A Framework for Developing Management Goals for Species at Risk with Examples from Military Installations in the United States
A decision framework for setting management goals for species at risk is presented. Species at risk are those whose potential future rarity is of concern. Listing these species as threatened or endangered could potentially result in significant restrictions to activities in resource management areas in order to maintain those species. The decision framework, designed to foster proactive management, has nine steps: identify species at risk on and near the management area, describe available information and potential information gaps for each species, determine the potential distribution of species and their habitat, select metrics for describing species status, assess the status of local population or metapopulation, conduct threat assessment, set and prioritize management goals, develop species management plans, and develop criteria for ending special species management where possible. This framework will aid resource managers in setting management goals that minimally impact human activities while reducing the likelihood that species at risk will become rare in the near future. The management areas in many of the examples are United States (US) military installations, which are concerned about potential restrictions to military training capacity if species at risk become regulated under the US Endangered Species Act. The benefits of the proactive management set forth in this formal decision framework are that it is impartial, provides a clear procedure, calls for identification of causal relationships that may not be obvious, provides a way to target the most urgent needs, reduces costs, enhances public confidence, and, most importantly, decreases the chance of species becoming more rare.
KeywordsEndangered species Military installations Species at risk Causal analysis Threats Rare species Trend analysis Threatened species Recovery Recovery goals
Species in decline or of low viability are at risk of becoming more rare—or even extinct—in the future. The International Union for Conservation of Nature (IUCN) assesses the conservation status of species on a global scale in order to highlight taxa threatened with extinction and includes them on its Red List. The United States Fish and Wildlife Service (USFWS) and the National Oceanic and Atmospheric Administration (NOAA) Marine Fisheries Service list certain rare plant and animal species as federally threatened or endangered. Additional species that are in decline or at low population abundances may be listed by these agencies in the future and are informally termed species at risk (NatureServe 2004; USGS 1998).
Environmental management goals must be developed and implemented in order to increase the viability of species in decline. The USFWS and NOAA specifically set forth recovery goals for listed species. Species at risk also benefit from targeted management programs, and resource managers benefit from early interventions. Yet, a large fraction of conservation management actions are not based on a formal examination of evidence; decisions are often based on common sense, personal experience, and discussions with stakeholders (Sutherland 2006). In reality the list of potential causes of declines in species abundance or viability is long; a location- and species-specific analysis must be conducted to evaluate these causes and formulate unique solutions for each species in a location. Scientifically based frameworks for species management are needed.
Few scientific frameworks exist for identifying and selecting conservation management actions. Frameworks related to setting extinction risk designations are more common (e.g., IUCN 2001). The Conservation Measures Partnership, a group of nongovernmental organizations including the Nature Conservancy, has developed “Open Standards for the Practice of Conservation.” This framework is written to apply to broad conservation planning and includes planning phases (including threat identification but not recommendations for how to identify most probable threats), implementation of monitoring and other actions, adaptive management, and education (CMP 2007). US State Wildlife Action Plans must include information on the abundance and distribution of species, descriptions of their habitats, threats, proposed conservation plans, monitoring plans, etc., but an analytical framework has not been developed for generating these elements.
Our objective was to develop a framework for setting management goals for species at risk. The framework is based on methods that are consistent with natural resources management practices and use the best available scientific knowledge in conservation biology. The use of this framework might prevent petitions to federally list particular species or might support findings that listing is “not warranted,” based on existing data.
The framework guides the resource manager through a set of questions, including where is the species in decline (on-base, in the region of the installation, or elsewhere); what are the probable, primary threats to the species; and how should management goals be designed to address these threats? We anticipate that the most uncertain aspect of the framework is identifying the greatest threats to the local population. This step adapts elements of the USEPA’s Stressor Identification Guidance, which was developed to identify probable causes of biological impairments in aquatic ecosystems (USEPA 2000).
Examples presented in the framework address species at risk on military lands in the United States. Many rare species reside on military installations because of the presence of large areas of undeveloped land, ordnance exclusion zones, and early successional, disturbed areas that may be preferred habitat (Tazik and Martin 2002; Efroymson and others 2009). Furthermore, residential development and other unfavorable land-use changes and activities on adjacent, privately held lands may not provide appropriate habitat (e.g., Mann and others 1996). Examples of how federally listed species, e.g., the desert tortoise (Gopherus agassizii) and subspecies, e.g., Sonoran pronghorn (Antilocapra americana sonorienisis), have interfered with military training are given in GAO (2002) and Efroymson and others (2005). Proactive management of species at risk onsite or offsite before the taxa become federally listed is more cost-effective and less disruptive to military training and testing than waiting until species abundance or viability is reduced to a critical threshold (Department of the Army 2006).
Species at Risk
Species at risk have been formally identified in many international, national, and state contexts. Thus, unless or until these rankings are harmonized, resource managers should identify which species status scheme they have used. Candidate species, as defined in the Endangered Species Act (ESA) are plants and animals for which there is sufficient information on their status and threats to propose them as threatened or endangered but which are of lower priority for listing than other species (USFWS 2007a). Listing priorities are based on the magnitude of threat(s), immediacy of threat(s), and taxonomic uniqueness (e.g., subspecies have lower priority than full species) (USFWS 2007a).
The nonprofit conservation data center NatureServe (2004) defines species at risk as plant and animal species that have not been federally listed as threatened or endangered under the ESA but that have been designated candidates for federal listing or are considered critically imperiled or imperiled throughout their ranges. The US Army adopts the NatureServe list of species at risk, potentially adding species for which there is a concern about ESA listing in the foreseeable future (Department of the Army 2006).
NatureServe (2004) assigns conservation status ranks to species using seven factors: (1) total number and condition of occurrences of species and infraspecies taxa (varieties, subspecies, or populations), (2) population size, (3) extent of range and area occupied by species or infraspecies taxa, (4) short-term and long-term trends in factors 1–3, (5) threats; (6) environmental specificity, and (7) fragility. Ranks are estimated at three geographic scales: global, national, and subnational, with the species-at-risk status reflecting the global rank. NatureServe’s Rank 1 (critically imperiled) and Rank 2 (imperiled) are the organization’s highest conservation status categories, coinciding with their definition of species at risk, which encompasses both federally listed and unlisted species under the ESA. Here we focus on those species that are not currently federally listed as threatened or endangered. Global conservation rankings are sometimes correlated with national rankings (e.g., IUCN) and may be used to define species at risk. But differences in data availability and interpretation may lead to different rankings of species by different organizations (Mehlman and others 2004).
