Four years (2016–2019) of active acoustic data were used to characterize temporal patterns in polar cod densities in the Chukchi Sea (Fig. 1). Acoustic backscatter data, a proxy for fish density, were collected using an ASL Acoustic Zoo- plankton Fish Profiler (http://www.aslenv.com/AZFP.html), deployed at 28–35 m depth (depending on year), looking upwards, and operating at 36, 125, 200, and 455 kHz.

Acoustic backscatter corresponding to fish was discrimi- nated from other sources of backscatter using differences in mean volume backscattering strength (MVBS; Madureira et al. 1993; Kang et al. 2002; Korneliussen and Ona 2003) between 125 and 38 kHz data (ΔMVBS125-38 kHz). Fish Sv values were integrated into hourly averages from January 1, 2016 to December 31, 2019. A detailed deｓcription of the CEO acoustic data collection and processing can be found in (Gonzalez et al. 2021).

Although no direct fish sampling was conducted in asso- ciation with acoustic measurements, we can rely on catch data from fisheries surveys carried out in the NE Chukchi Sea to attribute most of the observed fish backscatter to polar cod. Polar cod accounted for 81–90% of total fish biomass and abundance from bottom (Barber et al. 1997; Goddard et al. 2014; Sigler et al. 2017; Logerwell et al. 2018) and pelagic (Lowry and Frost 1981; De Robertis et al. 2017) trawl surveys conducted from spring through autumn, ice- free seasons. Polar cod constituted the majority of fish bio- mass (63–99%) and abundance (93–99%) in sample catches from four midwater trawls conducted on Hanna Shoal in close proximity to the CEO in summer of 2017 (Levine and De Robertis pers. comm). Other species caught near Hanna Shoal included capelin (Mallotus villosus), Lumpenus sp.,

staghorn sculpin (Gymnocanthus tricuspis), and Liparidae

snailfish. As further support of this backscatter classifica- tion, age-0 (i.e. < 12 months old) polar cod was the dominant contributor to 38 kHz backscatter in the northern region of the Chukchi Sea in acoustic-trawl surveys conducted in 2012 and 2013 as part of the Arctic Ecosystem integrated survey (De Robertis et al. 2017) and constituted > 85% of the catch per unit effort in a 2019 survey in the Chukchi Sea (Levine et al. 2021).

Our study spanned years with highest temperatures and lowest sea ice conditions on record but strong differences in temperature and sea ice conditions among sample years were observed. Years 2016 and 2018 (hereafter “cold” years) were characterized by later sea ice retreat, earlier advance, greater sea ice concentration, and lower water temperatures than 2017 and 2019 (hereafter “warm” years). Sea ice retreat occurred on July 13th 2016, June 3rd 2017, July 14th 2018, and May 12th 2019 (Fig. 2a). Sea ice advance occurred on November 21st 2016, December 5th 2017, November 23rd 2018, and December 7th 2019 (Fig. 2a). This resulted in longer open water periods (i.e. period between sea ice retreat and advance dates) in 2017 (185 days) and 2019 (209 days) than in 2016 (131 days) and 2018 (132 days).

原创作者：上海逐海仪器设备有限公司