Modelling the Sensitivity of Colonially Breeding Marine Birds to Oil Spills: Guillemot and Kittiwake Populations on the Pribilof Islands, Bering Sea

Open Access
Publisher Website
Google Scholar
Abstract
Simulation and analytic models were developed to estimate the responses of colonial breeding marine birds to environmental perturbations such as oil spills occurring within the foraging area of the populations. Short-term impacts were considered through a demographic submodel, which simulated changes in population size and age structure through a breeding season and a foraging submodel, which simulated the daily foraging activities of individuals and estimated their daily energy intake and mortality from direct contact with a spill. The demographic submodel partitioned adults and young into age-states. Transfers of individuals between states were produced by immigration, breeding, emigration, growth and mortality. Survivorship of young was modeled as a function of food delivery rate by the adults. The foraging model generated a distribution of the population at sea in accordance with optimization of individual trip times in relation to projected resource levels in areas surrounding the colony. This distribution was perturbed by effecting reductions in food supply or direct mortality of adults in areas affected by an oil spill and the rate of recovery of the population to the optimal distribution and the changes in survival of various age classes projected. Monte Carlo sensitivity tests indicated that model projections were relatively robust to variations in input values, standard errors of adult survivorship and chick fledging success ranging from 0.03 to 0.21. Other sensitivity tests suggested that the model system was relatively insensitive to variations in the probability of adult death following a spill encounter but was substantially more sensitive to variations in the rate at which the population responded to the occurrence of a perturbation by adjusting its foraging distribution and to large changes in the food availability of areas following perturbation. Applications of the model to populations of guillemots and kittiwakes breeding on the Pribilof Islands in the Bering Sea suggested that guillemot mortality was greatest following perturbations located in shallow inshore areas about the breeding islands, while kittiwakes were less severely affected by localized perturbations but were sensitive to spills occurring over a wider area about the islands. Model explorations employing chronic low-level pollution rather than 1-time perturbations suggested that the guillemot population breeding on St. George Island could tolerate an overall reduction of 10% in the food supply without suffering a major decline in fledging success but a reduction of 40% would lead to total reproductive failure in the colony. Specific scenarios of oil spills were made to explore the patterns of population mortality and recovery times projected by the models. Kittiwakes were less sensitive to perturbations, especially those occurring in the immediate vicinity of the breeding colonies. A simple population dynamics model was used in conjunction with the short-term model to estimate the long-term population consequences of 1-time and chronic perturbations. The effects of varying 1-time mortality on guillemots were more severe if adults rather than 1st-yr birds suffered the mortality. If a given level of 1-time mortality was superimposed on a chronic low-level change in survivorship or fecundity, recovery times were markedly longer. Analyses were hampered by a lack of field information on several critical model parameters. These features of seabird biology merit closer attention and their present uncertainty renders model projections general rather than precise.