US states list “species of greatest conservation need” based on a variety of sources. Few states conduct their own threat or rarity analyses. For example, the Delaware Wildlife Action Plan considers federal legal status, Natural Heritage state and global species ranks, the Mid-Atlantic Bird Conservation Initiative, the Wildlife Species of Conservation Concern from the Northeast Endangered Species and Wildlife Diversity Technical Committee, and other known Delaware populations with species significance or sensitivity (see http://www.dnrec.state.de.us/nhp/information/06_Species.pdf).
Using the NatureServe definition approximately 523 species at risk, mostly vascular plants, are present on about 30% of US military installations, including 47 candidates for federal listing, 136 critically imperiled species, and 340 imperiled species (NatureServe 2004; Department of Army 2006). A large fraction of species at risk are present on installations in the southeastern US (NatureServe 2004; Gregory and others 2006). Species at risk are considered to be present on particular military installations if individual plants or animals have been observed to reside on or within two kilometers of a military installation, partly because of the low resolution of some spatial data and partly because of the potential for migration onto the installation (NatureServe 2004).
Framework for Developing Management Goals for Species at Risk
Step 1: List Species at Risk
The first step in the framework is to develop a list of species of concern that are on or near the management area (Fig. 1). The list may be developed from existing lists of rare species at the site or the integration of national and state lists such as the USFWS candidate species for listing, NatureServe Rank 1 and Rank 2 species, state Species of Greatest Conservation Need, and field reconnaissance. For US military lands, the list can be derived from the Army’s list of 54 high-priority species at risk, generated by NatureServe for all DoD installations (Department of the Army 2006).
Step 2: Describe Available Information and Potential Information Gaps for the Species
Natural resource managers should collect and review information about each species at risk in order to develop hypotheses regarding habitat-related and other probable threats to the local or regional population. Information from local observations, field notes, and the scientific literature can be used to make a quick assessment regarding the sufficiency of data (Fig. 1). Information of value includes species habitat requirements (core habitat, migration habitat, reproduction habitat, unsuitable habitat), food requirements, predators, competitors, and disease. Ideally, spatial information on the variables that determine suitable habitat would be available for the management area and the region. Information about life stages and reproductive parameters is also useful. This assessment will lead to a determination about whether additional field studies are needed or information concerning surrogate species should be considered. The latter is only appropriate if supported by evidence that habitat, threats, life history, etc. are similar (see Andelman and Fagan 2000 for a criticism of the surrogate species approach to conservation).
Step 3: Determine the Spatial Extent of Species at Risk and Its Habitat
The spatial distribution of each species and its potential habitat on and near the management area should be determined (Fig. 1). Potential habitat is the full set of locations that are suitable for the species under ideal conditions, as compared to the realized habitat (a subset of potential that is occupied under prevailing land use and management). Potential habitat includes areas that were historically occupied and are still available or could be restored, as well as other areas that could be made suitable for the species. If land-use change outside of a management area such as a military installation is unfavorable, the remaining potential habitat may be in the management area. Because animals and plants do not observe the boundaries of management areas such as military installations (Department of the Army 2006; Beaty and others 2003; Efroymson and others 2005), larger areas should be considered.
It is also important to identify whether the spatial extent of populations influenced by the resource manager (e.g., the military) spans across lands managed by different owners. Maintaining viable populations of species at risk requires cooperative management goals on both sides of management borders (NatureServe 2004), and the US Army recommends that individual installations work with other interested organizations to conserve species at risk (Department of the Army 2006). For example, clusters of the federally endangered red-cockaded woodpecker (Picoides borealis) that are located off-base but demographically connected to on-base populations may be included in counts toward US Army Regional Recovery Goals (Beaty and others 2003) if the entire area is under federal control (e.g., via legal agreements with other landowners).
Numerous methods are available to identify potential habitat for species at risk. Most involve monitoring organisms; using geographic information systems (GIS) overlay techniques to discover correlations between organism locations and their biological, chemical and physical characteristics and to infer key habitat suitability variables; modeling of potential habitat using those variables; and verifying habitat suitability results. Example quantitative methods include binomial logistic regression to estimate the probability of gopher tortoise burrows at locations in the region of Fort Benning, GA (Baskaran and others 2006), several methods for predicting vegetation distributions that incorporate spatial dependence (Miller and others 2007), and Bayesian belief network models of habitat relationships and multiple stressors to predict the presence of 12 rare species in northwest forests (Marcot 2006). Many additional species distribution models are reviewed in Pearson (2007) and Austin (2007). Coarse habitat models are available through state and regional Gap Analysis Projects. If local monitoring information for the species of interest is not available, habitat boundaries may be estimated using habitat suitability information in the literature and appropriate GIS data. Regardless of the approach, ecologists familiar with the local landscape should approve any maps of habitat extent. Users of this framework should be wary of results from informal, unscientific surveys of stakeholders regarding the location of habitat.
Quantitative tools can be used to establish if the species exists in a metapopulation within the management area. Genetic surveys can be useful for identifying metapopulation structure via genetic distances among individuals from distinct populations. Understanding the spatial structure of metapopulations involves two types of models: (1) habitat models that relate species presence to habitat attributes, and (2) metapopulation models that represent the spatial connections among populations. Metapopulation models are useful for population viability analysis (PVA) when the species tends to form subpopulations with weak linkages (e.g., see Stevens and Baguette 2008). These models can evaluate the viability of species with different spatial arrangements of habitat and help resource managers make reasonable decisions about which parcels of land to protect from threats. For metapopulations, it is important to conserve some parcels of suitable habitat that are currently unoccupied (McElhany and others 2000) and to protect corridors that link subpopulations (Hargrove and others 2005). RAMAS Metapop is an example of a commercially available, generic metapopulation model that has been used for PVAs for plant, invertebrate, and vertebrate species (Applied Biomathematics 2003).
Step 4: Select Metric(s) for Assessing Species Status and Trends
Management progress can be tracked with consistent metrics that allow species at risk to be monitored through time. Several criteria for selecting good ecological indicators are pertinent. These include easy measurement, sensitivity to threats, predictable response to threats, prediction of responses that can be averted by management actions, and low variability in response (Dale and others 2004).
First, the natural resources manager should identify which criteria were used to designate the species at risk by the natural resources managers in Step 1. Knowing whether a designation was based on a downward trend in abundance, low numbers, threats, environmental specificity, fragility or other measures of population viability is key to selecting effective metrics of species status and trends. The rationale for the species-at-risk classification may point to an appropriate metric. For the example of military lands in the US, status of population size and extent of range and area occupied by the species or taxon [factors 2 and 3 from NatureServe (2004) described above] are criteria used to designate species at risk. However, these metrics may be measured at a spatial scale that is inappropriate for decisions at the local scale. If the rationale for the designation involved environmental specificity or fragility, abundance and area occupied might still be useful metrics of species status.
