A General Framework for Setting Quantitative Population Objectives for Wildlife Conservation

Quantitative population objectives are necessary to successfully achieve conservation goals of secure or robust wildlife populations. However, existing methods for setting quantitative population objectives commonly require extensive species-specific population viability data, which are often unavailable, or are based on estimates of historical population sizes, which may no longer represent feasible objectives. Conservation practitioners require an alternative, science-based method for setting long-term quantitative population objectives. We reviewed conservation biology literature to develop a general conceptual framework that represents conservation biology principles and identifies key milestones a population would be expected to pass in the process of becoming a recovered or robust population. We then synthesized recent research to propose general hypotheses for the orders of magnitude at which most populations would be expected to reach each milestone. The framework is structured as a hierarchy of four population sizes, ranging from very small populations at increased risk of inbreeding depression and extirpation ( < 1,000 adults) to large populations with minimized risk of extirpation ( > 50,000 adults), along with additional modifiers describing steeply declining and resilient populations. We also discuss the temporal and geographic scales at which this framework should be applied. To illustrate the application of this framework to conservation planning, we outline our use of the framework to set long-term population objectives for a multi-species regional conservation plan, and discuss additional considerations in applying this framework to other systems. This general framework provides a transparent, science-based method by which conservation practitioners and stakeholders can agree on long-term population objectives of an appropriate magnitude, particularly when the alternative approaches are not feasible. With initial population objectives determined, long-term conservation planning and implementation can get underway, while further refinement of the objectives still remains possible as the population’s response to conservation effort is monitored and new data become available.


INTRODUCTION
Conservation objectives are the specific, measurable changes that are necessary to achieve broad and visionary conservation goals (CMP 2013), such as improving ecological integrity (LCC 2014); enhancing ecosystem function, services, or resilience (UNEP 2010); and achieving recovered, secure, or robust wildlife populations (Rich et al. 2004;NOAA 2012). Quantitative conservation objectives determine when conservation goals have been reached, and measure progress toward these goals (Nicholson and Possingham 2006). Thus, generating clear and scientifically defensible conservation objectives is a critical component of successful conservation planning and implementation (Margules and Pressey 2000;Villard and Jonsson 2009). However, the process of establishing objectives is not straightforward and is subject to debate (Tear et al. 2005;Wilhere 2008).
When the conservation goal is to achieve recovered, secure, or robust wildlife populations, conservation objectives often take the form of a target population size. To establish these population objectives, one recommended approach is to estimate the minimum viable population (MVP) size (Himes Boor 2014; Doak et al. 2015), defined as the smallest population size at which the population has a high probability of persisting for a desired length of time (Morris and Doak 2002). Yet accurate estimates of MVP require extensive long-term demographic data that are often not available (Wolf et al. 2015), and analysis of inadequate data can have misleading results (Chapman et al. 2001;Patterson and Murray 2008). Few recovery plans for at-risk species have actually used population viability analyses to set recovery criteria Neel et al. 2012), and there is growing recognition that successful species conservation requires population sizes well above the MVP (Soulé et al. 2003;Sanderson 2006;Redford et al. 2011).
A second approach is to set population objectives based on estimates of historical population sizes or trends (Sanderson 2006;McClenachan et al. 2012). For example, the North America Waterfowl Management Plan originally set an objective of returning North American waterfowl populations to the sizes observed in the 1970s (NAWMP 1986), and Partners in Flight set an objective of returning North American landbird watch list species to population sizes observed in the 1960s (Rich et al. 2004). These objectives are designed to reverse population declines, guiding conservation plans toward conditions at a time when populations were likely more robust. Though historical population sizes can provide important context for understanding the magnitude of population declines (McClenachan et al. 2012), the time-period selected is arbitrary. If historical estimates are too recent or population objectives are too modest, once they are achieved it may be difficult to maintain support for additional conservation efforts (Petrie et al. 2011). If historical estimates are instead from a more distant time-period, they may be impossible to achieve in landscapes that have changed significantly as a result of urbanization, climate change, or other factors (Marsh et al. 2005;Lawler 2009).
Much has been written about general considerations that should go into setting population objectives (Sanderson 2006;Redford et al. 2011;Westwood et al. 2014), but aside from using historical or MVP size estimates, there is little guidance for choosing appropriate numbers. Science-based guidelines for choosing appropriate population objectives are needed in cases when population viability analyses are not feasible, when historical population size estimates are unavailable or inappropriate, or when the long-term goal is beyond recovery from recent declines. For example, regional or continental conservation plans that seek to keep common species common require an efficient method of setting population objectives for multiple species. In the absence of such guidelines, uncertainty about how to proceed, and fear of failure, can lead to MARCH 2017 https://doi.org/10.15447/sfews.2017v15iss1art8 delays in taking conservation action (Meek et al. 2015), while objectives that are instead driven by political considerations are often far lower than what is necessary to achieve conservation goals (Svancara et al. 2005;Noss et al. 2012). An ability to efficiently set long-term population objectives that are "in the ballpark" of what is necessary to achieve secure, robust populations would allow conservation planning to get underway while still allowing refinement of the objectives as additional data become available.
We examined the literature on setting population objectives and the relevant underlying principles of conservation biology to develop a general population objective-setting framework that could be easily adapted to many taxa. We then make specific recommendations for population objectives and the spatial scale at which they should be established. To illustrate the application of this framework, we also describe the process we used to set population objectives for a suite of focal bird species, and discuss considerations in applying this framework to other systems.

