Invasion by non-native plants into natural areas is an important component of
global change that threatens biodiversity and ecosystem structure and function.
Mountains are currently among the least invaded ecosystems, however, these
biodiversity hotspots are increasingly under threat of exotic plant invasion.
Evaluation of plant species distribution patterns in mountain ecosystems can provide
insight into dominant processes of plant invasion and inform management to reduce
the spread of non-native plants in mountains. The objectives of this study were to: 1)
evaluate the main drivers of plant invasions in the Wallowa Mountain Range of
northeastern Oregon, and the extent patterns of native species distributions and their
causes differed from non-native species along an elevation gradient; and 2)
investigate inter- and intra-annual understory vascular plant species variability along
an elevation gradient to inform long-term monitoring of plant invasion dynamics in
the Wallowa Mountains.
We sampled understory vascular plant communities in summer 2012 along
three forest roads in the Wallowa Mountains. Transects (n=20) were evenly stratified
by elevation (60 meters) along each road. Indicator species analysis (ISA) was used
to identify habitat tendencies for non-native species. Non-metric multidimensional
scaling (NMS) related community composition to environmental and species trait
factors. Predictors of non-native and native species abundance and richness were
evaluated using Non-parametric multiplicative regression (NPMR). Canopy
openness and elevation niche widths were modeled for all common species and
evaluated for evidence of high elevation or closed canopy specialization.
Non-native species richness decreased continuously with increasing elevation.
In contrast, native species richness displayed a unimodal distribution with maximum
richness at the mid-elevations. According to NMS and NPMR, elevation, canopy
openness, and disturbance were the strongest correlates of non-native species
abundance. Non-native species were concentrated in low elevation bunchgrass,
roadside, and open forest habitats with high canopy openness, and moderate to high
disturbance intensity. Overlays of species trait factors onto ordinations revealed that
the transition into the subalpine community and over-story canopy closure exceeding
approximately 60 percent were potential barriers to non-native species establishment.
Vegetation sampling was repeated three times (June, July, and August 2013)
during the growing season in summer 2013 for a subset (n=10) of our transects.
Blocked Multi Response Permutation Procedure (MRBP) was used to test the
hypothesis of no change in species composition between years (summer 2012 and
summer 2013) and between each sampling period (June, July, and August) within one
growing season. NMS related community composition at different times to
environmental and trait category factors. Blocked Indicator Species Analysis was
used to identify species with tendencies towards specific sampling periods. Important
predictors of vegetation change were identified and indicator species distributions
were related to NMS axes using NPMR.
Vegetation composition changed both between years and between each
sampling period within one growing season. Species richness increased with
sampling effort, dominant species were consistently found throughout all sampling
periods and rare species showed the highest rate of turnover between sampling
periods. The highest species richness and greatest number of indicator species were
associated with the early summer (June) sampling period. Approximately 25% of all
non-native species showed tendencies towards a specific sampling period where
greater than half favored the June sampling period within lowland bunchgrass
communities. Annual species including the most dominant non-native species within
the lowland and montane zones showed the highest inter- and intra-annual variability
compared to other trait categories. Total species richness was positively related to
turnover and was the most important predictor of relative abundance change between
all sampling periods. Species turnover was highest in species rich montane plots;
however, trait abundance and richness did not statistically differ in montane plots.
These data suggest that elevation, canopy, openness, and disturbance are
important factors structuring non-native plant distributions in the Wallowa
Mountains. Our findings that non-native species tended to be concentrated in low elevation,
early successional habitats coupled with a lack of high elevation or closed
canopy specialization indicates that non-native plants that occurr at the highest
elevations are likely generalist species that depend on disturbance (primarily open
over-story canopy) for spread into higher elevation sites until they are eventually
filtered out by harsh environmental conditions coincident with the subalpine
transition zone. Collectively these results suggest that interactions between
introduction pathways that favor human assisted dispersal into the low-elevations
where ruderal species are favored and environmental filters (elevation, canopy
openness, and disturbance) are the dominant factors structuring non-native plant
distributions in the Wallowa Mountains.
Our results also indicate that long-term monitoring of vegetation change in the
Wallowa Mountains will be improved through periodic sampling during each
growing season. Sampling once during the spring or early summer and then again
coincident with peak biomass should improve estimates of species distributions and
diversity patterns by capturing both vernal and late season species. Early season
sampling may be particularly important for monitoring non-native plants in the
Wallowa Mountains as most non-natives were concentrated in low elevation semiarid
bunchgrass communities where spring and early summer specialization is