The adjusted krill–salp log-density link was uncertain at 16 Palmer survey cells in 1993–2024

We tested whether larger salp catches were linked to smaller Antarctic krill catches after control for location, survey year and tow conditions. We used 750 standard January–February net tows at 16 repeated Palmer LTER cells west of the Antarctic Peninsula in 1993–2024. Each observed cell-year had equal total weight. The primary model used log10(density + 1) for both taxa. The pooled salp slope was −0.227. With cell effects, year effects, survey day, daylight and filtered volume, the slope was −0.099, with a 95% three-survey block interval of −0.229 to +0.022. The adjusted partial correlation was −0.079. Controls reduced the absolute point slope by 56%. Salps had zero catch in 383 tows. With an offset of 0.1, the adjusted slope interval excluded zero; with an offset of 10 it included zero. Among 367 positive salp catches, the slope was −0.005, with an interval of −0.131 to +0.120. The primary result permits a material negative relation and does not establish equivalence. Pooled catch relations alone do not establish local replacement or competition. Code, source hashes, all estimates, bootstrap draws, figures and independent checks are preserved. Original source files remain external because their metadata have use restrictions.

Antarctic krill · Salpa thompsoni · Palmer LTER · catch density · zero catches · fixed effects · time-block bootstrap · species distributions