Siscowet Lake Trout abundance modeling
Print
EmailEstimates on siscowet lake trout abundance in Lake Superior varies between agencies modeling fish population.
Michigan Sea Grant hosted a “Siscowet Lake Trout Abundance” workshop at the Lake Superior Technical Committee (LSTC) meeting in Duluth, MN in January 2012. The workshop helped answer the question on why estimates on siscowet lake trout abundance presented in a previous Michigan Sea Grant workshop had an estimate by U.S. Geological Survey (USGS) of 45 million pounds while the MSU Quantitative Fisheries Center data showed an estimate of 600 million pounds. Dan Yule, of USGS, provided a reminder of the results presented at the 2009 Baraga siscowet lake trout workshop. Siscowet lake trout densities were 4.2 times greater in day trawls versus night trawls. A wide range of sizes were captured with day and night bottom trawls, but only large siscowet lake trout captured in midwater trawls. The USGS is not certain where small fish go at night. Lakewide biomass estimates are based on 57 day bottom samples collected >80 m during spring from 2002-2006. Biomass (kg/ha) was estimated by weight/area swept by trawl. Estimated total biomass from areas >80 m were 2.34 kg/ha, with a total biomass 14,673 metric tons (range 8,134-21,291).
Jim Bence, of MSU Quantitative Fisheries Center, presented how his estimates were derived for the same workshop. Bence used ratios of lean to siscowet lake trout combined with assessment model estimates of lean lake trout abundance to derive estimates of siscowet lake trout biomass in Michigan waters of Lake Superior. There were a number of critical assumptions in his modeling, including the assumption that the CPUE of two lake trout forms was proportional to density. The resulting models were a function of depth and where possible incorporated year effects. The data from MI-2 through MI-7 indicate an estimate of " 40 siscowet lake trout per lean lake trout. Converting number to biomass, total biomass is 280,302 metric tons. Bence tried various ways to reduce numbers, but to no avail; the resulting estimates were still large. Overall, the Lake Superior Technical Committee is comfortable with ratio of lean to siscowet lake trout, which could be another important source of error in the modeling. Bence then posed whether the differences between trawl and gill net estimates could be because the assessment models provided overestimates of the number of leans. More data that is recent indicates that in a number of management zones (MI-5, MI-6, and MI-7), the estimates of leans are actually higher now, so differences, if recalculated today, would even be higher. Another potential factor raised was that the issue could be very low fishing rates, in comparison to assumed sea lamprey mortality. If sea lamprey mortality were actually lower than assumed, this would increase fishing mortality estimates and greatly reduce the number of leans estimated. This is the greatest uncertainty in the models.
The LSTC then discussed what it might like to see happen to improve confidence in the bottom trawls. It was suggested that further examination of the swimming performance of siscowet lake trout occur, either from an endurance study or from the literature. It was also raised that the LSTC is assuming gill nets catchability for lean and siscowet lake trout is the same and it may be easier to catch the different forms in gill nets. However, it was noted that the ratios of leans and siscowet lake trout in trawls and gill nets are similar. There were a number of suggestions made as to how the USGS might estimate catchability better, including performing mark-recapture studies, using different width trawls (which might capture larger fish) and using cameras on the trawl to estimate escapement. It was decided that flash photography was likely the best way to reduce light impacts if cameras are used to estimate escapement.
