Harvesting Control Rules that deal with Scientific Uncertainty

By using robustness methods we design HCRs that explicitly include scientific uncertainty. Under scientific uncertainty –when the perceived model can be generated by a nearby op- erating model– robust HCRs are designed assuming that the (inferred) operating model is more persistent than the perceived model. As a result, a robust HCR has a steeper ratio between fishing mortality and biomass than a non robust one. We prove that constant effort HCRs are not robust. Moreover, rather than decreasing fishing mortality reference points for exploitation, the optimal robust response to scientific uncertainty is to increases biomass precautionary limit points when knowledge about the stock status decreases. Finally, we show that robustness can be implemented if fishing mortality is increased faster than lin- early –by a factor of 2-fold– when a stock is assessed as above 0.5BMSY. We illustrate our findings by designing HCRs for 17 ICES stocks using this rule of thumb.