Validity of inferring size-selective mortality and a critical size limit in Pacific salmon from scale circulus spacing

Size-selective mortality owing to lack of energy reserves during the first marine winter has been suggested to be a result of juvenile salmon failing to reach a critical size or condition by the end of their first marine summer and not surviving the following winter due to this presumed energy deficit. This hypothesis implies strong size dependency of mortality, and is subject to empirical data support for acceptance. Scale circulus spacing has been interpreted as an index for body size, and we reviewed the effect of size-selective mortality with a knife-edge mortality function on descriptive statistics for a scale circulus spacing index (SCSI). In order to invoke size selection as an important driver of mortality during the first year of ocean rearing, it is necessary to demonstrate not only that size-selective mortality is directed towards the smaller members of the population, but that the selective nature of the mortality can account for a substantial portion of the observed mortality. If the assumption is made that a random sample of a single juvenile population has been obtained, then studies that employ a SCSI to infer size-selective mortality coupled with a critical size limit must demonstrate a shift toward larger values of the SCSI, but also a concomitant reduction in the variance and range of the SCSI and an increase in the skewness and kurtosis of the SCSI values. Through simulation we found that the percentage of adults that displayed a SCSI value greater than the maximum observed in the juvenile sample was highly dependent on the initial juvenile sample size and size-selective mortality rate. Geographical distributions of juvenile Pacific salmon can be stratified by size, with larger individuals migrating earlier from local ocean entry locations than smaller individuals, and thus differential timing migration of juveniles based upon body size prior to the collection of the marine juvenile sample may be a more plausible explanation of published trends in the SCSI, rather than invoking substantial size-selective mortality and a critical size limit.