Inferring Threat from Scientific Collections

Exact formulas for the probability of extinction or change in the conservation status of species are described for data based on frequency of sighting. These formulas generalize an expression previously described for the probability of extinction from (binary) sighting data. The formulas will be used in contexts where sightings are recorded as frequencies, such as when observations are aggregated in time. We argue that computing the probability of extinction or change in conservation status will be most useful for setting conservation priorities and in screening large data sets contained in museum and herbarium collections and in biological resource inventories.     

Optimal Allocation of Conservation Resources to Species That May be Extinct

Abstract: Statements of extinction will always be uncertain because of imperfect detection of species in the wild. Two errors can be made when declaring a species extinct. Extinction can be declared prematurely, with a resulting loss of protection and management intervention. Alternatively, limited conservation resources can be wasted attempting to protect a species that no longer exists. Rather than setting an arbitrary level of certainty at which to declare extinction, we argue that the decision must trade off the expected costs of both errors. Optimal decisions depend on the cost of continued intervention, the probability the species is extant, and the estimated value of management (the benefit of management times the value of the species). We illustrated our approach with three examples: the Dodo (Raphus cucullatus), the Ivory‐billed Woodpecker (U.S. subspecies Campephilus principalis principalis), and the mountain pygmy‐possum (Burramys parvus). The dodo was extremely unlikely to be extant, so managing and monitoring for it today would not be cost‐effective unless the value of management was extremely high. The probability the Ivory‐billed woodpecker is extant depended on whether recent controversial sightings were accepted. Without the recent controversial sightings, it was optimal to declare extinction of the species in 1965 at the latest. Accepting the recent controversial sightings, it was optimal to continue monitoring and managing until 2032 at the latest. The mountain pygmy‐possum is currently extant, with a rapidly declining sighting rate. It was optimal to conduct as many as 66 surveys without sighting before declaring the species extinct. The probability of persistence remained high even after many surveys without sighting because it was difficult to determine whether the species was extinct or undetected. If the value of management is high enough, continued intervention can be cost‐effective even if the species is likely to be extinct.     

Effects of Recent Environmental Change on Accuracy of Inferences of Extinction Status

Correctly classifying a species as extinct or extant is of critical importance if current rates of biodiversity loss are to be accurately quantified. Observing an extinction event is rare, so in many cases extinction status is inferred using methods based on the analysis of records of historic sighting events. The accuracy of such methods is difficult to test. However, results of recent experiments with microcosm communities suggest that the rate at which a population declines to extinction, potentially driven by varying environmental conditions, may alter one's ability accurately to infer extinction status. We tested how the rate of population decline, driven by historic environmental change, alters the accuracy of 6 commonly applied sighting‐based methods used to infer extinction. We used data from small‐scale experimental communities and recorded wild population extirpations. We assessed how accuracy of the different methods was affected by rate of population decline, search effort, and number of sighting events recorded. Rate of population decline and historic population size of the species affected the accuracy of inferred extinction dates; however, faster declines produced more accurate inferred dates of extinction, but only when population sizes were higher. Optimal linear estimation (OLE) offered the most reliable and robust estimates, though no single method performed best in all situations, and it may be appropriate to use a different method if information regarding historic search efforts is available. OLE provided the most accurate estimates of extinction when the number of sighting events used was >10, and future use of this method should take this into account. Data from experimental populations provide added insight into testing techniques to discern wild extirpation events. Care should be taken designing such experiments so that they mirror closely the abundance dynamics of populations affected by real‐world extirpation events. Efectos del Cambio Ambiental Reciente sobre la Precisión de las Inferencias sobre el Estado de Extinción     

Rates of Movement of Threatened Bird Species between IUCN Red List Categories and toward Extinction

