The High Value of Climate Mitigation in Limiting Global Fish Biomass Losses to the year 23

It is widely understood that future climate change will have large deleterious impacts on marine ecosystems. Large synthesis efforts conducted to date by FishMIP with higher trophic level (HTL) models have shown that under high emissions scenarios, future fish biomass losses of the order of 20% can be expected by 2100. However, the impact of anthropogenic climate change beyond 2100 remains largely unexplored, even though policy choices made now will have implications over multiple centuries. Here we identify to 2300 the degree to which Marine Ecosystem Models (MEMs) driven offline with ESMs under CMIP forcing protocols provide consistent projections of fish biomass changes, under an approximately 2°C climate stabilization scenario (SSP1-2.6), and we also assess the value of climate stabilization at 2°C relative to a high emissions future (SSP5-8.5). The value of mitigation is identified by defining mitigation gain as the difference in fish biomass losses between the low emissions (SSP1-2.6) and high emissions (SSP5-8.5) scenarios, using a total of 18 climate realizations under each scenario. We highlight two main findings. First, we identify that under approximately 2°C stabilization, by 2300, forced ensemble-mean (average of 18 realizations) changes in fish biomass are not regionally emergent above the 90% confidence interval defined by inter-model disagreement. This is important, as it means that the fish modeling community cannot yet estimate the regional vulnerability of fish communities to climate perturbations under a 2°C target that rise above model structural uncertainty, or by implication under a more stringent 1.5°C target. This is due to model structural uncertainties found in both ESMs and MEMs. Second, the models are nevertheless able at 90% confidence intervals to demonstrate mitigation gain (the difference between results under SSP1-2.6 and SSP5-8.5 forcing) under 2°C stabilization over more than half the global domain by 2300. This second main result offers potentially important value in informing efforts towards sustainable development goals. We discuss the broader implications of these findings.