Assessment of Pacific saury stock in the North Pacific using the non-equilibrium surplus production model

• Main Authors: Yu-Ling Yang, 楊育霖
• Other Authors: Wen-Bin Huang
• Format: Others
• Published: 2014
• Online Access: http://ndltd.ncl.edu.tw/handle/n373fd

碩士 === 國立東華大學 === 自然資源與環境學系 === 102 === In the Far East, Pacific saury Cololabis saira is one of the major economic species. It is mostly caught by Japan, Taiwan, Korea and Russia, with Taiwan becoming the largest harvesting country of saury in 2013. In this study, the generalized linear model (GLM) was applied to standardized CPUE data of Pacific saury for the North Pacific fishery of Taiwan. The factors considered in GLM include year, month, longitude, latitude, water temperature, and the interaction terms between month, longitude, latitude, and water temperature. The standardized CPUE generally showed a gradually rising trend. The non-equilibrium surplus production model was then applied to assess the stock status of the North Pacific saury. The model considered two hypothesis of catchability coefficient (q). In scenario 1, the catchability coefficient was considered to be constant. Scenario 2 considers a linear increase in catchability through time. The assessment result from scenario 1 showed that the maximum sustainable yield (MSY) was 66 thousand tons, and B2012/BMSY and F2012/FMSY were estimated to be 1.90 and 0.36. In scenario 2, the MSY was estimated to be 50.7 thousand tons, and the B2012/BMSY and F2012/FMSY were estimated to be 1.53 and 0.59. These two results both indicate that the Pacific saury in the North Pacific is not overfished, and appears to be moderately exploited. In order to estimate the uncertainty of the parameters, this study used likelihood profile, bootstrap percentile and Biased corrected bootstrap percentile (BC1 percentile) methods to estimate the 95% confidence intervals of the parameters. The results from scenario 1 showed the confidence interval produced by the BC1 percentile tended to be narrower than the other two methods. Scenario 2 showed that the width of the confidence interval produce by likelihood profile is the narrowest one. This study set effort level at the mean fishing effort of the previous three years to projection stock biomass. The result from scenario 1 showed that the stock biomass will rise to 456.8 thousand tons by 2022. The result from scenario 2 showed that the stock biomass will drop to 122.6 thousand tons. This study also evaluated impact of the uncertainty associated with the catch and CPUE data on the model estimations using Monte Carlo simulations. The results indicate that when incorporating observation error for catch and CPUE data, the accuracy of estimations deteriorated, particularly for the CPUE data.