ROBERT A. GARROTT

Professor        Fish and Wildlife Management

Ecology Department--Montana State University

Office Phone:  (406) 994-2270        FAX:  (406) 994-3190        e-mail:  rgarrott@montana.edu

 

 

 

 

 

Dr. Garrott's Home Page

 

 

 

 

 

Graduate Research

Current Projects

Recently Completed Theses

Prospective Student Information

Available Graduate Positions

Funding Partnerships

 

 

Mass estimation of Weddell seals using techniques of photogrammetry

Darren Ireland with Robert A. Garrott and Jay Rotella

 

This project represents the first step towards linking the population dynamics of Weddell seals in Erebus Bay to local and large scale environmental fluctuation. For over thirty years the population of Weddell seals in Eerebus bay has been studied using mark-recapture techniques. Trends in population vital rates have been identified but little information on factors which may influences these trends, such as resource availability and environmental characteristics, has been collected. Direct measurement of Weddell seal food resources is nearly impossible because the pelagic, mid-water fish species, Pleuragramma antarcticum, on which adult Weddell seals depend, are located hundreds of meters below the ice-covered ocean surface. As an alternative, a large sample of weights from individual animals may be used as an indicator of resource availability. Annual changes in resources should be reflected in the body mass of seals returning to breeding colonies each year.

 

In order to collect mass measurements on Weddell seals, previous researchers have restrained and anesthetized the animals, or coaxed them onto a livestock weighing platform. Both methods are laborious and risky which has limited samples sizes. Since the seals show little variation in overall body shape and do not have long pelage that distorts the outline of their body, morphometric measurements can be made from photographs. By collecting a set of images from individual seals and then weighing those animals immediately following photography, morphometric measurements can be regressed against the known weight of the seal to create a predictive equation of mass. This will allow the annual collection of mass estimates from a large sample of seals across the entire population. The mass information and other environmental data such as fast-ice extent, can then be related to population characteristics on an annual basis.

Three angles of seal photographs.

 

Figure 1.  Three angles of seal photographs.

 

Boom system used in seal photography.  Second boom system used in photographing seals.

           Figure 2.  Two boom systems used for photographing seals.

Digital photographs of seals are taken from three different angles: overhead, head/tail looking down the long axis of the seal, and side looking perpendicular to the long axis of the seal (Figure 1).  The photographs are taken using two boom systems which allow the images to be collected in a standardized manner to reduce the effects of image distortion on photogrammetric measurements (Figure 2).

 

Three measurements are then collected from each image (Figure 3). During the first field season, 32 adult females were successfully photographed and weighed along with 67 pups (Figure 4). Preliminary analysis of the adult data has shown that the body width measurement from overhead photographs is the best single predictor of mass (R2=0.83, N=23) (Figure 5). Models containing five and six predictors have explained almost 98% of the variation in mass. Effects of head and body position in photographs and other sources of variation will be analyzed to determine the optimal position of the seal for photography and the photographic angles necessary to capture the most significant morphometric measurements.

Three photogrammetric measurements taken on seals.

 

Figure 3.  Three photogrammetric measurements taken on seals.

 

 

 

Weighing of Weddell seal.

Figure 4.  Weighing of a Weddell seal.

Regression of weight vs. overhead width measurements

Figure 5.  Weight vs. Overhead Width measurements.

With a properly functioning scale, a much larger sample size of known mass animals and associated photographs will be collected in the coming field season (Oct – Dec 2003) to further refine the predictive equations for mass. Additionally, the technique will be applied to a sample of mother pup pairs that are photographed at parturition and again at weaning to measure the effect of adult female parturition weight on the rate of mass transfer to offspring and subsequent offspring weaning weight. With an accurate predictive equation, mass estimation through photogrammetry will then continue on an annual basis as a measure of food resource availability and correlations with population dynamics will be explored.

 

Funding:  National Science Foundation