Problems of selection of explanatory variables for the dynamics of generation numbers of Pacific salmon

The problems of choosing a model type when studying the causes of generation size variability, as well as the consequences of attempts to approximate multifactor processes with single-factor models, are considered using the example of sockeye salmon from the Okhotsk region. It is shown that single-factor models in a multifactor space give a distorted idea of cause-and-effect relationships. A model of generation size formation for sockeye salmon has been developed, including the number of parents, variability in precipitation, and air temperature in the freshwater phase of juvenile life as explanatory variables. The combined effect of parental size and weather factors explains more than 90% of the variance in generation size for sockeye salmon.