FISHERY BULLETIN: VOL. 75, NO. 1 



charge and temperature) which were included 

 because of their potential biological importance 

 will require a greater knowledge of spawning 

 intensities and a longer term data base. 



Overall, the model implies a predictive capabil- 

 ity for large year classes and for extremely poor 

 year classes. The model provides a satisfactory 

 indication of the general magnitude of a year class 

 prior to entering the fishery in 14 of the 16 yr. 



For initial model purposes, the survival index 

 was not computed beyond 1970 because the 1971 

 year class is still being harvested by the fishery, 

 and the total catch from that year class necessary 

 for verification of the number of recruits is not 

 known. Forecasting in real time can be ac- 

 complished by inserting the routinely available 

 environmental data into the survival index 

 equation. The expected number of recruits for a 

 given year class is obtained by determining age 

 structure and abundance of 2-yr-old and older fish 

 from fishery landings the previous fishing season, 

 estimating an exploitation and survival rate to 

 determine the number that will survive to spawn 

 the next year class, calculating the expected 

 number of eggs produced, and estimating the 

 expected number of recruits from the Ricker 

 function. Multiplying the expected number of 

 recruits by the predicted survival index gives the 

 predicted number of recruits. Estimates of the 

 number of recruits can be made as early as April of 

 the year-class year, and can be revised when ac- 

 tual exploitation rates are determined to allow 

 better estimates of the size of the spawning stock 

 which produces the year class. Thus, an initial 

 prediction of the number of recruits can be made 

 approximately 1 yr before they become available 

 to the fishery the following spring. 



DISCUSSION 



Refinement of the predictive capability of the 

 recruit-environment model is dependent on in- 

 creased knowledge of the biology of Atlantic 

 menhaden and on better understanding of the 

 effects of the many factors that influence dis- 

 tribution, abundance, and survival. The model is 

 concerned only with variation introduced into 

 year-class size during the relatively short life 

 phase in which larvae are oceanic and before 

 metamorphosis takes place. The model concen- 

 trates on those factors which influence larval 

 distribution and act as a mechanism to transport 

 larvae into the vicinity of estuarine nursery 



grounds, thereby increasing survival. Major 

 sources of variation such as food availability and 

 predation have not been directly considered. 

 However, since these factors are, to some extent, 

 influenced by the number of larvae produced by 

 the spawning stock, variations induced by them 

 should be partially accounted for by the density- 

 dependent Ricker function. The actual fluctuation 

 in availability of food could only be determined by 

 broad-scale surveys over the entire menhaden 

 spawning range and would require a continuous 

 time series for a number of years. Likewise, the 

 determination of predation and cannibalistic 

 influences would require extensive field surveys 

 and controlled laboratory experiments. 



Problems in determining the influence of 

 pertinent environmental factors are compounded 

 by the large geographic range of menhaden 

 spawning activities. The influence of any one 

 particular factor at a specific location could only be 

 determined if the amount of spawning at that 

 location was known. Comparison of environmen- 

 tal factors against a survival index for the entire 

 stock, as has been done in this study, requires the 

 selection of broad-scale factors having major 

 influence over large portions of the spawning 

 range, or the selection of representative data 

 which provide a generalized environmental index 

 for a selected factor. Localized variations may be 

 highly significant, but masked by overall survival 

 success or failure without knowledge of localized 

 spawning intensity. 



Cushing (1969, 1974) cited failures in attempts 

 by other authors to correlate year-class strength 

 and winds (or pressure gradients), and suggested 

 that variation in wind direction may be a greater 

 source of variation than the strength of winds from 

 a single direction. The U.S. east coast is composed 

 of an almost continuous series of bays and sounds, 

 which extend both north and south of the major 

 spawning region for Atlantic menhaden. Under 

 these circumstances, variations in wind direction 

 would probably influence the route of larval drift. 

 However, unless northward or southward larval 

 movement was extreme, larvae would not be 

 transported away from suitable nursery areas as 

 long as there was a significant onshore component 

 of wind-driven circulation. Thus wind direction 

 would be a significant factor only if that direction 

 reduced the westward component of Ekman 

 transport or if the normal seasonal wind pattern 

 reversed, generating eastward (offshore) trans- 

 port. 



36 