Abstract:  In recent centuries bird species have been deteriorating in status and becoming extinct at a rate that may be 2–3 orders of magnitude higher than in prehuman times. We examined extinction rates of bird species designated critically endangered in 1994 and the rate at which species have moved through the IUCN (World Conservation Union) Red List categories of extinction risk globally for the period 1988–2004 and regionally in Australia from 1750 to 2000. For Australia we drew on historical accounts of the extent and condition of species habitats, spread of invasive species, and changes in sighting frequencies. These data sets permitted comparison of observed rates of movement through the IUCN Red List categories with novel predictions based on the IUCN Red List criterion E, which relates to explicit extinction probabilities determined, for example, by population viability analysis. The comparison also tested whether species listed on the basis of other criteria face a similar probability of moving to a higher threat category as those listed under criterion E. For the rate at which species moved from vulnerable to endangered, there was a good match between observations and predictions, both worldwide and in Australia. Nevertheless, species have become extinct at a rate that, although historically high, is 2 (Australia) to 10 (globally) times lower than predicted. Although the extinction probability associated with the critically endangered category may be too high, the shortfall in realized extinctions can also be attributed to the beneficial impact of conservation intervention. These efforts may have reduced the number of global extinctions from 19 to 3 and substantially slowed the extinction trajectory of 33 additional critically endangered species. Our results suggest that current conservation action benefits species on the brink of extinction, but is less targeted at or has less effect on moderately threatened species.     

Identifying Anomalous Reports of Putatively Extinct Species and Why It Matters

Abstract:  As species become very rare and approach extinction, purported sightings can stir controversy, especially when scarce management resources are at stake. We used quantitative methods to identify reports that do not fit prior sighting patterns. We also examined the effects of including records that meet different evidentiary standards on quantitative extinction assessments for four charismatic bird species that might be extinct: Eskimo Curlew ( Numenius borealis ), Ivory-billed Woodpecker ( Campephilus principalis ), Nukupu`u ( Hemignathus lucidus ), and O`ahu `Alauahio ( Paroreomyza maculata ). For all four species the probability of there being a valid sighting today, given the past pattern of verified sightings, was estimated to be very low. The estimates of extinction dates and the chance of new sightings, however, differed considerably depending on the criteria used for data inclusion. When a historical sighting record lacked long periods without sightings, the likelihood of new sightings declined quickly with time since the last confirmed sighting. For species with this type of historical record, therefore, new reports should meet an especially high burden of proof to be acceptable. Such quantitative models could be incorporated into the International Union for Conservation of Nature's Red List criteria to set evidentiary standards required for unconfirmed sightings of "possibly extinct" species and to standardize extinction assessments across species.     

Identifying Sighting Clusters of Endangered Taxa with Historical Records

Abstract: The probability and time of extinction of taxa is often inferred from statistical analyses of historical records. Many of these analyses require the exclusion of multiple records within a unit of time (i.e., a month or a year). Nevertheless, spatially explicit, temporally aggregated data may be useful for identifying clusters of sightings (i.e., sighting clusters) in space and time. Identification of sighting clusters highlights changes in the historical recording of endangered taxa. I used two methods to identify sighting clusters in historical records: the Ederer–Myers–Mantel (EMM) test and the space–time permutation scan (STPS). I applied these methods to the spatially explicit sighting records of three species of orchids that are listed as endangered in the Republic of Ireland under the Wildlife Act (1976): Cephalanthera longifolia, Hammarbya paludosa, and Pseudorchis albida. Results with the EMM test were strongly affected by the choice of the time interval, and thus the number of temporal samples, used to examine the records. For example, sightings of P. albida clustered when the records were partitioned into 20‐year temporal samples, but not when they were partitioned into 22‐year temporal samples. Because the statistical power of EMM was low, it will not be useful when data are sparse. Nevertheless, the STPS identified regions that contained sighting clusters because it uses a flexible scanning window (defined by cylinders of varying size that move over the study area and evaluate the likelihood of clustering) to detect them, and it identified regions with high and regions with low rates of orchid sightings. The STPS analyses can be used to detect sighting clusters of endangered species that may be related to regions of extirpation and may assist in the categorization of threat status.     

