Some single-stratum removals reversed Bering Sea median latitude trend signs in 1982-2019 point estimates

We tested whether median marine latitude trend signs depend on one survey stratum. We used area weights, fixed taxon sets, two regions and two periods in public OceanAdapt records from 1982-2019. For 62 eastern Bering Sea taxa, omission ranges were -8.99 to +3.97 and -1.15 to +6.61 km per decade. Some removals reversed the full-footprint median sign in each period. For 50 Northeast US fall taxa, all omission medians stayed positive, with ranges +2.13 to +7.72 and +5.85 to +9.14. All four full-footprint 98.75 percent block intervals included zero. Thus, stable point signs did not establish positive population trends. We excluded 16 Bering and 13 Northeast taxa whose centroid failed under some removal. Omission changes the spatial target and cannot separate observation bias from real local change. We do not test climate or pollution causes. The package gives verified source evidence, raw retrieval details, code, uncertainty, checks and a figure.

species distributions · marine biodiversity · observation bias · OceanAdapt · stratum removal · trend sensitivity · uncertainty