Ecosystem Model of PWS Herring

Management and Restoration Tool

Bi-stable States

We have analyzed the Alaska Fish and Game’s Age Structure Analysis (ASA) Models for Prince William and Sitka Sounds of herring adult stocks and recruitment.  This was done because differences and similarities between the two time series revealed much information on both the internal dynamics and external forcing of fluctuations in recruitment and the size of the adult population. 

These patterns include the existence of bi-stable states for the adult populations in which the size of the populations are clustered in time in either a high density state or a low density state. 

In PWS the high density state of the 80s fell to the low density state in the early 90s and has remained there.  In Sitka Sound the low density state of the 60s and early 70s moved to the high density state at the end of the 70s and has remained there. 

The high density states in both regions were triggered by extraordinarily successful recruitments in the mid-1970s that most likely resulted from unique oceanographic conditions.  The bi-stable states are best explained in terms of thresholds that set maxima and minimum population size for each of the 2 states. 

These thresholds may exist at either the adult stage or pre-adult stage.  We note that much of these new insights are based upon the analysis by our subcontractor, Vince Patrick.  

Four Year Cycle

The patterns in annual recruitment that occurred in both Sounds between 1964 and 1988 is characterized by large variations in the rates of recruitment and a clearly defined 4 year cycle in which exceptionally high rates of recruitment are followed by 3 years of low recruitment.  Specifically, the mean rate of recruitment of 3 year olds/ spawner and its coefficient of variation were 0.83 and 1.1 for Sitka Sound and 0.51 and 1.1 for PWS. 

From 1989 to present the recruitment per spawner for both sounds remained the same, but the annual variation in recruitment per spawner decreased.  Specifically, the mean rate of recruitment of 3 year olds/ spawner and its coefficient of variation were 0.85 and 0.46 for SS and 0.47 and 0.86 for PWS.   The transition from a cyclic recruitment to a non-cyclic recruitment was likely caused by an oceanographic shift in 1992 and 1993 that decreased productivity in the two Sounds.   This decrease in productivity greatly depressed the expected high recruitment of 1992 and thus removed the 4 year harmonics of the previous period. 

In PWS this decline in productivity combined with excess fishing and stress from the oil spill drove the population into a low density state.  On the other hand in Sitka Sound the decline in 1992 recruitment was insufficient to displace the population from its high density state. 

This data suggests strongly that the 4 year cycle was not maintained by an external environmental cycle, but rather it was maintained by the “internal” biological dynamics of the population.  We believe that this cycle is best explained by density-dependent nonlinearities in dynamics at stages in the herring’s life cycle, and more specifically it is the result of the process called “cohort dominance”.  This process has been observed for selected species of fish in lakes, where either competition or cannibalism between year classes yields large annual variations between cohort populations.