Near real-time forecasting of recruitment from larval surveys: application to Alaska pollock

作者：Bailey Kevin M^*; Zhang Tianyang; Chan Kung sik; Porter Steven M; Dougherty Annette B

来源：Marine Ecology Progress Series, 2012, 452: 205-217.

DOI：10.3354/meps09614

• Population dynamics
• Fisheries
• Environment
• Ocean research
• Regime shift
• Ecological community

The fishery for Alaska pollock Theragra chalcogramma in the western Gulf of Alaska is dependent on the strength of incoming year classes. Methods were developed to measure late-larval stage abundance and sizes at-sea and to use such data in a near real-time model to forecast recruitment about 4 yr in advance. Larval abundances were weighted for size-and temperature-dependent mortality to project the number attaining a critical size of 15 mm; this weighting reflects the greater survival value of larger larvae to the population. The method was applied to catches in the 2007 and 2008 larval surveys to show the feasibility of attaining near real-time estimates. The accuracy of the method to forecast recruitment was evaluated from historical data. Adjusting larval abundance for size-and temperature-dependent mortality improved the trend with recruitment, but the relationship was not significant over the entire time series (1979 to 2007) because of an increasing trend in predators of juvenile stages, decoupling the parameters. A threshold-based generalized additive recruitment forecast model was developed to account for autocorrelative structure caused by inter-cohort interactions and mortality-related factors occurring after the larval period. The model closely tracks past recruitment trends. Based on low larval abundances and high numbers of predators, low to moderate abundances of age-4 recruits were predicted for 2007 to 2011, which matched observed levels reasonably well.

• 出版日期2012