A warmth link to lower spatial community difference was sensitive to analysis choices in Bering Sea trawls in 2001-2019

We tested whether warm years had more similar recorded communities across fixed trawl survey strata. We used 21,372 hauls in the eastern Bering Sea and Northeast US fall surveys from 1982 to 2019. We calculated the exact expected mean pairwise Sorensen difference for one haul per stratum. Four region-period models controlled linear year. Only late Bering had a negative interval under the main model: -0.285 index points per C, with a four-test interval of -0.543 to -0.027. One point is 0.01 on the zero-to-one index. That interval included zero with one covariance lag, year-to-year differences, or added sampling-location controls. Point estimates stayed negative in those checks. Other main intervals included zero. These results do not give robust repeated support for a common warmth link to spatial convergence. They do not establish unequal true slopes or causal warming effects. Detection, gear changes, restricted strata, coarse surface temperature and short annual series limit inference. Code, haul matrices, source hashes, all estimates, independent checks and two figures are included.

biodiversity · spatial community difference · beta diversity · trawl survey · sea-surface temperature · observation bias · sampling effort · Bering Sea · regional replication