Under climate change, model ensembles suggest that declines in phytoplankton biomass amplify into greater reductions at higher trophic levels, with serious implications for fisheries and carbon storage. However, the extent and mechanisms of this trophic amplification vary greatly among models, and validation is problematic. In situ size spectra offer a novel alternative, comparing biomass of small and larger organisms to quantify the net efficiency of energy transfer through natural food webs that are already challenged with multiple climate change stressors. Our global compilation of pelagic size spectrum slopes supports trophic amplification empirically, independently from model simulations. Thus, even a modest (16%) decline in phytoplankton this century would magnify into a 38% decline in supportable biomass of fish within the intensively-fished mid-latitude ocean. We also show that this amplification stems not from thermal controls on consumers, but mainly from temperature or nutrient controls that structure the phytoplankton baseline of the food web. The lack of evidence for direct thermal effects on size structure contrasts with most current thinking, based often on more acute stress experiments or shorter-timescale responses. Our synthesis of size spectra integrates these short-term dynamics, revealing the net efficiency of food webs acclimating and adapting to climatic stressors.
International audienceThe conversion of units is one of the difficulties of model parameterisation. Conversion errors may result not only from incorrect choices of conversion factors, but also from incorrect choices of the value itself. In biogeochemical models, mesozooplankton, is the highest trophic level of the food web, and it is very often reduced to a single variable generally considered as a representation of the copepod community, the dominant taxa in mesozooplankton. If this simplifies the information to be obtained for the stock, a correct parameterisation of the processes related to the copepod community is already a tricky task due to the wide range of copepod species, sizes, stages and behaviour. The goal of this paper is to improve the communication between experimentalists and modellers by giving indications for the conversion of copepod vital rates from experimental to biogeochemical model units. This includes the choice of values, conversion factors, terminology distinction and the scale transfer. To begin with, we briefly address the common problem of the conversion of a rate per individual to a rate per mass. Then, we focus on unit conversion problems for each specific rate and give recommendations. Finally, we discuss the problem of scale transfer between the level of organisation at which the rate value is measured at characteristic time and space-scales versus the level of representation of the corresponding process in the model, with its different characteristic time and space-scales