The appropriate metric for assessing status and trend should account for historical ranges of abundance of the species in question. In some cases, listed species have historically been rare. For example, the pallid sturgeon (Scaphirhynchus albus) has likely always been rare throughout the Mississippi River drainage (Kallemeyn 1983).
Metrics relating to stable, declining, or growing populations, such as minimum population size (over several years), age structure, productivity (e.g., fledging rate), female population size, population growth rate, and probability of extirpation by a specified future time might be appropriate for use in this framework. Metrics may be selected with potential management goals in mind (see Step 7 of the framework).
The result of this step is the selection of one or more metrics that relate to population abundance or production and trends. The selected metrics will be used in Step 5 to determine the status and trend of the population or metapopulation of interest and possibly in Step 7 to determine management goals.
Step 5: Assess Status of Local Population or Metapopulation
The fifth step of the framework is to assess the status and trend of the species at risk on and near the areas discovered in Step 3 using metrics from Step 4. Is this species in decline or at low numbers on or near the installation or elsewhere, or is it stable at low numbers across its range? (Fig. 1) Status may be determined based on a combination of monitoring and population modeling.
Because trends in population size and extent of range may be criteria for identifying species at risk (NatureServe 2004), resource managers should determine if existing data were already examined for trends. The resource manager should ask where and when monitoring of the subject population has occurred on and off the management area (e.g., military installation) and relate sample procedures and times to the life cycle of the species. The level of confidence in population status is related to life history traits and to possible sampling difficulty. For example the southern hog-nosed snake (Heterodon simus) is highly fossorial and cryptic even when above-ground (Tuberville and others 2000).
Tools are available to estimate abundance, production, and trends in these variables. The required data to estimate abundance can come from mark-and-recapture monitoring, transect surveys, bird song surveys, and surveys of breeding grounds or hibernacula of hibernating mammals. Methods are available and others are under development for natural resource managers to determine the appropriate monitoring effort and length of time required to detect population decline, stability, or growth. Researchers have previously created rough indices of population trend (e.g., Althoff and others 2004 at Fort Riley, Kansas, USA), and these may be adequate for making decisions about whether a species needs management in a particular location. For example, a military installation Natural Resources office can collect consistent data over time to establish trends in habitat suitability and population densities of targeted species.
Trend analysis can be used to indicate which life history parameters should be monitored to achieve the greatest statistical power. For example, Boulanger and others (2000) found that existing monitoring techniques were only capable of detecting reductions greater than 50% for marbled murrelets (Brachyramphus marmoratus), whereas monitoring of adult survival could detect a smaller (20%) decrease. Uncertainty factors for species that are known to be difficult to detect would be helpful for resource managers. Information on status and trend is used to make an initial judgment about whether the species needs local management. If the local population is stable or growing, the user of the framework does not need to conduct the threat assessment in Step 6 or set management goals in Step 7. However, the user may still consider managing less stable populations off-site to reduce the risk of listing.
Step 6: Conduct Threat Assessment
Whether or not the species-at-risk designation was precipitated by particular threats, management goals for species at risk will not be viable unless threats are well understood. For example, a management goal that is based on an assumption about insufficient habitat area at the range-wide scale will not be effective if the primary threat at the local scale is a toxicant. But identifying the role of specific threats can be challenging. For example, the prime cause of rarity of the endangered humpback chub (Gila cypha) is still debated as to whether it is attributable to voracious introduced predators, lack of sediments in the dammed river, or changes in water temperature (Coggins and others 2006; Converse and others 1998; Marsh and Douglas 1997; Paukert and Petersen 2007; Stone and others 2007). Similarly, the limited monitoring data available for southern hog-nosed snake, a species at risk that is present on military installations (NatureServe 2004), are insufficient to identify primary threats (Tuberville and others 2000). The threat assessment step is comprised of three sub-steps: listing candidate threats; evaluating the relative magnitude, probability, and immediacy of these potential threats; and determining susceptible life stages of the species.
Step 6a: Identify Potential Threats
A natural resources manager should compile a list of location-specific threats as part of this step, even if general threats are well known from species recovery plans (if species are listed in one area and at risk in another), State Wildlife Action Plans, NatureServe databases, or other scientific literature or were listed in Step 2. For example, toxic levels of a chemical may be present at a particular location, and it is useful to list several potential sources of contamination that might be the object of management actions.
Hoekstra and others (2002) describe threats from USFWS recovery plans, including construction (development), agriculture, resource use, water diversion, pollution, exotic species, species interactions, habitat dynamics, and other factors such as inbreeding and climate change. These are also candidate threats for species at risk.
Threats that were part of the rationale for the species-at-risk designation are typically candidate threats at particular locations. For example, Appendix 3 to NatureServe (2004) summarizes information about threats to some species at risk found at military installations, as well as providing a qualitative estimate of threat severity, immediacy of threat, and scope of threat for a minority of species at risk. The NatureServe database does not specify threats to local populations such as site-specific toxicants.
Habitat loss, which is known to be one of the most important threats to listed species (Wilcove and others 1998), is a potential threat for most species at risk; species with narrow habitat specificity and restricted geographic range are most susceptible (Rabinowitz 1981; also see Step 6c where susceptible life history traits are identified). Habitat loss may lead to local extinction at low population densities (“Allee effect”) because of the inability to find mates or breeding territories (Allee 1938). Ninety-two percent of species in randomly sampled recovery plans face habitat-related threats, including habitat destruction, degradation, or fragmentation (Campbell and others 2002). Thirty-four listed and one candidate species on or adjacent to 32 US western, arid military installations are threatened by habitat loss and degradation from various sources of mostly off-base land-use change and other stressors (Tazik and Martin 2002).
If habitat-related threats are suspected to be the primary threats, resource managers may need to determine if these relate to loss of forage area, loss of shelter, loss of breeding grounds, fragmentation of migratory corridors, etc., and to what extent these losses are reversible. Moreover, resource managers need to know the cause of the habitat loss. Activities that result in changes to soil or vegetation components of wildlife habitat such as controlled burns (Dale and others 2002; Quist and others 2003; Demarais and others 1999) may adversely affect or benefit species of concern (Rotstein 2003; Andow and others 1994). Habitat-related threats may be less important for some populations residing on large tracts of land such as military installations that support extensive natural areas than at other locations. Development and other disturbance outside of more pristine environments can cause rare species and their habitats to concentrate within these areas (NatureServe 2004). The drivers of off-site land-cover change are important to identify.