Inferring National and Regional Declines of Rare Orchid Species with Probabilistic Models

Abstract: Fragmentation of natural habitats can increase numbers of rare species. Conservation of rare species requires experts and resources, which may be lacking for many species. In the absence of regular surveys and expert knowledge, historical sighting records can provide data on the distribution of a species. Numerous models have been developed recently to make inferences regarding the threat status of a taxon on the basis of variation in trends of sightings over time. We applied 5 such models to national and regional (county) data on 3 red‐listed orchid species (Cephalanthera longifolia, Hammarbya paludosa, and Pseudorchis albida) and 1 species that has recently come to the attention of conservation authorities (Neotinea maculata) in the Republic of Ireland. In addition, we used an optimal linear estimate to calculate the time of extinction for each species overall and within each county. To account for bias in recording effort over time, we used rarefaction analysis. On the basis of sighting records, we inferred that these species are not threatened with extinction and, although there have been declines, there is no clear geographical pattern of decline in any species. Most counties where these orchid species occurred had a low number of sightings; hence, we were cautious in our interpretation of output from statistical models. We suggest the main drivers of decline in these species in Ireland are modification of habitats for increased agricultural production and lack of appropriate management. Our results show that the application of probabilistic models can be used even when sighting data are scarce, provided multiple models are used simultaneously and rarefaction is used to account for bias in recording effort among species over time. These models could be used frequently when making an initial conservation assessment of species in a region, particularly if there is a relatively constant recording rate and some knowledge of the underlying recording process. Regional‐scale analyses, such as ours, complement World Conservation Union criteria for assessment of the extinct category and are useful for highlighting areas of under recording and focusing conservation efforts of rare and endangered species.     

Understanding the drivers of expert opinion when classifying species as extinct

The criteria as laid out by the International Union for the Conservation of Nature (IUCN) Red List are the gold standard by which the extinction risk of a species is assessed and where appropriate biological extinctions are declared. However, unlike all other categories, the category of extinct lacks a quantitative framework for assigning this category. Given its subjective nature, we surveyed expert assessors working on a diversity of taxa to explore the attributes they used to declare a species extinct. Using a choice experiment approach, we surveyed 674 experts from the IUCN Species Survival Commission specialist groups and taskforces. Data availability, time from the last sighting, detectability, habitat availability, and population decline were all important attributes favored by assessors when inferring extinction. Respondents with red-listing experience assigned more importance to the attributes data availability, time from the last sighting, and detectability when considering a species extinction, whereas those respondents working with well-known taxa gave more importance to the time from the last sighting. Respondents with no red-listing experience and those working with more well-known taxa (i.e., mammals and birds) were overall less likely to consider species extinct. Our findings on the importance assessors place on attributes used to declare a species extinct provide a basis for informing the development of specific criteria for more accurately assessing species extinctions.     

Spatial Structure and Population Extinction: A Study with Drosophila Flies

Abstract: The total amount of habitat and also its distribution and subdivision affect the extinction probability of a resident population Two species of Drosophila are studied in spatial configurations of a single large habitat patch, single small habitat patches, and two small but connected habitat patches in which a low rate of migration, roughly one fly per generation, is possible. The single large habitat patch shows the lowest extinction rate lower than the combined rate of two small patches of the same total size. For one of the species, the "corridor" between the pair of small patches seems to produce a "rescue effect" that lowers extinction rates, probably due to a decrease in the coefficient of variation in fluctuations of the population sire in this coupled system. The systems seem to have been influenced by demographic stochasticity, based on the relationship of population size to extinction probability.     

Passive acoustic monitoring of the decline of Mexico's critically endangered vaquita

The vaquita (Phocoena sinus) is the world's most endangered marine mammal with approximately 245 individuals remaining in 2008. This species of porpoise is endemic to the northern Gulf of California, Mexico, and historically the population has declined because of unsustainable bycatch in gillnets. An illegal gillnet fishery for an endangered fish, the totoaba (Totoaba macdonaldi), has recently resurged throughout the vaquita's range. The secretive but lucrative wildlife trade with China for totoaba swim bladders has probably increased vaquita bycatch mortality by an unknown amount. Precise population monitoring by visual surveys is difficult because vaquitas are inherently hard to see and have now become so rare that sighting rates are very low. However, their echolocation clicks can be identified readily on specialized acoustic detectors. Acoustic detections on an array of 46 moored detectors indicated vaquita acoustic activity declined by 80% between 2011 and 2015 in the central part of the species’ range. Statistical models estimated an annual rate of decline of 34% (95% Bayesian credible interval –48% to –21%). Based on results from 2011 to 2014, the government of Mexico enacted and is enforcing an emergency 2‐year ban on gillnets throughout the species’ range to prevent extinction, at a cost of US$74 million to compensate fishers. Developing precise acoustic monitoring methods proved critical to exposing the severity of vaquitas’ decline and emphasizes the need for continual monitoring to effectively manage critically endangered species.     

