Gentoo penguin colony growth on the Antarctic Peninsula slowed after 2010 and the chinstrap decline continued unchanged in within-site tests of MAPPPD v5.0 nest counts, 1980–2025

Question: did the Antarctic Peninsula penguin trends of the period before 2011 (gentoo increase, chinstrap and Adelie decrease) continue in 2011-2025? Data: MAPPPD version 5.0 (released 21 October 2026, CC BY 4.0), nest counts with accuracy classes 1-3, CCAMLR subareas 48.1 and 48.2, seasons 1980-2025. Methods: within-site (site fixed effects) broken-stick log-linear trends with a knot at 2010, unweighted and weighted by colony size, with bootstrap over sites; early (1980-2010) and late (2011-2025) trends at the same colonies from separate counts; forecasts of counts after 2010 from earlier trends. We wrote the analysis plan before we fitted trends. Results: gentoo growth slowed. At 36 matched colonies, the mean site trend fell from +8.4%/yr to +2.2%/yr (29 of 36 slowed; Wilcoxon p = 0.0005). The size-weighted trend fell from +3.4%/yr (95% CI 2.4 to 4.9) to +1.1%/yr (-0.2 to 2.5; p = 0.024). The slowdown also occurred with knots at 2005 and 2015, with ground counts only, with seasons up to 2022 (before H5N1), and at established colonies. The size-weighted chinstrap decline did not change (-2.8%/yr before, -2.4%/yr after; p = 0.68), and 21 matched colonies gave -2.7 and -3.8%/yr. An unweighted model shows no chinstrap decline after 2010; many small colonies first counted after 2010 cause this result. The Adelie trend after 2010 on the Peninsula is not resolved (weighted 95% CI -5.5 to +7.3%/yr). In the Ross Sea, the Adelie increase continued (+1.4%/yr). Linear extrapolation of site trends from before 2011 did not forecast counts in 2011-2025 better than a no-change forecast for any species, and it predicted gentoo counts too high. Limitations: sparse and irregular counts (10, 21 and 36 matched colonies), non-random survey sites, more UAV counts after 2015, and no test of mechanisms such as density dependence or sea ice. New gentoo colonies are outside the within-site measure. Code, raw data and results are in the package.

penguins · Antarctic Peninsula · Pygoscelis papua · Pygoscelis antarcticus · Pygoscelis adeliae · population trends · MAPPPD · trend persistence · out-of-sample forecast · density dependence · Southern Ocean · seabirds