A list of candidate threats to populations at risk residing on or near military installations may add military activities and stressors to the typical list of species-specific and habitat-related threats. These include firing of munitions, foot traffic from soldiers, maneuvering of tracked vehicles, clearing of vegetation to maintain targets, fog oil from obscurants, aircraft overflights, and land-use change associated with base closure and realignment. (See Andersen and others 2004, for a fuller description.) More-specific stressors that may be associated with these activities include noise, fire, heat, water- and wind-eroded soils, compaction of soils, floods, forest management, pollution, and encroachment of invasive vegetation (Efroymson and others 2005, 2009). Military stressors are only credible, potential threats to species at risk if populations on military lands are in decline.
For some species, especially those that require disturbed habitat that has not attained a mature successional state, military activities help maintain the required habitat (Warren and others 2007). Therefore the analysis of threats to species at risk in the longleaf pine ecosystem should evaluate the loss of habitat due to the suppression of disturbances such as fire (van Lear and others 2005).
Noise is a potential stressor that overlaps spatially with contamination from explosives and clearing of range vegetation. Wildlife avoidance of blast or aircraft noise should be incorporated into habitat models if there is reason to believe this threat is significant. Noise was suspected to have significant adverse impacts on threatened bird species on Army installations, but after careful analyses, artillery noises and related military activities did not affect red-cockaded woodpecker reproductive success (Fort Bragg, Fort Stewart, and Fort Polk) (Delaney and others 2000; Doresky and others 2001) or population abundance of golden-cheeked warbler (Dendroica chrysoparia) (Fort Hood) (Anders and Dearborn 2004). Black bears (Ursus americanus) did not respond physically to weapons training exercises, except for small areas near firing positions (Telesco and van Manen 2006).
Many disturbances may quickly be excluded as potential threats to certain species at risk. Sources of disturbance on military installations are beneficial to particular species. For example, the Karner blue butterfly (Lycaecides melissa samuelis) occurs in habitat such as military impact areas where its disturbance-dependent food, wild lupine, is found (Smith and others 2002; Andow and others 1994), and amphibians prefer pools with high levels of ground disturbance caused by military activities (Warren and Buttner 2008).
Step 6b: Identify Threats of Highest Priority
Ranking threats is an important precursor to setting management priorities for species at risk. This recommendation has surfaced in many reviews and analyses of recovery goals for US federally listed species (Carroll and others 1996; Crouse and others 2002; Schultz and Gerber 2002; Lawler and others 2002). Step 6b of the framework is an analysis to identify the most significant threats to species at risk on or near the management area of interest (Fig. 1), considering magnitude of current threats, the immediacy of future threats, the spatial scale of threats, or other metrics. Hypotheses regarding greatest threats may be generated by military conservation staff and other ecologists or naturalists.
Significant threats to listed species are rarely obvious, nor are they identified using formal logic. Although some threats are easy to identify (as when species have narrow forage or shelter requirements), others are more complex. For example, factors causing the decline of the federally endangered Indiana bat (Myotis sodalis) may include climate change, pesticides, stream channelization and bank modification, disturbance of hibernacula, forest clearing, and agricultural development to a differing extent in different regions (USFWS 2007b; Sternberg and others 1998; Menzel and others 2005). NatureServe (2004) reviewed some of the scientific literature to summarize many of the known threats to species at risk.
Causal analysis or threat identification is an important component of any scientific process used to derive management goals for potentially imperiled species. Environmental causal analysis frameworks previously have been developed to determine causes of aquatic impairments (USEPA 2000), conservation priorities (CMP 2007), and more specific analyses such as determining most probable causes of coral reef decline (Fabricius and De’Ath 2004). The USEPA’s Stressor Identification Guidance Document (USEPA 2000) and related papers (Suter and others 2002; Cormier and others 2003; Cormier and others 2002; Norton and others 2002) provide principles for (1) determining causality of impairments in aquatic ecosystems and (2) supporting conclusions with evidence. Several of these factors are derived from the epidemiological principles of Hill (1965).
Step 6 of this framework builds upon the components of the USEPA causal analysis framework and involves: listing of candidate causes (threats) of population decline or low viability (Step 6a); organization and analysis of data with respect to associations that could support or refute the proposed causal relationships; and characterizing probable threats (USEPA 2000).
Organization and Analysis of Data. Data potentially useful for ranking threats may come from the resource management area itself, other areas within the range of the population of interest, or literature about the species of interest or related species. Three principal types of evidence include associations between measures of potential threats and population metrics at the management area, specific population effects elsewhere, or results of a mechanistic chain of events, and manipulations of threats to enhance or mitigate effects (USEPA 2000; Suter and others 2002). Unfortunately, existing data are sometimes not sufficient to support threat analysis. For example, a program in which species at risk were monitored at Camp Blanding, FL, was not designed to determine causality (Gregory and others 2006).
Characterization of Probable Threats. Three methods for characterizing causes (threats) from the Stressor Identification Guidance Document (USEPA 2000) are recommended. The first is that by eliminating alternatives, the primary threat will be identified. For example, one might eliminate a range of frequencies of sound as potential threats to Indiana bat (Myotis sodalis) given that they roost near airports, roads, and hog farms. The second is diagnosis—linking particular “symptoms” to particular causes or threats, whereby most evidence is associated with a causal mechanism (USEPA 2000). Although mechanistic relationships are important for establishing chemical contaminants as causes of aquatic impairments, they are probably less important for diagnosing habitat-related threats. An example of diagnosis might be linking crushed plants to tracked vehicles or soldier foot traffic.
Spatial co-occurrence. Is the population declining or of low viability in the same area where the putative threat occurs? If not, does the species migrate or have a long dispersal distance? Is the population stable where the threat is absent? The scale of evaluation may be important for determining if a threat is identified or if it goes unnoticed (Groves 2003). Many military installations have good geographic data for land cover and locations of military activities. However, data on the spatial distribution of organisms are sometimes chance observations, which typically don’t coincide with stressors from military activities, such as noise from explosions.
Temporal relationships. Has the presence of the putative threat preceded the species decline or other measure of low population viability? Is the population stable when the threat is absent? Temporal associations between management actions (e.g., burn frequency) and population decline or growth should also be considered. Some of these relationships may be obtained from the literature.
Biological gradient. Is the species decline or low population density more apparent in areas where the candidate threat is present in greatest area or intensity (e.g, high loss of habitat area or fragmentation)?
