Cisco Recruitment Dynamics in Lake Superior during 1978–2007

Abstract

Historically, the cisco (Coregonus artedi) was the most abundant fish species in the Great Lakes, but by the mid-1900s, cisco populations were greatly reduced throughout the basin. Over-fishing, habitat degradation, and interactions with exotic species caused cisco yield to decline by 80–99% in each lake. Declining yields forced commercial fishers to target other species and brought about new regulations designed to prevent further losses, but except for a few strong year-classes in the 1990s, cisco stocks failed to recover in the lower Great Lakes. Reduced commercial fishing pressure enabled cisco to recover in portions of Lake Superior, but historic stock structure was altered, and abundance is now driven by highly erratic age-1 recruitment and few year-classes of adults. Management agencies have begun exploring the feasibility of restoring cisco stocks throughout Lake Superior and the entire Great Lakes basin, but limited understanding of factors that drive recruitment variation and the spatial scale at which these factors operate remain barriers to establishing self-sustaining populations. Identifying major density-independent and density-dependent factors that regulate age-1 cisco recruitment dynamics in Lake Superior, and the spatial scale at which these factors operate, would be invaluable to cisco restoration and management efforts throughout Lake Superior and the entire Great Lakes basin. In Chapter 1, I used a Ricker stock-recruitment model to identify and quantify the appropriate spatial scale for modeling age-1 cisco recruitment dynamics in Lake Superior. I found that recruitment variation of cisco in Lake Superior was best described by an 8-parameter regional model with separate stock-recruitment relationships for western, southern, eastern, and northern stocks. The spatial scale for modeling was ~260 km (range = 230–290 km). I also found that the density-independent recruitment rate and the rate of compensatory density-dependence varied among regions at different rates. The density-independent recruitment rate varied 2-fold among regions (range = 2.4–4.9 age-1 recruits/spawner) and the rate of compensatory density-dependence varied 21-fold among regions (range = -0.2 to -3.4 spawners-1). Finally, I found that peak recruitment and the spawning stock size that produced peak recruitment varied among regions. Peak recruitment varied 10-fold among regions (range = 0.5–5.4 age-1 recruits/ha) and the spawning stock size that produced peak recruitment varied 21-fold among regions (range = 0.3–6.1 spawners/ha). My findings support the hypothesis that cisco recruitment is regulated within four different regions of Lake Superior, suggest that large-scale abiotic factors driving compensatory density-dependence are more important than small-scale biotic factors in regulating cisco recruitment in Lake Superior, and suggest that fishery managers throughout Lake Superior and the entire Great Lakes basin should address cisco restoration and management efforts on a regional scale in each lake. In Chapter 2, I used a generalized version of the Ricker stock-recruitment model to identify and quantify the effects of biotic and abiotic factors on age-1 cisco recruitment dynamics within four different regions of Lake Superior. I found that recruitment variation of cisco in Lake Superior was correlated to adult spawning stock size in all four regions, the density of juvenile cisco during the year prior to cisco hatching in three of four regions, average April air temperature during spring when ciscoes were 11–12 months of age in three of four regions, average April wind speed during spring when ciscoes were hatching in two of four regions, and the biomass of rainbow smelt during the year of cisco hatching in one of four regions. My findings support the hypothesis that different biotic and abiotic factors regulate cisco recruitment within different regions of Lake Superior, suggest that air temperature during spring when ciscoes are 11–12 months of age drives recruitment variation on a lake-wide scale, whereas adult spawning stock size, intraspecific interactions with juvenile cisco, wind speed during spring when ciscoes are hatching, and interspecific interactions with rainbow smelt regulate recruitment variation on a regional scale in Lake Superior, and suggest that fishery managers throughout Lake Superior and the entire Great Lakes basin should evaluate the potential effects of similar biotic and abiotic factors on recruitment prior to addressing cisco restoration and management efforts in each lake.