Tracking Fish Mortality with a New Model and Historic Data

Fellowship Awarded to VA Student Studying Behavior of Bluefin Tuna Anglers
August 4, 2015
New Initiative Helps Saltwater Aquarium Industry ‘Go Green’
August 11, 2015
Show all

Tracking Fish Mortality with a New Model and Historic Data

When Todd Gedamke was a graduate student at VIMS, he and his advisor John Hoenig developed a new model to track fish mortality. Now, Hoenig and his current student are extending the model to include additional types of data.

Monkfish drawing. ©NOAA Fishwatch

Monkfish drawing. ©NOAA Fishwatch

By Julia Robins, Staff Writer

A hundred years after the Swedish Institute of Marine Research began its annual fish survey, Todd Gedamke got a call. The institute wanted him to identify the historical trends in the abundance, size structure, and distribution of a Swedish flatfish over that entire period to improve long-term management. Using this data from the Swedish records, Gedamke, a Virginia Institute of Marine Science (VIMS) graduate, was able to test a new kind of model that could improve the way we manage fisheries. The model was first developed during Gedamke’s graduate study, when VIMS researcher John Hoenig was his advisor.

“It’s been known for a hundred years that if you fish hard there won’t be many big fish left,” says Hoenig. “The average size will go down as you fish and this will give you a way to estimate the mortality rate experienced by the population.”

Todd Gedamke.

Todd Gedamke holds a barndoor skate.

A key feature of Hoenig and Gedamke’s estimator is that it allows for a changing mortality rate over time. Previous models assumed that the mortality rate is constant over time. The problem is that under this assumption, the model has a lagging response. For instance, if fishing completely stops, the mortality rate will immediately drop because fewer fish are being removed from the population. Fish that remain will survive longer and grow longer, but it will take time before the longer average length is seen in the population. Since traditional models are based on average length, they won’t detect the new mortality rate for years.

Using the trawl survey in Sweden and the estimator, Gedamke calculated that the mortality rate had changed nine times over the past century. Anthropogenic (human-caused) events like the two world wars affected fish stocks. Average size decreased before World War I as the fishery grew larger, increased when the war began and fishing went down, rose again after the war, and followed a similar pattern during and after World War II.

“You can actually see all of that reflected in the average size of the fish,” says Hoenig, “So that was quite exciting.”

Hoenig and his current student Quang Huynh are extending the model to include additional types of data such as catch rate, fishing activity, and total catch.

The strength of Gedamke and Hoenig’s model is its versatility. The model not only eliminates the assumption that the mortality rate does not change over time, but also allows for the inclusion of additional types of data.

What’s more, Hoenig, Gedamke, and Huynh have found that the model works well with multiple species. They have already had promising results with barndoor skate in New England, various snappers in the Caribbean, and greater amberjack in the Gulf of Mexico. Researchers in France are also applying the model to tuna species.

“What I like is, the methods that we develop have some generality,” Hoenig says. This generality, coupled with the ability to incorporate additional data and combine with other methods, is making Gedamke and Hoenig’s model useful in Sweden, New England, the Caribbean, and beyond.

This Sea Grant project was funded by the NOAA National Marine Fisheries Service.