Complete exposure pathway. Does the candidate threat reach individuals? This question is important for threats that move or attenuate with distance from the source (e.g., sediments, chemical contaminants, noise).
Consistency of association. Are the candidate threat and declining species abundance or low viability observed together in different places and times across the species range or across the part of the range where a less widespread potential threat is present (e.g., military activities)? For example, declines in amphibian populations are geographically widespread and likely caused by a geographically widespread threat, be it ozone depletion or a new pathogen (Blaustein and Bancroft 2007). Likewise, declines in neotropical migrants likely stem from losses of habitat shared among species migrating to and from the same areas (Holmes 2007).
Experiment. Have controlled field manipulations demonstrated a relationship between the candidate threat and species decline or low viability? Experiments are conducted more often with surrogates than with rare species, although experiments are possible when one life stage is abundant (e.g., Trimble and others 2009) or when a potential threat can be manipulated (e.g., Whitehead and others 2008).
Plausibility. Is there a reasonable mechanism to explain the relationship between the putative threat and species decline?
Analogy. Is the hypothesized relationship between the putative threat and species decline similar to other, well-established cases involving related stressors and species? For example, sedimentation is known to be a significant cause of declines among mussels, which are experiencing higher rates of extinction than any other aquatic or terrestrial faunal group (Poole and Downing 2004). Consequently, this factor should be at the top of the list when considering threats to riverine mussel species.
Specificity of cause. Is the candidate threat consistently associated with specific measures or symptoms of species decline? Whereas effects of disturbances such as habitat loss are not easily traced to specific causes, some toxicants do cause lesions that are fairly specific. This line of evidence requires previous observations relating the potential threat to decline of the species. Therefore, it would not be expected to be useful for evaluating declines of relatively unstudied species, such as many of the 24 species at risk that are each restricted to one of twelve military installations (NatureServe 2004). All of these are flowering plants or soil or stream invertebrates, except for the white sands woodrat (Neotoma micropus leucophaea).
Predictive performance. Have local declines or low abundance of related species been observed that might be predicted by exposure to the putative threat?
Consistency of evidence. Is the hypothesized relationship between the putative threat and species decline or low viability consistent with all available evidence?
Coherence of evidence. Can apparent inconsistencies among lines of evidence be explained mechanistically? For example, an apparent conflict between a) fish lesions that are usually associated with toxicity of a particular chemical and b) non-toxic levels measured in the stream could be resolved if exposures are only episodic.
Step 6c: Identify Life History Traits and Life Stages Most Susceptible to Threat
Many life history traits (e.g., high trophic level, monogamy, social breeding, territoriality, narrow habitat preference, restricted geographic range, low mobility, large home range) make species more susceptible to extinction risk (Purvis and others 2000; Crooks 2002; Jager and Efroymson 2004; Jones and others 2003) and point to potentially effective management strategies. Susceptibility is a combination of exposure and sensitivity. For example, the importance of family structure in the demographics of red-cockaded woodpecker led biologists to understand the importance of managing nest cavities (Walters and others 2002). Providing artificial nest cavities allowed non-breeding, “helper” sub-adults to form their own colonies nearby, thereby increasing populations faster than would otherwise be possible. Other life history traits can influence the likelihood of management success, e.g., late age of maturity, biennial reproduction, dependence on pollinators, vector-dependent dispersal, intermittent juvenile recruitment, and rigid behavioral patterns (National Recovery Working Group 2005). Moreover, long-term monitoring can reveal species or genera that require environmental cues for life-cycle events (e.g., amphibians that may require minimum precipitation for emergence; Gregory and others 2006). Information on susceptibility also focuses attention on particular life stages that require management. However, many species become rare because their life histories are not well understood. The implementation of this step in the framework may require more studies of the species.
Quantitative Tools. Recent assessments of the use of science in endangered species management advocate increased use of PVA, because it can reveal the life stages or demographic processes that should be the focus of monitoring or management (Crouse and others 2002; NRC 1995; Carroll and others 1996; Morris and others 2002). Elasticity analysis with a PVA model can be used to identify the life stage with the greatest potential for influencing population recovery (e.g., Velez-Espino and others 2006). Note, however, that elasticities are reported as the response to a percentage change in survival, and that survival tends to be much higher for older life stages than for younger ones. It is important to incorporate knowledge of the actual scope for improving survival of different life stages via changes in management. Because juvenile survival has a larger scope for improvement, management actions focused on increasing juvenile survival are capable of producing larger responses in population size. Population modeling coupled with demographic sensitivity analysis (Caswell 2001) can allow resource managers to identify sensitive life stages that are not obvious from monitoring in the field. Researchers are using population and individual-based models to determine which parameters have the largest influence on population growth of listed species found at military installations (Grimm and Railsback 2005), and these findings may be useful for those managing species at risk.
Step 7: Set Management Goals
This step of the framework identifies potential management goals for species at risk in the management area. These goals should be defined with respect to a relevant spatial scale. For example, management goals that are intended to be implemented on a military installation will be different from those that are intended to address species declines at broader scales. If the local population of a species at risk is stable, according to Step 5, and not threatened by local stressors such as military activities, according to Step 6, it may require no special management (Fig. 1). If primary threats to a local population operate outside of the management area (e.g., decline in habitat quality for gopher tortoise due to agriculture; Baskaran and others 2006), then management goals should involve neighboring lands. Hamazaki and others (2003) emphasize the importance of considering the regional context in local management of rare species. Management goals for species at risk should be consistent with other goals and practices described in land or water management plans, such as military Installation Natural Resource Management Plans (INRMPs).
Management goals should be specific targets that are expected to reduce the likelihood of listing the species as federally threatened or endangered. Examples include numbers of individuals, areas of habitat, and population growth rates. Metrics selected in Step 4 may be modified for future monitoring in this step.
Recovery goals for endangered and threatened species in US Fish and Wildlife Service recovery plans
Type of recovery goal
Metric for recovery
Total population size
Number of subpopulations
Number of individuals in each subpopulation
Trends in total population size
Trends in number of subpopulations
Trends in number of individuals in each subpopulation
Age structure of population
Productivity and net recruitment
Total range (presence/absence)
Quality of habitat
Quantity of habitat
Securement of water rights
Acquisition of land
Existence of threats
Significance of threats
Various categories of management goals relate to avoiding species decline or low numbers or viability (Table 1). Statistically significant trends in population abundance (see Step 8) are direct measures of species decline; demographic and habitat-based extinction thresholds are more indirect measures (Table 1). Additional goals involve species or habitat properties that relate to trends that can be measured. Properties of stable populations include minimum population size; age structure; productivity (e.g., fledging rate); and amount, suitability, and arrangement of habitat that are required if habitat-based threats are limiting population abundance.
