Early latitude rates gave larger later errors than a common-rate rule in two US trawl surveys in 1982-2019

We tested whether taxon latitude rates in 1982-2000 carried over to 2001-2019. We used fixed survey strata, area weights and 59 eastern Bering Sea taxa and 62 Northeast US fall taxa. Early and late rates had correlations of -0.252 and -0.202. A rule that carried each early taxon rate forward gave larger squared errors than a rule that used the early regional mean rate for all taxa. Its scores were -2.637 and -1.158, where zero means equal error. Shared-year block intervals for the correlations were -0.500 to 0.149 and -0.362 to 0.061. Both included zero. This is a retrospective test of recorded rates in a fixed cohort. It does not identify true movement rates or climate causes. Taxon selection uses the full record. Short records, survey changes, catchability, regional dependence and missing taxa limit use. The package gives the annual panel, source hashes, raw-data build code, all estimates, a figure and arithmetic checks.

species distributions · temporal transfer · OceanAdapt · observation bias · trawl surveys · range shifts