116 ( ) = ( ) ∗ ( ) ( ) (2) 3.3.3 Statistical analysis of nesting site selection We have performed principal component analysis (PCA) to test the influence of the ‘islet’ and ‘habitat’ variable types on the number of nests within the GLS and to assess which variables were most influential over the nesting site selection. First, inter-site selection was investigated with a PCA relating the predicted seasonal number of nests per islet with the ‘islet’ variables. Since the islets’ geographical features and position are considered stable over time at the scale of this study, we used the mean value of each selected variable over the six distinct seasons, including for the number of nests. Second, intra-site selection was investigated with a PCA relating the number of nests predicted per sector with the ‘habitat’ variables. The first two dimensions are retained as explanatory variables in subsequent model evaluation (Berr 2023). All statistical analyses were conducted in R 4.2.2 (R Core Team, 2023), with particular use of the packages ade4 (Dray and Dufour 2007) for PCA construction. 4. Results 4.1 Spatiotemporal distribution of nesting activities Over the six nesting seasons covered by this study, 51 field surveys were conducted in the GLS nesting area, for a total effort of 98 days and over 330 hours of beach survey, leading to the census of 1,105 successful nests and 1,081 aborted nests over the 28 surveyed islets (Tableau 8). Tableau 8. Summary of seasonal monitoring effort and outputs over the six considered seasons within the GLS. Nesting season Number of field surveys Number of surveyed islets Number of islets with a confirmed nesting activity Number of successful nests Number of aborted nests 2017-18 5 11 11 156 99 2018-19 5 22 22 187 143 2019-20 9 21 18 131 252 2020-21 9 24 19 163 213 2021-22 11 24 21 247 177 2022-23 12 25 21 221 197
RkJQdWJsaXNoZXIy MjE1NDI=