REDUS has improved fisheries stock assessment

How much fish there is the ocean, and how much can be safely harvested without having a negative impact on fish stocks, are two vital questions for both the Norwegian fishing industry and marine scientists. 

“Because fish live in a vast, deep ocean, it is much harder to count them than land animals. So as marine scientists, we have to use indirect methods to estimate the size and resilience of fish stocks”, explains Erik Olsen, the research group leader who has managed the REDUS project. 

To monitor the state of fish stocks, fisheries scientists undertake scientific surveys and obtain catch data from fisheries. This information is then analysed using advanced mathematical models, to help researchers estimate how much fish there is in the ocean now, and how stocks are likely to develop in the coming years. 

Objective: to make advice on fisheries management and quotas more reliable 

However, estimates are never totally accurate. 

“There is always a certain amount of uncertainty associated with an estimate, and it is frustrating for both the fishing industry and the authorities when estimates don’t match subsequent observations”, says Olsen. 

That is why the REDUS project was initiated in 2016. REDUS is an acronym for Reduced Uncertainty in Stock Assessments. The project, which is now drawing to a close after five years, has led to several changes in the processes that underpin quota advice. 

A new tool to be used for the major fish stocks 

The scientists working on the REDUS project studied a wide range of topics and all stages in the process: right from collecting samples during surveys through to trawling methods, acoustic surveys, fish stock estimates and management strategies. 

“We have developed a ‘virtual workbench’ for analysing survey and catch data, which makes it possible to test how different models, user settings and data selections affect fish stock estimates. This tool allows us to measure the uncertainties at each stage of the process, right from sampling of fish during a survey through to the final advice”, the project manager explains. 

Herring and cod have been the “guinea pigs” for the new tool, but the results have been so good that the tool will now be rolled out much more widely. 

“As of 2021, the tool will be used for all of the major data-rich fish stocks that the IMR provides advice for.” 

More precise– and cheaper – survey estimates 

Developing new methods has been one of the key aims of the REDUS project. 

“Amongst other things, we have developed new methods for conducting surveys that both improve our estimates and save time, which in turn makes the surveys cheaper”, says Erik Olsen. 

“We now have a more standardised method for trawl surveys, which means there is less variation in our results than when we started out in 2016. We have also increased the precision of the biological sampling used to estimate the age structure of fish stocks, which helps us to produce good age-structured population models. “ 

The IMR has also worked with the Norwegian Computing Centre to further develop the models used to estimate the size of fish stocks. 

“In addition, we’ve developed a new model to cover the ‘holes’ in our surveys, which can arise if ice, bad weather or technical problems prevent us from covering part of the survey area in a given year.” 

Sonar: not as easy to use as had been hoped 

Many of the IMR’s surveys are so-called acoustic surveys, where scientists use data from echo sounders to estimate the quantity of fish in the water. 

The REDUS project has sought to improve acoustic survey methods, with a particular emphasis on if and how sonar can be used to estimate the quantity of fish in a given volume of water. 

“We haven’t finished our work on this yet, because it turns out that using sonar to count fish is not as straightforward as we initially thought”, says Olsen. 

“Herring lottery” produces more representative samples 

To find out the age of the fish that are harvested, it is important for the samples taken from fishery catches to be as representative as possible. The best way to ensure this is to take the samples from a random selection of fish. 

The scientists working on REDUS therefore partnered with the fishing industry to develop a “lottery” for taking random samples of catches. The “lottery” was initially used for herring, but now it has been extended to other pelagic fish stocks. 

“Together with the Norwegian Computing Centre we have developed a statistical tool that calculates the age-specific catch, including the uncertainty of the estimate. It is important to include this data in our fish stock models”, says Erik Olsen. 

Simulating the outcome of different strategies 

One important goal of fisheries management in Norway and the other countries around the North Atlantic is to come up with good harvesting strategies. These strategies should enable you to harvest as much fish as possible, without putting a population at risk of collapse. 

“In order to find out the optimal harvesting strategy for a fish stock, you have to test thousands of potential strategies using data simulation. In REDUS we have developed a tool to do this, which allows us to simulate strategies both for individual populations and for whole ecosystems. The latter option means we can see how a strategy for a specific fish stock may impact other species.