Conservation of endangered species - Ecological Applications: Toward a Sustainable World - page 124

population size, age-class structure and immunological status of individuals in the
population, so these need to be incorporated if simulation models are to be
worthwhile.
The Ethiopian wolf ( Canis simensis ) is critically endangered, now restricted in the
Ethiopian highlands to just seven fragmented and isolated populations ranging in
size from 15 to 250 individuals, with a grand total of about 500. The habitat patches
are surrounded by agricultural land where domestic dogs are a reservoir for rabies.
We know something about the risk associated with disease because a rabies out-
break in the early 1990s reduced the biggest of the wolf populations by two thirds
and only after 10 years did pack numbers approach pre-epidemic levels. As a result,
dog vaccination was instigated in 1996. The question of whether other disease man-
agement strategies are needed, and where these should be applied, was addressed
by Haydon et al. (2002) using a standard population simulation model approach but
with an additional epidemiological component.
The wolves live in close-knit, male-biased packs consisting of an average of six
adults, one to six yearlings, and one to seven pups. Only dominant females breed.
Dispersal is limited by the scarcity of unoccupied habitat: males stay in the pack
where they were born, but two thirds of females disperse at 2 years old and become
'fl oaters' until a breeding vacancy becomes available.
The simulation models were constructed on the basis of published information
or realistic biological assumptions. Carrying capacity of each occupied patch was
set at one individual per km 2 . Each pack (of which there were several in each habitat
patch) had a maximum size of 13 adults, of which only two could be female. Each
year packs with at least one male and one female gave birth with a probability of
0.63 to litters of pups of from one to six (with average probabilities for each litter
size ranging from 0.05 to 0.32). Age classes (pups 0-1 years, juveniles 1-2 years,
adults 2 years) were updated before breeding, and immediately prior to breeding

a pack could undergo realistic events such as loss of a randomly selected female to
the 'fl oater' pool, recruitment of a female from the fl oater pool or splitting of large
packs into two. Adults had an average 0.15 probability of dying each year but for
younger animals this probability was 0.45 for males and 0.55 for females. Together
with random elements (stochasticity) in the precise birth and death probabilities
applied, this is the normal bread and butter of individual-based population viability
models.
To account for disease, each individual was also subject to a susceptible-
infectious-recovered (S-I-R) process. Susceptible individuals may become infected
(and for a period be infectious), followed by death or recovery. The models incor-
porated realistic probabilities for all these epidemiological elements. Infection prob-
abilities from dog to wolf differed for pack members and fl oaters, and there were
different probabilities again from wolf to wolf among members of a pack, between
different packs and between pack members and fl oaters. Recovered individuals are
immune to further infection. Population simulations were performed for various
sized areas (25-250 km 2 ), to encompass the known sizes of habitat patches occupied
by the wolves, for a range of dog disease incidence (1-140% of a realistic baseline
incidence) and for a range of anti-rabies wolf vaccination efforts.
In disease-free simulations (1000 for each scenario), the probability of populations
persisting in the different sized patches is generally 100%; only in the case of the
smallest habitat patch (25 km 2 , with its correspondingly small population - carrying

>