Predicting the Risk of Extinction through Hybridization

Abstract: Natural hybridization threatens a substantial number of plant and animal species with extinction, but extinction risk has been difficult to evaluate in the absence of a quantitative assessment of risk factors. We investigated a number of ecological parameters likely to affect extinction risk, through an individual-based model simulating the life cycle of two hybridizing annual plant species. All parameters tested, ranging from population size to variance in pollen-tube growth rates, affected extinction risk. The sensitivity of each parameter varied dramatically across parameter sets, but, overall, the competitive ability, initial frequency, and selfing rate of the native taxon had the strongest effect on extinction. In addition, prezygotic reproductive barriers had a stronger influence on extinction rates than did postzygotic barriers. A stable hybrid zone was possible only when habitat differentiation was included in the model. When there was no habitat differentiation, either one of the parental species or the hybrids eventually displaced the other two taxa. The simulations demonstrated that hybridization is perhaps the most rapidly acting genetic threat to endangered species, with extinction often taking place in less than five generations. The simulation model was also applied to naturally hybridizing species pairs for which considerable genetic and ecological information is available. The predictions from these "worked examples" are in close agreement with observed outcomes and further suggest that an endemic cordgrass species is threatened by hybridization. These simulations provide guidance concerning the kinds of data required to evaluate extinction risk and possible conservation strategies.     

On Uncertain Sightings and Inference about Extinction

The extinction of many species can only be inferred from the record of sightings of individuals. Solow et al. (2012, Uncertain sightings and the extinction of the Ivory‐billed Woodpecker. Conservation Biology 26: 180–184) describe a Bayesian approach to such inference and apply it to a sighting record of the Ivory‐billed Woodpecker (Campephilus principalis). A feature of this sighting record is that all uncertain sightings occurred after the most recent certain sighting. However, this appears to be an artifact. We extended this earlier work in 2 ways. First, we allowed for overlap in time between certain and uncertain sightings. Second, we considered 2 plausible statistical models of a sighting record. In one of these models, certain and uncertain sightings that are valid arise from the same process whereas in the other they arise from independent processes. We applied both models to the case of the Ivory‐billed Woodpecker. The result from the first model did not favor extinction, whereas the result for the second model did. This underscores the importance, in applying tests for extinction, of understanding what could be called the natural history of the sighting record. Sobre Avistamientos Inciertos e Inferencia de la Extinción     

Comparison of conservation metrics in a case study of lemurs

Conservation planning is important to protect species from going extinct now that natural habitats are decreasing owing to human activity and climate change. However, there is considerable controversy in choosing appropriate metrics to weigh the value of species and geographic regions. For example, the added value of phylogenetic conservation‐selection criteria remains disputed because high correlations between them and the nonphylogenetic criteria of species richness have been reported. We evaluated the commonly used conservation metrics species richness, endemism, phylogenetic diversity (PD), and phylogenetic endemism (PE) in a case study on lemurs of Madagascar. This enabled us to identify the conservation target of each metric and consider how they may be used in future conservation planning. We also devised a novel metric that uses a phylogeny scaled according to the rate of phenotypic evolution as a proxy for a species’ ability to adapt to change. High rates of evolution may indicate generalization or specialization. Both specialization and low rates of evolution may result in an inability to adapt to changing environments. We examined conservation priorities by using the inverse of the rate of body mass evolution to account for species with low rates of evolution. In line with previous work, we found high correlations among species richness and PD (r = 0.96), and endemism and PE (r = 0.82) in Malagasy lemurs. Phylogenetic endemism in combination with rates of evolution and their inverse prioritized grid cells containing highly endemic and specialized lemurs at risk of extinction, such as Avahi occidentalis and Lepilemur edwardsi, 2 endangered lemurs with high rates of phenotypic evolution and low‐quality diets, and Hapalemur aureus, a critically endangered species with a low rate of body mass evolution and a diet consisting of very high doses of cyanide.     

