Photogrammetry as a tool for estimating size and condition in the Hawaiian monk seal (Monachus schauinslandi), the harbor seal (Phoca vitulina), and the northern elephant seal (Mirounga anguistirosis)

Abstract

A photographic technique was developed to indirectly estimate mass and body condition of the endangered Hawaiian monk seal (Monachus schauinslandi). This project differs from other photogrammetric studies because subjects were photographed under natural field conditions without the use of chemical immobilization. Captive northern elephant seals (Mirounga angustirostris) and harbor seals (Phoca vitulina) were also sampled to create a general predictive model of mass for phocid seals. Newly weaned Hawaiian monk seal pups were photographed and measured in the wild. Morphological variables (length, side area, anterior/posterior area, and perimeter) were measured from photographs to establish a predictive relationship with body mass. Photographic (i.e., distance from the seal to the camera) and morphological deviations in body position were tested in the field and did not reduce the precision in producing a photograph which accurately reflected body condition. Photographs of all three species were subjectively ranked in terms of quality, based on the degree to which the seal's body position deviated from the ideal body position used to obtain photographs. Results indicated that overall, deviations in body positioning (e.g., a seal rolled on its side) did not significantly affect photogrammetric values compared to those obtained in a standardized position. Both photographic and morphometric variables (together and separately) were used to create species-specific regression models to predict mass for the Hawaiian monk seal, harbor seal, and northern elephant seal. Regressions of mass created for all three species yielded useful, predictable models. A general phocid model was created using the Hawaiian monk seal and elephant seal populations. Although combining photogrammetric and sociometric variables produced the most reliable models (based on 95% confidence intervals), the most accurate models (based on 42 values) contained only morphometric variables. However, many practical and reliable models were created using only photogrammetric variables. This finding indicates that the use of photogrammetry alone to assess body condition may be suitable to indirectly estimate mass and body condition.