CONCEPTUAL DEFINITION AND GENERALIZED FRAMEWORK
To develop a conceptual definition of a recovered or robust population, we examined recent conservation biology literature. Four features were commonly recommended: genetically robust, self-sustaining, ecologically functional (or effective), and resilient (Conner 1988;Soulé et al. 2003;Sanderson 2006;Redford et al. 2011;Wolf et al. 2015).
A genetically robust population is defined as sufficiently abundant that inbreeding and genetic drift are minimized, and with sufficient gene flow among sub-populations that genetic diversity is maintained, maximizing the capacity to adapt to changes in environmental conditions (Redford et al. 2011;Frankham et al. 2014). A self-sustaining population is both stable or growing in abundance and demographically viable, meaning it is sufficiently abundant that the risk of extinction or extirpation from demographic or environmental stochasticity is minimized (Traill et al. 2010;Redford et al. 2011). An ecologically functional population is sufficiently abundant that key ecological interactions and functions are maintained (Conner 1988;Soulé et al. 2003), such as seed dispersal, nutrient transport, or its role in the food web (Peery et al. 2003;Estes et al. 2010;Galetti et al. 2013). Finally, a resilient population is one that can recover from disturbance, and is an important component of climate change adaptation strategies (Folke et al. 2004;Lawler 2009). A resilient population should comprise multiple genetically robust and selfsustaining sub-populations, encompassing the concepts of redundancy and representation of the genetic diversity across the species' range (Redford et al. 2011;Wolf et al. 2015). A resilient population can recover from an environmental catastrophe in one part of the range, and allows for shifts in distribution, which minimizes the risk of extinction or extirpation across the range.
Based on these conservation biology principles, we developed a generalized population status framework that describes the milestones a population would be expected to pass in the process of becoming a robust population ( Table 1). The framework is structured as a hierarchy of four population sizes: 1. Very small populations at increased risk of inbreeding depression 2. Small populations with reduced risk of inbreeding depression but that are still vulnerable to longterm loss of genetic diversity and extinction or extirpation through demographic and environmental stochasticity 3. Viable populations with reduced vulnerability to extinction or extirpation, and 4. Large populations with minimized vulnerability to extinction or extirpation and increased ability to maintain key ecological interactions and functions.
We also identified two modifiers to these population status categories that reflect considerations beyond population size: • Steeply declining populations that are at increased risk of extinction or extirpation no matter their population size (Stanton 2014), and • Resilient populations with increased ability to recover from disturbances. 4 VOLUME 15, ISSUE 1, ARTICLE 8 taxa, and the well-known 50 / 500 rule has been used to identify the threshold at which populations are at increased risk of inbreeding depression (Mills et al. 2005). Based partly on observations of lab and domesticated animals, an effective population size of at least 50 breeding adults is estimated as necessary to avoid inbreeding depression in the short-term (Franklin 1980). Assuming a ratio of actual population size to effective population size of 10 : 1 (Frankham 1995), this corresponds to a total population size of 500. However, Frankham et al. (2014) have proposed a revision upward to an effective population size of 100 and a corresponding total population size of 1,000, in part because wild populations are likely to be more susceptible to inbreeding depression than domesticated or lab animals. Thus, we propose that population sizes of fewer than 1,000 be considered very small populations at increased risk of inbreeding depression and local extirpation.
Above this threshold for very small populations, small populations may still be vulnerable to loss of long-term genetic diversity and extirpation from demographic or environmental stochasticity. To maintain evolutionary potential over the long-term, the 50 / 500 rule also proposes that an effective population size of more than 500 breeding adults is needed (Franklin 1980). As above, Frankham et al. (2014) recommend revising this estimate upward to an effective population size of 1,000,

