Approaches for estimating natural mortality: Application to summer flounder (Paralichthys dentatus) in the U.S. mid-Atlantic

We investigate several approaches to estimate natural mortality (M) for summer flounder Paralichthys dentatus. Historically, a value of 0.2 y−1 has been used for all ages and both males and females in stock assessments. Recently, M has been increased to 0.3 y−1 for males. A range of estimates of M are available from the different approaches. Methods based on maximum age are not appropriate due to exploitation history and the sampling design used to collect age data and estimates of M from relationships with life history parameters are too imprecise to be useful in the stock assessments. Estimates of M from other species are variable and many are unreliable. Modeling higher M for young individuals may be appropriate given the substantial number of 0-year old individuals caught in the fisheries and surveys. Methods based on M being inversely proportional to size may provide a useful age-structured M for young individuals, but the absolute level still needs to be estimated to scale the age-based curve. Simulation analysis suggests that, given the model assumptions and the type of data available, the stock assessment model is able to estimate both female and male M with moderate precision, but with some bias depending on the true values of M. The estimates of M from the assessment were 0.29 (0.23–0.34) y−1 and 0.54 (0.49–0.59) y−1 for females and males, respectively, which are considerably higher than the values used in the current assessment. However, these estimates are sensitive to model assumptions. The estimates of M for males are consistently higher than those for females. A well designed and implemented tagging program where the tagging data are integrated into the stock assessment may be the best approach to produce reliable estimates of M for summer flounder. In the meantime, estimating sex-specific M within the stock assessment model appears to be the most appropriate approach.

► Maximum age methods are inappropriate due to exploitation history and sampling design.
► Estimates of M from life history and maximum age relationships are imprecise.
► Estimates of M from other species are variable and may be unreliable.
► The stock assessment model is able to estimate M with moderate precision, but with some bias.
► Estimates of M were 0.29 (0.23–0.34) y−1 and 0.54 (0.49–0.59) y−1 for females and males.