The Worst Fishing Years Come When the Ocean Is Both Too Hot and Too Empty

Ocean heat extremes and spells of unusually low ocean productivity each raise the odds of a very poor fishing year in a given place, and the odds rise furthest when the two arrive together, according to a study published Sept. 1 in Nature Communications.

William Cheung of the University of British Columbia, Thomas Frölicher of the University of Bern and colleagues matched fisheries catch records against sea temperature and net primary production, the rate at which microscopic plants at the base of the food web build new tissue. The analysis covers 6,659 time-series for 1,246 exploited species across 254 coastal fishing zones from 1993 to 2019. The study defines an extreme low catch as the lowest tenth of a stock's catch anomalies.

Two consecutive years of extreme high temperature raised the odds of an extreme low catch by 40%, with a 95% confidence interval of 20% to 65%. Where the previous year's heat extreme coincided with the lowest productivity year in a stock's record, odds ratios reached 1.8, or nearly 80% higher odds than in non-extreme years. That figure describes an association measured across pooled catch records, not a demonstrated cause.

The strongest effects were in tropical and subtropical waters. Sharks and rays, small open-water fish (such as anchovies and sardines) and commercial invertebrates were the most sensitive groups, with median catch drops above 40% in extreme years. Peru's small open-water fisheries, dominated by anchoveta, showed more than a ninefold rise in the odds of an extreme low catch and median declines near 60%.

The team also ran climate simulation output through the same statistical framework, projecting that the share of stocks hitting extreme low catches in vulnerable zones will rise from about 11% in the early 1990s to almost 16% by the middle of this century, on similar paths under both a low-emissions and a high-emissions scenario.

Adding extreme fishing effort to the model as a control changed the temperature and productivity odds ratios by less than 3%.