A geometric morphometric approach for detecting different reproductive stages of a free-ranging killer whale Orcinus orca population.

The expansion of drone-based aerial imagery has facilitated an increase in data obtained from free-ranging marine mammal populations, in particular cetacean species. This non-invasive approach allows for body condition assessments, including nutritional and reproductive health. Yet, existing methods of image analysis are time-consuming and lack the granularity to determine early-stage pregnancies and miscarriage rates. In this study, we leveraged a four-year dataset of drone-based aerial imagery paired with known reproductive statuses (i.e., non-pregnant, early-stage pregnant, late-stage pregnant, and lactating) for killer whales (Orcinus orca) to develop a geometric morphometric-based protocol for detecting reproductive status. We demonstrate the significant separation of resulting shapefiles related to reproductive status between all statuses apart from lactating. This approach reliably detects early-stage pregnancy and highlights the morphological locations of major shape changes during the lactation period. We illustrate the applicability of our geometric morphometric protocol for rapid, robust determination of reproductive status in a free-ranging cetacean species. This work helps to satisfy the need for universal tools for non-invasively gleaning population demographic data from free-ranging cetaceans especially of populations which are experiencing prey-related reproductive failures, to understand miscarriage rates and trigger subsequential conservation actions.