Ogive characterisation in medium-term stock projections

Under prevailing stock assessment methodologies, stochasticity is introduced into the simulation of future fish-stock population dynamics via random variation about a) an hypothesised spawner-recruitment model (possibly constrained to emulate the historical time-series of recruitment), and b) the initial population size.  However, hindcast simulations suggest that the importance of appropriate characterisations of age-dependent measures of growth and reproductive potential may equal or exceed that of recruitment or starting numbers.  This paper proposes a methodology modification by which such concerns might be addressed, through vector autoregressive moving-average (VARMA) modelling of maturity and weights-at-age ogives in stock projections.  A number of technical points contingent on these modifications are raised.  North Sea cod (Gadus morhua L.) is used as a test case for the new method, and implications for the management of this and other stocks are discussed.