A Bayesian framework with implementation error to improve the management of the red octopus (Octopus maya) fishery off the Yucatán Peninsula

Abstract

The red octopus (Octopus maya) is an endemic species of the Yucatán Peninsula and its fishery is one of the most important along the Atlantic coast of Mexico. Commercial exploitation started in 1949. Since 2002 an index of abundance has been estimated, and this index was used to perform a stock assessment and decision analysis using the Schaefer model. A Bayesian approach was applied to estimate the model parameters and to project the species population under two management scenarios with a constant harvest rate and a positive implementation error. Results suggest that in 1995 the biomass corresponded to 23% of the population carrying capacity (K) and that the current stock is only 14% of K. The population may be depleted and a rebuilding plan might be necessary. In the decision analysis, when the implementation error was included, the Markov Chain Monte Carlo simulations suggested that the current level of exploitation (50% harvest rate) could produce a decreasing trend with the most probable biomass of 9679 t and an expected catch of 7920 t in 2018, and an expected probability of 0.82 of the population being less than 40% of K. On the contrary, a 30% harvest rate would raise the expected catch in 2018 (12,058 t), also reducing the probability of the population being smaller than 40% of K. The inclusion of the implementation error provides a more realistic scenario and represents a more conservative option; therefore, using this type of auxiliary data within a Bayesian framework is recommended for the decision making process. If adopted by Mexican fisheries managers, the approach used in this study could help improve the management of this resource and keep exploitation at sustainable levels.