712 



Fishery Bulletin 92(4), 1994 



ern anchovy were exacerbated by use of a compli- 

 cated model with insufficient data. The model we de- 

 veloped for northern anchovy was, therefore, simpler 

 and more parsimonious (SMPAR). 



The SMPAR model is a biomass dynamic model 

 designed to give more precise biomass estimates for 

 management of northern anchovy. SMPAR is a hy- 

 brid between simple surplus production and complex 

 age-structured models. It resembles a surplus pro- 

 duction model because age-composition data are not 

 used and fishing mortality rates are equal for all age 

 groups. The model is age structured, however, and 

 some rudimentary relationships between age-specific 

 abundance and abundance indices are assumed. 



As described above, the SMPAR model for north- 

 ern anchovy did not use any age-composition data, 

 although data were available for most fishing sea- 

 sons prior to 1991. We chose to exclude age-composi- 



tion data from our model because the data are diffi- 

 cult to interpret, require complex modeling ap- 

 proaches, and were not available for recent seasons. 

 In this paper, we describe the SMPAR model and 

 data for northern anchovy. Bias and variance in bio- 

 mass and recruitment estimates are assessed by us- 

 ing bootstrap techniques. Sensitivity analyses show 

 how model assumptions and contradictory trends in 

 the data affect biomass estimates from SMPAR. We 

 use simulation analysis to show how SMPAR would 

 perform under a wide range of fishing mortality and 

 recruitment conditions. 



Data and methods 



Fishing seasons were used to aggregate most of our 

 data including landings data and indices of abun- 

 dance (Table 1; Fig. 1). Fishing seasons for northern 



