Ocean surface winds drive local-scale movements within long-distance migrations of seabirds

Abstract

Long-distance migration is a major part of the life cycle of many seabirds. The main processes driving local movements within those long-distance migratory movements are essentially unknown. Here, we studied detailed patterns of the movements with respect to distance from land of the most abundant seabird species migrating across the northernmost part of the Strait of Gibraltar and analysed how ocean surface winds influence those patterns. We did this by using visual and S-band radar surveys. Our results show that seabirds followed lines of travel that were located nearer the coast than randomly expected. Re-sampling techniques and comparison with additional data from ship-based counts corroborated this pattern, which was not substantially affected by the decrease in detection at distances of up to 3,000 m. Wind direction and speed covaried with local patterns of flight trajectories in a general manner. All the seabirds responded to headwinds by approaching the coast in proportion to the magnitude of wind intensity. Such a change in flight patterns could be a strategy to reduce the effect of headwinds, by approaching the coast where wind intensity was reduced by orographic factors. Under tailwind conditions, seabirds tended to fly further from the coast, profiting from increasing winds further from shore. Our results imply that modification of off-shore distance in relation to conditions of ocean surface winds may be an energetically advantageous strategy for migrating seabirds. Off-shore distances were also dependent on global and local migratory behaviour of different species, but not on flight type.

Appendix: Ship transects surveys across the Strait of Gibraltar

From 2002 to 2006, within the project “Migres Marinas”, staff from the University of Cádiz and the Migres Foundation carried out ship transects across the Strait of Gibraltar from commercial ferries covering the line Algeciras–Tangier (unpublished data; Fig. 5). In all these surveys, the strip transects technique described by Tasker et al. (1984) was used. This method consists of continuous counts along the transect line, where all birds detected within a pre-set perpendicular distance (the strip width is 300 m to both sides of the line, 600 m in total) are recorded. Continuous counts were subdivided into 10-min sequences, with ship position being recorder at the beginning, the middle and the end of each sequence.

Fig. 5

Sketch of the Strait of Gibraltar with the commercial ferry route between Algeciras and Tangier from which strip transects were carried out (the northern route is the one that the ferry follows in the transect Algeciras to Tangier; the southern one is that from Tangier to Algeciras). From all the positions measured at the beginning, at the end and in the middle of each observation sequence, we were able to map the mean trajectory (broken line) and the limits of 90% variation around the line (dotted line). The transect piece is shown as a black line. The number of spotted birds at each latitudinal location of the ship was transformed into distance from our reference point at Tarifa Island (along the N–S line from Tarifa towards Africa). A small line shows where the 3,000 m distance from Tarifa is located, at which data for further analysis were truncated

For the purpose of comparing seabird distribution during autumn migration obtained from land counts and from ship transects, trips carried out in the months of October and November from 2002 to 2006 from Algeciras to Tangier were selected, constituting a total of 14 trips. Transect’s zones with an N–S direction were eliminated. The transect starting at the northernmost point of the NE–SW transect and stopping at Tarifa longitude was used (Fig. 5). The number of spotted birds during each 10-min sequence was pooled, and the geographic position measured in the middle of the corresponding sequence was taken as the position of this group of birds. The latitudes derived from the ship transects were transformed into distances south of the Tarifa reference point. To do it, the latitude at the transect’s northernmost point was consider as distance cero from Tarifa and the distance at Tarifa longitude as it correspondent distance to Tarifa coast (Fig. 5) and apply the follow equation:

$$ \begin{array}{*{20}c} {{ \min } . {\text{latt}} \to {\text{ransect}}\,{\text{start}}} \hfill \\ {{ \max } . {\text{lat}} \to {\text{transect}}\,{\text{end}}} \hfill \\ {{ \min } . {\text{distance}} \to {\text{distance}}\,{\text{start}}(0)} \hfill \\ {{ \max } . {\text{distance}} \to {\text{distance}}\,{\text{across}}\,{\text{Strait}}\,{\text{from}}\,{\text{bird}}\,{\text{observatory}}\,{\text{to}}\,{\text{transect}}\,{\text{end}}} \hfill \\ {{\text{lat}} = {\text{distance}}/({ \max } . {\text{distance}} - { \min } . {\text{distance}}) \times ({ \max } . {\text{lat}} - { \min } . {\text{lat}}) + { \min } . {\text{lat}}\,{\text{return}}({\text{distance}})} \hfill \\ \end{array} $$

In this way, position from coast on ferry route was translated into distance from bird observation station. Distances obtained were pooled into distance intervals of 500 m and were truncated at 3,000 m. This conversion could be done, because field evidence shows that the majority of seabirds maintain an east–west direction (either E–W or W–E) when crossing the Strait, thus crossing the NE–SW part of the ferry transect at an angle of about 10–20°. The vast majority of the tracks recorded in the Tarifa Island by the radar followed the same E–W or W–E direction (Mateos 2009). Also, the birds/flocks visually observed from the Tarifa Island throughout the Migres Marinas program followed E–W directions. There are no indications of birds maintaining their off-shore distances and following the NE–SW direction of the coastline to the E of Tarifa. Obviously, birds/flocks approaching the reference point from E or W while flying further north than the latitude of the Isla de Tarifa have to redirect their flight course slightly southward to avoid flights across land. Our suspicion is that they use the Tarifa island as a geographical reference point, while taking up their previous E–W or W–E course after passage.

For the comparison of land-bound observations and ship transects, we included the total number of birds spotted within each interval both from land and from ship, standardising the numbers with the available observation time per distance interval. The number of birds per interval, standardised by the observation time in that interval and given as a percentage of the standardised total within the N–S distance of 3,000 m from the reference point were compared between land and ship surveys by using χ ² test.

To determine whether the probability to detect birds further away from Tarifa decreases with increasing wind in the Strait of Gibraltar, by using Distance Sampling procedure (as explained in “Methods” section) all ship transects were used, both those form Algeciras to Tangier and from Tangier to Algeciras during the whole year (46 trips from Algeciras to Tangier and 46 from Tangier to Algeciras).