Site‐dependent regulation of breeding success: Evidence for the buffer effect in the common guillemot, a colonially breeding seabird

1. Density‐dependent regulation can offer resilience to wild populations experiencing fluctuations in environmental conditions because, at lower population sizes, the average quality of habitats or resources is predicted to increase. Site‐dependent regulation is a mechanism whereby individuals breed at the highest quality, most successful, sites, leaving poorer quality, less successful sites vacant. As population size increases, higher quality sites become limiting but when populations decline, lower quality sites are vacated first, offering resilience. This process is known as the ‘buffer effect’. However, few studies have tested whether such regulation operates in populations experiencing changes in size and trend. 2. We used data from a population of common guillemots Uria aalge, a colonially breeding seabird, to investigate the relationship between site occupancy probability, site quality and population size and trend. These data were collected at five sub‐colonies spanning a 38‐year period (1981–2018) comprising phases of population increase, decrease and recovery. 3. We first tested whether site quality and population size in sub‐colonies explained which sites were occupied for breeding, and if this was robust to changes in sub‐colony trend. We then investigated whether disproportionate use of higher quality sites drove average site quality and breeding success across sub‐colony sizes and trends. Finally, we tested whether individuals consistently occupied higher quality sites during periods of decline and recovery. 4. Higher quality sites were disproportionality used when sub‐colony size was smaller, resulting in higher average site quality and breeding success at lower population sizes. This relationship was unaffected by changes in sub‐colony trend. However, contrary to the predictions of the buffer effect, new sites were established at a similar rate to historically occupied sites during sub‐colony decline and recovery despite being of lower quality. 5. Our results provide support for the buffer effect conferring resilience to populations, such that average breeding success was consistently higher at lower population size during all phases of population change. However, this process was tempered by the continued establishment of new, lower quality, sites which could act to slow population recovery after periods when colony size was low.