Assessing uncertainty in sighting records: an example of the Barbary lion

Abstract

As species become rare and approach extinction, purported sightings can be controversial, especially when scarce management resources are at stake. We consider the probability that each individual sighting of a series is valid. Obtaining these probabilities requires a strict framework to ensure that they are as accurately representative as possible. We used a process, which has proven to provide accurate estimates from a group of experts, to obtain probabilities for the validation of 32 sightings of the Barbary lion. We consider the scenario where experts are simply asked whether a sighting was valid, as well as asking them to score the sighting based on distinguishablity, observer competence, and verifiability. We find that asking experts to provide scores for these three aspects resulted in each sighting being considered more individually, meaning that this new questioning method provides very different estimated probabilities that a sighting is valid, which greatly affects the outcome from an extinction model. We consider linear opinion pooling and logarithm opinion pooling to combine the three scores, and also to combine opinions on each sighting. We find the two methods produce similar outcomes, allowing the user to focus on chosen features of each method, such as satisfying the marginalisation property or being externally Bayesian.

A fast re‐sampling method for using reliability ratings of sightings with extinction‐date estimators

“…The species examples we use represent (1) four with well‐known sighting records (baiji dolphin [ Lipotes vexillifer ], Barbary lion [ Pathera leo ], Ivory‐billed Woodpecker [ Campephilus principalis ], and Oʻahu nukupuʻu [ Hemignathus lucidus ]), containing a mix of unambiguous (reliability = 1) and ambiguous (reliability < 1) records, and which have been used frequently in previous papers that have proposed or evaluated methods for estimating extinction from sighting records (e.g., Jarić and Roberts , Lee et al. , , Boakes et al. ); (2) two (the Night Parrot [ Pezoporus occidentalis ] and Noisy Scrub Bird [ Atrichornis clamosus ]) with past sightings (of mixed certainty) spread across multiple years, followed by an extended interval (typically of many decades) of apparent absence, after which they were rediscovered (“Lazarus species”); and (3) two (Bramble Cay melomys [ Melomys rubicola ] and Alaotra Grebe [ Tachybaptus rufolavatus ]) that were regularly reported with a mix of sighting uncertainty, and might persist, except that other evidence (e.g., exhaustive searches across their range, or a complete loss of suitable habitat) implies that extinction is “definite.”…”

Are extinction opinions extinct?

“… Lee et al (2015) provided distributions for 32 alleged sightings of the Barbary lion ( Panthera l. leo ) which occurred between 1895 and 1956 in Algeria and Morocco. In this paper we use the individual quality score provided by Lee et al (2015) . We also examine the importance of the expert’s prior of the lion being extinct on the results.…”

“…The work of Lee et al (2015) provides several distributions for each lion sighting. One method considers the expert estimates for three different questions relating to the distinguishability of the species, observer competence and verifiability, and pools across experts and questions linearly, while another pools them logarithmically.…”

Untitled

“…The elicitation in Lee et al (2015) was not carried out explicitly under a belief of extinction or non-extinction. Five experts offered a best estimate and lower/upper bounds for three different aspects of sighting quality (in an un-blinded manner) for each sighting at time t. Lee et al (2015) use the most straightforward way to represent these three points as a probability density, that is as a triangle density. For simplicity we treat experts as exchangeable, ignore any correlation between the best and lower/upper estimates, and also ignore any correlation between the j questions (the differential weights of expert competency does something to adjust for inter-expert correlation as does the exhaustive group elicitation process).…”

“…Lee et al (2015) provided distributions for 32 alleged sightings of the Barbary lion (Panthera l. leo) which occurred between 1895 and 1956 in Algeria and Morocco. In this paper we use the individual quality score provided by Lee et al (2015). We also examine the importance of the expert's prior of the lion being extinct on the results.…”