Trophic amplification: A model intercomparison of climate driven changes in marine food webs

amplification (in 55% of the global ocean on average over the seven MEMs), but also with neg- ative inversion (27%) and negative attenuation (4%) (S3 Fig). Where NPP is projected to decrease (61% of the ocean surface), negative amplification is projected in 89% of these areas on average of the seven MEMs. In contrast, where NPP is projected to increase (39% of the ocean surface), a negative inversion is projected in 69% of the ocean waters, yet with 15% and 16% positive attenuation and positive amplification projected, respectively (S3 Fig). Under IPSL forcing, total consumer biomass is projected to decrease by 76% ± 10.5% of the ocean surface (across nine MEMs). This biomass decrease is in majority dominated by a nega- tive inversion response (46%, S4 Fig). Where NPP is projected to decrease (31% of the ocean surface), negative amplification is predominantly projected (78% ± 20% of the ocean surface). Otherwise, when NPP is projected to increase (69% of the ocean surface), similar contrasts between MEMs and qualitative results appeared, compared to GFDL (S4 and S5 Figs). The magnitude of the biomass changes to NPP changes ratio (R) is dependent on biomass response types and MEMs (ANOVA, p_value<0.05), yet not on ESMs (ANOVA, p_value>0.05) (Figs 5 and S4B). Where an amplification response is projected, biomass decreases on average three times more than NPP decreases (Negative amplification, R = 3 ± 0.3) and increases two and a half more than NPP increases (Positive amplification, R = 2.5 ± 0.8). Where the NPP climate signal is negatively or positively attenuated through the

food web, biomass decreases or increases by half of NPP (R = 0.5±0.1) in both cases. Where the NPP climate signal is inverted (negatively or positively), consumer biomass changes twice as much as NPP changes in the opposite direction (R = -2.2 ± 0.7). Finally, MEMs forced by lower trophic level biomass project lower magnitude R than lower trophic level production- forced MEMs (mainly DBPM and EcoOcean). 3.3 MEM agreement With the GFDL forced SSP5-8.5 scenario, MEM projections strongly agree, e.g., exhibit the same biomass response type for at least six models, between 40˚S and 50˚N latitude. Fig 3. Ensemble projections of NPP and total consumer biomass changes, relative to 1995–2014, under SSP5-8.5 and for all the considered MEMs. (a) based on GFDL/MEMs combinations; (b) based on IPSL/MEMs combinations. All values are relative to the standardised reference period of 1995–2014. Vertical grey line indicates the last year of the historical period. https://doi.org/10.1371/journal.pone.0287570.g003 PLOS ONE Trophic amplification revealed PLOS ONE | https://doi.org/10.1371/journal.pone.0287570 August 23, 2023 8 / 23