Investigating spatial variation and temperature effects on maturity of female winter flounder (Pseudopleuronectes americanus) using generalized additive models

Abstract

Generalized additive models were used to investigate fine-scale spatial variation in female maturity across the three United States’ winter flounder (Pseudopleuronectes americanus) stocks. The effect of temperature on maturity was also investigated. Maturity models explicitly incorporating spatial structure performed better than “traditional” methods incorporating spatial effects by aggregating data according to predefined stock boundaries. Models including temperature explained more of the variability in maturity than those based only on fish size or age but did not improve fit over models incorporating spatial structure. Based on the size- and age-at-maturity estimates from the spatially explicit models, distinct subareas were objectively identified using a spatially constrained clustering algorithm. The results suggested greater variation in size- and age-at-maturity within than between existing stock areas. The approach outlined here provides a method for identifying areas with different vital rates without the need to presume subjective boundaries.