POPULATION SIZE RECOMMENDATIONS
For this general framework to be useful in setting population objectives, the criteria defining each population status must be quantified, and the current status of populations of interest must be estimated. Although a population objective can be set without knowledge of the current population size an estimate is necessary to measure progress toward achieving that population objective, and likely also to estimate the extent of conservation actions that will be required to achieve the objective. The growing number of methods for estimating abundance or relevant surrogates of abundance (e.g., Royle and Nichols 2003), along with the growing availability of large citizen science data sets (e.g., Sullivan et al. 2014; Barlow et al. 2015) and data repositories (Iliff et al. 2008;Michener et al. 2012), is making population size and trend estimation more feasible for many species. In contrast, speciesspecific genetics, demography, and population dynamics data that could help identify the sizes at which an individual population is expected to reach each population status are not available for most species, and the cost of obtaining such estimates may be prohibitive. Instead, we turned to the recent conservation biology literature to recommend numerical thresholds for each population status based on the best available information (Table 1).
For very small populations, inbreeding is known to have consistently deleterious effects on fitness across Table 1 Population status framework describing the milestones a population would pass in the recovery process, shown with proposed thresholds for the number of adult individuals in the population, additional modifiers, and numerical criteria

Population status Description
Proposed thresholds

Very small
Expected to be well below minimum viable population size (MVP), and at increased risk of inbreeding depression in the short term. < 1,000

Small
May be below MVP and vulnerable to extirpation through environmental and demographic stochasticity and long-term loss of genetic diversity. < 10,000

Viable
Expected to meet or exceed MVP, reducing vulnerability to environmental and demographic stochasticity and preserving genetic diversity. > 10,000

Large
Expected to be well above MVP, minimizing vulnerability to environmental and demographic stochasticity, preserving genetic diversity, and improving ability to maintain key ecological interactions and functions.

Steeply declining
Increased risk of extinction or extirpation until the causes of the decline are addressed, no matter the population size.

