Progress report on genetic assignment of humpback whales from the California-Oregon feeding aggregation to the mainland Mexico and Central America wintering grounds
Karen K. Martien, Brittany L. Hancock-Hanser, Morgane Lauf, Barbara L. Taylor, Eric I. Archer, Jorge Urbán, Debbie Steel, C. Scott Baker, and John Calambokidis
International Whaling Commission
Humpback whales undertake annual migrations from low-latitude wintering grounds to higher latitude feeding grounds. Individuals learn their migratory routes from their mothers and exhibit strong fidelity to both feeding and wintering grounds. This migratory behavior results in a population structure best characterized as migratory herds, with members of the same herd sharing the same wintering and feeding grounds and facing the same environmental conditions and threats throughout the entire year. Published genetic studies have focused on feeding and wintering grounds as the strata of interest. However, understanding the relationships between herds is fundamental to understanding the population structure, dynamics, and life history of humpback whales, and to appropriate management. We use mitochondrial sequence data to genetically characterize the two migratory herds that use the California/Oregon (CA/OR) feeding ground and compare them to each other and to the wintering aggregations of which they are a part. We generated two new data sets for this study. The first consists of full mitochondrial genome sequences (16,384 basepairs) from the that feeds off of CA/OR and winters in Central America (henceforth the CentAm-CA/OR herd; n=65), and the herd that also feeds off of CA/OR but winters off of mainland Mexico (henceforth the MMex-CA/OR herd; n=50). The second new data set consists of mitochondrial control region sequences (389 basepairs) from humpback whales sampled off of CA/OR during the 2018 California Current Ecosystems Studies (CCES) survey (n=227). We compare these new data sets to published (Baker et al. 2013) mitochondrial control region data sets collected from the mainland Mexico wintering aggregation (n=62), the CA/OR feeding aggregation in 2004 (n=123), and the CA/OR feeding aggregation in 1988-89 (n=49). Our results show that the CentAm-CA/OR herd and MMex-CA/OR herds are significantly genetically differentiated from each other (FST = 0.054 and 0.044 for full mitogenome and control region sequences, respectively). However, the herds share a high proportion of haplotypes, even when using full mitogenome sequences. Consequently, many individuals cannot be reliably assigned to a herd using only mitochondrial data. The MMexCA/OR herd is significantly differentiated from the mainland Mexico wintering aggregation (FST = 0.044), confirming that wintering ground haplotype frequencies cannot be used as a proxy for haplotype frequencies in the herds that utilize a wintering ground. The 2018 CCES samples did not differ significantly from samples collected from the same feeding aggregation in 1988-89 (FST = 0.015) and 2004 (FST = 0.009), indicating that the genetic composition of the herds occupying the feeding ground has not changed significantly despite a near two-fold increase in abundance during the timespan of the two data sets. Finally, we found evidence that animals that are photographically identified as being part of the MMex-CA/OR herd are not fully representative of that herd. This bias could be due to Central America animals being photographed as they migrate past mainland Mexico and mistakenly identified as mainland Mexico animals, or it could result from spatial or temporal segregation of mainland Mexico animals on the wintering grounds such that some portion of the wintering aggregation is being over- or under-represented in the photographic catalog. Future work should investigate the level of nuclear genetic differentiation between herds, which may allow improved assignment of animals to migratory herd and yield insight to the potential sources of bias apparent in the MMex-CA/OR herd.