Uncertain sightings and the extinction of the Ivory-billed Woodpecker

The extinction of a species can be inferred from a record of its sightings. Existing methods for doing so assume that all sightings in the record are valid. Often, however, there are sightings of uncertain validity. To date, uncertain sightings have been treated in an ad hoc way, either excluding them from the record or including them as if they were certain. We developed a Bayesian method that formally accounts for such uncertain sightings. The method assumes that valid and invalid sightings follow independent Poisson processes and use noninformative prior distributions for the rate of valid sightings and for a measure of the quality of uncertain sightings. We applied the method to a recently published record of sightings of the Ivory-billed Woodpecker (Campephilus principalis). This record covers the period 1897-2010 and contains 39 sightings classified as certain and 29 classified as uncertain. The Bayes factor in favor of extinction was 4.03, which constitutes substantial support for extinction. The posterior distribution of the time of extinction has 3 main modes in 1944, 1952, and 1988. The method can be applied to sighting records of other purportedly extinct species.     

An occupancy‐based quantification of the highly imperiled status of desert fishes of the southwestern United States

Desert fishes are some of the most imperiled vertebrates worldwide due to their low economic worth and because they compete with humans for water. An ecological complex of fishes, 2 suckers (Catostomus latipinnis, Catostomus discobolus) and a chub (Gila robusta) (collectively managed as the so‐called three species) are endemic to the U.S. Colorado River Basin, are affected by multiple stressors, and have allegedly declined dramatically. We built a series of occupancy models to determine relationships between trends in occupancy, local extinction, and local colonization rates, identify potential limiting factors, and evaluate the suitability of managing the 3 species collectively. For a historical period (1889–2011), top performing models (AICc) included a positive time trend in local extinction probability and a negative trend in local colonization probability. As flood frequency decreased post‐development local extinction probability increased. By the end of the time series, 47% (95% CI 34‐61) and 15% (95% CI 6‐33) of sites remained occupied by the suckers and the chub, respectively, and models with the 2 species of sucker as one group and the chub as the other performed best. For a contemporary period (2001?2011), top performing (based on AIC_c) models included peak annual discharge. As peak discharge increased, local extinction probability decreased and local colonization probability increased. For the contemporary period, results of models that split all 3 species into separate groups were similar to results of models that combined the 2 suckers but not the chub. Collectively, these results confirmed that declines in these fishes were strongly associated with water development and that relative to their historic distribution all 3 species have declined dramatically. Further, the chub was distinct in that it declined the most dramatically and therefore may need to be managed separately. Our modeling approach may be useful in other situations in which targeted data are sparse and conservation status and best management approach for multiple species are uncertain.     

Population viability analysis for several populations using multivariate state-space models

The International Union for the Conservation of Nature and Natural Resources (IUCN), the world's largest and most important global conservation network, has listed approximately 16,000 species worldwide as threatened. The most important tool for recognizing and listing species as threatened is population viability analysis (PVA), which estimates the probability of extinction of a population or species over a specified time horizon. The most common PVA approach is to apply it to single time series of population abundance. This approach to population viability analysis ignores covariability of local populations. Covariability can be important because high synchrony of local populations reduces the effective number of local populations and leads to greater extinction risk. Needed is a way of extending PVA to model correlation structure among multiple local populations. Multivariate state-space modeling is applied to this problem and alternative estimation methods are compared. The multivariate state-space technique is applied to endangered populations of pacific salmon, USA. Simulations demonstrated that the correlation structure can strongly influence population viability and is best estimated using restricted maximum likelihood instead of maximum likelihood.     