Resilient
Multiple viable or large populations to hedge against environmental catastrophes At least two viable populations (> 10,000)

MARCH 2017
https://doi.org/10.15447/sfews.2017v15iss1art8 corresponding to a total population size of 10,000. Similarly, to minimize the risk of local extirpation from demographic or environmental stochasticity, studies have estimated average cross-taxa MVPs of a comparable magnitude. Reed et al. (2003) estimated MVPs for 102 vertebrate species, primarily birds and mammals, and identified a mean MVP of 7,316 (+/-562) breeding adults to achieve a 99% probability of persisting 40 generations, with the frequency distribution peaking in the 4,900-8,100 range. Traill et al. (2007) conducted a meta-analysis of MVPs across 287 populations of 212 species, of which 47% were vertebrates, and identified a median MVP of 4,169 (95% CI: 3,577-5,129), with the frequency distribution peaking near 10,000 (Traill et al. 2007). Thus, from both a genetics and population dynamics standpoint, we propose that populations larger than 10,000 breeding adults are more likely to be above MVP and capable of persisting and maintaining genetic diversity over the long-term (i.e., more likely to be viable populations).
Generalizing MVP estimates across species has been criticized because the MVP for any individual population may vary widely with species and context, such as the variance in environmental conditions and the kinds of threats different species face (Brook et al. 2011;Flather et al. 2011). Yet, even among the critics of this approach, there is a general agreement that MVPs are likely to be in the thousands, not hundreds (Mills et al. 2005;Brook et al. 2011;Flather et al. 2011;Jamieson and Allendorf 2012). While we agree that there is no universal MVP threshold and that this remains an area of active research, we propose that 10,000 is a reasonable starting point for setting long-term population objectives intended to achieve long-term viability. This estimate is larger than the population size of 5,000 that has previously been recommended as a general rule of thumb, based roughly on median MVP estimates across taxa (Traill et al. 2010). However, given the finding that shortterm data sets may under-estimate MVPs (Reed et al. 2003), the broader peak in the frequency distribution of MVP estimates (Reed et al. 2003;Traill et al. 2007), and the potential consequences of underestimating the population size needed, we prefer to err on the side of over-estimating the population size necessary. Thus, we have recommended a population size on the high side of the median MVP estimates (Noss et al. 2012).
Although meeting the MVP size is necessary for long-term persistence, populations should be well above MVP to maximize the likelihood that they are able to maintain key ecological interactions and functions (Conner 1988;Soulé et al. 2003;Sanderson 2006;Redford et al. 2011). Functional extinctions of species occur when species are not abundant enough to be ecologically effective, which can lead to cascading effects on ecosystems (Pyare and Berger 2003;Ripple et al. 2013), including declines of dependent species (Anderson et al. 2011), changes in evolutionary trajectory (Galetti et al. 2013), and phase shifts in ecosystem state (Estes et al. 2010). The point at which a population becomes ecologically effective is not well-studied, but, for example, the population size of salmon necessary to maintain its role in landscape nutrient dynamics was ten times the MVP (Peery et al. 2003). Depending on the ecological function in question, it may be that achieving a minimum density is more important than a specific population size, such as the minimum density of otters required to maintain kelp forests (Estes et al. 2010). Thus, other than "well above MVP," there are no rules of thumb for ecological effectiveness yet available, and this, too, remains an area of active research.
In addition to providing for ecological function, populations should also be well above MVP to allow time for declines to be detected and reversed before a population falls below MVP. Here, we defined a steeply declining population based on the International Union for Conservation of Nature (IUCN) Red List criteria of an observed or projected decline of greater than 30% over 10 years or generation length, whichever is longer (Mace et al. 2008). Thus, a steeply declining population that starts at a population size of 15,000 could drop below 10,000 (i.e., the threshold for viable) before the steeply declining status is recognized. Indeed, Frankham et al. (2014) have suggested that declining populations of fewer than 20,000 adults should be considered vulnerable to extinction on the IUCN Red List. In lieu of better information about the population sizes or densities required to maintain key ecological functions, and in favor of over-estimating the population size necessary, we propose aiming for large populations of more than 50,000 individuals, well above the threshold for viable, and providing time to detect and intervene in even precipitous population declines.
Finally, we defined a resilient population as having at least two viable or large sub-populations, presumably with at least some connectivity among them. Having more than one genetically robust, self-sustaining, and ecologically functional population improves representation of the full array of genetic diversity across the species range, and builds in redundancy, which hedges against disturbances or disasters in one part of the range (Redford et al. 2011;Wolf et al. 2015). Thus, to achieve a goal of supporting a resilient, genetically robust, and self-sustaining population, our framework provides a starting point for defining science-based population objectives that would require more than one sub-population, each of which is at least viable (>10,000 individuals), preferably large (> 50,000), and stable or increasing (i.e., not steeply declining).

