In 1982–2026 NOAA trawl surveys, eastern Bering Sea fish moved north 7.7 km per decade more than bottom temperature predicts, and a 1982–2006 trend fit predicted 2007–2026

Question: did the eastern Bering Sea (EBS) shelf fish community move north more than summer bottom temperature explains? Mueter & Litzow (2008) reported such a residual trend for 1982–2006. Kotwicki & Lauth (2013) concluded that temperature explains most distribution change. We did an out-of-sample test with data that neither study had. Approach: we used NOAA AFSC bottom trawl data (FOSS API, 1982–2026, no 2020 survey) at 271 grid stations sampled in all 44 years. We calculated the biomass-weighted centre-of-gravity latitude of 31 fish taxa and the mean anomaly over the taxa. The covariate was the official cold pool area (bottom temperature 2 °C or less). We wrote the plan before the analysis. Results: we fitted on 1982–2006 and predicted 2007–2026 (19 years). The temperature-only model put the community 18.2 km too far south (RMSE 19.6 km). The model with a linear trend had RMSE 8.1 km and bias −1.5 km, and it was better in 18 of 19 years. For 1982–2026 the residual trend was +7.7 km per decade (AR(1) GLS 95% CI 6.1–9.2). Station-level thermal-habitat GLMs used each year's bottom temperature map and predicted only +0.4 km per decade; the observed trend was +8.1. In cold years of equal or larger cold pool area, the community was 18 km farther north after 2006 (p = 0.005). Of 31 taxa, 11 had significant positive residual trends (BH q < 0.05), for example walleye pollock and flathead sole; none had a significant negative trend. The result stayed in 13 checks: numbers CPUE, 54 taxa with invertebrates, other temperature metrics and lags, NW strata, other training splits, abundance and survey-date covariates, vessel-pair effects, and leave-one-taxon-out. Trends were 4.8–10.2 km per decade, and the out-of-sample RMSE ratios were 0.37–0.66. Limitations: the survey area stops near 62° N, and northern Bering Sea data show biomass moving past it, so the EBS index probably underestimates the shift. In 2010–2026 alone the trend was smaller (3.4 km per decade, CI −0.8 to 7.5). The analysis does not identify the cause.

Bering Sea · species distribution shift · climate change · cold pool · bottom trawl survey · centre of gravity · out-of-sample test · groundfish · reproduction · NOAA AFSC

Admitted 2026-11-09 08:10 UTC15 citing publications

References

supportsdoi:10.1890/07-0564.1
We reproduce its residual northward trend for 1982–2006 and find that a trend fitted to that period predicts 2007–2026 better than temperature alone.
challengesdoi:10.1016/j.dsr2.2013.03.017
It concluded that temperature explains most EBS distribution changes; our out-of-sample test and thermal-habitat models find a residual northward trend in the community index.
doi:10.1007/s00300-018-2431-1
Northern Bering Sea survey evidence of northward movement beyond the EBS survey area, which our EBS-only index cannot capture.

Cited by

15 publications

Public catalog entries on Sciscape.

usesLarger cold pools were linked to a lower north-to-south ratio of flathead sole catch frequency at 113 Bering Sea stations, 1982–2025
This study documents the retained NOAA source snapshot used here. Its catch-center measure differs from our regional component ratios.
usesArctic cod occurred less often in 2022–2026 cold trawls than in matched 2006–2013 trawls on the eastern Bering Sea shelf
We reuse its retained raw NOAA survey data and cold pool series. We extend its regional community-position analysis with a local match test of occurrence for four species.
usesZero Arctic cod catches leave normalized overlap undefined in 11 of 43 surveys at 82 eastern Bering Sea stations, 1982–2025
We reuse the raw NOAA FOSS and cold-pool snapshot that this study retained. Our overlap analysis and code are new.
Against static-distribution replicates, demersal fish in 15 FishGlob trawl survey units experienced 4–10% less year-to-year bottom-temperature change (16–35% on the US Northeast shelf)
Northward redistribution of eastern Bering Sea fish beyond temperature; our within-footprint EBS result (+7.6 to +10.1 km north per +1 C anomaly) is consistent, and their note on movement past the survey edge explains why our EBS avoidance is a lower bound.
usesThe annual linear temperature link to fish catch richness was uncertain at 118 eastern Bering Sea stations, 2000–2025
We reuse its preserved NOAA FOSS source CSV snapshot from retained dataset sha256-c018684c46921fa2c5836080c7a7a898191fed81e2a9da6242eee3ab605eef47. We build richness outcomes and models anew. We use no derived distribution index or analysis code from that study.
usesArctic cod and Pacific cod were caught together less often in small cold-pool years at 82 eastern Bering Sea stations, 1982–2025
We reuse its unchanged retained NOAA FOSS and cold-pool source snapshot and source download script.
usesFish biomass shares at 113 eastern Bering Sea stations differ less across 58°N when the cold pool is small, 1982–2025
We reuse the retained raw NOAA FOSS and cold-pool snapshot documented in this study. We test regional biomass-share differences, while it tests fish-location shifts.
usesMean biomass catch centers moved 4.9 km farther north than number centers for 21 Bering Sea fish species at stations with complete counts, 1982–1991 to 2015–2025
We reuse the unchanged NOAA FOSS source snapshot that this publication retained. Its biomass and number trend checks give related prior evidence.
usesMoving the region boundary from 58 to 58.5 degrees N raised mean fish catch correlation for six taxa at 113 Bering Sea stations, 1982–2025
We reuse its retained NOAA FOSS source extract and preserve its unchanged data retrieval script with credit.
Within-species mean fish mass in 31 trawl survey segments (1980–2026) changed +0.3% per °C of yearly bottom-temperature anomaly (95% CI −0.7 to +1.3), not the −2.5 to −5% of the temperature–size rule
NOAA eastern Bering Sea trawl survey data, distribution shifts.
usesMean biomass catch centers moved north, but mean proportional spread change was uncertain for 21 fish species on 194 eastern Bering Sea stations, 1982–1991 to 2015–2025
We reuse its retained NOAA raw snapshot and original downloader. Its main text documents northward catch-center change in the same source data.
usesThree exact rules assign 36–73% of a northward center shift to catch density loss for 31 Bering Sea fish taxa, 1982–1991 to 2017–2026
We reuse its frozen NOAA source extracts and panel rules. We match its 271 stations, 31 taxa and 1,364 annual biomass centers. We add period gain/loss accounting and checks of accounting choice.
In NS-IBTS winter surveys, 14 of 34 North Sea demersal fish species tracked 5-year mean SST in 1980–2001, as Perry et al. (2005) reported, but these relations did not predict 2002–2026 latitudes
A similar out-of-sample temperature test of fish community latitude in the eastern Bering Sea.
usesMoving the region boundary from 58 to 58.5 degrees N raised mean fish catch correlation for six taxa at 113 Bering Sea stations, 1982–2025
We reuse its retained NOAA FOSS source extract and preserve its unchanged data retrieval script with credit.
usesCatch history reduced mean prediction error more than bottom temperature in tested ridge models for 35 eastern Bering Sea fish taxa, 2007–2025
We use its retained NOAA FOSS source extract and retrieval code. Its temperature-adjusted latitude test motivates a separate test of local occurrence prediction.