"Harvest mortality" usually refers to the proportion of a population that is killed by humans for economic purposes (e.g. birds or terrestrial mammals that are shot by hunters, fish caught in nets or on lines, trees cut down by loggers, etc.). For a specified population and time period, it will be something like (number harvested/total number alive at time 0).
The first hit of this Google Scholar search for "harvest mortality" is by Solberg et al. They describe a complicated procedure for estimating the number of moose of a given age and sex alive in a given year. Basically, they add up the reported number of moose shot by age and sex in each year, combine it with estimates of the non-hunting mortality by age and sex, and do the accounting to back-calculate how many animals of a given age/sex must have been alive in a given year (they do the book-keeping up to a maximum age beyond which few animals survive).
Once they've done the hard part, estimating the total number of moose of a given age that were alive at the beginning of a given year,
The age-specific harvest mortality was calculated
as the annual number of moose harvested divided by
the annual number of moose present in the respective
Solberg, Erling Johan, Anne Loison, Bernt-Erik Sther, and Olav Strand. “Age-Specific Harvest Mortality in a Norwegian Moose Alces Alces Population.” Wildlife Biology 6, no. 1 (2000): 41–52. https://doi.org/10.2981/wlb.2000.036.