GEOGRAPHIC AND TEMPORAL SCALES
This framework can be flexibly applied to define population objectives for any geographic or temporal scale. However, the appropriate scale for any individual case will depend on both the conservation goals and the biology of the species of interest. As examples, we discuss three cases: (Case 1) a species with a single population of limited distribution, (Case 2) a species with several discrete populations, and (Case 3) a wide-ranging species with a continuous distribution throughout its range. In any of these cases, the largest scale on which this framework should be applied is the spatial extent of a single biological population ( Figure 1A), defined as a group of individuals of one species that is largely reproductively isolated such that changes in the population's size are driven primarily by births and deaths rather than movement (Berryman 2002). For example, for a conservation goal of achieving a genetically robust and self-sustaining population, a population objective of viable or large could be set for the entire population of the species in Case 1 or Case 3, or each of the individual isolated populations in Case 2; the sizes of each of the populations in Case 2 could not be summed to reach the combined status of viable if each one is reproductively isolated.
On the other hand, it may be desirable to apply the framework to smaller segments of a population, such as to meet a conservation goal of maintaining representation of genetic diversity across the population's range, preserving local adaptations, and improving resilience (Tear et al. 2005). To achieve this goal in any of the cases above, a population objective of viable or large could be set for each of several sub-populations to achieve a resilient population overall ( Figure 1B). As an example for Case 3, population objectives could be set on the scale of ecoregions (e.g., Olson et al. 2001), conservation regions (e.g., NABCI 2000), or joint ventures (e.g., NAWMP 1998). At a still smaller scale, the managers of any one of these individual conservation planning units could define a conservation goal of achieving multiple genetically robust and self-sustaining sub-populations within their own borders, and set population objectives of viable or large for each of several subunits to achieve a resilient population overall ( Figure 1C).
Because conservation plans are often developed on spatial scales that are smaller than the full extent of a biological population, managers may want to apply the framework to individual conservation planning units and sub-units. However, the minimum spatial scale that should be considered depends on the behavior and natural history of the species of interest. For example, estimates of population density or territory size can inform the minimum area required to achieve a viable or large population. Similarly, estimates of historical population sizes or habitat extent can help estimate the historical capacity of a conservation planning unit or subunit to support a viable or large population. Potential capacity could also be estimated, such as by estimating the historical habitat extent and subtracting the amount permanently lost to urbanization. If a single conservation planning unit or sub-unit is too small in area to support viable or large populations of the species of interest, it may be more reasonable to redefine the conservation goals to set population objectives for groups of adjacent planning units or sub-units. For species with particularly low densities or large home ranges, conservation groups may need to coordinate with each other on a common population objective over a large area (Sanderson et al. 2008). Whichever scale is selected, the population objectives can always be downscaled by assigning a percentage of the objective to individual planning units (Figure 1).
For migratory species, another consideration is how to set objectives for non-breeding populations.
If there were simple, direct relationships between breeding and non-breeding populations of the same individuals, managers in each area could coordinate on a common population objective to be able to support the population throughout the full annual cycle. More often, migration patterns are complex, with individuals from multiple breeding populations mixing during the non-breeding season and vice versa. In the case of individuals from a single breeding population dispersing to contribute to multiple non-breeding populations, the non-breeding population objective could represent just a fraction of the breeding population objective. However, just as for breeding populations, establishing or maintaining multiple viable or large non-breeding populations or sub-populations would improve resilience and ecological function during the non-breeding season, and hedge against the risk of environmental catastrophes in one part of the non-breeding season range. In the case of individuals from several separate breeding populations converging to form a single non-breeding population, the non-breeding population objective may need to be much higher to be able to support multiple self-sustaining breeding populations. Further, non-breeding population objectives may also need to reflect within-season movements (e.g., Cormier et al. 2013) and the ebb and flow of individuals during key periods, for example at important migration staging areas (Myers 1983).
The temporal scale on which this framework should be applied is flexible, but we intend population objectives of viable or large to represent long-term conservation objectives (e.g., 100 years). If the focus is instead on the short-term (e.g., 10 years), it can be difficult to imagine ever achieving viable or large populations of many species because of current political and economic feasibility, logistical difficulties, or potential controversy (Manning et al. 2006;Sanderson et al. 2008). Yet there are many examples of successful conservation projects that were once considered impossible (Svancara et al. 2005;Manning et al. 2006). Over the long term, all these circumstances can change, and thus current obstacles should not limit the long-term conservation vision (Sanderson et al. 2008). We recommend viewing population objectives as long-term conservation endpoints at which we expect to reach the goal of secure or robust populations, while shortterm milestones or checkpoints can break the longterm objectives down into stages and help progress to be evaluated (Manning et al. 2006;Villard and Jonsson 2009). For example, a short-term population objective may be to move a very small population to small, to stabilize a population that is steeply declining, or simply to maintain the status quo and limit the effects of any new threats.

