Paired gentoo and chinstrap nest-count changes had a correlation of 0.06 at 20 Antarctic sites in 1980–2025

We tested whether gentoo and chinstrap nest counts changed together at the same Antarctic sites. We used MAPPPD version 5.0 nest counts from CCAMLR subareas 48.1 and 48.2. The primary sample had 225 intervals at 20 sites in 1980–2025, with gaps from one to ten years. We calculated annual changes in log(count + 1), removed each site mean, and gave each site equal total weight. The correlation was 0.060. The nominal 95% whole-site resampling range was −0.120 to 0.230. Ground counts gave 0.080, and accuracy classes 1–2 gave 0.031. Requiring all four endpoint counts to be at least 20 gave 0.272 at 12 sites. Those same sites gave 0.199 before the count threshold, so site selection accounted for much of the difference. Complete-date and longer-gap checks gave negative estimates with wide site ranges. Only five spatial blocks were available, and uncertainty depended on the resampling method. The main estimate was weak and did not retain its direction across all planned checks. The records do not establish a common direction of local change or identify competition. The package has unchanged sources, source-row links, all panels and draws, figures, code, provenance, and offline reproduction instructions. Independent calculations passed. A second source-based run matched all 81 result and figure files.

penguins · nest counts · MAPPPD · population change · covariance · survey coverage · reproducibility