Whenever a nest is found, the team member places a small coloured marker on the nest. During the first pass, the team found 142 nests. The number of nests on the ground is larger than 142 because not all nests are visible and they can be easily missed. After lunch, the team members change their search routes and again look for nests. If a nest is found on the second pass, the colored marker indicates that the nest was also found on the first pass. The following data were obtained:

Nests found on the first pass: 142.

Nests found on the second pass: 137.

Nests found on both passes: 95 These three pieces of information can be used to estimate the number of nests not seen on either pass using the Lincoln-Petersen estimator. The reasoning proceeds as follows: Of the 142 known nests from the first pass, only 95 were found again. This implies that the detection efficiency of the team is 95/142=67% on each pass. Because 137 nests were found on the second pass, the estimated total number of nests on the island is 137/67%=205. This can be organized into a simple expression:

Estimated total number of nests = (number found on pass 1)(number found on pass 2)/(number found on both passes) = (142)(137)/95 = 205 nests.

As with any statistic, a measure of precision can be attached to this estimate. Using standard statistical methods, we are 95% confident that the number of nests on the island is somewhere between 193 and 219 nests.

This simple method can be extended to more complex situations. There can be more than two sampling occasions; animals can leave or enter the study population between sampling occasions; etc.

This methodology has been extended to study fish (estimating the number of salmon returning to spawn in a stream), fowl (what is the yearly survival rate of migratory waterfowl?), flora (how many rare and cryptic plants are there in a region?), and other wildlife (how many injecting drug users are there in a urban area?).

For further information, please contact Carl Schwarz (cschwarz@stat.sfu.ca), Department of Statistics and Actuarial Science, SFU.