Effects of competitive pressure and habitat heterogeneity on niche partitioning between Arctic and boreal congeners

The rapidly changing climate in the Arctic is expected to have a major impact on the foraging ecology of seabirds, owing to changes in the distribution and abundance of their prey but also that of competitors (e.g. southerly species expanding their range into the Arctic). Species can respond to interspecific competition by segregating along different niche axes. Here, we studied spatial, temporal and habitat segregation between two closely related seabird species: common guillemot Uria aalge (a temperate species) and Brünnich’s guillemot Uria lomvia (a true Arctic species), at two sympatric sites in Iceland that differ in their total population sizes and the availability of marine habitats. We deployed GPS and temperature-depth recorders to describe foraging locations and behaviour of incubating and chick-rearing adults. We found similar evidence of spatial segregation at the two sites (i.e. independent of population sizes), although segregation in environmental space was only evident at the site with a strong habitat gradient. Unexpectedly, temporal (and, to a limited extent, vertical) segregation appeared only at the least populated site. Overall, our results show complex relationships between the levels of inferred competition and that of segregation.

colony from the GPS data (see below) was used in conjunction with TDR data to confirm classifications.
Sea surface temperatures at diving locations For each section identified as "on the sea surface", separated by segments of flight, sections corresponding to diving were removed and the average surface temperature recorded by the TDR was calculated. Dives were assigned to their corresponding "on the sea surface" period to analyse dive patterns in relation to SST.

Tests of interspecific overlap along each axis
Whether the two species overlapped less than by chance was tested following Geange et al.
(2011), i.e. by permuting species labels across individuals and re-calculating the overlap indices, repeating the operation 1,000 times and estimating the pseudo-p-value as the proportion of simulated overlap values that are lower than the observed. A similar permutation method was used by [5] to test for the significance of overlap values derived from auk foraging data. At Látrabjarg, the same permutation approach was used to contrast the use of the ice edge by the two species, in this case contrasting the observed difference between BG and CG in proportions of dives < 5 km away from the ice edge, with its simulated equivalent.

Test of differences in overlap between sites (bootstrap approach)
Overlap values were compared between sites or breeding stages using a bootstrap procedure.
Each iteration worked as follows: for each site (or breeding stage) and species, individuals were sampled with replacement, maintaining the observed sample size (e.g. 33 sampled CG at Látrabjarg) and calculated the overlap indices for each horizontal, vertical, temporal, and habitat axis as described above. After 1000 iterations, for each axis, 95% confidence intervals (CI) were calculated using the 2.5 th and 97.5 th percentiles of the overlap indices at each site or breeding stage and of the differences between sites or breeding stages. If the 95% CI of the inter-site (or inter-stage) difference did not overlap with zero, the difference was considered as significant.

Drivers of temporal segregation
Temporal segregation appeared between the studied BG and CG only at Grímsey. This segregation could be driven by different prey preferences, as diel activity patterns are usually related to foraging on prey performing diel vertical migration (DVM), as was found in both CG and BG at lower latitudes [6], [7]. However, Grímsey is far enough north to experience the midnight sun in midsummer and whether diel vertical migration continues in these conditions is debated [8], [9].
Alternatively, heterogeneous diel activity patterns can persist in BG even in the high Arctic in the form of sex-specific inverted rhythms [10]. If our sampling was biased toward one sex for this species at this site, these temporal segregation patterns might reflect more sexual segregation that interspecific segregation. Unfortunately, the sex of the tracked birds is unknown. Subtle morphological differences between the sexes exist [11], but the overlap between the sexes remains quite high and variations between colonies can be larger than within-colony variations between the sexes [11], preventing us from exploiting morphometric data to this end.
The distribution of deployment times at the two sites could suggest some possible bias towards one sex. Indeed, temporal segregation between the sexes in BG is often related to one sex attending the colony mostly at night while the other attends the colony mostly during the day [10], [12]. Deployment times at Grímsey are more narrowly concentrated than at Látrabjarg ( Fig. S8), which could suggest a potential bias towards one sex, but only molecular sexing could allow us to be sure.
Nevertheless, when repeating the analyses at Látrabjarg with only the subset of individuals that were captured during the same hours of the day as birds at Grímsey, the results did not change.
Temporal overlap between BG and CG was 0.81, and was still not significantly lower than expected by chance (p-value = 0.138). If the temporal segregation observed at Grímsey was only due to biased sampling towards one sex, we would have expected temporal segregation to appear at Látrabjarg when biasing the sampling in the same way, but we did not observe this.
This suggests that the sex of the birds is likely not the only factor responsible for the difference between the two sites.

Arnarfjӧrður
Arnarfjӧrður is a popular fjord for fisheries targeting the Northern shrimp Pandalus borealis [13], even though the stocks have recently declined [14]. The species is a known prey of Brünnich's guillemots in Svalbard [15], so BG from Látrabjarg could be targeting this species in Arnarfjӧrður. Note that another fjord accessible from Látrabjarg, Ísafjarðardjúp, is also known for its Northern shrimp stock [16]. However, conspecific seabirds at sea tend to segregate according to their colony of origin, and Ísafjarðardjúp is within foraging range of the very large colonies in the Hornstrandir [17], which could explain why Arnarfjӧrður is the only fjord used by birds from Látrabjarg.