Population Genomics of Lake Sturgeon (Acipenser fulvescens) in the Churchill, Hayes, and Nelson Rivers. A draft report prepared for Manitoba Hydro by Université Laval and North/South Consultants Inc.

A report prepared for Manitoba Hydro by Université Laval and North/South Consultants Inc. Manitoba Hydro has been investigating the feasibility of developing a hydroelectric generating station at the Conawapa site, located approximately 29 km downstream of the Limestone Generating Station (GS) on the Nelson River. An Environmental Studies Program was developed to provide information that will: - assist in applying the principles of sustainable development in designing and planning the project; and - form the baseline for an environmental impact assessment to meet licensing requirements should a decision be made to proceed with the project. The program includes terrestrial, wildlife, archaeology, and aquatic components and has been developed and conducted with the participation of local First Nations. The Conawapa aquatic studies program was developed to provide information on the interrelated components of the lower Nelson River aquatic ecosystem. The program includes studies of the physical habitat, water quality, detritus, algae, aquatic macrophytes, aquatic invertebrates, fish and marine mammals. Individual reports are being prepared that focus on specific aquatic components and locations. This report presents results of a Lake Sturgeon population genetics study using Genotype-by-Sequencing (GBS) to investigate relatedness among groups of fish resident in the Churchill, Hayes, Fox, Gods, and Nelson rivers. The primary objectives of the study were as follows: - Develop a high-resolution genetics toolkit for Lake Sturgeon, which would enable population assignment tests, parentage, and parentage/sibship inference analyses; and - Test the methods using samples collected from known parents and hatchery reared progeny. Using the developed genetic markers, the secondary objectives of this study were to: - Assess population structure of Lake Sturgeon along the Nelson River, between the Jenpeg GS and Hudson Bay (>500 km) i.e., identify populations and determine the amount of gene flow via population assignment tests; and - Focusing on juveniles (and where possible a single cohort), use sibship inference to determine the number of contributing females and the degree of relatedness amongst progeny, and further, determine if siblings of upstream residents exist in downstream locations that are separated by rapids or a generating station. Previous Lake Sturgeon genetic studies focused on microsatellite markers, which use several relatively long repeat sequences of neutral DNA (i.e., segments of the genetic code not under positive or balancing selection). However, in Lake Sturgeon, microsatellite markers identified to date exhibit a low level of differentiation, which means that the quantity of information that each marker provides in relation to population structuring and gene flow is low. Single nucleotide polymorphisms (SNPs) provide even less information per marker, but they are far easier to identify in a given genome, meaning that the relative amount of information that can be obtained via a GBS approach far exceeds that which can be obtained by microsatellites in this species. Lake Sturgeon tissue samples were collected from adult Lake Sturgeon captured at known or suspected spawning sites in the Nelson, Fox, Gods, Hayes, and Churchill rivers, as well as from juvenile Lake Sturgeon captured in the various locations. A total of 1,178 Lake Sturgeon were sequenced. The results revealed genetic differences among many of the putative ’populations‘. Given generation time of Lake Sturgeon (~35 years) relative to the timescale of Nelson River hydroelectric development (first station built ~50 years ago), it was clear that structured populations existed along the Nelson River long before the construction of hydroelectric dams; if there was any historical movement of fish past major historical falls/rapids e.g., Grand Rapid [Kelsey GS] and Kettle Rapids [Kettle GS], this would not occur. A variety of clustering methods were employed to assess population structure, but all were consistent in indicating a lack of genetic structure among Weir, Angling and Lower Limestone Rapids spawning groups, suggesting one inter-mixed population. Based on the inclusion of larval fish produced in the hatchery, sibship analyses were able to resolve known control relationships, thereby validating the method. Sibship results for the cohort from Sea Falls included in the analysis were consistent with expectations of numerous full-sibling relationships; this cohort had previously been generated following rearing the eggs of a single female at the Grand Rapids hatchery. Sibship results also provide information on the quantity of contemporary gene-flow occurring among population within the study area. Specifically, in the context of the secondary objectives, results indicated that 2008 cohort juveniles in Stephens Lake and Long Spruce Reservoir were likely not spawned in those locations, as each individual sequenced was determined to have at least one sibling located further upstream in Gull Lake. Downstream redistribution of larval Lake Sturgeon must have occurred following spawning at an upstream location. In conclusion, results of the current study dramatically improve the understanding of both historical and contemporary Lake Sturgeon population structure. Population units can essentially be defined by the presence of contiguous spawn-drift-settle-establish habitat sequences, but the distance over which these units occur varies even along the length of the Nelson River. The upper reaches of the Nelson River flow through the Boreal Shield ecozone where riverine habitats were historically fragmented by repeated rapid/fall-river- lake sequences, and results showed that populations existed historically in relatively small stretches of river, with only one-way (downstream) gene flow occurring. Downstream of Kettle Rapids (Kettle GS), the lower Nelson River flows through the Hudson Plain ecozone and the nature of habitat changes dramatically, which explains the presence of a single inter-mixed population downstream of the Limestone GS, despite utilization of multiple spawning locations