Age-specific in situ recruitment of female King Eiders estimated with mark-recapture

In addition to estimating survival probability of adult birds, estimating recruitment of new individuals into avian breeding populations is fundamental to understanding rates of population change. Notions about mean recruitment age can lead to erroneous conclusions about population projections if the probability of capture is ignored. We calculated the mean recruitment age of King Eiders (Somateria spectabilis) using two methods: (1) a naive estimate based strictly on observed age at first recapture of marked ducklings as nesting females; and (2) reversed capture histories, which incorporate probability of capture into estimates. From 1996 to 2009, we marked 2,390 King Eider ducklings, 53 of which were recaptured from 2007 to 2010 as females nesting at Karrak Lake, Nunavut, in Canada's Central Arctic region. The naive approach estimated mean (+/- 95% CL) recruitment age as 4.58 +/- 0.42 yr, whereas reversed capture histories estimated mean recruitment age as 4.08 +/- 0.34 yr. We illustrate the influence of recruitment age (range: 3-9 yr) on the predicted annual rate of population change. We fit numerous ecological covariates to test for cohort effects, phenology of vernal thaw, absolute and relative nesting phenology of mothers, maternal body size, density dependence, and relative clutch size on age-specific recruitment probability. There was good support for a negative effect of relative initiation date of nests that produced ducklings, and equivocal support for an additive negative influence of vernal thaw at the age that ducklings were recruited as breeders. We discuss the implications of variation in female recruitment age for King Eider population biology and fitness. More broadly, we reiterate previous advice (e.g., Pradel et al. 1997, Schwarz and Arnason 2000), against calculation of mean recruitment age from age at first capture, regardless of study species, particularly when detection probability of recruits is low.