J. Forcada, Can population surveys show if the Mediterranean monk seal colony at Cap Blanc is declining in abundance?, J APPL ECOL, 37(1), 2000, pp. 171-181
1. The Mediterranean monk seal is an endangered marine mammal with only a f
ew populations thought to be viable. The largest aggregation of this specie
s is found at the Peninsula of Cap Blanc, on the western Sahara, where a co
lonial structure is maintained. The size of this colony is estimated by ann
ual capture-recapture surveys, but it is unclear if the attained level of p
recision is enough to detect changes in numbers in the short term.
2. In this study, simulated capture-recapture experiments were used to inve
stigate the relationship between population estimates and abundance. The st
atistical power necessary to detect a population decline with ongoing surve
ys was estimated, and the requirements for an improvement in precision were
evaluated. Additionally, changes in abundance over consecutive surveys wer
e examined for comparison with an alternative method of Bayesian inference.
3. Results indicated that the power to detect trends from the actual survey
s was lower than should be required for the conservation of a small populat
ion, such as the colony at Cap Blanc. A 5% annual rate of decline in abunda
nce would only be detected with high power after 12 years, with a populatio
n of 317 seals being reduced to 172. If the annual rate of decline was 10%,
8 years would be required to detect a reduction of 317 to 136 seals.
4. Capture-recapture surveys can produce reliable data but do not identify
moderate or low population declines. The increase in precision of surveys i
s costly and time consuming, and other monitoring methods are needed to det
ect early signs of a decline. Bayesian methods cannot provide a better resu
lt because they are not robust to small sample size and heterogeneity in ca
pture probabilities. Despite the need for alternative methods for earlier d
etection, capture-recapture surveys are still required to estimate the size
of the colony, and precision in population estimates can be evaluated in r
elation to cost effectiveness.