Realistic fisheries management reforms could mitigate the impacts of climate change in most countries

data required to assess current status and forecast future distributions. In this analysis, we eval- uated only the 779 single-species stocks, because the spatial distributions of the mixed-species stocks could not be projected by Gaines et al. [19] and therefore could not be spatially allocated into national jurisdictions. Projections began in 2012 with initial biomasses, fishing mortali- ties, and conditions (i.e., B/BMSY) determined by aggregating values from Costello et al. [16] (S1 Fig) and initial distributions determined by AquaMaps [22]. Projections were made through 2100 using the following general procedure: (1) distributions were updated based on a modified version of the Garcı´a Molinos et al. [23] species distribution model (see below); (2) carrying capacities were assumed to change in proportion to changes in range size, i.e., a 10% increase in range size results in a 10% increase in carrying capacity; and (3) biomass, catch, and profits were then updated based on a modified version of the Costello et al. [16] bioeco- nomic model and the selected management scenario. We provide brief descriptions of the species distribution and bioeconomic models below, but see Gaines et al. [19] and the original references for more details. Species distribution model The modified Garcı´a Molinos et al. [23] species distribution model (SDM) is a bioclimatic envelope model that uses information on species depth preferences, thermal tolerances, and the direction and speed of thermal change, i.e., climate velocity, to project changes in species

distributions under warming. AquaMaps species distribution maps [22] were used as the starting point (i.e., 2012) for the projections. In each subsequent time step, the SDM calculated the relocation of the distribution (thermal envelope) of each species as dictated by the spatial direction and rate of change of local (1˚ resolution) climate velocities based on sea surface tem- peratures under the selected emissions scenario (RCPs 4.5, 6.0, and 8.5; S1 Table). Range pro- jections are restricted by species’ thermal tolerances and depth preferences [19]. Bioeconomic model and management scenarios The modified Costello et al. [16] bioeconomic model uses a Pella-Tomlinson [24] surplus production model to forecast fish population dynamics under five management scenarios (Table 1). The Pella-Tomlinson production model requires four input parameters for each stock: the initial biomass, carrying capacity (K), intrinsic growth rate (g), and a shape parame- ter (ϕ) that determines the proportion of carrying capacity at which production is maximized. Parameters were developed for species-stocks following the procedure detailed in Gaines et al. [19] and are based on individual stock parameters [16] sourced from a combination of produc- tion models fit to the RAM Legacy Database [25] and catch-MSY models [26] fit to the FAO Catch Database [1]. The shape parameter is fixed at the meta-analytic average for fish [27], which maximizes productivity at 40% of carrying capacity. Carrying capacity is updated each

year based on the resulting changes in range size from the SDM assuming a 1:1 proportional change (see [19] for a detailed justification of this assumption). A correlated change in range size and population abundance is predicted by the basin model hypothesis [28] which posits that range size will co-vary with population abundance through density-dependent habitat selection. Empirical support for this correlation has been shown for many marine species [29– 34] and the proportionality of this correlation was confirmed by an analysis of the distribution and relative abundance of 11 unfished species in Gaines et al. [19]. Although habitat degrada- tion, resource limitation, disease, and other factors may also impact carrying capacity, we do not consider them here. The harvest rate is based on the following five management scenarios: business-as-usual (i.e., no adaptation), productivity shift adaptation only, range shift adaptation only, full PLOS ONE Realistic fisheries management reforms could mitigate the impacts of climate change in most countries PLOS ONE | https://doi.org/10.1371/journal.pone.0224347 March 5, 2020 4 / 21