APPLICATION TO A MULTI-SPECIES REGIONAL CONSERVATION PLAN
We used this population status framework to set long-term population objectives for a suite of breeding riparian bird species in the Central Valley of California, as part of the Central Valley Joint Venture (CVJV), a collaboration of 20 agencies and organizations (Dybala et al. 2017, this volume). Here, we illustrate how this framework can be applied to a multi-species regional conservation plan by briefly outlining the process we undertook to agree on initial long-term population objectives. We then recommend several additional factors to consider in applying this framework to other conservation plans.
First, we identified our long-term conservation goal: for the Central Valley to have riparian ecosystems capable of supporting genetically robust, selfsustaining, ecologically functional, and resilient wildlife populations. We identified a suite of breeding riparian landbird focal species that represented a range of life history characteristics, specific vegetation associations, and current population sizes, which we assumed would be good indicators of the condition of riparian ecosystems (Chase and Geupel 2005). We then examined the current status of each species within each of four planning regions. Although there is likely to be considerable connectivity among individuals in each planning region, we treated the individuals in each region as though they were separate populations. Synthesizing recent bird survey data across the Central Valley, we estimated the current density of each regional focal species population, and, based on estimates of the current extent of riparian vegetation in the Central Valley, we estimated the current size of each regional population. These ranged from very small to large, but over half were very small or small. In contrast, using estimates of the historical extent of riparian vegetation, we estimated that each region easily had the historical capacity to support viable or large populations of all focal species. Even after accounting for permanent losses of historical riparian vegetation to current and projected urbanization, we estimated that each region still had the potential to support viable or large populations of all focal species. Thus, we assumed that long-term objectives of viable or large populations were reasonable conservation endpoints for each region.
With this groundwork laid, we then defined longterm population objectives designed to meet our long-term conservation goal of riparian ecosystems capable of supporting resilient wildlife populations within the Central Valley, requiring multiple viable or large regional populations of each focal species. Therefore, we defined objectives that would meet the following criteria: 1) within each region, all focal species populations are viable (>10,000); 2) within each region, most (70%) focal species populations are large (> 50,000); and 3) across all regions, each focal species has at least one large (> 50,000) population ( Figure 2). We took this approach, rather than setting all large objectives, because our primary interest in these species was as indicators of ecosystem condition, and we recognized that some regions were likely better suited to some focal species than others. Thus, we did not expect to be able to achieve large populations of all focal species in all regions, but we aimed to ensure each species would have at least one large population within the Central Valley.
With these population objectives defined, conservation strategies and management plans for achieving the objectives can be developed, implemented, and evaluated (CMP 2013 we assumed that habitat availability is the primary limiting factor for most species. Thus, we also set riparian habitat objectives for each region, based on estimates of how much additional riparian vegetation would be required to meet the population objectives. These habitat objectives are useful estimates of the magnitude of riparian habitat restoration that will be needed, and of how much funding will need to be secured for those projects. Monitoring plans will be essential to measuring the actual contribution of restoration projects toward reaching the longterm population objectives, identifying the most successful practices, and informing any necessary revisions. When habitat quantity may not be the primary limiting factor, region-wide monitoring and/ or specific research effort will be needed to identify other conservation priorities, such as efforts to identify the causes of steeply declining populations.
In applying this framework to our system, we identified several considerations that might suggest revisions to the general population status framework when it is applied to other systems: 1. Beyond large, resilient populations, are there additional goals or stakeholder interests that should be included in the conservation planning process? For example, if one of the goals was to accommodate recreational hunting, the population objective might need to be substantially higher than the numbers proposed here (Sanderson 2006 3. Are any future changes anticipated, such as effects of climate change on the ecosystem or populations of interest? For example, the potential capacity of the planning unit to support a species of interest may change if the distribution of the species is expected to shift into or out of the area, or if sea level rise or changing weather patterns will limit potential habitat (Galbraith et al. 2002;Stralberg et al. 2009). In these cases, the conservation goals, the species of interest, and/or the geographical scale of the conservation plan may require adjustment to accommodate these projected shifts.
Answering these questions is not simple and will require decision-making despite many uncertainties. Thus, we recommend considering the longterm population objectives as hypotheses for the population sizes required to meet the conservation goals. As such, they should be reviewed regularly and refined to incorporate new information (Armstrong and Wittmer 2009).

