Landscape genetics reveal broad and fine‐scale population structure due to landscape features and climate history in the northern leopard frog (Rana pipiens) in North Dakota

Authors

Justin M. Waraniak, North Dakota State University--FargoFollow
Justin D. L. Fisher, North Dakota State University
Kevin Purcell, North Dakota State University--Fargo
David M. Mushet, Northern Prairie Wildlife Research Center
Craig A. Stockwell, North Dakota State University--Fargo

ORCID IDs

https://orcid.org/0000-0002-6653-2021

https://orcid.org/0000-0001-7882-8674

Date of this Version

10-30-2018

Citation

Ecology and Evolution. 2019;9:1041–1060. www.ecolevol.org 1041-1061

Comments

U.S. Government

Abstract

Prehistoric climate and landscape features play large roles structuring wildlife populations. The amphibians of the northern Great Plains of North America present an opportunity to investigate how these factors affect colonization, migration, and current population genetic structure. This study used 11 microsatellite loci to genotype 1,230 northern leopard frogs (Rana pipiens) from 41 wetlands (30 samples/wetland) across North Dakota. Genetic structure of the sampled frogs was evaluated using Bayesian and multivariate clustering methods. All analyses produced concordant results, identifying a major east–west split between two R. pipiens population clusters separated by the Missouri River. Substructuring within the two major identified population clusters was also found. Spatial principal component analysis (sPCA) and variance partitioning analysis identified distance, river basins, and the Missouri River as the most important landscape factors differentiating R. pipiens populations across the state. Bayesian reconstruction of coalescence times suggested the major east– west split occurred ~13–18 kya during a period of glacial retreat in the northern Great Plains and substructuring largely occurred ~5–11 kya during a period of extreme drought cycles. A range‐wide species distribution model (SDM) for R. pipiens was developed and applied to prehistoric climate conditions during the Last Glacial Maximum (21 kya) and the mid‐Holocene (6 kya) from the CCSM4 climate model to identify potential refugia. The SDM indicated potential refugia existed in South Dakota or further south in Nebraska. The ancestral populations of R. pipiens in North Dakota may have inhabited these refugia, but more sampling outside the state is needed to reconstruct the route of colonization. Using microsatellite genotype data, this study determined that colonization from glacial refugia, drought dynamics in the northern Great Plains, and major rivers acting as barriers to gene flow were the defining forces shaping the regional population structure of R. pipiens in North Dakota.