Estimating natural mortality (M) is one of the most difficult and critical elements of a stock assessment (Hewitt et al. 2007).

There are two basic approaches used to estimate natural mortality: tagging studies and growth parameters.

Tagging studies are used in the Brownies Model, where multiyear tagging studies are used to estimate natural mortality based on recaptures:

The Paulys Model (using growth parameters) is an indirect way of estimating natural mortality.

This group also works on demography, and IHME's work on estimating mortality has been at times controversial.

Green and Flinders reported a maximum estimated annual adult mortality of 88% in Idaho.

Using a different set of predictive models to estimate mortality, a 2012 study determined the number of documented and undocumented deaths in 2010 to be 1.24 million.

Studies estimate mortality at between 38% and 75%.

These model life tables both established new empirical regularities and proved invaluable in the development of later techniques for estimating mortality and fertility in populations with inaccurate or incomplete data.