, Rasmussen, Ding, Bennett, Blackburn, Gaston, Gittleman and Owens2006; Brum et al., Reference Brum, Graham, Costa, Hedges, Penone, Radeloff, Rondinini, Loyola and Davidson2017). The production of these asset maps rests on the assumption that protecting these places from the threats to species persistence in these locations will result in the best outcomes for biodiversity conservation at large. While such maps may highlight the uneven distribution of biodiversity and areas that could yield returns from investment, since the identification of these places is not situated in a properly formulated problem (with an objective, constraints, etc., as outlined above) it is not possible to discern whether these places are indeed conservation priorities, nor the relative priority of one place compared to another.

Broader contextual considerations and challenges

Prioritising species to prevent extinctions needs to consider not only extinction risk, but a range of enabling conditions that influence the likelihood of success of conservation investments such as financial, cultural, logistical, ethical, human livelihoods and social factors (Miller et al., Reference Miller, Rodriguez, Aniskowicz-Fowler, Bambaradeniya, Boles, Eaton, Gardenfors, Keller, Molur, Walker and Pollock2006; Fitzpatrick et al., Reference Fitzpatrick, Murray, Paxton and Brown2007; Moir and Brennan, Reference Moir and Brennan2020). Furthermore, there are often conflicting objectives (Simmons et al., Reference Simmons, Nolte and McGowan2021), differing value judgements (Latombe et al., Reference Latombe, Lenzner, Schertler, Dullinger, Glaser, Jaric, Pauchard, Wilson and Essl2022), varying risk preferences and tolerances (Tulloch et al., Reference Tulloch, Maloney, Joseph, Bennett, Di Fonzo, Probert, O’Connor, Densem and Possingham2015) and sources of uncertainty. Given all these complexities, it is important to utilise a variety of theories and methods from the decision science toolbox (Hemming et al., Reference Hemming, Camaclang, Adams, Burgman, Carbeck, Carwardine, Chades, Chalifour, Converse, Davidson, Garrard, Finn, Fleri, Huard, Mayfield, Madden, Naujokaitis-Lewis, Possingham, Rumpff, Runge, Stewart, Tulloch, Walshe and Martin2022).

Knowledge constraints are often posed as an argument against prioritisation: to not prioritise species reduces the risk of misallocating effort. There is indeed uneven knowledge of threatened species, with a research bias towards larger bodied species and charismatic or economically valued species (Allek et al., Reference Allek, Assis, Eiras, Amaral, Williams, Butt, Renwick, Bennett and Beyer2018); within taxonomic groups, there is often a bias towards species occurring in developed nations (Buechley et al., Reference Buechley, Santangeli, Girardello, Neate-Clegg, Oleyar, McClure and Sekercioglu2019). Building on the pre-cautionary principle the use of the best available knowledge and prioritising further research as well as implementation of conservation activities is key. Importantly, the very act of determining the most appropriate conservation action (or suite of actions) might be considered as a control variable if these actions are unknown at the time of prioritisation (Game et al., Reference Game, Kareiva and Possingham2013; Raymond et al., Reference Raymond, Wen, Cooke and Bennett2018). Decision science methods based on the value-of-information theory to determine and appraise the relative value of further data collection versus managing species at risk of extinction to determine optimal strategy are also available (Bennett et al., Reference Bennett, Maxwell, Martin, Chades, Fahrig and Gilbert2018). Similarly, proactively progressing activities to prevent species from being at risk of extinction in the first place is critical (Walls, Reference Walls2018).

As knowledge improves or as values and perceptions change, priorities are likely to shift. Furthermore, funded projects may underperform or new projects may require investment when additional funding is not available immediately (Gerber, Reference Gerber2016). Reallocation of ongoing investments in response to shifting priorities (i.e., reprioritisation) would need to be balanced by the likely transaction costs, lost opportunities and the risks associated with not honouring the needs of ongoing project commitments (Wu et al., Reference Wu, Dodd, Hauser and McCarthy2021).