Genetic, demographic, and ecological effects of habitat fragmentation

Habitat fragmentation is a leading cause of the loss of biodiversity. The direct and indirect ecological effects of habitat fragmentation are poorly understood. I used a combination of genetic, demographic, and experimental tools to investigate the effects of habitat fragmentation on the California red-backed voles (Clethrionomys californicus) and deer mice (Peromyscus maniculatus) in forests of southwestern Oregon.; Demographic and genetic data provided insights into vole population ecology that would not have been possible with only one type of data. Vole abundances fluctuated greatly at sizes below 50 individuals per fragment. I analyzed variation in mtDNA and five nuclear microsatellite loci in vole samples collected from two forest fragments and two unfragmented control sites in 1998 and 1999. Fragments had significantly lower mtDNA allelic diversity than controls, but not nuclear heterozygosity or numbers of alleles. The use of only trapping and mtDNA marker data would imply that fragment populations are at least partially isolated and vulnerable to inbreeding depression. In contrast, the combined abundance estimates and all genetic data show that small fragment populations must be linked to nearby forests by high rates of gene flow, which is probably male-biased. This immigration is likely important to the persistence of small vole populations on forest fragments. However, experimental evidence suggests that the ability of dispersing voles to inoculate seedlings in clearcuts surrounding fragments with ectomycorrhizal fungi, which are essential for forest regeneration, may be limited by abiotic conditions in clearcuts. I found no evidence of reduced fitness (adult survival) on fragments relative to controls.; I found positive effects of fragmentation on deer mice and negative effects of mice on trillium (Trillium ovatum), an understory plant that shows reduced recruitment and increased extinction risk in fragmented forests. Multi-strata and closed population mark-recapture models showed that mouse survival was highest in clearcuts, intermediate on forest fragments, and lowest in contiguous forests. Matrix projection models suggested this should result in drastic differences in habitat-specific population growth rates. In agreement with this, mouse densities were higher in fragmented than in unfragmented forest sites. Predation on trillium seeds was higher in areas of higher mouse relative abundance.; These results provide evidence of subtle, yet important population and community level impacts of fragmentation. For California red-backed voles, there was no detectable fitness effects of habitat fragmentation, but their ecological role as ectomycorrhizae dispersers appears to be compromised in clearcuts. There also appears to be mechanism for a human-caused trophic cascade in which positive effects of fragmentation on deer mouse demography lead to increased predation rates and extinction risk for trillium populations.