Incorporating Climate and Ocean Change into Extinction Risk Assessments for 82 Coral Species

Many marine invertebrate species facing potential extinction have uncertain taxonomies and poorly known demographic and ecological traits. Uncertainties are compounded when potential extinction drivers are climate and ocean changes whose effects on even widespread and abundant species are only partially understood. The U.S. Endangered Species Act mandates conservation management decisions founded on the extinction risk to species based on the best available science at the time of consideration—requiring prompt action rather than awaiting better information. We developed an expert‐opinion threat‐based approach that entails a structured voting system to assess extinction risk from climate and ocean changes and other threats to 82 coral species for which population status and threat response information was limited. Such methods are urgently needed because constrained budgets and manpower will continue to hinder the availability of desired data for many potentially vulnerable marine species. Significant species‐specific information gaps and uncertainties precluded quantitative assessments of habitat loss or population declines and necessitated increased reliance on demographic characteristics and threat vulnerabilities at genus or family levels. Adapting some methods (e.g., a structured voting system) used during other assessments and developing some new approaches (e.g., integrated assessment of threats and demographic characteristics), we rated the importance of threats contributing to coral extinction risk and assessed those threats against population status and trend information to evaluate each species’ extinction risk over the 21st century. This qualitative assessment resulted in a ranking with an uncertainty range for each species according to their estimated likelihood of extinction. We offer guidance on approaches for future biological extinction risk assessments, especially in cases of data‐limited species likely to be affected by global‐scale threats. Incorporación del Cambio Climático y Oceánico en Estudios de Riesgo de Extinción para 82 Especies de Coral     

Extinction of Mammal Populations in Western North American National Parks

Patterns of local extinction of mammal populations in western North American parks were examined in relation to current biogeographic and population lifetime models. The analysis was based on species sighting records as of 1989. While western North American parks are obviously not true isolates, patterns of mammal extinction in them are nonetheless consistent with two predictions of the land-bridge island hypothesis. First, the number of extinctions has exceeded the number of colonizations since park establishment, and, second, the rate of extinction is inversely related to park area. Factors influencing the lifetime of mammal populations were evaluated using a stepwise multivariate survival analysis procedure for censored data. Survival time for mammal populations was positively related to estimated initial population size. After accounting for population size, species within the order Lagomorpha were particularly prone to extinction. Finally, after controlling for population size and taxon variation, survival time was positively related to age of maturity, indicating that species with longer generation times—age of maturity and generation time are highly correlated in mammals—persist longer in absolute time.     

Island Extinction Rates from Regular Censuses

Regular censuses conducted over a long time allow the calculation of both extinction and immigration rates. We present formulae for estimation of those rates. We use them on bird censuses of three British islands. These formulae improve on previous estimators of extinction but reaffirm that smaller populations have a higher probability of becoming extinct. On the other hand, they suggest no correlation between extinction rate and either body size or migratory status among birds.     

Determinants of local extinction and turnover rates in urban bird communities.

Studying the effects of urbanization on the dynamics of communities has become a priority for biodiversity conservation. The consequences of urbanization are mainly an increased fragmentation of the original landscapes associated with a decrease in the amount of favorable habitats and an increased pressure of human activities on the remaining patches suitable for wildlife. Patterns of bird species richness have been studied at different levels of urbanization, but little is known about the temporal dynamics of animal communities in urban landscapes. In particular, urbanization is expected to have stronger negative effects on migratory breeding bird communities than on sedentary ones, which should lead to different patterns of change in composition. Using an estimation method accounting for heterogeneity in species detection probability and data collected between 2001 and 2003 within a suburban area near the city of Paris, France, we tested whether these communities differ in their local extinction and turnover rates. We considered the potential effects of patch size and distance to Paris' center as a measure of the degree of urbanization around the patches. As expected, local rates of extinction and turnover were higher for migratory than for sedentary species, and they were negatively related to patch size for migratory species. Mean species richness of the sedentary species increased during the study period and their local turnover rate was negatively related to the distance to the urban core, showing a trend to colonize the most urban patches. These results highlight the very dynamic nature of the composition of some local bird communities in fragmented habitats and help to identify factors affecting colonization and extinction.