COMMENTARY Forecasting fisheries collapse Steven D. Gaines 1 and Christopher Costello Bren School of Environmental Science and Management, University of California, Santa Barbara, CA 93106 When sustainably managed, wild harvests from the sea provide livelihoods and a vital source of protein for hundreds of millions of people (1, 2). If managed well, these benefits are perpetually renewable. However, numer- ous challenges can compromise the sustain- ability of this natural bounty. Ironically, a significant challenge arises because of the diversity of species caught in nearly all fish- eries. Although healthy diverse ecosystems can provide resilience against some perturba- tions, fishing pressure in these ecosystems is almost always geared to the strongest, most economically valuable stocks. In the process of fishing, many other species are incidentally caught including both (i ) species that nobody wants to catch but are costly to avoid, com- monly called bycatch, and (ii ) species with substantially lower economic value than tar- geted, highline stocks. Even when the target stock is perfectly managed, these incidentally caught stocks can collapse if their life history traits make them more susceptible to a given level of fishing effort than target species. Without regular monitoring and strong man- agement of all species, these weak stocks may force the closure of otherwise highly profit- able fisheries. One solution is to expand sci- entific assessments to all species affected by each fishery, but this is prohibitively expen- sive. Fish are expensive to monitor, and as a result, even the majority of target species are currently not adequately assessed world- wide (3). Solving the weak stock problem requires new innovation. In PNAS, Burgess et al. (4) offer a unique forecasting tool to innovatively address this global challenge. If the looming collapse of overfished weak stocks could be foreseen well in advance, the environmental and economic costs of elimi- nating the problem before it occurs may be far less than the costs of restoring overharvested species after the fact. Overharvesting weak stocks or bycatch species often triggers a reg- ulatory shutdown of the entire fishery to allow rebuilding (5). This rebuilding can last for many decades, because the life history traits that put the species at risk also mean that their recovery will be exceedingly slow. For exam- ple, weak stocks such as yelloweye rockfish in the highly diverse rockfish assemblage along the west coast of the United States only stands a 50% chance of recovery by 2047, despite significant reductions in fishing pressure (6). Burgess et al. (4) provide a simple, but effective, means of forecasting the stocks for which incidental catch presents a significant future risk. The forecasts can project risk dec- ades into the future. Rather than requiring major new monitoring efforts, these forecasts can be made with readily available informa- tion. The clever logic underlying the unique tool is that future mortality from fishing will be driven by fishing pressure focused on tar- get species. These key species commonly have far more extensive monitoring data on current and past catch than do the weak stocks. The authors derive a measure of fu- ture risk from incidental catch called the T-score. This threat measure couples projec- tions of future fishing mortality with differ- ences in life history traits between target and other species. Species with a T-score <1 are not at risk from incidental fishing. As the T-score increases to >1, the weak stock is projected to become overfished, and if T > 2, incidental fishing mortality is forecast to put the species at risk for extinction. Because the T- score can be forecasted well into the future, it could trigger proactive management actions that would substantially reduce the economic and ecological harm that occurs from the alter- native of collapse and rebuilding cycles. In addition to providing a valuable early warning system, we believe the framework developed by Burgess et al. has several broader implications. First, the clear articulation of the multispecies catch problem may help explain some previously unexplained patterns. For example, Costello et al. (3) produced a global analysis of a large number of previously unas- sessed fisheries. One of the intriguing results was that large fisheries (in terms of landings) tended to be in much better shape than small fisheries. Although several factors likely contribute to this discrepancy, the paper of Burgess et al. (4) provides one possible expla- nation: large fisheries are far more likely to be target stocks, whereas small fisheries may in- clude many more species that are along for the ride and are often harvested incidentally. It is quite possible that with methods like this, interventions (such as those described below) could have been undertaken early and those species would not have been reduced to such low levels. Second, this forecasting tool could have important implications for the design of institutions to manage fisheries. If manage- ment is constrained to merely regulate the fishery as a common pool resource, then, on learning a species has a high T-score, managers are left with no choice but to decrease fishing mortality on all species. The larger the discrepancy in life histories, the larger the costs of such reductions. However, under some other innovative management institutions, promising alternatives exist. Two key ones are as follows. (i ) Marine pro- tected areas (MPAs) could be implemented to protect the weaker species from extirpa- tion and continue permitting a higher fish- ing mortality outside the protected area. For this alternative to be really effective, there has to be some geographical difference be- tween the species that allows the MPA to protect the weak stock while incurring smaller costs on the harvest of the targeted species. (ii ) Multispecies individual trans- ferable quotas (ITQs): this rights-based approach provides a market signal to avoid bycatch. The idea here is that species with a high T-score get a lower quota, which rai- ses the price of catching those species. This creates a strong financial incentive for fish- ermen to innovate and find more selective ways to fish. Without a price signal, there is no particular reason fishermen would try to avoid catching these weak stocks. This ap- proach has worked well in the West Coast Groundfish fishery in the United States, where changes in fishing gear from trawling to hook and line and avoidance of key weak stock hab- itats have dramatically reduced incidental catch of previously threatened species (7). Third, this technology should significantly increase the value of information in a stock assessment. It is commonly believed that assessments must be done stock by stock. A typical assessment costs hundreds of thou- sands of dollars per species, which is pro- hibitive for most fisheries. The Burgess et al. (4) approach allows us to predict the fate of many species from an assessment of only one species. Thus, the value of the stock assess- ment increases dramatically. Fishermen who benefit from harvesting one group of species may, with this method, benefit from a stock assessment on other species. Thus, they may Author contributions: S.D.G. and C.C. wrote the paper. The authors declare no conflict of interest. See companion article on page 15943. 1 To whom correspondence should be addressed. E-mail: gaines@ bren.ucsb.edu. www.pnas.org/cgi/doi/10.1073/pnas.1315109110 PNAS | October 1, 2013 | vol. 110 | no. 40 | 15859–15860 COMMENTARY