Fig 1.
Schematic representation of migrant source and recipient population models.
(A) The recipient population was allowed to stabilize for 100 years with a carrying capacity of 1000 individuals and was then subjected to a 50-year population decline to simulate environmental change; this population size decline happened over 10 years, and the population was held at the population minima for 40 years at three levels representing IUCN (International Union for Conservation of Nature) and state agency classifications: vulnerable for species that have a > 10% decline in 10 years, KB = 700; endangered with a 50–70% decline in 10 years, KB = 300; and critically endangered with a 80–90% decline in 10 years, KB = 100. After this reduced population period, we simulated habitat restoration by allowing the population to grow again, up to the historical carrying capacity. Additional simulations for populations that did not face population decline were used as a control. (B) Migrants entered the recipient population at a set frequency: 1 migrant per generation, 100 individuals at a single time period (e.g., 100 individuals in year 151) and 25 individuals at four time periods (e.g., 25 individuals in years 151, 165, 181, 195). As a basis for comparison, we also included scenarios where migration was completely absent. (C) We tracked the demographic and evolutionary response of the recipient populations across 350 years.
Table 1.
Description of parameters and values tested.
Fig 2.
Population genetic and demographic responses in populations compared among extinction risk categories in the absence of migration.
The census population size (A) is illustrated for 100 replicates of the simulation runs, where each line represents a single iteration. Here, the census size is depicted as less than the carrying capacity to match the point in the model when population response parameters were quantified (i.e., after Allee effects, reproduction, and subsequent mortality). Population response is summarized across these replicates as observed heterozygosity (B), divergence from the historical recipient populations over time (C), and divergence between the recipient and historical migrant source populations at each year (D). In the absence of migration, a drop in census size can push populations on a new evolutionary trajectory with different allele frequencies as a result of the loss of genetic variants that occurred due to drift. Lines in B-D represent mean values across the 100 replicate runs and polygons represent confidence intervals scaled to compare evolutionary outcomes between each parameter set, assuming alpha = 0.05 (i.e., 95% confidence intervals).
Fig 3.
Genetic and demographic responses to migration in populations with different starting minor allele frequencies and a population crash to 70% of the historical population size.
Solid lines indicate a single migrant per generation and dotted lines depict scenarios without migration. Starting minor allele frequencies were either low (0.05–0.15) in both the migrant source and recipient populations (cyan), high (0.4–0.5) in both populations (purple), high in the migrant source populations and low in the receiving populations (pink), or low in the migrant source populations and high in the receiving populations (blue). Lines represent mean values across 100 replicate runs and polygons represent confidence intervals scaled to compare evolutionary outcomes between each parameter set, assuming alpha = 0.05 (i.e., 95% confidence intervals). The proportion of migrant ancestry (A), observed heterozygosity (B), divergence from the historical recipient populations over time (C), and divergence each year from the historical migrant source populations (D) are depicted in each panel. These outcomes suggest that the minor allele frequencies influence evolutionary trajectory of populations connected by migration when migrations occur during a population crash, such that these recipient populations with migrants from populations with higher minor allele frequencies are less diverged from the migrant source populations compared to recipient populations with migrants with lower minor allele frequencies.
Fig 4.
Population genetic and demographic outcomes of incorporating migration into population crash recovery.
Populations were shrunk (years 100–150) to sizes reflecting common extinction risk classifications (critically endangered, 90% reduction; endangered, 70% reduction; vulnerable, 30% reduction; no population reduction), and movement of a single individual per year (solid line) was used to support and bolster these populations through remediation (year 151) and population recovery. These trends were compared to the same demographic patterns but without migration (dotted lines). The proportion of migrant ancestry present in the recipient populations (A), observed heterozygosity (B), divergence of the recipient populations from the historical populations over time (C), and divergence of the recipient populations from the migrant source populations each year (D) illustrate the new evolutionary trends resulting from these migration decisions. Lines represent mean values across 100 replicates and shaded areas represent the confidence intervals needed to compare among parameter sets assuming alpha = 0.05 (i.e., 95% confidence intervals).
Fig 5.
Population genetic and demographic outcomes in populations compared among extinction risk categories while incorporating migration into population crash recovery.
Populations were shrunk (years 100–150) to sizes reflecting common extinction risk classifications (critically endangered, 90% reduction; endangered, 70% reduction; vulnerable 30% reduction; no population reduction), and movement of a single individual per year (solid line), burst migration of 100 individuals once (year 151; dashed line), and four pulse migrations of 25 individuals (years 151, 165, 181, 195; dashed and dotted line) was used to support and bolster these populations through remediation (year 151) and population recovery. These trends were compared to the same demographic patterns but without migration (dotted line) and in the absence of a population crash (grey lines). The proportion of migrant ancestry present in the recipient populations (A-C), observed heterozygosity (D-F), divergence of the recipient populations from the historical population over time (G-I), and divergence of the recipient populations from the historical migrant source populations each year (J-L) illustrate the new evolutionary trends resulting from these management decisions every 50 years. Lines represent mean values across 100 replicates and error bars represent the confidence intervals needed to compare among parameter sets assuming alpha = 0.05 (i.e., 95% confidence intervals).
Fig 6.
Population genetic and demographic responses with a population crash to 70% of the starting population with different timings of management intervention.
Assisted migrations were either burst translocations (100 individuals once) or pulsed (25 individuals 4 times). Light grey lines depict assisted migrations implemented concurrently with habitat restoration (year 125 for burst; years 125, 140, 155, and 170 for pulse) and black lines indicate migrations that occurred after habitat restoration was completed (year 151 for burst; years 151, 165, 181, and 195 for pulse). Lines represent mean values across 100 replicate runs and polygons represent confidence intervals scaled to compare evolutionary outcomes between each parameter set, assuming alpha = 0.05 (i.e., 95% confidence intervals). The proportion of migrant ancestry (A), observed heterozygosity (B), divergence from the historical recipient populations over time (C), and divergence each year from the historical migrant source populations (D) are depicted in each panel. These outcomes suggest that implementing a pulse migration after restoration and with population growth will supplement populations with similar levels of increased genetic diversity and with less influence of alleles with migrant ancestry.