Effect of type and extent of vegetative cover on terrestrial small mammals in the semi-arid Okanagan Valley, British Columbia

The intermontane grasslands and shrub-steppe rangelands of the Okanagan Valley of southern British Columbia, Canada have a long history of anthropogenic disturbances. Livestock grazing, urbanization and agriculture have all played roles in altering the historic habitat mosaic. Less than 10% of semi-arid grasslands remain in a natural state and are occupied by several species of terrestrial small mammals. These species may provide useful insight into the effects of habitat alteration because they occur in sufficient abundance for sampling, select habitat at multiple scales, and are strongly influenced by habitat structures. We investigated a wide range of semi-arid habitats, most with a history of livestock grazing, and predicted that habitats with a high percentage cover of native grasses and forbs, and some degree of vertical structure, would support the highest overall abundance and species diversity of small mammals. We studied small mammal communities from 1994 to 1997 on 24 sites selected to represent a wide variety of habitat conditions that spanned a range of plant species assemblages. The 24 sites covered a range of levels of mineral soil exposure with no plant cover ranging from 1.4% to 27.2%. Mean cover of major shrubs were bitterbrush (Purshia tridentata) that ranged from 0% to 29.8% and sagebrush (Artemisia tridentata) that ranged from 0% to 16.2%. Mean cover of grass-forbs ranged from 21.6% to 69.5%. The Great Basin pocket mouse (Perognathus parvus) was the dominant species numerically (mean of 29–44 mice/ha in high density years) and showed a strong positive relationship to mean cover of total shrubs, at least above 20% ground cover. The deer mouse (Peromyscus maniculatus) was the second most abundant species (mean of 23–26/ha in high density years) with a negative correlation of mean cover of grasses and forbs. The western harvest mouse (Reithrodontomys megalotis) reached mean densities of 5–6/ha and was positively correlated with grass-forb sites but not Artemisia-dominated sites. The consistently low numbers of montane voles (Microtus montanus) (2–3/ha, up to 6–10/ha at high densities) and lack of any relationship with cover of grasses and forbs reflected the poor habitat conditions for this keystone species. Mean abundance and species diversity of total mammals were positively correlated with a measure of vegetation cover density, thereby suggesting that some degree of protection from predators was perceived.