Hidden Hippos: Using Photogrammetry and Multiple Imputation to Determine the Age, Sex, and Body Condition of an Animal Often Partially Submerged

Abstract

Demographic Information on threatened species is important to plan conservation actions. Due to their aquatic lifestyle, the subtle nature of hippo sexual dimorphism, and their occurrence in inaccessible areas, it is difficult to visually determine hippo ages and sexes. Previously, hippo body lengths have been measured from drone images and used to estimate age. However, due to hippos’ propensity to be partially submerged, it is often difficult to obtain the required measurements. We used the novel technique of multiple imputation to estimate missing body measurements. Further, we explored if male and female hippos could be differentiated in drone images based on body proportions, also examining body condition indices and how these varied seasonally. Multiple imputation increased the number of hippos that we aged threefold, and the body lengths we obtained fell within the range provided in literature, supporting their validity. We provide one of the first age structure breakdowns of a hippo population not from culled hippos. Accounting for overall size, males had wider necks and snouts than females. Hippo body condition varied seasonally, indicating responses to resources and reproduction. We provide a new technique and demonstrate the utility of drones to determine age and sex structures of hippo populations.

Keywords

• photogrammetry
• multiple imputation through chained equations
• missing data
• hippos
• demographics
• drone