Modelling Stock–Recruitment Processes in Crustacea: Some Practical and Theoretical Perspectives

Abstract

Several approaches toward development of stock and recruitment models for exploited crustacean populations are reviewed. Such approaches include modifications of classical stock–recruitment models, or ones more directly related to crustacean biology. The latter are believed to offer the most promise for practical application. Standard yield per recruit models using continuous growth functions have been applied to crustacean stocks, but moult increment – frequency versions of yield per recruit calculations better reflect the discrete growth in crustaceans and changes in growth on maturity. They can be extrapolated easily to calculate fecundity per recruit and assess the impact of exploitation on spawning potential. Simple, semiquantitative approaches, such as life history tables, promote investigation of hypotheses of growth, mortality, maturity/fecundity, and harvesting strategy on management, but like yield per recruit models, cannot easily take into account density-dependent recruitment. We may look forward to the development of models that take into account the nature of crustacean life histories, reflecting the need for cross-scheduling of growth and reproduction in environmentally limiting conditions. Modelling life history processes in biological time units related to moult cycle duration, and cross-converting to real time for consideration of the fisheries component, should offer a notable simplification of the modelling process. The existence of several "choices" for an individual crustacean at different points in the moulting/reproduction cycles makes cohort models cumbersome and seems to require the adoption of a stochastic approach, for instance Markov-related processes, which better take into account complexities of biology and fishery-related processes. For many crustaceans, recruitment is believed to be subject to a "bottleneck" somewhere subsequent to the early larval stages, and identification of the species niche for postlarval stages could be of great practical importance for management and stock enhancement. The concept of the fractal surface as a postlarval and juvenile habitat is suggested as a promising approach, and an expression for natural mortality at size is derived for obligate crevice dwellers on a fractal surface. In relatively few circumstances for Crustacea have density-dependent factors been demonstrated in the field as affecting spawning success and the survival to recruitment of postlarval and juvenile stages. Recruit survival appears to be dominated by environmental conditions that vary significantly; seasonal timing of larval release depending on environmental change from year to year. Short-cutting the investigation of precise impacts of stock density, fishing effort, and environment on recruitment can be achieved using production models, delayed recruitment models, models with autoregressive terms, or production models using mortality rates, where effort definition is difficult and catchability a function of behaviour and environnment. The overriding influence of environment on recruitment success is illustrated for both short- and long-lived species in the tropics and northern latitudes, and this is especially true for high unit value Crustacea resources whose heavily exploited fisheries generally operate at low spawning stock sizes. Fluctuating predator density, or other multispecies interactions, affect recruitment and available number of niches, and modelling of trophic relationships has promise. The range of possible models corresponds to various degrees of refinement of the data base, and the importance of biotic, abiotic, and geographical factors in controlling crustacean recruitment is stressed. Also of fundamental importance are the economics of exploitation of species with a high and elastic demand, which results in high actual and latent levels of effort, fisheries heavily dependent on incoming year-classes, and serious problems in maintaining exploitation rates at reasonable levels.