Models provide clues to finding yellow perch in Lake Michigan

Understanding where and when fishes will be present is a key for angling success but it is also important to fish ecologists and fisheries scientists.  Researchers are using new statistical modeling techniques to predict increases or decreases in local abundance based on changes in various factors.  Among the many factors that impact the abundance and distribution of fish are the abundance and distribution of predators, competitors and conspecifics, water quality and physical habitat.  Bacheler et al. (2011) created vary-coefficient generalized additive models (GAMs) to identify which factors influence abundance and distribution of yellow perch Perca flavescens in the Bay of Green Bay (just Green Bay hereafter).

The authors used their models to look at how various factors effected catch per unit effort (CPUE; number of fish per min) of age-0 and age-1+ (age-1 and older) yellow perch trawls in southern Green Bay.  CPUE data for the model came from otter trawls conducted by the WDNR as part of their annual fishery-independent survey of the yellow perch population.  The factors the authors used as predictor variables for local yellow perch CPUE were: overall global changes in indices of abundance, double-crested cormorant Phalacrocorax auritus abundance, water temperature, dissolved oxygen, water clarity, bottom depth and position.  Otter trawls were deployed from a research vessel at predetermined sites in August or September.  For these analyses only data from surveys conducted from 1985 to 2009 were used.  Yellow perch less than 100 mm total length (TL) were considered age-0, fish bigger than 100 mm TL were classified as age-1+.  Global age-1+ indices of abundance were estimated by statistical catch-at-age model based on commercial effort and harvest, recreational effort and harvest and trawl survey data.  Global age-0 index of abundance was estimated from the annual trawl survey.  Cormorant abundance was estimated from nest counts conducted by the U.S. Department of Agriculture, Wildlife Services.  Water quality data came from the Green Bay Metropolitan Sewerage District records.

The vary-coefficient GAM showed that a combination of cormorants, dissolved oxygen, water clarity, bottom depth and global yellow perch indices of abundance affected abundance and distribution of age-0 and age-1+ yellow perch in Green Bay.  Local age-0 and age-1+ yellow perch CPUE decreased with increased cormorant abundance with noticeably lower CPUE near cormorant nesting islands.  Local age-0 and age-1+ yellow perch CPUE was higher in nearshore areas during years with high dissolved oxygen levels.  Local age-0 and age-1+ yellow perch CPUE was higher near the mouth of the Fox River during years with higher water clarity.  Local age-0 yellow perch CPUE decreased with increased depth.  Local age-1+ CPUE was maximized at 13 m of depth.  Age-0 yellow perch spread to deeper, offshore waters when global CPUE is high and pulled back to shallow waters of southwest Green Bay when global CPUE is low.  Age-1+ CPUE was not affected by global CPUE.

Implications for anglers

            The authors point out that the model explained 63-65% of the variation in yellow perch CPUE.  They admit that variables were chosen based on the availability of reliable data.  Reliable data on predators like walleye Sander vitreus and American white pelicans Pelecanus erythrorhynchos, competitors like alewives Alosa pseudoharengus and round gobies Neogobius melanostomus and invasive species like zebra mussels Dreissena polymorpha were not available.  The authors had a large amount of data: 195,545 age-0 yellow perch and 91,510 age-1+ yellow perch were collected from 1,808 trawls over 25 years.  The authors argue that because they had a large number of samples and a high degree of consistency between the age-0 and age1+ models, the patterns they observed are sound.

            For experienced anglers, this study likely confirms patterns you’ve already noted.  For the novice, this study can help improve your yellow perch fishing not just in Green Bay but all along the Lake Michigan coast.  The information regarding dissolved oxygen is hard to incorporate into your fishing strategy.  The best way to use this information would be to carry an oxygen sensor with your fishing gear and monitor dissolved oxygen levels.  The sensors are expensive (the cheapest one I found was $150) and, therefore, might not be worth adding to your arsenal.  You can find records online but they are usually delayed a few years.  The other variables are pretty easy to incorporate into your fishing strategy.  Water clarity can be judged visually, simply look for good water clarity and fish there.  Avoid areas with poor water clarity and days when the rivers are spitting out a lot of mud.  Boaters with depth finders should look for fish in waters 13 m deep (42-43 feet).  Shore anglers will have difficulties accessing water 13 m deep.  Anglers should also avoid areas frequented by cormorants.  Since cormorants tend to move around it may be difficult to know if you’re favorite fishing spot is also a cormorant feeding grounds.  If you fish often enough, you should be able to get some idea about the cormorant presence in your area.  Please share any ideas on chasing away cormorants that won’t attract the attention of law enforcement.

Bacheler, N.M., T.J. Paoli and G.M. Schecht. 2011. Controls on abundance and distribution of yellow perch: predator, water quality and density dependent effect.  Transaction of the American Fisheries Society. 140:989-1000.

Selected definitions

Conspecifics: individuals of the same species

Distribution: the geographic range of an organism

Fishery-independent: in the absence of any fishing activity

Trawl:  a method of fishing involving boats pulling a net through the water