Metapopulation structure should be considered when setting management goals, and metapopulations at risk may require a set of integrated goals. NOAA Fisheries restoration goals for Pacific salmon include preservation of basic historical metapopulation structure, including genetic exchange across populations, the opportunity for neighboring populations to serve as sources in the event of local extirpations, and a spatial distribution of populations, such that not all are susceptible to localized catastrophic events (McElhany and others 2000). Four criteria considered include abundance, productivity, spatial distribution, and diversity. Listed Pacific salmon stocks will be considered for delisting if (1) at least half of the historical populations meet viability standards (very large, large- and intermediate-sized population with current abundance at or above historical abundance), (2) at least one “highly viable” population remains, (3) all major life history strategies are represented among viable populations, and (4) non-viable populations have sufficient productivity that overall metapopulation growth remains positive.
Most management of species at risk will likely be adaptive. Goals may be adjusted as hypotheses about worst threats or most sensitive life stages are tested. Bakker and Doak (2009) recommend “population viability management” as a formal way to identify combinations of monitoring and management of rare species that may be adjusted through time, using PVA.
Although specific management goals may be set incrementally (e.g., 20% more individuals than last year) or qualitatively, the goals that are most effective at preventing listing or local extinction will be determined scientifically. For example, PVA can be used to identify extinction thresholds such as minimum viable population size and minimum viable habitat area (Jager and others 2006), and the management goal may be set to exceed these thresholds by an uncertainty factor. Spatially explicit population models are useful for PVA in spatially heterogeneous landscapes (Dunning and others 1995), although they have rarely been used in setting species recovery goals (Morris and others 2002). Spatial models are particularly useful when representing metapopulation dynamics, movements among habitats, and threats that influence organisms in some places more than in others. A process for evaluating effects of land use and management using models in a spatial context (Dale and others 1998) has been useful in determining management goals.
As in Step 6c, population models can help identify which parts of the life cycle are most affected by threats and therefore should be the subject of management options (Good and others 2007). Once a critical life stage is identified, management priorities should focus on restoring habitat required by this life stage or removing threats during this period (e.g., minimize noise during breeding season). For example, efforts to restore sea turtles have typically focused on protecting beaches serving as nesting areas to increase the chances that juveniles will survive. However, Crouse and others (1987) used a stage-based model to determine that population growth was most sensitive to adult survival, and, consequently, management now focuses on requiring marine commercial fisheries to use turtle-excluder devices.
Step 8: Develop Species Management Plans and Prioritize Implementation
Step 8a: Develop Species Management Plan
Meeting species management goals entails management of both natural resources and human activities on the land or in the water. Most goals will be achieved through habitat management or threat reduction. For example, the management of endangered red-cockaded woodpecker in the southeastern US is essentially the management of longleaf pine habitat (e.g., James and others 2001; van Lear and others 2005). On military installations, activity is restricted only near woodpecker clusters and mature trees. “Systemic control,” in which the threat is managed (Lessard and others 2005) is likely to be more effective than “symptomatic control,” in which the numbers of individuals are controlled by harvesting or culling competitors and predators (Lessard and others 2005). Whether the acquisition of habitat is a systemic or symptomatic control depends on whether habitat loss was the proximate threat.
If more than one rare species, including species at risk, requires focused management, then actions that achieve the maximum joint benefit (as defined by conservation experts) should be favored (Joseph and others 2008). Multispecies management is a logical strategy for meeting conservation goals in many regions.
Habitat Restoration. One of the principal options for managing species at risk that are threatened by habitat loss is restoring habitat. In some ecosystems this may entail restoring the historical fire regime. Resource managers should include potential risks from restoration activities in their decision making, but these activities generally pose less risk to the species than long-term habitat loss (van Lear and others 2005). Instead, effects of restoration may be in another sector, as when burning to maintain habitat impacts air quality and soil sustainability (Garten 2006).
Land Acquisition and Management. Management actions may include the protection of populations on lands that adjoin the management area (NatureServe 2004). If actions offsite such as purchases of property development rights are feasible, then these acquisition options should be explored, especially if restoration of habitat onsite is not feasible. For example, the DoD Private Lands Initiative involves cooperative agreements between the US Army and non-governmental organizations to purchase land titles or easements for conservation or training buffer purposes (USAEC 2003). Rare species on military lands have also benefitted from Safe Harbor Agreements, i.e., voluntary arrangements between the USFWS or NOAA-Fisheries and private landowners that allow the latter to manage listed species under the assurance that no future regulatory restrictions will be imposed (USFWS 2004).
Monitoring. Species at risk will require continued monitoring to reassess their status and to determine the effectiveness of management actions (Gregory and others 2006). Adding monitoring to the management plan requires a return to Step 5. The metric selected in Step 4, as well as the any new measures defined in Step 7, should be used for continued monitoring so that the management plan may be altered as the new information is obtained (i.e., adaptive management). For example, when spatial analysis and modeling identified potential habitat of Karner blue butterfly on Fort McCoy (Dale and others 2000), significant cost savings resulted from narrowing the monitoring area. Moreover, monitoring plans should be designed to provide evidence of causality (Gregory and others 2006) that may be used to improve management.
If an increasing trend in population size is a management goal, it is important to specify a change that can, in practice, be detected through species monitoring. When several years of species monitoring data are available, Gerrodette’s (1987, 1991) trend detection-power analysis may be useful to determine the monitoring effort and time required to detect trends in population abundance or other metrics. This empirical approach evaluates spatial and temporal variation in estimates of relative abundance from methods in Step 5 and then simulates this variability to determine the magnitude of trend that is detectable within a fixed period of time. Rigorous methods of trend analysis to set species management goals are a state-of-the-art pursuit in conservation biology (Hatch 2003; Boulanger and others 2000; Freilich and others 2005; Maxwell and Jennings 2005).
Management Plans. A framework for setting management goals for species at risk is only useful for a management area if these goals are integrated in natural resources management practice and coordinated with wildlife agency professionals and stakeholders such as neighboring land owners and conservation groups. For US military installations, the appropriate mechanism is the INRMP. INRMPs describe training or testing mission goals, environmental goals and timeframes, land management goals and schedules, recommended actions and their expected costs, legal requirements, and the ecoregional context of the installation’s resources.
Analysis of Alternatives. At this point in the framework, the user of the framework should have enough information to complete a conceptual model of the relationship of alternative management options to anticipated impacts on the future of the species at risk. This model may aid natural resource managers in evaluating anticipated consequences of alternative management.
