The Herring Research and Monitoring (HRM) program goal is to improve the ability to predict the herring stocks, which requires a model of some sort. During the first phase of the HRM program (2012-2017) we focused on adapting the existing age-structure-analysis model used by Alaska Department of Fish and Game to one using Bayesian statistics. The Bayesian version of the model was then used to examine the value of different historic inputs. Information from herring stocks around the world was collected to determine how the herring crash in PWS compared to declines and recovery of other herring stocks. In the current phase of the program (2017-2021) the Bayesian model will be used to predict herring population levels in PWS. We are also examining how to incorporate new information about disease and other herring observations. Environmental data will be examined to see if it can help improve the prediction of recruitment and ultimately we will evaluate alternative management strategies to examine the impact on the population.
We will use the Bayesian age-structure-analysis (BASA) model to predict the Prince William Sound herring stock using data collected by the various monitoring projects in the HRM program and to examine the impacts of alternative management strategies. We will look at how adding different types of information (environmental, or herring observations) improve the ability of the model to predict herring stocks. The incorporation of new disease information will be examined to determine a more appropriate way to use disease measurements to predict mortality. We will continue to collect information on herring stocks around the world to put the changes in catch, biomass, and recruitment of herring in PWS in context of other stocks.
In the first phase, we were able to rebuild the ASA model into a Bayesian structure that provides more information about the probability range of the prediction. The model was then used to examine the value of different inputs, and it was found that the egg deposition work was necessary to create an initial anchor of the biomass estimate, and the disease information to allow for the large mortality required to simulate the crash were the two most important pieces of information. When the crash was put into context of fluctuations in worldwide herring populations, it was found that both the magnitude of decline and the length the population has been low in PWS are highly unusual. In the current phase of the project, we expect to be able to understand better if environmental information can be used to predict changes in herring populations. We hope that new disease-related measurements can be used to determine likely mortality by disease better. We will also see if alternative management rules could be implemented to allow a fishery without harming the population.