Figure 2
Example of population objectives set for multiple species (1-6) in multiple planning regions (A-D) within a conservation planning unit. These objectives meet the criteria of viable populations for all species within each region, large populations for most species (>70%) within each region, and at least one large regional population for each species. The objectives allow for variation in the suitability of each region for each species. For example, species 3 is less suited to region B and naturally has lower densities there, making it harder to achieve a large population, whereas species 2 naturally has higher densities in region B. Species 1 has low densities everywhere, but it would be easiest to achieve a large population in region A with the most existing habitat. By achieving these objectives, each species would have more than one viable or large regional population, to reach the goal of resilient populations within the larger conservation planning unit.

DISCUSSION
Our framework provides transparent, science-based guidelines for setting population objectives designed to meet long-term conservation goals. The population size thresholds we have proposed represent general hypotheses for the orders of magnitude necessary for most populations to reach each population status. This framework is not a replacement for detailed species-or population-specific genetic, demographic, or population dynamics information, particularly for threatened and endangered species. Where these data are available or can be obtained, they should always be preferred. However, for most species and populations, these data are lacking and would be prohibitively time-and resource-intensive to obtain, especially for entire suites of focal species without special legal status. In these cases, our framework represents a useful set of initial hypotheses, allowing conservation plans to get underway while still allowing the objectives to be further tailored to individual situations and revised as additional information becomes available.
Even with this framework, the process of setting population objectives is not strictly scientific. The appropriate spatial scale at which the framework should be applied will depend on the ultimate conservation goals, the species of interest, and the historical or potential capacity of the conservation planning unit, while the population objectives may require adjustments to reflect additional stakeholder values or anticipated effects of climate change. However, the complexity of the problem makes these guidelines an invaluable starting point for setting long-term population objectives of an appropriate magnitude to achieve conservation goals of genetically robust, self-sustaining, ecologically functional, and resilient populations.