Report of a cetacean survey in the Western Approaches of the ...

Measuring the effect of attractionResponsive movement is a major issue for common dolphin surveys in general. The very narrow halfstrip widths, 294m and 235m from 2004 and 2005 respectively suggested that further investigation wasnecessary.We pooled data for primary sightings of common dolphins in sea state ≤3 (to make sure that higher seastates were not affecting the data), excluding re-sightings and sightings aft of the beam during fastspeed mode (following predetermined transects or when following straight lines) for the differentsurvey years. The perpendicular distance plots (see Fig. 9) show a substantial peak in the first bin (lessthan 100m). This peak is much more pronounced than would normally be expected for a sightingsurvey of this type and is consistent with substantial responsive movement. Distance plots from thesecondary platform (see Fig. 11), where one would expect to sight animals at greater distances andbefore responsive movement, show a substantial peak in the first bin (less than 100m) for theperpendicular data, whereas this is no longer evident for the radial distance plot (Fig. 11).We explored responsive movement further by using a vector component of the dolphin’s velocity awayfrom the vessels (i.e. the cosine of the difference between bearing and heading, where a value of ‘1’indicates a heading directly away from the vessel, ‘-1’ directly towards and ‘0’ perpendicular).The results for the primary platform (2005) and trackers platform (2005), depicted in Figures 12a&b,show a distinct large peak close to ‘-1’, i.e. the majority of sightings (excluding re-sightings) areapproaching the vessel. When sightings closest to the survey vessel (< 100m radial distance), which areexpected to be most strongly attracted, are excluded from the primary data, the effect is still apparent(see Fig. 13a). When only sightings with a distance in the 25 percentile (> 400m) furthest from the boatare included in this analysis, the effect for the primary platform remains (see Fig. 13b), whilst it nolonger apparent for the trackers platform for distances > 1,100m or > 1,500m (see Figures 14a&b).Similarly, findings for primary 2004 sightings, with a distance in the 25 percentile farthest from theboat are also shown (SLOW+TS (slow) >420m and T+S (fast) >650m). Findings for harbour porpoiseprimary sightings made during the 2004 survey are also shown (Fig.16) these were used as a indicatorof the reliability of observed swimming directions (something that is notoriously difficult to do at sea)and were found to be comparable to other studies (e.g. Palka and Hammond).Different survey speeds are thought to influence the degree of responsive movement. From the 2004primary data, it appeared that there are two effects when comparing the two survey speeds (De Boer etal., 2005). One is a ‘movement’ effect and the other is a ‘sightability’ effect. There is no evidence,from the heading data for fast speed mode in 2004 to suggest that dolphins were detected beforeresponsive movement. This was also found in 2005 (fast speed data only). For the slow speed data in2004, however, there is an indication that dolphins were detected before responsive movement. Theradial distance plot is more peaked (see Fig. 17) compared to fast speed mode, and, stemming from theidea that surfacing behaviour closer to the survey vessel changes your ability to sight an animal (DeBoer et al., 2005), we concluded that this is probably due to an observational/detectability effect.Indeed, it could well be that dolphins approaching a fast moving vessel are more likely to surface in the‘middle class’ of distances (around 200-300m) whereas dolphins that are approaching a slow movingvessel tend to surface much closer (