Quantitative Tools. In some cases, life-table response experiments can be used to quantify how demographic parameters (survival, fecundity) change in response to management (Caswell 2001). In other cases, enough is known about the mechanistic relationships between threats, proposed actions, and species to use more complex models. For example, Jager and others (1997) used known relationships between water temperature and Chinook salmon development, growth, and survival in a spatially explicit individual-based model to predict how recruitment would respond to changes in flow regime. Given an adequate population model, it becomes possible to evaluate management alternatives in a spatial context. A rigorous scientific approach is to formulate a spatial optimization problem. The solution will allocate different resource areas to management alternatives such that specified management goals are met. Note that in previous steps, we have identified such quantitative metrics and goals for management. Formulating a spatial optimization problem forces the user to clarify and articulate the problem in a quantifiable manner. Solving the spatial optimization produces an answer to the question of how best to reach management goals via spatial decisions (Önal and Briers 2006).
Step 8b: Determine Priorities for Implementing Species Management Plan
The resource manager needs to state clearly the criteria that are used to set species management goals. The USFWS prioritizes its allocation of resources partly based on the degree of threat, potential for recovery, and taxonomic uniqueness (USFWS 2004). Costs, logistics, chance of success, acceptability of potential listing, tolerance of uncertainty, and ongoing management activities are additional factors to consider during this step (IUCN 2001; Good and others 2007). The US Army focuses on species at risk whose listing would adversely impact necessary Army missions (Department of the Army 2006). Also, if a significant fraction of the individuals of a species at risk reside on DoD installations and adequate habitat exists to support the species, then the species will get priority consideration (Department of the Army 2006). Clearly, species at risk that are only found at one management site and nowhere else should have priority for proactive management.
Quantitative Tools. Quantitative tools are available for ranking and prioritizing species for management. Criteria, such as those listed above, are used to assign values to species on a fixed scale. In many applications, each stakeholder assigns weight to each criterion, depending on his or her priorities. It is also possible to use the same procedure with only one resource manager assigning values and weights. The project prioritization protocol of Joseph and others (2008) demonstrated that including cost and probability of success as criteria can increase the number of species managed, compared to considering level of threat and metrics of species value (i.e., evolutionary distinctiveness, ecological importance, social distinctiveness) alone.
Step 9: Ending Special Management of Species at Risk
The final component of the framework for setting management goals for species at risk is guidance concerning the termination of special management of species at risk. Stable species require less management. Clearly, funding that is used to manage one species is not available to manage another with different needs.
Step 9a: Develop Criteria for Removing Species at Risk from Enhanced Management Program
Criteria for terminating special species management (the counterpart to delisting under ESA) were specified as a part of setting management goals in Step 7. These criteria may include an upward trend in abundance, a target number of individuals, a target number of nests or burrows, a target area of restored ecosystem, etc. These criteria are not subject to the approval of fish and wildlife management agencies, because species at risk are not regulated. However, it is recommended that there be coordination with these agencies through the management plans (e.g., the INRMP development process). To address this step conservation staff need to answer the question—what constitutes population “stability?” This is just as challenging a question as the analogous question for listed species–what constitutes “recovery?” Although ESA requires clear, measurable criteria for delisting, only about half of recovery plans for federally listed species specify the number of years that a population must meet recovery goals before it is considered ‘recovered’ (Schultz and Gerber 2002).
For many species, terminating special management may never be advisable. Scott and others (2005) argue that many species are “conservation-reliant.” These species are susceptible to “pervasive and recurrent” threats that put species at risk of extinction in the absence of ongoing conservation management. Natural resource managers cannot complete Step 8 for species that are deemed conservation-reliant.
Species at risk can greatly benefit from management attention before they become very rare. This decision framework for setting management goals for species at risk encourages resource managers to explicitly follow nine steps: listing the species at risk at the management area of concern, describing available information and potential information gaps for each species, determining species distribution and potential habitat, selecting metrics for describing species status, assessing status of local population or metapopulation, conducting threat assessment, setting and prioritizing management goals, developing species management plans, and developing criteria for ending special species management. The proactive management set forth in this formal decision framework has several benefits. It is impartial to specific stakeholders. It provides a procedure for and calls for identification of causal relationships that may not be obvious (Suter and others 2002). By providing a way to target the most pressing needs, the framework reduces costs while decreasing the chance of species becoming rare. Finally, the existence of such a framework can improve confidence in the decisions of natural resource managers and regulatory agencies.
We thank the US Army Corps’ Engineer Research and Development Center for funding this work through the program on Habitat-centric Species at Risk (SAR) Research to Avoid Future Training Restrictions and the US Army Species at Risk Program. Dr. Tiomothy Hayden was the program manager. The project was conducted by Oak Ridge National Laboratory (ORNL), which is managed by the UT-Battelle, LLC, for the US Department of Energy under contract DE-AC05-00OR22725. Some of the information cited herein was provided by NatureServe (http://www.natureserve.org/) and its natural heritage member programs, a leading source of information about rare and endangered species, and threatened ecosystems.
- Allee WC (1938) The social life of animals. WW Norton & Co., New York, p 293Google Scholar
- Andow DA, Baker RJ, Lane CP (eds) (1994) Karner blue butterfly: a symbol of a vanishing landscape, Minnesota Agriculture Experiment Station, Miscellaneous Publication, University of Minnesota, St. Paul, 84-1994Google Scholar
- Applied Biomathematics (2003) Ramas software. Copyright 1982–2007, Setauket, NY. http://www.ramas.com/. Accessed 19 August 2008
- Boulanger J, Martin K, Kaiser G, Derocher AE (2000) Evaluating uncertainty in estimating population trends for research and conservation of marbled murrelets. In: Biology and conservation of forest birds. Society of Canadian Ornithologists Special Publication No. 2, pp 53–63Google Scholar
- Caswell H (2001) Matrix population models, 2nd edn. Sinauer Associates, Inc., Sunderland, MAGoogle Scholar
- CMP (2007) Open standards for the practice of conservation, version 2.0. http://conservationmeasures.org/CMP/Site_Docs/CMP_Open_Standards_Version_2.0.pdf. Conservation Measures Partnership.
