History and References

Outside of a handful of research articles, much of the ongoing work in CKMR is either yet unpublished, or in grey literature that may be hard to find. Our intention here is to provide a list of CKMR-related research, with links where possible.

CKMR was first formulated by Hans Skaug in the early 2000s [1], although Hans credits Tore Schweder for proposing the initial idea. A related approach for estimating paternal abundance under more restrictive conditions (that maternal genotypes are known) was published in the same year [2]. However, CKMR really took off after multiple refinements that first appeared in print 15 years later [3] (not to mention rapid advances in genetic sequencing). The work of Bravington et al. [3] differs from that of Skaug [1, 4] in that kinship probabilities are written as a function of expected relative reproductive output (ERRO) of parents.

In its simplest form, where sampling occurs once and the frequency of parent-offspring pairs is used to estimate abundance, CKMR can be viewed as a Lincoln-Petersen type estimator. One of the earliest published examples of CKMR (written before CKMR was an acronym) used parent-offspring pairs from spawning and juvenile Chinook salmon to estimate abundnace of spawners in exactly this fashion [5]. Extensions to the Lincoln-Petersen estimator that involve multiple years and cohorts were also recently used to estimate the abundance of Brook trout populations [6] which could be compared with alternative removal-sampling estimates of abundance as a form of validation.

However, to date the largest successes in applying CKMR to natural populations have undoubtedly been centered at the CSIRO marine lab in Hobart, Tasmania.

There, a team of statisticians and geneticists (including Mark Bravington) have been extremely successful in applying and extending CKMR to high profile tuna [8] and shark populations [10], including many yet-to-be published examples. They have also conducted scoping studies for additional species, including statistical [e.g., 11, 12] and genetic requirements [13].

Additional work has been done relating close-kin estimates of abundance to effective population size [14] and to investigate potential bias of naive (non-spatial) CKMR estimators under incomplete mixing (limited dispersal) coupled with spatially biased sampling [15]. There is also a number of published or ongoing applications of CKMR to elasmobranch populations [16] and seals.