ASSESSING THE VULNERABILITY OF SEABIRDS TO OIL POLLUTION - SENSITIVITY TO SPATIAL SCALE

Area Vulnerability Scores (AVSs) are a measure of the potential impact that oil pollution in an area of sea might have on seabird population s. They are used in the United Kingdom and the Republic of Ireland as a conservation tool, principally to inform the decision making process es related to the licensing of oil exploration at sea and to responses to oil pollution incidents. AVSs are currently calculated and mapped at a scale of 15' latitude x 30' longitude grid squares. Patterns of s eabird distribution are known to be scale dependent and the same is al so true for oil pollution. This suggests that the physical interaction s between oil pollution and seabirds may also be dependent on spatial scale. We consider the sensitivity to spatial scale of the vulnerabili ty that is determined by the interactions between oil pollution and se abirds and examine the scale dependent variation in vulnerability, as defined by AVS. The results demonstrate scale dependence in AVS and so me features of its spatial distribution. These are decreasing heteroge neity in AVS with increasing spatial scale, and the presence of some s patial structure at a scale 20-30 km which could not be detected when estimating AVS at larger scales. These patterns in vulnerability are a consequence of scale dependence in the spatial distribution of seabir d density In addition an increase in scale resulted in a tendency to o verestimate AVS which also lead to an apparent expansion in size of th e higher vulnerabilities such as those around the coast. These results were an artifact of the calculation of AVS. The results of this study have implications for the way in which we use AVS and other similar m easures to assess vulnerability They suggest that calculation of AVS s hould be done on a fine spatial scale, however, bird density data are not available at the appropriate spatial resolution without combining data over an extended temporal scale which runs the risk of distorting the spatial patterns. These effects may be mitigated by calculating A VS on a spatially stratified basis, at high resolution in the areas of high density and coverage, and at low resolution in areas of low dens ities and low coverage.