Three-year cluster intervals crossed zero for temperature slopes of a copepod contrast in four US shelf samples in 1977-2023

We tested whether warmer samples had more Centropages typicus relative to Calanus finmarchicus. We used 22,364 EcoMon tows in two latitude bands and two periods from 1977 to 2023. Models controlled cell by month, year, bottom depth, day within month and time of day. All four slopes were positive, from 0.021 to 0.378 log contrast units per degree C. With year clusters and correction for four tests, only the early northern interval was above zero. With adjacent three-year clusters, all four intervals crossed zero. Salinity control, finer cells, shorter seasonal bins, cruise effects and a larger abundance offset did not give clear repeated evidence across all four samples. An independent fit reproduced the primary slopes. This is a test of local temperature associations, not a causal estimate of climate change. One shared survey, coarse habitat controls, missing temperature and low numbers of time clusters limit inference. Air pollution and extreme heat duration were not tested.