- Dale VH, King AW, Mann LK, Ashwood TL (2000) Contributions of spatial information and models to management of rare and declining species. In: Hill MJ, Aspinall RJ (eds) Spatial Information for land management. Gordon and Breach Science Publishers, Lausanne, Switzerland, pp 159–172Google Scholar
- Dale VH, Mulholland PJ, Olsen LM, Feminella JW, Maloney KO, White DC, Peacock A, Foster T (2004) Selecting a suite of ecological indicators for resource management. In: Kapustka L, Galbraith H, Luxon M, Biddinger GR (eds) Landscape ecology and wildlife habitat evaluation: critical information for ecological risk assessment, land-use management activities, and biodiversity enhancement practices, ASTM STP 1458. ASTM International, West Conshohocken, PA, pp 3–17CrossRefGoogle Scholar
- Delaney DK, Pater LL, Hayden TJ, Swindell L, Beaty T, Carlile L, Spadgenske, E (2000) Assessment of training noise impacts on the red-cockaded woodpecker: 1999 results, ERDC/CERL TR-00-13. Construction Engineering Research Laboratory, Champaign, ILGoogle Scholar
- Demarais S, Tazik DJ, Guertin PJ, Jorgensen EE (1999) Disturbance associated with military exercises. In: Walker LR (ed) Ecosystems of disturbed ground. Elsevier, Amsterdam, pp 385–396Google Scholar
- Department of the Army (2006) Memorandum from JB Balocki, Director Environmental Programs. Army species at risk policy and implementing guidance. Washington, DC, September 15, 2006Google Scholar
- Doresky J, Morgan K, Ragsdale L, Townsend H (2001) Effects of military activity on reproductive success of red-cockaded woodpeckers. Journal of Field Ornithology 72:305–311Google Scholar
- Efroymson RA, Morrill VA, Dale VH, Jenkins TF, Giffen NR (2009) Habitat disturbance at explosives-contaminated ranges. In: Sunahara G, Hawari J, Lotufo G, Kuperman R (eds) Ecotoxicology of explosives and unexploded ordnance. CRC Press, Boca Raton, FL, pp 253–276Google Scholar
- GAO (Government Accountability Office) (June 2002) Military training: DoD lacks a comprehensive plan to manage encroachment on training ranges. Report to Congressional Requestors, GAO-02-614. Washington, DCGoogle Scholar
- Grimm V, Railsback SF (2005) Individual-based modeling and ecology. Princeton University Press, Princeton, NJGoogle Scholar
- Groves CR (2003) Drafting a conservation blueprint: a practitioner’s guide to planning for biodiversity. The Nature Conservancy and Island Press, Washington, DCGoogle Scholar
- Hill AB (1965) The environment and disease: association or causation? Proceedings of the Royal Society of Medicine 58:295–300Google Scholar
- IUCN (2001) IUCN red list categories and criteria: version 3.1. IUCN Species Survival Commission. International Union for Conservation of Nature and Natural Resources, Gland, SwitzerlandGoogle Scholar
- Jager HI, Efroymson RA (2004) Spatial life history influences the risks associated with habitat loss and fragmentation. International Association. 19th annual symposium of the United States Regional Association of the International Association for Landscape Ecology. Transdisciplinary challenges in landscape ecology, Las Vegas, NVGoogle Scholar
- Lessard RB, Martell SJD, Walters CJ, Essington TE, Kitchell JF (2005) Should ecosystem management involve active control of species abundances? Ecology and Society 10(2):1 [online]Google Scholar
- Marcot BG (2006) Characterizing species at risk I: modeling rare species under the Northwest Forest Plan. Ecology and Society 11(2):10 [online]. http://www.ecologyandsociety.org/vol11/iss2/art10/
- McElhany P, Ruckelshaus MH, Ford MJ, Wainwright TC, Bjorkstedt EP (2000) Viable salmon populations and the recovery of evolutionarily significant units, NOAA Technical Memorandum NMFS-NWFSC-42. US Department of Commerce, 156 ppGoogle Scholar
- NatureServe (2004) Species at risk on Department of Defense installations. Prepared for Department of Defense, US Fish and Wildlife Service, Arlington, VAGoogle Scholar
- NRC (National Research Council) (1995) Science and the Endangered Species Act. National Academy Press, Washington, DC, USAGoogle Scholar
- Pearson RG (2007) Species’ distribution modeling for conservation educators and practitioners. Synthesis. American Museum of Natural History, New York. Available at http://biodiversityinformatics.amnh.org/files/SpeciesDistModelingSYN_1-16-08.pdf. Accessed 19 August 2008
- Rabinowitz D (1981) Seven forms of rarity. In: Synge H (ed) The biological aspects of rare plant conservation. Wiley, New York, pp 205–217Google Scholar
- Rotstein AH (October 16, 2003) Military coexists with endangered pronghorn on bombing range, The Associated Press State and Local Wire. http://www.globalsecurity.org/org/news/2003/031019-goldwater-range.htm. Accessed 19 August 2008
- Sternberg JE, Hays J, Sanborn S, McFarland L, Loring H, Sietman B (1998) Fauna, flora, and sensitive habitat on Fort Leonard Wood, MO. USA/CERL Special Report 98/95. US Army Corps of Engineers Construction Engineering Research Laboratory, Champaign, ILGoogle Scholar
- Tuberville TD, Bodie JR, Jensen JB, LaClaire L, Gibbons JW (2000) Apparent decline of the southern hognose snake, Heterodon simus. Journal of the Elisha Mitchell Scientific Society 116:19–40Google Scholar
- USAEC (U.S. Army Environmental Center) (2003) Private Lands Initiative (PLI). Fact sheet. Available at http://aec.army.mil/usaec/natural/pli-factsheet.pdf. Accessed 19 August 2008
- USEPA (2000) Stressor identification guidance document, EPA/822/B-00/25. US Environmental Protection Agency, Office of Water, Washington, DCGoogle Scholar
- USFWS (2004) Safe harbor agreements for private landowners. Fact sheet. U.S. Fish and Wildlife Service, Arlington, VA. http://www.fws.gov/endangered/factsheets/harborqa.pdf. Accessed 19 August 2008
- USFWS (2007a) Candidate species. Section 4 of the Endangered Species Act. U.S. Fish and Wildlife Service, Arlington, VA. http://www.fws.gov/endangered/pdfs/factsheets/candidate_species.pdf. Accessed 19 August 2008
- USFWS (2007b) In: Pruitt L, TeWinkel L (eds) Indiana bat (Myotis sodalis) draft recovery plan. First Revision. US Fish and Wildlife Service, Ft. Snelling, MNGoogle Scholar
- USGS (1998) Species at risk program. Department of the Interior, US Geological Survey, Biological Resources Division, FR 63(189):52284–52285Google Scholar
- Warren SD, Holbrook SW, Dale DA, Whelan NL, Elyn M, Grimm W, Jentsch A (2007) Biodiversity and the heterogeneous disturbance regime on military training lands. Restoration Ecology 15:606–612Google Scholar