Population genetics and geographic origins of mallards harvested in northwestern Ohio

Abstract

The genetic composition of mallards in eastern North America has been changed by release of domestically-raised, game-farm mallards to supplement wild populations for hunting. We sampled 296 hatch-year mallards harvested in northwestern Ohio, October-December 2019. The aim was to determine their genetic ancestry and geographic origin to understand the geographic extent of game-farm mallard introgression into wild populations in more westward regions of North America. We used molecular analysis to detect that 35% of samples were pure wild mallard, 12% were early generation hybrids between wild and game-farm mallards (i.e., F1-F3), and the remaining 53% of samples were assigned as part of a hybrid swarm. Percentage of individuals in our study with some form of hybridization with game-farm mallard (65%) was greater than previously detected farther south in the mid-continent (~4%), but less than the Atlantic coast of North America (~ 92%). Stable isotope analysis using δ2Hf suggested that pure wild mallards originated from areas farther north and west than hybrid mallards. More specifically, 17% of all Ohio samples had δ2Hf consistent with more western origins in the prairies, parkland, or boreal regions of the mid-continent of North America, with 55%, 35%, and 10% of these being genetically wild, hybrid swarm, and F3, respectively. We conclude that continued game-farm introgression into wild mallards is not isolated to the eastern population of mallards in North America, and may be increasing and more widespread than previously detected. Mallards in our study had greater incidence of game-farm hybridization than other locales in the mid-continent but less than eastern North American regions suggesting further need to understand game-farm mallard genetic variation and movement across the continent.

Fig 1. Waterfowl survey areas currently assigned to the eastern and mid-continent stocks of mallards as identified by the United States Fish and Wildlife Service (USFWS) for management purposes.

Spatial layers of populations were provided by the USFWS and base outline political boundary layers are free, open-source from ESRI. WBPHS, Eastern Survey, and Great Lakes shapefiles were provided by USFWS. U.S. political boundaries were obtained through the U.S. Census Bureau (https://www.census.gov/geographies/mapping-files/time-series/geo/carto-boundary-file.html). Canadian province political boundaries were obtained through Open Canada census Bureau (https://open.canada.ca/data/en/dataset/a883eb14-0c0e-45c4-b8c4-b54c4a819edb). Except for the shapefiles listed above, no external data was used for the creation of the figure. ESRI ArcMap 10.8 was used for the creation of the figure.

Fig 2

Population structure analyses based on 22,136 independent bi-allelic ddRAD-seq SNPs assayed across Ohio (OH) samples, as well as reference wild (WMA) and game-farm (GFM) mallards, including (A) a plot of the first two principal components of the principal components analysis (PCA) and (B) ADMIXTURE assignment probabilities attained under an optimum K populations of 2 (S2B Fig). Finally, we provide the (C) expected average and range of assignment probabilities across simulated generations (left), as well as the proportion of Ohio samples falling into those generations (right).

Fig 3. Geographic depiction of probable origins of hatch-year mallards harvested in Ottawa and Sandusky counties, Ohio, USA, from October to December 2019.

Legend numbers correspond to the number of individuals assigned to each pixel based on a 2:1 odds ratio. A) all samples in our study without application of a banding data informed weighted spatial mask, B) origins constrained by a banding data informed weighted spatial mask, C) Origins of pure wild mallards without application of a banding data informed weighted spatial mask, and D) Origins of hybrid-swarm mallards without application of a banding data informed weighted spatial mask. U.S. and Canadian political boundaries were obtained through the R package ‘rnaturalearth’. Atlantic and Mississippi flyway shapefiles were obtained from the metadata of Weltzin et al. 2018 (https://www.sciencebase.gov/catalog/item/5b75d10ee4b0f5d5787fea6f). Atlantic and Mississippi flyway boundaries and data were used from Weltzin et al. 2018 (https://www.sciencebase.gov/catalog/item/5b75d10ee4b0f5d5787fea6f). The calibration equation used to produce the assignment to origin was obtained from van Dijk et al. 2014. Except for the shapefiles listed above, no external data was used for the creation of the figure. Figures were created using the R statistical framework (version 4.0.2, R Core Team 2018).