The importance of observer effort on the accuracy of breeding success estimates in the Common Guillemot Uria aalge

• Published in: Bird study 2020-01, Vol.67 (1), p.93-103
• Main Authors: Harris, Mike P., Heubeck, Martin, Bogdanova, Maria I., Newell, Mark A., Wanless, Sarah, Daunt, Francis
• Format: Article
• Language: English
• Subjects: Age; Breeding success; Chicks; Colonies; Juveniles; Monitoring; Sampling; Sampling methods; Seabirds; Statistical analysis; Success; Survival; Thinning; Uria aalge
• ISSN: 0006-3657; 1944-6705
• DOI: 10.1080/00063657.2020.1779654

Capsule: A data-thinning approach was used to assess the effects of reducing the frequency of nest-checks on estimates of breeding success of Common Guillemots Uria aalge. Inter-year and inter-colony differences in fledging age and their implications for setting a minimum age after which a chick could be assumed to have fledged were evaluated. Aims: To assess the consequences of reducing sampling frequency on the estimation of breeding success, and on the robustness of the assumption that breeding has been successful if a chick survives to 15 days old. Methods: Breeding success, ages at fledging and loss of chicks were estimated from daily checks at two Scottish colonies over a six-year period. Data-thinning was used to assess the consequences of reducing checks from daily to every two or three days. Breeding success was recalculated assuming that all chicks surviving to 15 days fledged. Results: Reducing the frequency of checks from daily to every second or third day resulted in a small but statistically significant increase in the estimate of success. Between 20% and 25% of chick losses occurred when chicks were ≥15 days old. Assuming that these chicks had fledged resulted in significant increases in breeding success. Conclusion: Assumptions about whether or not a chick fledged had a greater impact on the estimate of breeding success than reducing the frequency of nest-checks from daily to every second or third day. There was no threshold age after which a chick could be assumed to have fledged. These findings are relevant to other monitoring schemes where there is pressure to reduce input. Sampling methods used in monitoring schemes need to be clearly stated and changes in protocols documented so that sampling effects can be incorporated into future analyses.