Date of Award


Level of Access Assigned by Author

Open-Access Thesis

Degree Name

Master of Science (MS)


Wildlife Ecology and Wildlife Conservation


James R. Gilbert

Second Committee Member

William B. Krohn

Third Committee Member

Joseph T. Kelley


Aerial survey data from 198 1 to 1997 of an increasing harbor seal (Phoca vitulina concolor) population in Penobscot and Blue Hill bays, Maine, were used to evaluate spatial and temporal patterns of use of pupping sites by mother-pup pairs. Pupping sites refer to haul-out sites where pups were observed during surveys, and are assumed to be used consistently from birth to weaning. Sites with pups were spatially clustered to remove spatial auto-correlation, reduce temporal variability and provide biologically cohesive and independent sample units. Spatial, temporal and habitat analyses were completed for 2 spatial scales: individual sites and clusters of sites. The temporal and spatial distribution of the population increase was examined for individual sites (individual site scale) and clusters of sites (cluster scale) to predict behavioral and environmental characteristics that may effect pupping site use at different population densities. On the site scale, Taylor Power Law (TPL) and Analysis of Variance (ANOVA) results revealed that numbers of pups at all sites increased at an ' equal rate providing no strong evidence for aggregation or repulsion behavior. However, increased use of new sites at high densities decreased spatial pattern, implying some degree of spatial limitation. Clusters of sites that were used during all survey years had the highest number of pups and increased steadily with the overall increase, while clusters used less frequently (< all survey years) maintained low numbers of pups and had higher temporal variability. Logistic and multiple regression were used to relate frequency of pupping site use and numbers of pups per site and cluster to physical characteristics of sites. Initial site selection was for clusters of sites with a high availability of haul-out space, access to close alternative haul-outs, and seclusion from humans (significant variables (α = 0.05): area of intertidal zone, number of sites in a cluster, and minimum distance to an site ≥ 2000 ha). Individual site use, however, was poorly related to physical site characteristics, and is likely more dependent on unmeasured dynamic characteristics such as wind direction, surf, tidal phase, human use patterns, harbor seal behavior, and food availability. Physical characteristics of sites are poor to moderate indicators of pup production and distribution on the individual site and cluster scales, respectively. Spatial auto-correlation in pup counts was incorporated into multiple regression models using trend surface analyses. Spatial auto-correlation accounted for a moderate amount of variation in pup counts (r2 ≤ 0.35) on both the individual site and cluster scales. On the individual site scale, this likely represents species aggregation (patchy spatial pattern), while on the cluster scale it likely represents a larger scale spatial gradient across the study area resulting from an unmeasured environmental gradient in resources.